In article , Patrick Turner
wrote:

John Byrns wrote:

In article , "craigm"
wrote:
If the discharge current is constant (does not change due to the
voltage on
the capacitor), then this approach leads to ripple that does not vary with
the voltage on the capacitor.

With one exception, see below, the ripple should be more or less constant
with this scheme. At one time I was enamored with this scheme, but gave
it up for reasons I no longer remember. A reason that comes to mind now,
but was not the one I can't remember, is that while the ripple may be
relatively constant with modulation, the frequency at which tangential
clipping sets in will be dependent on the carrier amplitude, since the
slope of the discharge is independent of carrier level.

The "tangential clipping" also becomes more likely
if the audio modulation F is higher

Didn't I just I just get through saying that in my last paragraph above?

and the audio voltage is high.

I'm not sure what you might mean when you say "and the audio voltage is
high"? It is certainly true that tangential clipping gets worse as the
modulation level increases, if that is what you mean, but the "audio
voltage" is an imprecise way to talk about the modulation level.

That's because of the discharge rate of the cap charged by the diode.
Its worse with a detector which does not have a biased diode
and constant current source.

Not completely true, with a current source the discharge current doesn't
track the carrier level as it does with the common diode envelope
detector, so as the carrier level increases, tangential clipping sets in
at lower and lower audio frequencies, and lower modulation levels. Use of
a current source restricts the normally wide dynamic range of a diode
envelope detector, in terms of the carrier level.

In my detector, the output voltage cannot ever go to to zero, or 0V.

I didn't notice anyone saying the output voltage of your detector actually
went to zero, and how is that relevant?

Its at 55 volts even with no RF input,

And I will bet that there is no ripple with no RF input.

and then when a carrier of 10v p-p appears,

you get a rise in the direct voltage voltage across the 270pF of 5 volts
to +60v.
Then if there was 100% modulation, you get a varying voltage at the 270
pF between
+55v and +65v.
For low frequency audio, there is almost no change in the ripple voltage
at any
part of the audio
wave form.

I assume you put the "almost no change" part in because obviously the
ripple voltage must go to zero on negative modulation peaks of 100%.

But at 20 kHz modulation F, the ripple voltage is less on the negative going
slopes of the sine wave.

That may be a significant observation, I suspect that this is exactly what
you would expect when you are approaching the point of tangential
clipping, assuming by "negative going slopes of the sine wave" you mean
the down slope headed towards maximum negative modulation, which with your
detector is the same as negative going voltage, unlike the normal diode
detector.

I also believe it is a lot easier to do this in the solid state world than
in the tube world.

There are high voltages available in vacuum tube circuits which in
combination with a large resistor would make a pretty good approximation
to a current source.

If you use my detector circuit, you could easily establish a -200v
voltage source,

and take a 4.7M resistor from the 270 pF, to give a closer approximation to a
CCS.

If you don't bother with the cathode follower, then there is no need for a
-200 volt supply, and the already existing +200 volt supply will do the
job.

But speaking of the cathode follower, I may have found a justification for
the cathode follower in your circuit. It appears that when you apply bias
to the detector diode, at least a bias beyond that required to overcome
that caused by a non unity AC/DC load ratio, as in your circuit, that the
bias will cause the diode to present a nonlinear load to the source
driving it, over the audio cycle, beyond any non linearity due to non
ideal diode characteristics. I could be all wet about this, I need to
further corroborate my math, but if it is right, and diode bias does
result in a nonlinear load on the driving stage, then your cathode
follower to drive the diode does serve a useful purpose in your circuit.

If this is correct, the lesson I would take is to avoid using bias on the
diode, and instead make sure that the AC/DC load ratio is as close to
unity as practicable, either by using a cathode follower after the
detector, or by taping the audio output way down on the detector load
resistor, as the QUAD AM tuner does. Also for lowest distortion when bias
is used on the diode, the bias should track the signal level, rather than
be fixed as your bias is. This is especially important in a design like
yours where the AGC circuit provides rather loose control of the carrier
level at the detector.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , "craigm"
wrote:
If the discharge current is constant (does not change due to the
voltage on
the capacitor), then this approach leads to ripple that does not vary with
the voltage on the capacitor.

With one exception, see below, the ripple should be more or less constant
with this scheme. At one time I was enamored with this scheme, but gave
it up for reasons I no longer remember. A reason that comes to mind now,
but was not the one I can't remember, is that while the ripple may be
relatively constant with modulation, the frequency at which tangential
clipping sets in will be dependent on the carrier amplitude, since the
slope of the discharge is independent of carrier level.

The "tangential clipping" also becomes more likely
if the audio modulation F is higher

Didn't I just I just get through saying that in my last paragraph above?

and the audio voltage is high.

I'm not sure what you might mean when you say "and the audio voltage is
high"? It is certainly true that tangential clipping gets worse as the
modulation level increases, if that is what you mean, but the "audio
voltage" is an imprecise way to talk about the modulation level.

One could have 30% modulation, and have a
high audio output, and yes, the tangent Dn could be high,
and the same as if you had 100% modulation but with a lower carrier voltage.



That's because of the discharge rate of the cap charged by the diode.
Its worse with a detector which does not have a biased diode
and constant current source.

Not completely true, with a current source the discharge current doesn't
track the carrier level as it does with the common diode envelope
detector, so as the carrier level increases, tangential clipping sets in
at lower and lower audio frequencies, and lower modulation levels. Use of
a current source restricts the normally wide dynamic range of a diode
envelope detector, in terms of the carrier level.

I think you are wrong here.

The CCS allows a straight line cap discharge.
Using a cap discharge R which has a much wider voltage change across it
causes a discharge rate to vary, and you will see typical
curved discharge line, so that the time constant action is seen immediately.

If only you spent 30 minutes in your workshop with a CRO, and built my detector,
and observed carefully with your CRO, you'd see what I meant.



In my detector, the output voltage cannot ever go to to zero, or 0V.

I didn't notice anyone saying the output voltage of your detector actually
went to zero, and how is that relevant?

Well I though you implied that.
Its relevant that my detector output voltage never goes to 0V;
the audio voltage is the change in the output voltage of around
+/- 5 peak volts, so the current change in the 1M which discharges the 270 pF
doesn't vary much.



Its at 55 volts even with no RF input,

And I will bet that there is no ripple with no RF input.

That is correct.




and then when a carrier of 10v p-p appears,

you get a rise in the direct voltage voltage across the 270pF of 5 volts
to +60v.
Then if there was 100% modulation, you get a varying voltage at the 270
pF between
+55v and +65v.
For low frequency audio, there is almost no change in the ripple voltage
at any
part of the audio
wave form.

I assume you put the "almost no change" part in because obviously the
ripple voltage must go to zero on negative modulation peaks of 100%.

Sure, but the ripple is still there at 99% modulation.

Since 99% of what we hear comes from the first 95% of the modulation,
then the tiny tiny amount of flattening in the wave form
between 99% and 100% modulation is utterly irrelevant.

Build my detector, and see for yourself!



But at 20 kHz modulation F, the ripple voltage is less on the negative going
slopes of the sine wave.

That may be a significant observation, I suspect that this is exactly what
you would expect when you are approaching the point of tangential
clipping, assuming by "negative going slopes of the sine wave" you mean
the down slope headed towards maximum negative modulation, which with your
detector is the same as negative going voltage, unlike the normal diode
detector.

The variation in ripple voltages on the -ve going slopes of 20 kHz sine waves
in other types of detectors is usually worse than in my detector.

Try it, you'll like it, and its better than the rest!


I also believe it is a lot easier to do this in the solid state world than
in the tube world.

There are high voltages available in vacuum tube circuits which in
combination with a large resistor would make a pretty good approximation
to a current source.

If you use my detector circuit, you could easily establish a -200v
voltage source,

and take a 4.7M resistor from the 270 pF, to give a closer approximation to a
CCS.

If you don't bother with the cathode follower, then there is no need for a
-200 volt supply, and the already existing +200 volt supply will do the
job.

But I think I need to bother with a CF.

If you are so sure the CF can be dispensed with in some way, and get
better overall performance than my design as I posted it, then proove it.
Design something, test it, and post the results like I have.



But speaking of the cathode follower, I may have found a justification for
the cathode follower in your circuit. It appears that when you apply bias
to the detector diode, at least a bias beyond that required to overcome
that caused by a non unity AC/DC load ratio, as in your circuit, that the
bias will cause the diode to present a nonlinear load to the source
driving it, over the audio cycle, beyond any non linearity due to non
ideal diode characteristics. I could be all wet about this, I need to
further corroborate my math, but if it is right, and diode bias does
result in a nonlinear load on the driving stage, then your cathode
follower to drive the diode does serve a useful purpose in your circuit.

Phew, anyone understand all that?
The diode nonlinearity during its conduction time in the 455 kHz cycles
does not much contribute to any thd in any ausio detected.

Try my circuit, rather than try to comprehend it in mathematical
or conceptual terms. Reality is what counts.
See the wave forms under various conditions.



If this is correct, the lesson I would take is to avoid using bias on the
diode, and instead make sure that the AC/DC load ratio is as close to
unity as practicable, either by using a cathode follower after the
detector, or by taping the audio output way down on the detector load
resistor, as the QUAD AM tuner does.

There is *no* AC coupled load in my dual CF detector as posted.

Taking the audio output from a tap on the discharge resistor
is the best one can do with a conventional detector circuit.

Also for lowest distortion when bias
is used on the diode, the bias should track the signal level, rather than
be fixed as your bias is.

It does track the signal level, and you'd see all about this if you built my
circuit,
and observed the wave forms and measured the voltages.

This is especially important in a design like
yours where the AGC circuit provides rather loose control of the carrier
level at the detector.

True, but my detector has such a wide dynamic range at low distortion that the main
concern is the
limitation of the amplification by the IF amp tube, and to keep this tube working
in a nice linear manner.


Patrick Turner.



Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

Robert Casey wrote in message ...
John Stewart wrote:

This one is in RDH4.


Saw it. In a radio where I already split off the AVC function off the
audio detector,
I'm trying a fixed positive bias of 300mV on the "bottom" of the volume
control.
Simulations indicate lower distortion on many various levels of carrier,
and it seems
borne out in an actual set. But the only cap I'm working against is the
coupling cap
between the volume control and the 12AV6 grid. Which is essentially a
fixed value.


Tried another approach: Just jumper out the 12AV6 grid coupling cap. The
detected audio has an average negative voltage on it (what most AA5s filter
to get AVC voltage) and I use that to bias the 12AV6 grid. Stronger signals
can be handled by turning the volume control down. Just like you did with a
standard AA5.

------
Having to post thru google as earthlink's news servers are not accessable up
here in the NYC area...

Patrick, before I get into a point by point rebuttal to your comments, let
me briefly summarize my current understanding of the difference between
the traditional RC diode peak detector load network with its exponential
decay, and the current source load you use which has a constant linear
sawtooth decay.

The following discussion does not consider problems due to the effects of
a non ideal diode, or those due to an AC/DC load ratio that isn't unity.

It appears that the current source load is the best from the point of view
of tracking the modulation wave form, however with a fixed current source
operation is only optimal at one fixed carrier level, if a wide dynamic
range is to be achieved for the detector, then means must be provided to
cause the current source to track the carrier level. The current source
load does get into trouble at high negative modulation levels, although
perhaps not as seriously as with the RC load and its exponential decay.
Putting a finger to the wind, it looks like that with a 455 kHz IF and 20
kHz audio bandwidth, trouble starts at approximately 95% negative
modulation. If the carrier level is less than optimum for the amount of
bias applied then the problems with negative going modulation will occur
sooner, at lower percentages of negative modulation. Also if the carrier
level increases above the optimum for the amount of bias applied
tangential clipping will begin to appear in the high slope parts of the
wave form. This means that a detector with fixed bias, as is used in the
Turner detector, has its dynamic range caught in a vice between these two
effects, and must be operated at a fixed and carefully controlled carrier
level if the full advantage is to be taken of the diode bias.

The above is all a pretty straight forward and obvious result of the
circuit's operation. A less obvious point, which is only my theory only
at this point, is that the application of bias to a diode detector, beyond
that needed to compensate the AC/DC load ratio, causes the diode detector
to present a load to the driving source which varies over the audio
cycle. I have worked out a proof of this, but I am not yet completely
happy that the proof is entirely correct. There are also the issues of
the ifs ands and buts to work out, or in other words what are the
conditions under which diode bias will cause this non linear loading
effect, and when is it OK. For example it appears that diode bias of the
amount necessary to compensate the AC/DC load ratio is OK.

The current source load used in the Turner detector is effectively a diode
bias scheme, and as a result one might expect nonlinear loading on the
driving stage over the audio cycle, this may be the reason why the Turner
detector requires a cathode follower to drive it.

On to the ordinary RC diode load, what are its plusses and minuses? The
advantage of the RC load is that provides a wide dynamic range for the
carrier level, unlike the Turner detector with its fixed bias and hence
fixed discharge slope. The RC loaded diode peak detector adjusts
automatically to all carrier levels, as the carrier level goes up, so does
the slope of the discharge, so that tangential clipping will not occur in
the high slope area at any carrier level. The RC network diode load does
appear to cause a problem at high negative modulation levels which is
frequency dependent, and is not mentioned in the text books which
generally seem to base their design rules on the high slope region only.
The effects of this distortion of the negative peaks, which is a different
effect than the negative peak clipping caused by a poor AC/DC load ratio,
is hard to judge relative to negative peak distortion of the Turner
detector, because while it has an earlier onset at higher frequencies, it
is modulating frequency dependent, while the negative peak distortion of
the Turner detector appear to be independent of the modulating frequency.

One has to balance these various effects and make a choice, the Turner
constant current detector and the traditional RC loaded peak detector both
have problems. It appears that the best compromise might be a "constant"
current detector where the "constant" current is made to track the average
current level. Given the various tradeoffs and issues I throw my hat in
the traditional RC load ring, but others may find the faults of the Turner
detector preferable.

A lot of theoretical analysis remains to be done on these issues, and
there are decisions to be made, and metal to be cut to create prototypes
for testing these ideas, enough to keep me busy for many months, so I am
going to make an exception to my usual practice, and not proof read what I
have written, and just post this straight away typos and all, I expect I
may regret it, but I am tired of typing for a few days.

On to the specific points Patrick has raised.

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , "craigm"
wrote:
If the discharge current is constant (does not change due to the
voltage on
the capacitor), then this approach leads to ripple that does not
vary with
the voltage on the capacitor.

With one exception, see below, the ripple should be more or less
constant
with this scheme. At one time I was enamored with this scheme, but gave
it up for reasons I no longer remember. A reason that comes to
mind now,
but was not the one I can't remember, is that while the ripple may be
relatively constant with modulation, the frequency at which tangential
clipping sets in will be dependent on the carrier amplitude, since the
slope of the discharge is independent of carrier level.

The "tangential clipping" also becomes more likely
if the audio modulation F is higher

Didn't I just I just get through saying that in my last paragraph above?

and the audio voltage is high.

I'm not sure what you might mean when you say "and the audio voltage is
high"? It is certainly true that tangential clipping gets worse as the
modulation level increases, if that is what you mean, but the "audio
voltage" is an imprecise way to talk about the modulation level.

One could have 30% modulation, and have a
high audio output, and yes, the tangent Dn could be high,
and the same as if you had 100% modulation but with a lower carrier voltage.

Yes, and this is a problem for the Turner detector, but the traditional
diode peak detector automatically adjusts to it without causing any
problem, the traditional detector circuit loves 30% modulation at any
carrier level and produces very low distortion at those levels.
Irrespective of carrier level the traditional detector always operates
like the first Watt from a SE amplifier if I may make an audiophile
analogy, while the Turner detector gets into trouble when the carrier
level is not as expected.

That's because of the discharge rate of the cap charged by the diode.
Its worse with a detector which does not have a biased diode
and constant current source.

Not completely true, with a current source the discharge current doesn't
track the carrier level as it does with the common diode envelope
detector, so as the carrier level increases, tangential clipping sets in
at lower and lower audio frequencies, and lower modulation levels. Use of
a current source restricts the normally wide dynamic range of a diode
envelope detector, in terms of the carrier level.

I think you are wrong here.

No, I am right, and you essentially prove it for me in your next paragraph.

The CCS allows a straight line cap discharge.
Using a cap discharge R which has a much wider voltage change across it
causes a discharge rate to vary, and you will see typical
curved discharge line, so that the time constant action is seen immediately.

Exactly, the Turner detector has a "straight line cap discharge", which
means that if the carrier level is higher than the design point, the slope
of the modulation in the high slope region will exceed the "straight line
cap discharge". In the traditional diode envelope detector, the cap
discharges at a faster rate, higher slope, when the carrier is stronger.

If only you spent 30 minutes in your workshop with a CRO, and built my
detector,
and observed carefully with your CRO, you'd see what I meant.

I'm just as stubborn as you, I have looked at the ripple on the
traditional diode peak detector output, and its fairly obvious what the
nature of the Turner detector ripple is, I have decided your design is not
for me, and have judged that it isn't worth even 30 minutes time in the
workshop, obviously others may feel differently. I am working on the
design of a simple radio which will include the ability evaluate the
traditional diode peak detector, a diode peak detector driving a cathode
follower to eliminate the AC/DC load problem, the reflex or "infinite
impedance" detector, and the Selsted & Smith detector. For those so
inclined I suppose the Turner detector would also be easy enough to
accommodate.

In my detector, the output voltage cannot ever go to to zero, or 0V.

I didn't notice anyone saying the output voltage of your detector actually
went to zero, and how is that relevant?

Well I though you implied that.
Its relevant that my detector output voltage never goes to 0V;
the audio voltage is the change in the output voltage of around
+/- 5 peak volts, so the current change in the 1M which discharges the 270 pF
doesn't vary much.

The DC output voltage is almost totally irrelevant in my opinion, it is
just an artifact of the design of the detector. The only way in which it
is relevant is to the extent that it reflects an important operating
characteristic of the basic detector, but in absolute terms it has no
significance.

Its at 55 volts even with no RF input,

And I will bet that there is no ripple with no RF input.

That is correct.

and then when a carrier of 10v p-p appears,

you get a rise in the direct voltage voltage across the 270pF of 5 volts
to +60v.
Then if there was 100% modulation, you get a varying voltage at the 270
pF between
+55v and +65v.
For low frequency audio, there is almost no change in the ripple voltage
at any
part of the audio
wave form.

I assume you put the "almost no change" part in because obviously the
ripple voltage must go to zero on negative modulation peaks of 100%.

Sure, but the ripple is still there at 99% modulation.

I will bet that the ripple in the Turner detector actually starts dropping
off somewhat before that, if it really handles a 20 kHz audio bandwidth.

Since 99% of what we hear comes from the first 95% of the modulation,
then the tiny tiny amount of flattening in the wave form
between 99% and 100% modulation is utterly irrelevant.

The fact that the amplitude of the ripple varies over the audio cycle
doesn't necessarily imply distortion, what is important is that the "area
under the curve" is linear and two ways this can be achieved are with a
fixed amplitude sawtooth ripple voltage, or with a sawtooth ripple voltage
which tracks the instantaneous level of the modulated carrier. I am not
making any claims about particular detectors with this statement, I am
just trying to indicate that a varying ripple voltage doesn't have to
imply distortion as you like to suggest.

But at 20 kHz modulation F, the ripple voltage is less on the
negative going
slopes of the sine wave.

That may be a significant observation, I suspect that this is exactly what
you would expect when you are approaching the point of tangential
clipping, assuming by "negative going slopes of the sine wave" you mean
the down slope headed towards maximum negative modulation, which with your
detector is the same as negative going voltage, unlike the normal diode
detector.

The variation in ripple voltages on the -ve going slopes of 20 kHz sine waves
in other types of detectors is usually worse than in my detector.

It's not obvious why that should be the case, I suspect that you are not
making an Apples vs. Apples comparison here. In other words you may have
compared your detector with other detectors not designed to handle the
same audio bandwidth. If both detectors are designed to go into slope
limiting at the same frequency, on the high slope part of the modulating
wave form, then the ripple on both detectors should be essentially zero at
that point of incipient slope limiting, or tangential clipping. Note that
we are talking about the high slope areas here, not the problems at high
negative modulation peaks.

I also believe it is a lot easier to do this in the solid state
world than
in the tube world.

There are high voltages available in vacuum tube circuits which in
combination with a large resistor would make a pretty good approximation
to a current source.

If you use my detector circuit, you could easily establish a -200v
voltage source,

and take a 4.7M resistor from the 270 pF, to give a closer
approximation to a
CCS.

If you don't bother with the cathode follower, then there is no need for a
-200 volt supply, and the already existing +200 volt supply will do the
job.

But I think I need to bother with a CF.

If you are so sure the CF can be dispensed with in some way, and get
better overall performance than my design as I posted it, then proove it.
Design something, test it, and post the results like I have.

Haven't I already come around to the view that the CF is important in the
Turner detector design, and provided a hint of the theoretical reasons
why? The traditional diode peak detector doesn't have this same
requirement however.

But speaking of the cathode follower, I may have found a justification for
the cathode follower in your circuit. It appears that when you apply bias
to the detector diode, at least a bias beyond that required to overcome
that caused by a non unity AC/DC load ratio, as in your circuit, that the
bias will cause the diode to present a nonlinear load to the source
driving it, over the audio cycle, beyond any non linearity due to non
ideal diode characteristics. I could be all wet about this, I need to
further corroborate my math, but if it is right, and diode bias does
result in a nonlinear load on the driving stage, then your cathode
follower to drive the diode does serve a useful purpose in your circuit.

Phew, anyone understand all that?
The diode nonlinearity during its conduction time in the 455 kHz cycles
does not much contribute to any thd in any ausio detected.

Try my circuit, rather than try to comprehend it in mathematical
or conceptual terms. Reality is what counts.
See the wave forms under various conditions.

Why, I believe that it is already pretty obvious what its theoretical
performance is, let those who find it attractive try it, I will cast my
net in other directions.

All in all this has been a most fruitful discussion, I have learned four
new things about diode detectors, basically things that I was given to
understand were true but hadn't taken the time to evaluate for myself, or
effects that I hadn't realized the full import of, some of which I have
not seen mentioned in the traditional text books.

If this is correct, the lesson I would take is to avoid using bias on the
diode, and instead make sure that the AC/DC load ratio is as close to
unity as practicable, either by using a cathode follower after the
detector, or by taping the audio output way down on the detector load
resistor, as the QUAD AM tuner does.

There is *no* AC coupled load in my dual CF detector as posted.

I haven't looked closely at your "dual CF detector as posted" but I have
no reason to believe it has an "AC coupled load", however the schematic
that you posted for the "Turner AM Tuner" did have an AC coupled load of
serious proportions, IIRC.

Taking the audio output from a tap on the discharge resistor
is the best one can do with a conventional detector circuit.

What about a cathode follower, isn't that just as good or even better,
especially a DC coupled cathode follower as in the "Byrns AM Tuner"
design?

Also for lowest distortion when bias
is used on the diode, the bias should track the signal level, rather than
be fixed as your bias is.

It does track the signal level, and you'd see all about this if you built my
circuit,
and observed the wave forms and measured the voltages.

Can you explain the mechanism which causes the bias to track the average
carrier level in your detector? The schematic you posted shows an
approximation of a simple current source using a resistor and a fixed
voltage source, hence a fixed bias current, what am I missing?

This is especially important in a design like
yours where the AGC circuit provides rather loose control of the carrier
level at the detector.

True, but my detector has such a wide dynamic range at low distortion
that the main
concern is the
limitation of the amplification by the IF amp tube, and to keep this
tube working
in a nice linear manner.

"Wide dynamic range"? Dynamic range is the major fault in your design, if
the carrier level increases from the design level, then the
frequency/modulation depth at which tangential clipping sets in decreases
in inverse proportion to the increased carrier level. If the carrier
level decreases, bad things start happening on the negative peaks at lower
modulation levels. Your design puts the signal in a vice of limited
dynamic range, it may be fine at the designed carrier level for which you
optimized it, but it will not be as tolerant of other carrier levels as
the traditional diode detector.

All of which brings up the issue of your AGC circuit with only one
controlled stage, and how constant it keeps the carrier level at the
detector, over a reasonably expect range of signal strengths. I will have
to dig out the curves of transconductance vs. bias voltage for some
typical AM radio tubes, and see if I can get an estimate of the
effectiveness of the AGC system in your tuner.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

John Byrns wrote:

Patrick, before I get into a point by point rebuttal to your comments, let
me briefly summarize my current understanding of the difference between
the traditional RC diode peak detector load network with its exponential
decay, and the current source load you use which has a constant linear
sawtooth decay.

The following discussion does not consider problems due to the effects of
a non ideal diode, or those due to an AC/DC load ratio that isn't unity.

It appears that the current source load is the best from the point of view
of tracking the modulation wave form, however with a fixed current source
operation is only optimal at one fixed carrier level, if a wide dynamic
range is to be achieved for the detector, then means must be provided to
cause the current source to track the carrier level.

No, this isn't so. The current source is used precisely because it does
provide the same good performance regardless carrier or modulation level.

The current source
load does get into trouble at high negative modulation levels, although
perhaps not as seriously as with the RC load and its exponential decay.

But my circuit tracks and detects the positive going side of the envelope.

If the negative side of the envelope is to be tracked, then some alternative circuit
with a CCS
would need to be developed.

Putting a finger to the wind, it looks like that with a 455 kHz IF and 20
kHz audio bandwidth, trouble starts at approximately 95% negative
modulation.

Nothing is perfect.

If the carrier level is less than optimum for the amount of
bias applied then the problems with negative going modulation will occur
sooner, at lower percentages of negative modulation. Also if the carrier
level increases above the optimum for the amount of bias applied
tangential clipping will begin to appear in the high slope parts of the
wave form. This means that a detector with fixed bias, as is used in the
Turner detector, has its dynamic range caught in a vice between these two
effects, and must be operated at a fixed and carefully controlled carrier
level if the full advantage is to be taken of the diode bias.

I feel my only answer is to say again that you should
proceed directly to your wonderful workbench, and turn on the soldering iron,
and built the detector, and examine it, and you should find that the -ve going
audio sine wave slopes don't suffer too badly from tangential distortions
caused by the slopes of decay in the time constant between the 1M and 270 pF,
bearing in mind that most audio signals have a much declining % of amplitudes
of signals as F rises above 1 kHz..




The above is all a pretty straight forward and obvious result of the
circuit's operation. A less obvious point, which is only my theory only
at this point, is that the application of bias to a diode detector, beyond
that needed to compensate the AC/DC load ratio, causes the diode detector
to present a load to the driving source which varies over the audio
cycle. I have worked out a proof of this, but I am not yet completely
happy that the proof is entirely correct. There are also the issues of
the ifs ands and buts to work out, or in other words what are the
conditions under which diode bias will cause this non linear loading
effect, and when is it OK. For example it appears that diode bias of the
amount necessary to compensate the AC/DC load ratio is OK.

The current source load used in the Turner detector is effectively a diode
bias scheme, and as a result one might expect nonlinear loading on the
driving stage over the audio cycle, this may be the reason why the Turner
detector requires a cathode follower to drive it.

I avoid the non linear efects caused by using a detector powered by a current
source.
I use a voltage source.
Try my circuit, and the proof, and the "ifs and buts" will all become self evident.



On to the ordinary RC diode load, what are its plusses and minuses?

The typical circuit from an IFT LC using a tube diode and CRC filter to feed a
tapped resistor
to produce the audio to a cap coupled volume control
has less linearity than my circuit when the carrier is tiny, or huge.
But at all levels its worse than mine, so when I rebuild a complete radio, like what
I am presently doing
with a client's Stromberg Carlson which had severe corrosion throughout,
I add the detector of mine to improove the sound.

The
advantage of the RC load is that provides a wide dynamic range for the
carrier level, unlike the Turner detector with its fixed bias and hence
fixed discharge slope.

????

The RC loaded diode peak detector adjusts
automatically to all carrier levels, as the carrier level goes up, so does
the slope of the discharge, so that tangential clipping will not occur in
the high slope area at any carrier level.

Tangential slope distortion is so bad using the old way of doing things.
There is nothing except its cheapness to recommend it....

The RC network diode load does
appear to cause a problem at high negative modulation levels which is
frequency dependent, and is not mentioned in the text books which
generally seem to base their design rules on the high slope region only.
The effects of this distortion of the negative peaks, which is a different
effect than the negative peak clipping caused by a poor AC/DC load ratio,
is hard to judge relative to negative peak distortion of the Turner
detector, because while it has an earlier onset at higher frequencies, it
is modulating frequency dependent, while the negative peak distortion of
the Turner detector appear to be independent of the modulating frequency.

One has to balance these various effects and make a choice, the Turner
constant current detector and the traditional RC loaded peak detector both
have problems.

Nothing is perfect, but I garantee my detector has less inherent operarational
imperfections compared with any other diode detector you like to refer me to.


It appears that the best compromise might be a "constant"
current detector where the "constant" current is made to track the average
current level. Given the various tradeoffs and issues I throw my hat in
the traditional RC load ring, but others may find the faults of the Turner
detector preferable.

A lot of theoretical analysis remains to be done on these issues, and
there are decisions to be made, and metal to be cut to create prototypes
for testing these ideas, enough to keep me busy for many months, so I am
going to make an exception to my usual practice, and not proof read what I
have written, and just post this straight away typos and all, I expect I
may regret it, but I am tired of typing for a few days.

On to the specific points Patrick has raised.

Gee, and I thought you were concluding....



In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , "craigm"
wrote:
If the discharge current is constant (does not change due to the
voltage on
the capacitor), then this approach leads to ripple that does not
vary with
the voltage on the capacitor.

With one exception, see below, the ripple should be more or less
constant
with this scheme. At one time I was enamored with this scheme, but gave
it up for reasons I no longer remember. A reason that comes to
mind now,
but was not the one I can't remember, is that while the ripple may be
relatively constant with modulation, the frequency at which tangential
clipping sets in will be dependent on the carrier amplitude, since the
slope of the discharge is independent of carrier level.

The "tangential clipping" also becomes more likely
if the audio modulation F is higher

Didn't I just I just get through saying that in my last paragraph above?

and the audio voltage is high.

I'm not sure what you might mean when you say "and the audio voltage is
high"? It is certainly true that tangential clipping gets worse as the
modulation level increases, if that is what you mean, but the "audio
voltage" is an imprecise way to talk about the modulation level.

One could have 30% modulation, and have a
high audio output, and yes, the tangent Dn could be high,
and the same as if you had 100% modulation but with a lower carrier voltage.

Yes, and this is a problem for the Turner detector, but the traditional
diode peak detector automatically adjusts to it without causing any
problem, the traditional detector circuit loves 30% modulation at any
carrier level and produces very low distortion at those levels.
Irrespective of carrier level the traditional detector always operates
like the first Watt from a SE amplifier if I may make an audiophile
analogy, while the Turner detector gets into trouble when the carrier
level is not as expected.

That's because of the discharge rate of the cap charged by the diode.
Its worse with a detector which does not have a biased diode
and constant current source.

Not completely true, with a current source the discharge current doesn't
track the carrier level as it does with the common diode envelope
detector, so as the carrier level increases, tangential clipping sets in
at lower and lower audio frequencies, and lower modulation levels. Use of
a current source restricts the normally wide dynamic range of a diode
envelope detector, in terms of the carrier level.

I think you are wrong here.

No, I am right, and you essentially prove it for me in your next paragraph.

The CCS allows a straight line cap discharge.
Using a cap discharge R which has a much wider voltage change across it
causes a discharge rate to vary, and you will see typical
curved discharge line, so that the time constant action is seen immediately.

Exactly, the Turner detector has a "straight line cap discharge", which
means that if the carrier level is higher than the design point, the slope
of the modulation in the high slope region will exceed the "straight line
cap discharge". In the traditional diode envelope detector, the cap
discharges at a faster rate, higher slope, when the carrier is stronger.

If only you spent 30 minutes in your workshop with a CRO, and built my
detector,
and observed carefully with your CRO, you'd see what I meant.

I'm just as stubborn as you, I have looked at the ripple on the
traditional diode peak detector output, and its fairly obvious what the
nature of the Turner detector ripple is, I have decided your design is not
for me, and have judged that it isn't worth even 30 minutes time in the
workshop, obviously others may feel differently.

I am *not* as stubborn as yourself as you claim.

I have looked at *both* the old traditional way of doing things which you
seem to support, and some new way, which I found worked better, after I invented it
for myself.

Stubborn is such an awkward word.

I think you are irrational if you say you mustn't try my circuit.

You have not yet given me sufficiently plausible reasons why you
shouldn't rush right out to your work bench and build my circuit.


I am working on the
design of a simple radio which will include the ability evaluate the
traditional diode peak detector, a diode peak detector driving a cathode
follower to eliminate the AC/DC load problem, the reflex or "infinite
impedance" detector, and the Selsted & Smith detector. For those so
inclined I suppose the Turner detector would also be easy enough to
accommodate.

I believe my detector works better than Selstead and Smith's because the
IFT output is *entirely* buffered from any effect that a diode
could have.



In my detector, the output voltage cannot ever go to to zero, or 0V.

I didn't notice anyone saying the output voltage of your detector actually
went to zero, and how is that relevant?

Well I though you implied that.
Its relevant that my detector output voltage never goes to 0V;
the audio voltage is the change in the output voltage of around
+/- 5 peak volts, so the current change in the 1M which discharges the 270 pF
doesn't vary much.

The DC output voltage is almost totally irrelevant in my opinion, it is
just an artifact of the design of the detector. The only way in which it
is relevant is to the extent that it reflects an important operating
characteristic of the basic detector, but in absolute terms it has no
significance.

The +55v DV output voltage is very relevant, because it determines the
diode bias current of 0.055 mA, enough, to have the germanium diode turn on,
and maintain an "on" voltage across the diode which stays nearly constant
during diode conduction with or without a carrier.



Its at 55 volts even with no RF input,

And I will bet that there is no ripple with no RF input.

That is correct.

and then when a carrier of 10v p-p appears,

you get a rise in the direct voltage voltage across the 270pF of 5 volts
to +60v.
Then if there was 100% modulation, you get a varying voltage at the 270
pF between
+55v and +65v.
For low frequency audio, there is almost no change in the ripple voltage
at any
part of the audio
wave form.

I assume you put the "almost no change" part in because obviously the
ripple voltage must go to zero on negative modulation peaks of 100%.

Sure, but the ripple is still there at 99% modulation.

I will bet that the ripple in the Turner detector actually starts dropping
off somewhat before that, if it really handles a 20 kHz audio bandwidth.

Don't place bets on what might be the performance.

Build the circuit and measure it against with another based on anything else you
like.



Since 99% of what we hear comes from the first 95% of the modulation,
then the tiny tiny amount of flattening in the wave form
between 99% and 100% modulation is utterly irrelevant.

The fact that the amplitude of the ripple varies over the audio cycle
doesn't necessarily imply distortion, what is important is that the "area
under the curve" is linear and two ways this can be achieved are with a
fixed amplitude sawtooth ripple voltage, or with a sawtooth ripple voltage
which tracks the instantaneous level of the modulated carrier. I am not
making any claims about particular detectors with this statement, I am
just trying to indicate that a varying ripple voltage doesn't have to
imply distortion as you like to suggest.

But at 20 kHz modulation F, the ripple voltage is less on the
negative going
slopes of the sine wave.

That may be a significant observation, I suspect that this is exactly what
you would expect when you are approaching the point of tangential
clipping, assuming by "negative going slopes of the sine wave" you mean
the down slope headed towards maximum negative modulation, which with your
detector is the same as negative going voltage, unlike the normal diode
detector.

The variation in ripple voltages on the -ve going slopes of 20 kHz sine waves
in other types of detectors is usually worse than in my detector.

It's not obvious why that should be the case, I suspect that you are not
making an Apples vs. Apples comparison here. In other words you may have
compared your detector with other detectors not designed to handle the
same audio bandwidth. If both detectors are designed to go into slope
limiting at the same frequency, on the high slope part of the modulating
wave form, then the ripple on both detectors should be essentially zero at
that point of incipient slope limiting, or tangential clipping. Note that
we are talking about the high slope areas here, not the problems at high
negative modulation peaks.

I also believe it is a lot easier to do this in the solid state
world than
in the tube world.

There are high voltages available in vacuum tube circuits which in
combination with a large resistor would make a pretty good approximation
to a current source.

If you use my detector circuit, you could easily establish a -200v
voltage source,

and take a 4.7M resistor from the 270 pF, to give a closer
approximation to a
CCS.

If you don't bother with the cathode follower, then there is no need for a
-200 volt supply, and the already existing +200 volt supply will do the
job.

But I think I need to bother with a CF.

If you are so sure the CF can be dispensed with in some way, and get
better overall performance than my design as I posted it, then proove it.
Design something, test it, and post the results like I have.

Haven't I already come around to the view that the CF is important in the
Turner detector design, and provided a hint of the theoretical reasons
why? The traditional diode peak detector doesn't have this same
requirement however.

But speaking of the cathode follower, I may have found a justification for
the cathode follower in your circuit. It appears that when you apply bias
to the detector diode, at least a bias beyond that required to overcome
that caused by a non unity AC/DC load ratio, as in your circuit, that the
bias will cause the diode to present a nonlinear load to the source
driving it, over the audio cycle, beyond any non linearity due to non
ideal diode characteristics. I could be all wet about this, I need to
further corroborate my math, but if it is right, and diode bias does
result in a nonlinear load on the driving stage, then your cathode
follower to drive the diode does serve a useful purpose in your circuit.

Phew, anyone understand all that?
The diode nonlinearity during its conduction time in the 455 kHz cycles
does not much contribute to any thd in any ausio detected.

Try my circuit, rather than try to comprehend it in mathematical
or conceptual terms. Reality is what counts.
See the wave forms under various conditions.

Why, I believe that it is already pretty obvious what its theoretical
performance is, let those who find it attractive try it, I will cast my
net in other directions.

All in all this has been a most fruitful discussion, I have learned four
new things about diode detectors, basically things that I was given to
understand were true but hadn't taken the time to evaluate for myself, or
effects that I hadn't realized the full import of, some of which I have
not seen mentioned in the traditional text books.

If this is correct, the lesson I would take is to avoid using bias on the
diode, and instead make sure that the AC/DC load ratio is as close to
unity as practicable, either by using a cathode follower after the
detector, or by taping the audio output way down on the detector load
resistor, as the QUAD AM tuner does.

There is *no* AC coupled load in my dual CF detector as posted.

I haven't looked closely at your "dual CF detector as posted" but I have
no reason to believe it has an "AC coupled load", however the schematic
that you posted for the "Turner AM Tuner" did have an AC coupled load of
serious proportions, IIRC.

I suggest you build the dual CF detector as shown recently in a post.
This one is the best.

The one shown in my Turner AM radio does have some AC loading, but
still delivers what is asked from it, ie, low distortion audio, and lower
than any simple old fashioned detector method....



Taking the audio output from a tap on the discharge resistor
is the best one can do with a conventional detector circuit.

What about a cathode follower, isn't that just as good or even better,
especially a DC coupled cathode follower as in the "Byrns AM Tuner"
design?

The dual CF Turner Detector uses a direct coupled CF output buffer.
But a direct coupled gain triode could also be used, with the miller effect
exploited to afford some
filtering of the 455 kHz ripple.




Also for lowest distortion when bias
is used on the diode, the bias should track the signal level, rather than
be fixed as your bias is.

It does track the signal level, and you'd see all about this if you built my
circuit,
and observed the wave forms and measured the voltages.

Can you explain the mechanism which causes the bias to track the average
carrier level in your detector? The schematic you posted shows an
approximation of a simple current source using a resistor and a fixed
voltage source, hence a fixed bias current, what am I missing?

There is 55v staic idle condition voltage applied across a 1M resistor.
Thus 0.055 mA of DC flows.
Under conditions where local stations are detected, the typical audio
level detected is 3 v rms, so you have +/- 5 peak volt variation across the
1M C discharge resistor, so the current change in the 1M is
around +/- 10%, so in effect the current is virtually constant.
Any further effort to make the current change lower, or more constant
would not have any substantial benefit.



This is especially important in a design like
yours where the AGC circuit provides rather loose control of the carrier
level at the detector.

True, but my detector has such a wide dynamic range at low distortion
that the main
concern is the
limitation of the amplification by the IF amp tube, and to keep this
tube working
in a nice linear manner.

"Wide dynamic range"? Dynamic range is the major fault in your design, if
the carrier level increases from the design level, then the
frequency/modulation depth at which tangential clipping sets in decreases
in inverse proportion to the increased carrier level.

Wrong, the dynamic range of my detector is second to none.
I suggest you build and test it to find out about it.

If the carrier
level decreases, bad things start happening on the negative peaks at lower
modulation levels.

No, they don't.

Your design puts the signal in a vice of limited
dynamic range, it may be fine at the designed carrier level for which you
optimized it, but it will not be as tolerant of other carrier levels as
the traditional diode detector.

All of which brings up the issue of your AGC circuit with only one
controlled stage, and how constant it keeps the carrier level at the
detector, over a reasonably expect range of signal strengths. I will have
to dig out the curves of transconductance vs. bias voltage for some
typical AM radio tubes, and see if I can get an estimate of the
effectiveness of the AGC system in your tuner.

There was never any intention to have superlative AGC control in this radio of mine.

Some of the convenience benefits of the traditional AGC philosofy were traded away
for a more linear performance of the mixer and IF tubes, which work with fixed bias.

I suggest you build and test the dual CF detector circuit of mine
before returning to the podium to try to convince us all why you mustn't try
something
new and inovative, and of lasting benefit to anyone's listening experience.

I have tried the old fashioned detector methods, and the Selstead & Smith, and I
have moved on.

Patrick Turner.




Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

In article , Patrick Turner
wrote:

John Byrns wrote:

Patrick, before I get into a point by point rebuttal to your comments, let
me briefly summarize my current understanding of the difference between
the traditional RC diode peak detector load network with its exponential
decay, and the current source load you use which has a constant linear
sawtooth decay.

The following discussion does not consider problems due to the effects of
a non ideal diode, or those due to an AC/DC load ratio that isn't unity.

It appears that the current source load is the best from the point of view
of tracking the modulation wave form, however with a fixed current source
operation is only optimal at one fixed carrier level, if a wide dynamic
range is to be achieved for the detector, then means must be provided to
cause the current source to track the carrier level.

No, this isn't so. The current source is used precisely because it does
provide the same good performance regardless carrier or modulation level.

Patrick, this is a joke right? This is the really simple stuff, this
isn't one of the more complex and subtle points that we can all get
wrong. I am going to try and keep things simple by just sticking to this
one point. It is easily demonstrated that your claim is wrong with
respect to increasing carrier levels. Your circuit discharges the peak
hold capacitor with what is a reasonable approximation of a current
source, which means that the discharge is at a fixed rate of volts/sec.
Assuming a given fixed modulating frequency, and depth of modulation, the
maximum slope of the modulation that must be tracked by the voltage on the
peak hold capacitor is proportional to the average carrier amplitude.
That implies that if the carrier level is increased by say 6 dB, then the
slope in volts/sec that must be tracked increases by a factor of two,
while the discharge slope of your constant current circuit remains fixed,
ultimately leading to tangential clipping at some carrier amplitude. On
the other hand, while the traditional RC circuit has its problems, it is
not affected by the average level of the carrier that is feed to it. If
the average carrier increases by 6 dB, the peak modulation slope that must
be tracked increases by a factor of two as before, but since the discharge
current is not fixed, and varies in proportion to the carrier level, the
discharge slope also increases by a factor of two, and there will be no
additional tangential clipping with the traditional circuit when the
average carrier level is increased.

The bottom line is that the traditional circuit can handle any carrier
level no matter how large, without an increase in tangential clipping,
while the tangential clipping in your circuit, with a fixed discharge
rate, increases as the carrier level increases above the design point,
hence a poor dynamic range.

Now specifically what is wrong with what I have just said, where is my
error? This is the simple part of the problem, it is not even the complex
stuff where we all go wrong from time to time, yet you don't seem to be
able to grasp it.

There was never any intention to have superlative AGC control in this radio
of mine.

I never assumed that was your intention, but since your detector is
sensitive to the average carrier level, it is relevant, but that's not the
real reason I wondered out loud about your AGC circuit. The real reason
was that I was simply curious about the performance of an AGC system with
a single controlled stage, given that most radios use a minimum of two
controlled stages.

Some of the convenience benefits of the traditional AGC philosofy were
traded away for a more linear performance of the mixer and IF tubes, which
work with fixed bias.

That's fine as far as it goes, but it creates a problem for a detector
with a fixed discharge current source, and the consequent sensitivity to
overload that implies.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

"John Byrns" wrote in message
...
In article , Patrick Turner
wrote:

John Byrns wrote:

Patrick, before I get into a point by point rebuttal to your comments,
let
me briefly summarize my current understanding of the difference
between
the traditional RC diode peak detector load network with its
exponential
decay, and the current source load you use which has a constant linear
sawtooth decay.

The following discussion does not consider problems due to the effects
of
a non ideal diode, or those due to an AC/DC load ratio that isn't
unity.

It appears that the current source load is the best from the point of
view
of tracking the modulation wave form, however with a fixed current
source
operation is only optimal at one fixed carrier level, if a wide
dynamic
range is to be achieved for the detector, then means must be provided
to
cause the current source to track the carrier level.

No, this isn't so. The current source is used precisely because it does
provide the same good performance regardless carrier or modulation
level.

Patrick, this is a joke right? This is the really simple stuff, this
isn't one of the more complex and subtle points that we can all get
wrong. I am going to try and keep things simple by just sticking to this
one point. It is easily demonstrated that your claim is wrong with
respect to increasing carrier levels. Your circuit discharges the peak
hold capacitor with what is a reasonable approximation of a current
source, which means that the discharge is at a fixed rate of volts/sec.
Assuming a given fixed modulating frequency, and depth of modulation, the
maximum slope of the modulation that must be tracked by the voltage on the
peak hold capacitor is proportional to the average carrier amplitude.
That implies that if the carrier level is increased by say 6 dB, then the
slope in volts/sec that must be tracked increases by a factor of two,
while the discharge slope of your constant current circuit remains fixed,
ultimately leading to tangential clipping at some carrier amplitude. On
the other hand, while the traditional RC circuit has its problems, it is
not affected by the average level of the carrier that is feed to it. If
the average carrier increases by 6 dB, the peak modulation slope that must
be tracked increases by a factor of two as before, but since the discharge
current is not fixed, and varies in proportion to the carrier level, the
discharge slope also increases by a factor of two, and there will be no
additional tangential clipping with the traditional circuit when the
average carrier level is increased.

The bottom line is that the traditional circuit can handle any carrier
level no matter how large, without an increase in tangential clipping,
while the tangential clipping in your circuit, with a fixed discharge
rate, increases as the carrier level increases above the design point,
hence a poor dynamic range.

Now specifically what is wrong with what I have just said, where is my
error? This is the simple part of the problem, it is not even the complex
stuff where we all go wrong from time to time, yet you don't seem to be
able to grasp it.


John,

You seem to be limiting your considerations to the 'tangential clipping' and
not to other distortions that will occur. With the simple RC circuit the
decay of the signal differs from positive to negative peaks in the
modulation. This imparts an assymetry to the recovered signal. You are
trading one type of distortion for another. Using a constant current to
drain the capacitor provides a more linear output and one where slew rate
limiting can be easily computed as a function of frequency and amplitude.
Using a resistor to drain the capacitor provides an output where slew rate
limiting is more a function of frequency and less of amplitude.

However, if you know the maximum amplitude and modulating frequency of the
signal you are trying to detect, then for either detector one can determine
the proper component values for the desired result. These are the tradeoffs
that go into every design.

I suggest that those who are interested and/or following this discussion
would be well served by doing some modeling of the two proposals and
consider the results and how they are affected by changes in the input
signal. For a simple approach you could consider an ideal diode and signal
source, a capacitor and either a current source or a resistor. Try various
amplitudes and modulation levels.

Both circuit approaches work within their limitations. The question is 'what
are the limitations?'.

Suggestion: Consider a triangle wave for the modulation source. The math is
a lot easier.

Hint: Linear is good.

Once you understand the limitations of each circuit topology, then you can
understand how it interacts with the rest of the radio, or what requirements
each places on the rest of the radio.

Have fun, I'll be watching,

craigm


There was never any intention to have superlative AGC control in this
radio
of mine.

I never assumed that was your intention, but since your detector is
sensitive to the average carrier level, it is relevant, but that's not the
real reason I wondered out loud about your AGC circuit. The real reason
was that I was simply curious about the performance of an AGC system with
a single controlled stage, given that most radios use a minimum of two
controlled stages.

Some of the convenience benefits of the traditional AGC philosofy were
traded away for a more linear performance of the mixer and IF tubes,
which
work with fixed bias.

That's fine as far as it goes, but it creates a problem for a detector
with a fixed discharge current source, and the consequent sensitivity to
overload that implies.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

In article , "craigm"
wrote:

"John Byrns" wrote in message
...

Patrick, this is a joke right? This is the really simple stuff, this
isn't one of the more complex and subtle points that we can all get
wrong. I am going to try and keep things simple by just sticking to this
one point. It is easily demonstrated that your claim is wrong with
respect to increasing carrier levels. Your circuit discharges the peak
hold capacitor with what is a reasonable approximation of a current
source, which means that the discharge is at a fixed rate of volts/sec.
Assuming a given fixed modulating frequency, and depth of modulation, the
maximum slope of the modulation that must be tracked by the voltage on the
peak hold capacitor is proportional to the average carrier amplitude.
That implies that if the carrier level is increased by say 6 dB, then the
slope in volts/sec that must be tracked increases by a factor of two,
while the discharge slope of your constant current circuit remains fixed,
ultimately leading to tangential clipping at some carrier amplitude. On
the other hand, while the traditional RC circuit has its problems, it is
not affected by the average level of the carrier that is feed to it. If
the average carrier increases by 6 dB, the peak modulation slope that must
be tracked increases by a factor of two as before, but since the discharge
current is not fixed, and varies in proportion to the carrier level, the
discharge slope also increases by a factor of two, and there will be no
additional tangential clipping with the traditional circuit when the
average carrier level is increased.

The bottom line is that the traditional circuit can handle any carrier
level no matter how large, without an increase in tangential clipping,
while the tangential clipping in your circuit, with a fixed discharge
rate, increases as the carrier level increases above the design point,
hence a poor dynamic range.

Now specifically what is wrong with what I have just said, where is my
error? This is the simple part of the problem, it is not even the complex
stuff where we all go wrong from time to time, yet you don't seem to be
able to grasp it.


John,

You seem to be limiting your considerations to the 'tangential clipping' and
not to other distortions that will occur.

On the contrary I alluded to the other problems in my post above where I
said "while the traditional RC circuit has its problems". The problem is
that Patrick has a very thick head, so I am trying to keep things simple
so he might get the point.

With the simple RC circuit the
decay of the signal differs from positive to negative peaks in the
modulation. This imparts an assymetry to the recovered signal. You are
trading one type of distortion for another.

Yes I think I discussed this in another recent post, although I didn't
call it asymmetry, I called it a problem with the negative peaks, at least
I hope I actually posted that bit, and that it didn't get edited out, I
will have to check back in the message archive. In the interest of full
disclosure, I will also own up to having talked in a previous message like
the negative peak problem didn't exist in the traditional circuit, this
seems to be the tack most text books take, concentrating on the point of
maximum slope. It may be justifiable, as just like the tangential
clipping in the high slope area, the negative peak problem is related to
both modulating frequency and modulation depth, creating a problem only
with heavy modulation at high frequencies. For the reader who may be
confused by this negative peak asymmetry problem, it should be noted that
this is not the same clipping phenomenon caused by a poor AC/DC load
ratio, and that the negative peak clipping caused by a poor AC/DC load
ratio is not frequency dependent. All in all I don't believe the
asymmetry you are talking about is a serious problem in practice, but that
is an individual matter of judgment.

Using a constant current to
drain the capacitor provides a more linear output and one where slew rate
limiting can be easily computed as a function of frequency and amplitude.

If you look closely at the operation ot the constant current "drain", you
will see that it too has a distortion problem at the negative peaks, and
it affects all modulating frequencies, unlike the traditional circuit
where the asymmetry tends to disappear at lower modulating frequencies.

Using a resistor to drain the capacitor provides an output where slew rate
limiting is more a function of frequency and less of amplitude.

I wouldn't say that.

However, if you know the maximum amplitude and modulating frequency of the
signal you are trying to detect, then for either detector one can determine
the proper component values for the desired result. These are the tradeoffs
that go into every design.

Quite true.

I suggest that those who are interested and/or following this discussion
would be well served by doing some modeling of the two proposals and
consider the results and how they are affected by changes in the input
signal. For a simple approach you could consider an ideal diode and signal
source, a capacitor and either a current source or a resistor. Try various
amplitudes and modulation levels.

Both circuit approaches work within their limitations. The question is 'what
are the limitations?'.

I don't think Patrick will approve this sort of activity.

Suggestion: Consider a triangle wave for the modulation source. The math is
a lot easier.

Hint: Linear is good.

Once you understand the limitations of each circuit topology, then you can
understand how it interacts with the rest of the radio, or what requirements
each places on the rest of the radio.

Have fun, I'll be watching,

This is all way too complex, I was trying to keep things simple for
Patrick, and get him to try on one small bite sized piece at a time. The
first piece is how the modulating frequency and depth of modulation affect
the tangential clipping in the high slope part of the wave form with the
two circuits.

Once that is grasped, then he can move on to the negative peak asymmetries
both circuits have, but that is considerably harder to understand.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

Patrick, before I get into a point by point rebuttal to your comments, let
me briefly summarize my current understanding of the difference between
the traditional RC diode peak detector load network with its exponential
decay, and the current source load you use which has a constant linear
sawtooth decay.

The following discussion does not consider problems due to the effects of
a non ideal diode, or those due to an AC/DC load ratio that isn't unity.

It appears that the current source load is the best from the point of view
of tracking the modulation wave form, however with a fixed current source
operation is only optimal at one fixed carrier level, if a wide dynamic
range is to be achieved for the detector, then means must be provided to
cause the current source to track the carrier level.

No, this isn't so. The current source is used precisely because it does
provide the same good performance regardless carrier or modulation level.

Patrick, this is a joke right? This is the really simple stuff, this
isn't one of the more complex and subtle points that we can all get
wrong.

I am not joking when I said that the best performance under all conditions is
possible with a CCS, or a large value R with a fairly large 55v across it
to approximate a CCS.

Using CCS is new age stuff, and one never saw any CCS anyplace in domestic
electronics because there was always a cheaper crummier way to get the job done.

I am going to try and keep things simple by just sticking to this
one point. It is easily demonstrated that your claim is wrong with
respect to increasing carrier levels. Your circuit discharges the peak
hold capacitor with what is a reasonable approximation of a current
source, which means that the discharge is at a fixed rate of volts/sec.

Thank goodness the rate of discharge is fairly constant, regardless of AM%
and for the lower Audio F.


Assuming a given fixed modulating frequency, and depth of modulation, the
maximum slope of the modulation that must be tracked by the voltage on the
peak hold capacitor is proportional to the average carrier amplitude.
That implies that if the carrier level is increased by say 6 dB, then the
slope in volts/sec that must be tracked increases by a factor of two,
while the discharge slope of your constant current circuit remains fixed,
ultimately leading to tangential clipping at some carrier amplitude.

The innitial part of the discharge slope from the 270 pF and 1M that I use is
quite steep,
but quite fast enough to allow a few volts of 10 kHz audio AM, without tangential
distortion, where the discharge slope cuts off part of the negative going sine
wave.

Try my circuit on a breadboard, and you will see that all's well!!!!!!

How many times must I suggest you try something new for a change!!

On
the other hand, while the traditional RC circuit has its problems, it is
not affected by the average level of the carrier that is feed to it. If
the average carrier increases by 6 dB, the peak modulation slope that must
be tracked increases by a factor of two as before, but since the discharge
current is not fixed, and varies in proportion to the carrier level, the
discharge slope also increases by a factor of two, and there will be no
additional tangential clipping with the traditional circuit when the
average carrier level is increased.

The bottom line is that the traditional circuit can handle any carrier
level no matter how large, without an increase in tangential clipping,
while the tangential clipping in your circuit, with a fixed discharge
rate, increases as the carrier level increases above the design point,
hence a poor dynamic range.

I found the typical traditional circuit suffered from tangential distortion just
like any other.

Whatever the the diode detector circuit is, it should be set up
to be able to produce a large enough voltage without tangential or other
distortions
and I believe my circuit produces the least compared to the traditional.

When you try my circuit instead of wasting an enormous amount of time on
discussions,
you will see the superiority of the circuit I have posted.



Now specifically what is wrong with what I have just said, where is my
error? This is the simple part of the problem, it is not even the complex
stuff where we all go wrong from time to time, yet you don't seem to be
able to grasp it.

I am able to grasp it, but I have limited time to discuss something so trivial,
and and you won't even give the idea of mine a try, so WTF do you know about
my idea if you have not tried it?

I have tried yours, and a pile of other variations.



There was never any intention to have superlative AGC control in this radio
of mine.

I never assumed that was your intention, but since your detector is
sensitive to the average carrier level, it is relevant, but that's not the
real reason I wondered out loud about your AGC circuit. The real reason
was that I was simply curious about the performance of an AGC system with
a single controlled stage, given that most radios use a minimum of two
controlled stages.

There is adequate AGC for local stations.
The mixer and IF amps are working in their linear regions,
and the main purpose of the AGC is to prevent IF overload.

The detector is quite linear regardless of whatever level of IF is present
up to about 10vrm of output, but I ask only 3vrm of audio output.
There is SFA distortions from my detector.





Some of the convenience benefits of the traditional AGC philosofy were
traded away for a more linear performance of the mixer and IF tubes, which
work with fixed bias.

That's fine as far as it goes, but it creates a problem for a detector
with a fixed discharge current source, and the consequent sensitivity to
overload that implies.

I dson't think so.

BTW, I tried shunting the secondary of an IFT while monitoring the signal at the
primary.
As I predicted, the primary signal reduced about 1 dB.

You have said the primary signal will increase when the seconday is shunted, ie
gain will increase, but I saw no sign of that.

Patrick Turner.





Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

John,

You seem to be limiting your considerations to the 'tangential clipping' and
not to other distortions that will occur.

On the contrary I alluded to the other problems in my post above where I
said "while the traditional RC circuit has its problems". The problem is
that Patrick has a very thick head, so I am trying to keep things simple
so he might get the point.

Unfortunately, you have not yet done a proper comparison measurement
of a traditional detector driven off an IFT secondary,
and compared the results to what I have proposed and posted using two CF tubes.

I don't have the time to spend on discussions that get nowhere with someone who
hasn't the time
to connect a handful of parts on a bench, and do some real work, instead of
endlessly talking around the subject, and making incorrect statements about skull
bone thickness.

Patrick Turner.

craigm wrote:
"John Byrns" wrote in message
...

In article , Patrick Turner
wrote:


John Byrns wrote:


Patrick, before I get into a point by point rebuttal to your comments,

let

me briefly summarize my current understanding of the difference

between

the traditional RC diode peak detector load network with its

exponential

decay, and the current source load you use which has a constant linear
sawtooth decay.

The following discussion does not consider problems due to the effects

of

a non ideal diode, or those due to an AC/DC load ratio that isn't

unity.

It appears that the current source load is the best from the point of

view

of tracking the modulation wave form, however with a fixed current

source

operation is only optimal at one fixed carrier level, if a wide

dynamic

range is to be achieved for the detector, then means must be provided

to

cause the current source to track the carrier level.

No, this isn't so. The current source is used precisely because it does
provide the same good performance regardless carrier or modulation

level.

Patrick, this is a joke right? This is the really simple stuff, this
isn't one of the more complex and subtle points that we can all get
wrong. I am going to try and keep things simple by just sticking to this
one point. It is easily demonstrated that your claim is wrong with
respect to increasing carrier levels. Your circuit discharges the peak
hold capacitor with what is a reasonable approximation of a current
source, which means that the discharge is at a fixed rate of volts/sec.
Assuming a given fixed modulating frequency, and depth of modulation, the
maximum slope of the modulation that must be tracked by the voltage on the
peak hold capacitor is proportional to the average carrier amplitude.
That implies that if the carrier level is increased by say 6 dB, then the
slope in volts/sec that must be tracked increases by a factor of two,
while the discharge slope of your constant current circuit remains fixed,
ultimately leading to tangential clipping at some carrier amplitude. On
the other hand, while the traditional RC circuit has its problems, it is
not affected by the average level of the carrier that is feed to it. If
the average carrier increases by 6 dB, the peak modulation slope that must
be tracked increases by a factor of two as before, but since the discharge
current is not fixed, and varies in proportion to the carrier level, the
discharge slope also increases by a factor of two, and there will be no
additional tangential clipping with the traditional circuit when the
average carrier level is increased.

The bottom line is that the traditional circuit can handle any carrier
level no matter how large, without an increase in tangential clipping,
while the tangential clipping in your circuit, with a fixed discharge
rate, increases as the carrier level increases above the design point,
hence a poor dynamic range.

Now specifically what is wrong with what I have just said, where is my
error? This is the simple part of the problem, it is not even the complex
stuff where we all go wrong from time to time, yet you don't seem to be
able to grasp it.



John,

You seem to be limiting your considerations to the 'tangential clipping' and
not to other distortions that will occur. With the simple RC circuit the
decay of the signal differs from positive to negative peaks in the
modulation. This imparts an assymetry to the recovered signal. You are
trading one type of distortion for another. Using a constant current to
drain the capacitor provides a more linear output and one where slew rate
limiting can be easily computed as a function of frequency and amplitude.
Using a resistor to drain the capacitor provides an output where slew rate
limiting is more a function of frequency and less of amplitude.

However, if you know the maximum amplitude and modulating frequency of the
signal you are trying to detect, then for either detector one can determine
the proper component values for the desired result. These are the tradeoffs
that go into every design.

I suggest that those who are interested and/or following this discussion
would be well served by doing some modeling of the two proposals and
consider the results and how they are affected by changes in the input
signal. For a simple approach you could consider an ideal diode and signal
source, a capacitor and either a current source or a resistor. Try various
amplitudes and modulation levels.

Both circuit approaches work within their limitations. The question is 'what
are the limitations?'.



The constant current idea is rather sensitive to signal level, but can
yield good results in a small sweet spot. When outside the sweet spot
you get mostly 2nd harmonic in the distortion products. Small amounts
of 2nd harmonic can sound pleasent, so someone messing with actual
hardware on the bench and listening with ears will think it works
"better" than it "is". ... But then again, people listen with ears
and not spectrum analyzers. :-)

Maybe some smart combination of the two schemes can be done...

Perhaps we can get back to detectors again eventually, but for now I have
had quite enough of them, so I will bite my tongue and refrain from
further comment on specific detectors for the moment.

In article , Patrick Turner
wrote:

There is adequate AGC for local stations.
The mixer and IF amps are working in their linear regions,
and the main purpose of the AGC is to prevent IF overload.

The term "local stations" as used above has also been used recently in
several other threads. I am curious what the readers of this forum would
consider to be a "local station"? What would be the range of received
field strengths that would define a "local station"? I am curious what
the range of field strengths might be that a receiver designed for "local
stations" would have to cope with?

Here in the US IIRC the FCC considers the "primary" service area of a
station to be defined by the 2 mV/M field strength contour. Also IIRC the
FCC requires a field strength of 5 uV/M for coverage of the primary City,
and recommends 20 uV/M for City coverage. One of these numbers, 2 mV/M, 5
uV/M, or 20 uV/M could probably be considered to be the lowest field
strength a "local station" would have. I will arbitrarily pick the 5 uV/M
City grade coverage number as the lower limit. At the upper end of the
scale here in the US the FCC requires 50 kW class A stations to have a
minimum unattenuated field strength of 2.56 Volts/Meter at 1 kM.
Depending on how close one wants to be to such a station and receive it
without overload, we might expect a received field strength of as much as
1,000 mV/M. Without checking the actual numbers in my notebook, which are
based on crude measurements and calculations using the local ground
conductivity and FCC propagation charts, I think the field strength of the
local 50 kW stations is about 300 uV/M in my workshop, so I will again
arbitrarily pick that number as my upper limit.

I would therefor consider a "local station" to be one having a received
field strength somewhere in the range of 5 mV/M to 300 mV/M, requiring a
receiver with total dynamic range of at least 36 dB. Does anyone have any
alternate thoughts on the meaning of "local station" as applied to
receivers?

BTW, I tried shunting the secondary of an IFT while monitoring the signal
at the primary.
As I predicted, the primary signal reduced about 1 dB.

You have said the primary signal will increase when the seconday is
shunted, ie gain will increase, but I saw no sign of that.

It will, something was screwed up in your experiment.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

In article , Patrick Turner
wrote:

John,

You seem to be limiting your considerations to the 'tangential clipping'
and not to other distortions that will occur.

On the contrary I alluded to the other problems in my post above where I
said "while the traditional RC circuit has its problems". The problem is
that Patrick has a very thick head, so I am trying to keep things simple
so he might get the point.

Unfortunately, you have not yet done a proper comparison measurement
of a traditional detector driven off an IFT secondary, and compared
the results to what I have proposed and posted using two CF tubes.

I'm not sure what "proper comparison measurements", or "CF tubes", have to
do with the theoretical aspects of tangential clipping? Unfortunately I
don't possess an AM generator that is adequate for making these
measurements, that is a generator that will do 100% negative modulation,
or anywhere near it, with low distortion. It isn't clear that you possess
such a generator either, and you seem to have engaged in a certain amount
of shucking and jiving with respect to the actual performance of your
detector.

Why don't you describe the AM generator you are using, how close it gets
to 100%, and what the distortion is at that point? I know you are using a
CRO to check for distortion, rather than a distortion analyzer, but still
if the generator has serious distortion at the extremes of modulation it
could mask some of the faults in a detector.

I don't have the time to spend on discussions that get nowhere with
someone who hasn't the time to connect a handful of parts on a bench, and
do some real work, instead of endlessly talking around the subject,

I am working on such a project, but first I must solve the AM generator
problem. I am beginning to get an idea or two as to how I can overcome
the AM generator problem. You have nearly pushed me to the point of
action in my workshop, as Danger Dave did a few years back with respect to
the workability of an amplifier design I had been contemplating for
several years. Once I built it, Danger Dave was quickly proved to be full
of it, and I expect a similar result again, once you have motivated me
sufficiently. But first lets hear more about the AM generator you are
using for your tests?

and making incorrect statements about skull bone thickness.

As they say, "if the shoe fits wear it"! I remember the "thick as a
brick" thread from earlier this year, where you clearly demonstrated the
thickness of your skull. For those don't remember, that adventure might
have been called the "octave" matter. It was related to the slope of the
attenuation curve of an RF tank circuit, IFT, or other similar circuit.
Workshops, simulations, and what not didn't enter into the matter because
you had conveniently measured, plotted, and posted the response curves for
an AM aerial circuit which made a perfect example for discussion. The
trouble started when you and your fellow countryman Phil Allison claimed
that the slope of the attenuation curve of a tank circuit was stepper
close to resonance and that the slope of the attenuation progressively
became less steep as you moved away resonance. I had been under the
impression that the slope of the attenuation curve actually increased as
you moved away from resonance, and asymptotically approached a slope
determined by the order of the filter. After a few back and forths it
became obvious to me that the problem was one of the different frequency
reference points we were using, you and Phil were using Zero frequency as
your reference, while I was using the center frequency of the filter as my
reference. At that point I said I completely agreed with your
conclusions, given your frame of reference, but you refused to accept my
concept of using the filter center frequency as an alternative view of the
situation, and told me it just wasn't valid. That is a perfect example of
a thick skull, since generally there are alternate definitions for things,
and as long as they are consistent with the facts, in that case your
measured and posted results, they are just as valid as what you may
consider to be a more conventional viewpoint, although in the case of the
"octave" matter I am not entirely sure yours was the conventional
viewpoint, but the bottom line was they both worked, and you denied that
my approach had validity.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

"John Byrns" wrote in message
...

The term "local stations" as used above has also been used recently in
several other threads. I am curious what the readers of this forum would
consider to be a "local station"?

I have no idea what they use to describe them now... but I remember that
AM's used to be basically classed as either Local (frequencies such as
1230.. 1 KW or less daytime power), Regional (frequencies such as 620... 5
KW-50KW daytime power) and Clear channel (frequencies such as 1160.. 50 KW
daytime and nighttime power). Locals were generally required to go off the
air at dusk unless they had specific nighttime authorization, which
generally required drastic power reduction and/or directional antenna system
(most with nighttime authorization would run at 250 watts, but I know of
some that run as little as 10 watts nighttime... as in why bother??)

Brenda Ann Dyer wrote:
"John Byrns" wrote in message
...

The term "local stations" as used above has also been used recently in
several other threads. I am curious what the readers of this forum would
consider to be a "local station"?


I have no idea what they use to describe them now...


The local "graveyard" stations are on 1230, 1240, 1340, 1400, 1450 and
1490. Nearly all are 1000 Watts day and night.

--
Mike Westfall, N6KUY, WDX6O
Los Alamos, NM (DM65uv)
Online logbooks at http://dxlogbook.gentoo.net
Remove the Reptile to Reply

In article , "Brenda Ann Dyer"
wrote:

"John Byrns" wrote in message
...

The term "local stations" as used above has also been used recently in
several other threads. I am curious what the readers of this forum would
consider to be a "local station"?

I have no idea what they use to describe them now... but I remember that
AM's used to be basically classed as either Local (frequencies such as
1230.. 1 KW or less daytime power), Regional (frequencies such as 620... 5
KW-50KW daytime power) and Clear channel (frequencies such as 1160.. 50 KW
daytime and nighttime power). Locals were generally required to go off the
air at dusk unless they had specific nighttime authorization, which
generally required drastic power reduction and/or directional antenna system
(most with nighttime authorization would run at 250 watts, but I know of
some that run as little as 10 watts nighttime... as in why bother??)

Well yes, a "Local" station is an obsolete FCC term for what are now
called class C stations. Class C stations operate with full power both day
and night by the way. That isn't exactly what I was asking about though,
considering the FCC doesn't have dominion over Australia, which is where
the post originated that used the term.

I was using the term in the context of a receiver designed to receive only
"local stations", and not intended for receiving distant stations, as for
example the old AA4 radios that didn't have an IF amplifier stage, where
the converter drove the detector directly.

What I was wondering was what range of field strengths those "local
station" receivers might have been intended to receive?


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

In article ,
(John Byrns) wrote:

In article , Patrick Turner
wrote:

John,

You seem to be limiting your considerations to the 'tangential clipping'
and not to other distortions that will occur.

On the contrary I alluded to the other problems in my post above where I
said "while the traditional RC circuit has its problems". The problem is
that Patrick has a very thick head, so I am trying to keep things simple
so he might get the point.

Oh, thank God the dueling banjos of rec.audio.tubes are back with more
of their lethally boring detector talk.

You don't often have a Troll this technical although it is getting more
goofy by the post.

--
Telamon
Ventura, California

I suggest that those who are interested and/or following this discussion
would be well served by doing some modeling of the two proposals and
consider the results and how they are affected by changes in the input
signal. For a simple approach you could consider an ideal diode and signal
source, a capacitor and either a current source or a resistor. Try various
amplitudes and modulation levels.

Both circuit approaches work within their limitations. The question is 'what
are the limitations?'.



The constant current idea is rather sensitive to signal level, but can
yield good results in a small sweet spot. When outside the sweet spot
you get mostly 2nd harmonic in the distortion products. Small amounts
of 2nd harmonic can sound pleasent, so someone messing with actual
hardware on the bench and listening with ears will think it works
"better" than it "is". ... But then again, people listen with ears
and not spectrum analyzers. :-)

Maybe some smart combination of the two schemes can be done...

The use of the CCS as a means of discharging the peak and hold cap
after the diode in a detector makes ther detector *less sensitive*
to signal level, and any signal level up to about 10vrms from the circuit I posted
using two CFs
from any % of modulation will be equally distortion free, compared to the
traditional
method which always will have more distortion for the same signal conditions.

Obviously, there are limits to how much signal can be accepted by any
detector circuit before there is a limit beyond which some form of thd
occurs which ruins the music.

But the circuit of mine will easily handle 10vrms of carrier modulated by 10
vrms of 1 kHz modulation signal.
If you had 90 vrms of carrier, and 10 vrms of modulation, obviously there would be
problems.

In practice, I try to set up an AM radio so an average strong local station
has enough IF signal at the secondary of the 2nd IFT
to make about 3vrms AF from a local station signal, so there maybe somewhere
about -5v of AGC bias generated.
The size of the signal to be detected should not be too high, or the IF tube
and perhaps mixer are working too hard, making signal voltages
which are going to cause a lot of thd on the recovered audio.

A triangular AF modulation wave should be able to be detected
so that the triangular wave has dead straights when viewed on a CRO.

But so often this plainly isn't the case, due to the non-linearities of the IF
amp, detector, and mixer,
in about that order.

It is important that the difference between the input RF LC tuned circuit
frequency
and the oscillator frequency always remain at the IF frequency
for all parts of the tunable band, ie, the two variable tuned circuits,
( or three if there is an RF amp ) all track each other within a a few kHz at
least.

Whilst totally re-engineering the Stromberg Carlson I am working on now,
I have been forced to wind the input and oscillator coils to achieve all this
since the original radio only tuned up to 1,300 kHz.

Patrick Turner.

John Byrns wrote:

In article , Patrick Turner
wrote:

John,

You seem to be limiting your considerations to the 'tangential clipping'
and not to other distortions that will occur.

On the contrary I alluded to the other problems in my post above where I
said "while the traditional RC circuit has its problems". The problem is
that Patrick has a very thick head, so I am trying to keep things simple
so he might get the point.

Unfortunately, you have not yet done a proper comparison measurement
of a traditional detector driven off an IFT secondary, and compared
the results to what I have proposed and posted using two CF tubes.

I'm not sure what "proper comparison measurements", or "CF tubes", have to
do with the theoretical aspects of tangential clipping?

It means to build samples of two comparable circuits, and
thoroughly measure and observe the workings of each,
and make your conclusions.

That isn't too hard now surely?

Unfortunately I
don't possess an AM generator that is adequate for making these
measurements, that is a generator that will do 100% negative modulation,
or anywhere near it, with low distortion. It isn't clear that you possess
such a generator either, and you seem to have engaged in a certain amount
of shucking and jiving with respect to the actual performance of your
detector.

I have two such generators, a old Topward, which uses chips to easily get
100% modulation of any signal between 2 Hz and 2 Mhz with a 400 Hz tone.

The other is a tube one I built, which is also capable of 100% modulation,
but the thd in the AF envelope is around 3% at the onset of 100% modulation.



Why don't you describe the AM generator you are using, how close it gets
to 100%, and what the distortion is at that point? I know you are using a
CRO to check for distortion, rather than a distortion analyzer, but still
if the generator has serious distortion at the extremes of modulation it
could mask some of the faults in a detector.

If you have a dual trace CRO, you should be able to view the
input gene RF signal on one trace, and the recovered audio from the detector
on the other trace.

Each trace can be overlaid, and at low % of modulation, the recovered
audio trace should perfectly outline the input RF signal modulation,
so if there is any thd in the input modulation shape, it doesn't matter.
The two traces should remain locked close when the amplitude of
the input RF and or modulation % is increased at least to an equivalent
level of signal to a strong local station when its received using 10 metres of
wire
as an antenna, and perhaps -8 volts of AGC is being generated.



I don't have the time to spend on discussions that get nowhere with
someone who hasn't the time to connect a handful of parts on a bench, and
do some real work, instead of endlessly talking around the subject,

I am working on such a project, but first I must solve the AM generator
problem. I am beginning to get an idea or two as to how I can overcome
the AM generator problem. You have nearly pushed me to the point of
action in my workshop, as Danger Dave did a few years back with respect to
the workability of an amplifier design I had been contemplating for
several years. Once I built it, Danger Dave was quickly proved to be full
of it, and I expect a similar result again, once you have motivated me
sufficiently. But first lets hear more about the AM generator you are
using for your tests?

Its a simple triode oscillator with a grid LC, with tap on the LC for the cathode
current.

This feeds a 6BX6 RF amp, which has an LC in its anode circuit.

The AF is fed into the 6BX6 grid circuit to alter the anode current at AF, and
modulate the output at the anode.
The anode LC has a secondary winding to reduce the output impedance.

One two gang tuning cap from an old radio is used.

It took about a fortnight to build, and a fortnight to de-bug, and to get the
sawtooth oscillator working, so when switched to 455 kHz,
the Fo could be wobulated 40 kHz each side to display the IF bandpass contour.

I used about 10 x 68v zener diodes operating at a DV lower than the zener voltage
to make a varicap diode to give a high enough C shift to cause the wanted F
deviation.
Some wobulators use a spinning tuning cap driven by a motor, but I wanted
no mechanical parts which could wear out.

and making incorrect statements about skull bone thickness.

The process taught me all about my own cerebral bone thickness.
I took aim as to what I wanted from thre test gear, and I didn't stop
until I achieved it.

Probably just as well I never had any involvement with discussion groups
back in 1993 to 2000, I was learning by doing, and I wasn't able to load all my
dumb
questions, about 20 per day, onto any news group.

I answered many questions myself.




As they say, "if the shoe fits wear it"! I remember the "thick as a
brick" thread from earlier this year, where you clearly demonstrated the
thickness of your skull. For those don't remember, that adventure might
have been called the "octave" matter. It was related to the slope of the
attenuation curve of an RF tank circuit, IFT, or other similar circuit.

Phil Allison and I were quite correct in our assessment about
attenuation rates in RF tank circuits, and I was the one to measure a typical
LC taken from an old radio and post the results at the binaries groups,
to prove and define what I was saying, leaving no room for any doubt, or BS.


Workshops, simulations, and what not didn't enter into the matter because
you had conveniently measured, plotted, and posted the response curves for
an AM aerial circuit which made a perfect example for discussion. The
trouble started when you and your fellow countryman Phil Allison claimed
that the slope of the attenuation curve of a tank circuit was stepper
close to resonance and that the slope of the attenuation progressively
became less steep as you moved away resonance.

Well indeed the rate of attenuation is steeper near Fo, and then becomes less.

There is only 6 dB /octave attenuation when you are 20 dB or more away from Fo.
Say you have a tuned LC with Fo = 1,200 kHz, then at 600 kHz, the rate of
attenuation
is 6 dB /octave, so that between 600 kZ and 300 kHz, there is only 6 dB of
attenuation.

But close to Fo, within a few kHz, and if the Q of the LC is say 50,
the rate of attenuation is far far greater than 6 dB / octave with regard to RF
frequencies.
The rate of audio F carried by a modulated carrier follows the RF attenuation
shape.


I had been under the
impression that the slope of the attenuation curve actually increased as
you moved away from resonance, and asymptotically approached a slope
determined by the order of the filter. After a few back and forths it
became obvious to me that the problem was one of the different frequency
reference points we were using, you and Phil were using Zero frequency as
your reference, while I was using the center frequency of the filter as my
reference. At that point I said I completely agreed with your
conclusions, given your frame of reference, but you refused to accept my
concept of using the filter center frequency as an alternative view of the
situation, and told me it just wasn't valid. That is a perfect example of
a thick skull, since generally there are alternate definitions for things,
and as long as they are consistent with the facts, in that case your
measured and posted results, they are just as valid as what you may
consider to be a more conventional viewpoint, although in the case of the
"octave" matter I am not entirely sure yours was the conventional
viewpoint, but the bottom line was they both worked, and you denied that
my approach had validity.

I have seen no reference of your interpretive methodology in any text books,
and the text book methods to which I adhere to explain it all nicely,
and I don't have any intention of going right through all that long and tortuous
discussion again.
I been there, done that, and I have moved on.

Meanwhile, I am probably wasting far too much time as it is re-building an
old Stromberg Carlson, giving it a real good Turnerization, with singing lessons
included.

But slowly the work looks promising....

It will never be as good as my other radio circuit which I posted last week
because
the IF tube is a vari mu 6G8, and I have AGC applied......


Patrick Turner.




Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

The constant current idea is rather sensitive to signal level, but can
yield good results in a small sweet spot. When outside the sweet spot
you get mostly 2nd harmonic in the distortion products. Small amounts
of 2nd harmonic can sound pleasent, so someone messing with actual
hardware on the bench and listening with ears will think it works
"better" than it "is". ... But then again, people listen with ears
and not spectrum analyzers. :-)

Maybe some smart combination of the two schemes can be done...


The use of the CCS as a means of discharging the peak and hold cap
after the diode in a detector makes ther detector *less sensitive*
to signal level, and any signal level up to about 10vrms from the circuit I posted
using two CFs
from any % of modulation will be equally distortion free, compared to the
traditional
method which always will have more distortion for the same signal conditions.



I may have a mistake in my simulation circuit. Diode and signal
source connected to produce the negative going waveform. 50pF to
ground. Then pass thru a 50K resistor, and then another 50pF
to ground, a pi circuit. On the output of the pi circuit I
connected a very high value multi-megohm resistor to B+ of 120V
(which would look like a CCS). Make the CCS current draw too low
and you get tangent distortion. Too low and you get severe waveform
top crushing. Obviously I got something wrong...

Oh, thank God the dueling banjos of rec.audio.tubes are back with more
of their lethally boring detector talk.

You don't often have a Troll this technical although it is getting more
goofy by the post.


Not a complete troll, as we are attempting to figure out a way to
get decent sound reception from AM radio stations. We believe that it
can be done, and we're hashing out the details. And adding a bit of
flame war spice to liven things up... ;-)

In article , Robert Casey
wrote:

The constant current idea is rather sensitive to signal level, but can
yield good results in a small sweet spot. When outside the sweet spot
you get mostly 2nd harmonic in the distortion products. Small amounts
of 2nd harmonic can sound pleasent, so someone messing with actual
hardware on the bench and listening with ears will think it works
"better" than it "is". ... But then again, people listen with ears
and not spectrum analyzers. :-)

Maybe some smart combination of the two schemes can be done...

Yes, the major problems with using a fixed current source to discharge the
peak hold capacitor can be avoided by making the discharge current track
the average carrier level. This should be duck soup with transistors, but
would be a little trickier to do with vacuum tubes.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John,

You seem to be limiting your considerations to the 'tangential
clipping'
and not to other distortions that will occur.

On the contrary I alluded to the other problems in my post above where I
said "while the traditional RC circuit has its problems". The
problem is
that Patrick has a very thick head, so I am trying to keep things simple
so he might get the point.

Unfortunately, you have not yet done a proper comparison measurement
of a traditional detector driven off an IFT secondary, and compared
the results to what I have proposed and posted using two CF tubes.

I'm not sure what "proper comparison measurements", or "CF tubes", have to
do with the theoretical aspects of tangential clipping?

It means to build samples of two comparable circuits, and
thoroughly measure and observe the workings of each,
and make your conclusions.

That isn't too hard now surely?

That is not responsive to my question as asked, and I am not talking about
answering a question with a question.

Unfortunately I
don't possess an AM generator that is adequate for making these
measurements, that is a generator that will do 100% negative modulation,
or anywhere near it, with low distortion. It isn't clear that you possess
such a generator either, and you seem to have engaged in a certain amount
of shucking and jiving with respect to the actual performance of your
detector.

I have two such generators, a old Topward, which uses chips to easily get
100% modulation of any signal between 2 Hz and 2 Mhz with a 400 Hz tone.

Is "2 Hz" a typo? If not the resulting wave form at 100% modulation with
a 400 Hz tone would be interesting to observe on the CRO. Do you know
what the distortion for this "old Topward" is at 100% modulation? It
would be nice to be able to get away form subjective CRO measurements and
use a distortion analyzer instead.

The other is a tube one I built, which is also capable of 100% modulation,
but the thd in the AF envelope is around 3% at the onset of 100% modulation.

Its a simple triode oscillator with a grid LC, with tap on the LC for
the cathode
current.

This feeds a 6BX6 RF amp, which has an LC in its anode circuit.

The AF is fed into the 6BX6 grid circuit to alter the anode current at AF, and
modulate the output at the anode.
The anode LC has a secondary winding to reduce the output impedance.

One two gang tuning cap from an old radio is used.

It took about a fortnight to build, and a fortnight to de-bug, and to get the
sawtooth oscillator working, so when switched to 455 kHz,
the Fo could be wobulated 40 kHz each side to display the IF bandpass contour.

I used about 10 x 68v zener diodes operating at a DV lower than the
zener voltage
to make a varicap diode to give a high enough C shift to cause the wanted F
deviation.
Some wobulators use a spinning tuning cap driven by a motor, but I wanted
no mechanical parts which could wear out.

Thanks for the description of your AM generator, or AM/FM generator as it
sounds like it actually is.


As they say, "if the shoe fits wear it"! I remember the "thick as a
brick" thread from earlier this year, where you clearly demonstrated the
thickness of your skull. For those don't remember, that adventure might
have been called the "octave" matter. It was related to the slope of the
attenuation curve of an RF tank circuit, IFT, or other similar circuit.

Phil Allison and I were quite correct in our assessment about
attenuation rates in RF tank circuits, and I was the one to measure a typical
LC taken from an old radio and post the results at the binaries groups,
to prove and define what I was saying, leaving no room for any doubt, or BS.

But that was my point, you were quite correct using your frame of
reference, on the other hand my assessment of the attenuation rates in RF
tank circuits was also correct, and also perfectly described your measured
data, even though it used a different frame of reference. Your position
was, and still seems to be that anyone who takes a different perspective
on a matter is of necessity wrong, even if the alternate perspective
explains the data as well, or even better than your perspective does, you
need to learn to think outside the box, and be more creative as it were.

Workshops, simulations, and what not didn't enter into the matter because
you had conveniently measured, plotted, and posted the response curves for
an AM aerial circuit which made a perfect example for discussion. The
trouble started when you and your fellow countryman Phil Allison claimed
that the slope of the attenuation curve of a tank circuit was stepper
close to resonance and that the slope of the attenuation progressively
became less steep as you moved away resonance.

Well indeed the rate of attenuation is steeper near Fo, and then becomes less.

There is only 6 dB /octave attenuation when you are 20 dB or more away
from Fo.
Say you have a tuned LC with Fo = 1,200 kHz, then at 600 kHz, the rate of
attenuation
is 6 dB /octave, so that between 600 kZ and 300 kHz, there is only 6 dB of
attenuation.

But close to Fo, within a few kHz, and if the Q of the LC is say 50,
the rate of attenuation is far far greater than 6 dB / octave with
regard to RF
frequencies.

Again that is one perspective, but it doesn't mean there aren't over
equally valid perspectives, all that matters is that a view correctly
describes the measured data, which mine does, and IIRC yours does also.

The rate of audio F carried by a modulated carrier follows the RF attenuation
shape.

This statement is a little ambiguous, could you clarify it? It almost
sounds like you are claiming the rate of attenuation of the audio
recovered from a single tuned tank circuit will be greater than 6dB/Octave
near the corner frequency? Well actually now that I think about it that
probably is what you are trying to say.

I had been under the
impression that the slope of the attenuation curve actually increased as
you moved away from resonance, and asymptotically approached a slope
determined by the order of the filter. After a few back and forths it
became obvious to me that the problem was one of the different frequency
reference points we were using, you and Phil were using Zero frequency as
your reference, while I was using the center frequency of the filter as my
reference. At that point I said I completely agreed with your
conclusions, given your frame of reference, but you refused to accept my
concept of using the filter center frequency as an alternative view of the
situation, and told me it just wasn't valid. That is a perfect example of
a thick skull, since generally there are alternate definitions for things,
and as long as they are consistent with the facts, in that case your
measured and posted results, they are just as valid as what you may
consider to be a more conventional viewpoint, although in the case of the
"octave" matter I am not entirely sure yours was the conventional
viewpoint, but the bottom line was they both worked, and you denied that
my approach had validity.

I have seen no reference of your interpretive methodology in any text books,
and the text book methods to which I adhere to explain it all nicely,
and I don't have any intention of going right through all that long and
tortuous
discussion again.

And I wouldn't ask you to, if you notice I am not disputing your method, I
am simply disputing your apparent claim that my method is invalid. I
would ask you for one favor though, could you cite some of the textbooks
that explain your method so nicely? I have suddenly realized that in the
previous furor that you and Phil raised about my method not being in
textbooks, no one asked if your method was actually presented in any
textbooks, and I have suddenly realized that I have never seen it in a
textbook, although that certainly doesn't mean it isn't. Does the RDH4
describe your method? As far as my method not being in text books, I
think that notion was thoroughly debunked in the earlier thread when both
myself and another poster provided references to textbooks that explained
my method. In fact my method is the basis of several filter design
techniques. I suspect the problem is that you are restricting your
reading to radio design textbooks, when you should perhaps be looking in
filter design textbooks for this information, which is where you can
easily find it. I don't remember seeing your method explained in any
radio design textbooks, and I have a whole shelf full, but that doesn't
mean it isn't in one somewhere because I certainly haven't read every page
of each one, and that is why I asked for a citation, so that I can become
more familiar with your approach. My reaction to your approach is the
same as yours to mine, namely assuming it is equally valid, it does not
appear to have any practical application, and tends to confuse the issue
of what is really going on.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

"Robert Casey" wrote in message
...


Not a complete troll, as we are attempting to figure out a way to
get decent sound reception from AM radio stations.

I know it's subjective, but I think most AM radios sound decent. Some sound
even better. Authoritative sources say the distortion of the typical diode
detector is under 2%, which makes sense. There's been some unsupported
ideas posted on rec.radio.shortwave that diode detectors normally have much
higher distortion levels.

Most of my radios are tube radios. Many transistor radios sound horrible on
AM. I tend to think poor AM audio from transistor radios is caused by
running the diode detector at low voltages into a low impedance load, but
I've never checked it out.


We believe that it
can be done, and we're hashing out the details.

Yeah, but how much distortion do we get from transmitter modulation?
Selective fading? A zero distortion AM detector may not be a hell of alot
better than a properly designed diode detector.

And adding a bit of
flame war spice to liven things up... ;-)


Godwin got it wrong. Threads can go on indefinitely without mentioning
Nazis. But if they run long they will get personal.

Frank Dresser

In article , Robert Casey
wrote:

I may have a mistake in my simulation circuit. Diode and signal
source connected to produce the negative going waveform. 50pF to
ground. Then pass thru a 50K resistor, and then another 50pF
to ground, a pi circuit. On the output of the pi circuit I
connected a very high value multi-megohm resistor to B+ of 120V
(which would look like a CCS). Make the CCS current draw too low
and you get tangent distortion. Too low and you get severe waveform
top crushing.

Did you actually mean "Too low" in that last sentence, or was it a typo
and you meant "Too high"? I observed "severe waveform top crushing" when
the current was too high relative to the carrier level, although the
meaning of "top" is a little ambiguous.

Obviously I got something wrong...

Obviously, maybe the problem is that you have everything connected "bass
ackwards" as they say, and don't have the output resting at 55 volts so it
can't go anywhere near zero. You need to reverse your diode, bias your RF
source up to around 56 volts DC, and connect the resistor to ground, then
you will have the "real deal".


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

Robert Casey wrote:




The constant current idea is rather sensitive to signal level, but can
yield good results in a small sweet spot. When outside the sweet spot
you get mostly 2nd harmonic in the distortion products. Small amounts
of 2nd harmonic can sound pleasent, so someone messing with actual
hardware on the bench and listening with ears will think it works
"better" than it "is". ... But then again, people listen with ears
and not spectrum analyzers. :-)

Maybe some smart combination of the two schemes can be done...


The use of the CCS as a means of discharging the peak and hold cap
after the diode in a detector makes ther detector *less sensitive*
to signal level, and any signal level up to about 10vrms from the circuit I posted
using two CFs
from any % of modulation will be equally distortion free, compared to the
traditional
method which always will have more distortion for the same signal conditions.


I may have a mistake in my simulation circuit. Diode and signal
source connected to produce the negative going waveform. 50pF to
ground.

None of us know exactly what was your test circuit.

Many mistakes about electronics concepts are made in news group
discussions because ppl stray from being disciplined about exactly what
schematic they are using, and how they are measuring something.
And its very easy to make mistakes measuring things happening at RF,
because of stray capacitances, and the effect of a probe
on the circuit being measured.

Unfortunately, we don't have the ability to post our schematics here, but they can be
posted
at ABSE and ABPR, but not all of us will see them, since the binaries often just don't
make it around to distant corners of the globe like Oz.


Then pass thru a 50K resistor, and then another 50pF
to ground, a pi circuit. On the output of the pi circuit I
connected a very high value multi-megohm resistor to B+ of 120V
(which would look like a CCS). Make the CCS current draw too low
and you get tangent distortion. Too low and you get severe waveform
top crushing. Obviously I got something wrong...

If you tried my circuit which I did post and which uses tw CF stages,
SS diodes and generates a +ve going audio wave form,
perhaps you'd have found that around 20vrms of audio was possible
from a 20 vrms carrier with 100% AM.

But the dreaded slew rate limiting distortion starts at about 2kHz, and
the realistic levels of AM with no visible
distortion at 10 kHz is 4 vrms, which is plenty.


I have just managed to get an completely revised Stromberg Carlson
going the way it should, with ECC35 mixer and 6G8 IF amp, and
it makes 20vrms + on strong stations with a 3M wire antenna and with the AGC set up
so that when the input signal needed to generate -20v AGC voltage is reduced by
30 dBv, the audio level drops only 5 dB.
The problem is that there is too much gain, even with 150k strapped across
each of the 4 IFT LC circuits.
I am getting 6.5 kHz audio bw, -3 dB, which can be increased to 9 kHz
with an audio step network.
The tone control will then do the rest.

Trimming the AGC circuit to get least audio distortion was critical,
and I still will have to fiddle around with IFT loadings to reduce gain,
and still have the IF tube loaded just right.
Also critical is to keep the mixer and IF amp anode supply voltage at around the +200v
level,
or else the screen current becomes too high, and the distortion is far worse.

My 6G8 IF amp is a vari mu tube, and contributes thd to the audio signal,
and its important to minimize the thd.

Although a BCB AM radio looks simple, try building one from
start to finish without somebody else's circuit values.
Its not easy to get text book performances.

I tried to wind my own oscillator coil using a former with a slug
and 0.3mm solid wire.
There was no way I could get oscillation to begin, so any coils
should be litz wire, preferably with 7 strands.

I finally got the tuning range from 550 Khz to 1650 kHz,
and managed to get the RF tuning to track the oscillator tuning.
If this isn't correct, the set will still work, but the thd in the recovered audio
will be bad.

Cans around diy coils must be non ferrous, but after placing a can shield
over a coil on a chassis the inductance will be reduced a bit, maybe 10%,
so any tests with coils and tuning of them must be done with cans in place.

It should be easy to visualise that the sloped "tangential distortion"
will occur at a lower F and / or if the voltage is high.

The 455 kHz ripple voltage at the peak and hold cap
will have a discharge slope which when fully drawn out looks like the text book
depictions of the voltage reduction over time for a given C being discharged by an R.
All RC considered have a time constant, where R x C = the time it takes for the voltage
across the C
to reach 0.37 x the voltage before discharge begins.

In the case of a 270 pF and 1M resistor, the TC = 1,000,000 ohms x
270/1,000,000,000,000 Farads
= 270 uS.

The curve for the decay of voltage can easily be drawn, and the decay means that
after every 270 uS, the voltage has reduced by a factor of 0.37, out to infinity....
This is assuming the voltage was some finite value above 0V, and the R was connected to
the charged
cap and charge drained to 0V.
So consider a 270 pF cap charged to +10v.
It would reach 3.7v after 270 uS, when the current discharge
would have reduced from an initial maximum of 0.01 mA to 0.0037 mA, and so on.


To get a nice looking discharge curve, the part of the drawn line will have maximum
slope
in the time just after initial discharge.
We can plot a sine wave on the TC drawing, and see if the slope of the sine wave exceeds
the
TC slope, and if it does, the sine wave will be distorted, and part of the sine wave
gets cut off.

But if we have our 270 pF charged at +50v,
and we have the same 1M taken to 0V then the TC curve will still
show that the voltage will drop from 50v to 18.5v in 270 uS,
which is 31.5 in 270 uS, or an average of 0.117 v / uS.

If we could keep the rate of current discharge constant,
we would find that it would take about 135 uS for the voltage to fall 50v,
which would be easily drawn on a graph.
The voltage of a sine wave and or its F could thus be greater
if drawn on our TC diagram, without seeing distortion.

The first 4 volts of current discharge using the 1M to 0V from a
270 pF cap charged to +50v would indeed be a substantially
straight line sloping at a rate of 0.37 v per which means the voltage drop
wil be approx 2.6 uS in the 4 uS.

You can draw such a sloped line from a +4v point on a verical graph axis,
and down to a point at 2.6 uSalong a horizontal axis.
Then you can superimpose a drawing of a sine wave with the slope of the voltage decay
being the same as the slope of the sine wave, which will have a total time for one cycle

being 2.6 uS.
Your drawing should show that the sine wave has a maximum peak voltage of
about 1 volt, and the F = 1,000,000 / 2.6 = 38 kHz.

If the sine wave F was halved to 19 kHz, the sine wave amplitude could be 2 peakV.
If F was 9.5 kHz, we could have 4 pk volts.

Hence if we had a diode detector with a 270 pF cap, and with 1M taken to
0V and the start voltage was +50v, we could have an undistorted AF sine wave
at 9.5 kHz which 4 pk volts if the discharge was a perfect CCS,
but it isn't, in my detector, but still I can get about 3.5pk V of 9.5 kHz of
undistorted audio,
which is 2.5 vrms, which is plenty because the HF content
rapidly rolls off after 1 kHz in music, and in any case,
I like to only get about 3 vrms average value of audio from my detector
at any F, which is enough to achieve a great SNR, and low thd in the IF amp, even if we
have a
variable U type of IF tube.

This figure allows me to apply some tone compensation to be able to boost the treble
by 6 dB, and still have around 1v to apply
to an audio power power amp using a 12AX7 and EL34 in triode with some loop NFB.

Unless you have a clear plan for any radio project, with all the gains and voltage
values
optimally worked out, the result will be sub optimal,
and sound bad compared to what is possible.

Patrick Turner.

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John,

You seem to be limiting your considerations to the 'tangential
clipping'
and not to other distortions that will occur.

On the contrary I alluded to the other problems in my post above where I
said "while the traditional RC circuit has its problems". The
problem is
that Patrick has a very thick head, so I am trying to keep things simple
so he might get the point.

Unfortunately, you have not yet done a proper comparison measurement
of a traditional detector driven off an IFT secondary, and compared
the results to what I have proposed and posted using two CF tubes.

I'm not sure what "proper comparison measurements", or "CF tubes", have to
do with the theoretical aspects of tangential clipping?

It means to build samples of two comparable circuits, and
thoroughly measure and observe the workings of each,
and make your conclusions.

That isn't too hard now surely?

That is not responsive to my question as asked, and I am not talking about
answering a question with a question.

Unfortunately I
don't possess an AM generator that is adequate for making these
measurements, that is a generator that will do 100% negative modulation,
or anywhere near it, with low distortion. It isn't clear that you possess
such a generator either, and you seem to have engaged in a certain amount
of shucking and jiving with respect to the actual performance of your
detector.

I have two such generators, a old Topward, which uses chips to easily get
100% modulation of any signal between 2 Hz and 2 Mhz with a 400 Hz tone.

Is "2 Hz" a typo?

No.

There are six decade ranges of F on my gene.
The 400 Hz AM can be switched on in any range.
Interesting waves result when trying to modulate a 2Hz carrier with
400 Hz audio tone, or from an external source of any F, or wave shape.

If not the resulting wave form at 100% modulation with
a 400 Hz tone would be interesting to observe on the CRO. Do you know
what the distortion for this "old Topward" is at 100% modulation? It
would be nice to be able to get away form subjective CRO measurements and
use a distortion analyzer instead.

The thd of the envelope would be below 1%, because the Topward has a synthesised
sine wave maker chip,
where the two halves of a sine wave are joined at the crests by some
incomprehensible workins inside a chip
with lots of pins.
The topward thd of ordinary 1 kHz tones is about 0.3 %; that's what I measured;
pretty awful really, but what the overall features are good.
I can also switch from modulating any carrier F with AM of any % to FM, with a nice
wide
maximum +/- 10% F deviation.

If you have a spare 6 weeks, you could build your own,
but its easier to just buy a decent function generator, there are plenty for sale
in second hand electronics dealers because everyone has gone to
digital, and PC techniques, etc.


The other is a tube one I built, which is also capable of 100% modulation,
but the thd in the AF envelope is around 3% at the onset of 100% modulation.

Its a simple triode oscillator with a grid LC, with tap on the LC for
the cathode
current.

This feeds a 6BX6 RF amp, which has an LC in its anode circuit.

The AF is fed into the 6BX6 grid circuit to alter the anode current at AF, and
modulate the output at the anode.
The anode LC has a secondary winding to reduce the output impedance.

One two gang tuning cap from an old radio is used.

It took about a fortnight to build, and a fortnight to de-bug, and to get the
sawtooth oscillator working, so when switched to 455 kHz,
the Fo could be wobulated 40 kHz each side to display the IF bandpass contour.

I used about 10 x 68v zener diodes operating at a DV lower than the
zener voltage
to make a varicap diode to give a high enough C shift to cause the wanted F
deviation.
Some wobulators use a spinning tuning cap driven by a motor, but I wanted
no mechanical parts which could wear out.

Thanks for the description of your AM generator, or AM/FM generator as it
sounds like it actually is.

There were quite a few commercial examples of what I built.
I couldn't find any when I wanted one, so I decided to make one.



As they say, "if the shoe fits wear it"! I remember the "thick as a
brick" thread from earlier this year, where you clearly demonstrated the
thickness of your skull. For those don't remember, that adventure might
have been called the "octave" matter. It was related to the slope of the
attenuation curve of an RF tank circuit, IFT, or other similar circuit.

Phil Allison and I were quite correct in our assessment about
attenuation rates in RF tank circuits, and I was the one to measure a typical
LC taken from an old radio and post the results at the binaries groups,
to prove and define what I was saying, leaving no room for any doubt, or BS.

But that was my point, you were quite correct using your frame of
reference, on the other hand my assessment of the attenuation rates in RF
tank circuits was also correct, and also perfectly described your measured
data, even though it used a different frame of reference. Your position
was, and still seems to be that anyone who takes a different perspective
on a matter is of necessity wrong, even if the alternate perspective
explains the data as well, or even better than your perspective does, you
need to learn to think outside the box, and be more creative as it were.

No, not wrong. You could be right. I simply didn't bother to disprove what you were
saying,
since could se no need. I already had a system which works for me,
and its found in the text books.

Where is your method also found in texbooks?



Workshops, simulations, and what not didn't enter into the matter because
you had conveniently measured, plotted, and posted the response curves for
an AM aerial circuit which made a perfect example for discussion. The
trouble started when you and your fellow countryman Phil Allison claimed
that the slope of the attenuation curve of a tank circuit was stepper
close to resonance and that the slope of the attenuation progressively
became less steep as you moved away resonance.

Well indeed the rate of attenuation is steeper near Fo, and then becomes less.

There is only 6 dB /octave attenuation when you are 20 dB or more away
from Fo.
Say you have a tuned LC with Fo = 1,200 kHz, then at 600 kHz, the rate of
attenuation
is 6 dB /octave, so that between 600 kZ and 300 kHz, there is only 6 dB of
attenuation.

But close to Fo, within a few kHz, and if the Q of the LC is say 50,
the rate of attenuation is far far greater than 6 dB / octave with
regard to RF
frequencies.

Again that is one perspective, but it doesn't mean there aren't over
equally valid perspectives, all that matters is that a view correctly
describes the measured data, which mine does, and IIRC yours does also.

The rate of audio F carried by a modulated carrier follows the RF attenuation
shape.

This statement is a little ambiguous, could you clarify it?

Pretty simple.
Say you have a tuned circuit at 1 MHz with a Q = 50.
The f1 and f2 -3 dB points will be at 10 kHz each side of 1 MHz.
If the carrier has been modulated with a 10 kHz tone, and you have a detector
attached to demodulate the audio, the recovered audio at 10 khz will be
3 dB lower than it would be if the modulation was 100 Hz.

The 1,000 kHz carrier modulated with a 10kHz tone has 3 components,
first is the 1,000 kHz carrier, and the other two are sidebands at
1010 kHz and 990 kHz, and these sidebands are attenuated by the repsonse shape of
any filter.



It almost
sounds like you are claiming the rate of attenuation of the audio
recovered from a single tuned tank circuit will be greater than 6dB/Octave
near the corner frequency? Well actually now that I think about it that
probably is what you are trying to say.

What I am trying to say is to regurgitate what is in the textbooks,
so read them if you want to know what I think.



I had been under the
impression that the slope of the attenuation curve actually increased as
you moved away from resonance, and asymptotically approached a slope
determined by the order of the filter. After a few back and forths it
became obvious to me that the problem was one of the different frequency
reference points we were using, you and Phil were using Zero frequency as
your reference, while I was using the center frequency of the filter as my
reference. At that point I said I completely agreed with your
conclusions, given your frame of reference, but you refused to accept my
concept of using the filter center frequency as an alternative view of the
situation, and told me it just wasn't valid. That is a perfect example of
a thick skull, since generally there are alternate definitions for things,
and as long as they are consistent with the facts, in that case your
measured and posted results, they are just as valid as what you may
consider to be a more conventional viewpoint, although in the case of the
"octave" matter I am not entirely sure yours was the conventional
viewpoint, but the bottom line was they both worked, and you denied that
my approach had validity.

I have seen no reference of your interpretive methodology in any text books,
and the text book methods to which I adhere to explain it all nicely,
and I don't have any intention of going right through all that long and
tortuous
discussion again.

And I wouldn't ask you to, if you notice I am not disputing your method, I
am simply disputing your apparent claim that my method is invalid. I
would ask you for one favor though, could you cite some of the textbooks
that explain your method so nicely?

I would have a dozen on my shelf which explain radio theory sufficiently well,
including RDH4, and 11 others.

I have suddenly realized that in the
previous furor that you and Phil raised about my method not being in
textbooks, no one asked if your method was actually presented in any
textbooks, and I have suddenly realized that I have never seen it in a
textbook, although that certainly doesn't mean it isn't. Does the RDH4
describe your method?

Perhaps weapons of mass destruction do exist in Iraq, they just ain't been found
yet.


As far as my method not being in text books, I
think that notion was thoroughly debunked in the earlier thread when both
myself and another poster provided references to textbooks that explained
my method. In fact my method is the basis of several filter design
techniques. I suspect the problem is that you are restricting your
reading to radio design textbooks, when you should perhaps be looking in
filter design textbooks for this information, which is where you can
easily find it. I don't remember seeing your method explained in any
radio design textbooks, and I have a whole shelf full, but that doesn't
mean it isn't in one somewhere because I certainly haven't read every page
of each one, and that is why I asked for a citation, so that I can become
more familiar with your approach. My reaction to your approach is the
same as yours to mine, namely assuming it is equally valid, it does not
appear to have any practical application, and tends to confuse the issue
of what is really going on.

Regards,

John Byrns

I leave you to your perceptions which I cannot fully fathom.

Can anyone else work out precisely what John is trying to say?

Patrick Turner.



Surf my web pages at, http://users.rcn.com/jbyrns/

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

As they say, "if the shoe fits wear it"! I remember the "thick as a
brick" thread from earlier this year, where you clearly demonstrated the
thickness of your skull. For those don't remember, that adventure might
have been called the "octave" matter. It was related to the slope
of the
attenuation curve of an RF tank circuit, IFT, or other similar circuit.

Phil Allison and I were quite correct in our assessment about
attenuation rates in RF tank circuits, and I was the one to measure
a typical
LC taken from an old radio and post the results at the binaries groups,
to prove and define what I was saying, leaving no room for any
doubt, or BS.

But that was my point, you were quite correct using your frame of
reference, on the other hand my assessment of the attenuation rates in RF
tank circuits was also correct, and also perfectly described your measured
data, even though it used a different frame of reference. Your position
was, and still seems to be that anyone who takes a different perspective
on a matter is of necessity wrong, even if the alternate perspective
explains the data as well, or even better than your perspective does, you
need to learn to think outside the box, and be more creative as it were.

No, not wrong. You could be right. I simply didn't bother to disprove
what you were
saying,
since could se no need. I already had a system which works for me,
and its found in the text books.

Where is your method also found in texbooks?

You should be able to locate it with Google, back around January of this
year, if you can't find it I will post a reference again if you do me the
courtesy of citing a textbook, with page number, where a description of
your system can be found. Is it in the RDH4, what page number?

I have seen no reference of your interpretive methodology in any
text books,
and the text book methods to which I adhere to explain it all nicely,
and I don't have any intention of going right through all that long and
tortuous
discussion again.

And I wouldn't ask you to, if you notice I am not disputing your method, I
am simply disputing your apparent claim that my method is invalid. I
would ask you for one favor though, could you cite some of the textbooks
that explain your method so nicely?

I would have a dozen on my shelf which explain radio theory sufficiently well,
including RDH4, and 11 others.

Ahh, the old weapons of mass destruction excuse, you have the RDH4 and 11
other text books, and yet you can't come up with a citation for your
method?


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

As they say, "if the shoe fits wear it"! I remember the "thick as a
brick" thread from earlier this year, where you clearly demonstrated the
thickness of your skull. For those don't remember, that adventure might
have been called the "octave" matter. It was related to the slope
of the
attenuation curve of an RF tank circuit, IFT, or other similar circuit.

Phil Allison and I were quite correct in our assessment about
attenuation rates in RF tank circuits, and I was the one to measure
a typical
LC taken from an old radio and post the results at the binaries groups,
to prove and define what I was saying, leaving no room for any
doubt, or BS.

But that was my point, you were quite correct using your frame of
reference, on the other hand my assessment of the attenuation rates in RF
tank circuits was also correct, and also perfectly described your measured
data, even though it used a different frame of reference. Your position
was, and still seems to be that anyone who takes a different perspective
on a matter is of necessity wrong, even if the alternate perspective
explains the data as well, or even better than your perspective does, you
need to learn to think outside the box, and be more creative as it were.

No, not wrong. You could be right. I simply didn't bother to disprove
what you were
saying,
since could se no need. I already had a system which works for me,
and its found in the text books.

Where is your method also found in texbooks?

You should be able to locate it with Google, back around January of this
year, if you can't find it I will post a reference again if you do me the
courtesy of citing a textbook, with page number, where a description of
your system can be found. Is it in the RDH4, what page number?

I have seen no reference of your interpretive methodology in any
text books,
and the text book methods to which I adhere to explain it all nicely,
and I don't have any intention of going right through all that long and
tortuous
discussion again.

And I wouldn't ask you to, if you notice I am not disputing your method, I
am simply disputing your apparent claim that my method is invalid. I
would ask you for one favor though, could you cite some of the textbooks
that explain your method so nicely?

I would have a dozen on my shelf which explain radio theory sufficiently well,
including RDH4, and 11 others.

Ahh, the old weapons of mass destruction excuse, you have the RDH4 and 11
other text books, and yet you can't come up with a citation for your
method?

I won't have the same books as you have, but apart from RDH4,
I have Terman's Radio Engineering,
about 6 different dated copies of ARRL,
The british Communications Handbook, 5th Ed,
Phillips Radio Paractice, Essentials of Radio by Sluurzberb&Osterfield,
Applied Electronics by the staff of the Dept of Massacgusets insitute of Technology,

Electrical and Electronic Engineering by John D Ryder,
and I am too lazy to copy out the titles of the other approximate 10 books I have
read
on old fashioned electronics which all describe filters the same way, but not the
way you do.

If you wanna uphold your methods, go write a book.

All the books backing up what I am saying are on the shelves for you to read.

Patrick Turner.



Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

I finally got the revised Stromberg Carlson radio tuner section running about as
well as it ever will.

It includes the dual cathode follower stage for which I posted a schematic at abse and
abpr last week.

The only changes I made was to use IN914 diodes, and 200pF and 1M for the audio RC
peak&hold
network, as well as use a 12AT7 for the CF parts.

The idle DV at the cathode of V1 is at +48 v without a carrier, but
with a test signal carrier a little larger than the strongest station here,
The DV at the C after D1 was +103v, indicating a carrier of
55 pk volts, and the audio at 400 Hz at 90% modulation was 36 vrms, or
51v pk.

I measured the thd with a 1 kHz low distortion af signal,
and got 3% at this level of signal, which is about 2 dB short of total
overload if the IF amp.
At 30% modulation, thd was about 0.2%.

Then when I reduced the RF input by 30 dB, the audio output and carrier level
all fell by 10 dB, due to AGC action, and remeasured and got the same thd figures.

The distortion is so low in the receiver including the detector that its thd
cannot be measured because it is dominated by the thd in the RF test gene,
which measures similarly when I measured it alone.

The RF gene can achieve about 96% mod but the thd becomes quite high at about 7%,
because the pentode RF tube used does not cut off at a linear rate, and
I really should have used a pair of PP tubes with a NFB loop to make the RF modulated
signal
have far less thd in the AF envelope shape.

So the conclusion is that the radio I have just got running
does not produce the buckets of thd like so many other radios I have tested,
and anyone is welcome to use the design I had in my posted schematic.

It was of some importance to get the AGC application correct.
Too much directly applied AGC will virtually cut off the IF vari mu pentode IF amp,
so that with 28vrms of audio from as big a carrier which will support that,
you may only have 0.5 mA of anode current, and when you plot
the load line, it just isn't quite right.
Better to make sure that with extreme levels of carrier, the anode current is
over 1.5 mA, and thd I expect is a little lower.

For lower levels of carrier, tube current will increase to a max of 5 mA
at no carrier at all.
There should be some method of applying at least about -1.5 v to IF and mixer tube grids
because
such tubes go a bit beserko when biased close to 0V.

I have the AGC generated by a 33pF from the V1 cathode taken to a IN914 with its cathode
grounded.
The negative voltage generated at the anode of the diode goes to a 0.05 uF via 2.2M
for the IF amp G1, and then 1M to another 0.05 to the G1 of the mixer, then 2.2M to
the -1.5v from a back bias R in the PSU.
Anyway, it works OK, and lots of other value changes didn't seem to work as well.

Audio bw was 7 kHz.
Both IFTs had their coils moved closer together, just short of causing a twin peaked
response.
After moving them and testing them in their cans, and connecting
the right value of fixed capacitance to each coil to allow the easy adjustment of the
adjust caps for 455 kHz, I rewaxed the coils in a vat.

I really don't like this type of IFT, with a ceramic base bolted to the top of a 60mm
dia al can
which is difficult to quickly inspect and modify,
and which were a complete pita to work on, but my patience paid off,
and the set was easy and reliable to align, for a symetrical attenuation slope
each side of 455 kHz.

Each of the four LC circuit are loaded with 150k, and loading with 100k
would be quite acceptable, and probably give an audio bw of 8 kHz instead of the 7 kHz.

The second CF cathode has a 39 k load to 0V, with about 2.2 mA of Ik at high levels of
carrier, and perhaps 1.2 mA at low levels.
The higher the carrier, the higher the CF idle current, which assures their linear
operation
with increasing carrier and audio output signal.

The audio is fed from CF2 cathode via 0.1 uF cap to 100k which is in series with a 100k
log volume pot.
A 390 pF across the fixed 100k slightly compensates the audio bw.

A tone baxandal passive 100k linear tone control pot will be fitted to give about
a further 5 dB treble boost or 6 dB cut at 3.4 kHz, necessary, because nearly all the
radio stations have compressed and quite bright sounding programme material,
and the mobile phone interviews and over compressed programmes are too bright.

Its quite a nice sounding tuner when connected to a 25 watt UL amp and one of my monitor

speakers in the workshop, which are well revised old Kefs, but nevertheless quite
revealing.
Bass respons eof the tuner is down to around 8 Hz due to the 0.1 uF and 200k AC load on
the output
of the second CF, which is direct coupled to the first CF, since there is no AC couple
load.

The 1M plus 200 pF could be altered to to 1M and 50 pF, which would increase the
ripple voltage about 4 times, but it will still stay substantially the same value with
audio
output signal of say 10vrms, when a low length wire antenna is used.
But the discharge rate of the 50 pF will be four times faster and the onset of
slew rate limiting will occur at a higher F and output voltage, the cost being
slightly lower detector efficiency.
The low pass ripple RC filter of 100k and 39 pF following the peak and hold 200pF and 1M

has a pole above 40 kHz so the attenuation at 45 kHz is around 22 dB,
and the level of RF finding its way into the audio amp and to the speaker would be
negligible,
and not cause any problems.
But folks could use a pair of 100k to feed the second CF, and have a pair of 39 pF caps
arranged in a feedback filter on the CF giving twice the attenuation of RF ripple.

It would be possible to construct one's own IFTs, with separate coils to be distance
adjusted, but
to get the required Q without ferrite cores which are hard for the diyer to make and
fit,
seven core fine litz wire is needed, and when one buys any of that today I don't know.
Its a pita to work with, since tinning each fine strand is difficult.
And when winding a coil, the traversing has to be done back and forth with a special
guided winder,
which hardly anyone would have, so they must make a lathe to generate the final wind up
which gives low self capacitance, and they must also have a method of makng the wound
wires adhere
to each other as the winding is done, ie, some sort of varnish.

Temperatures have fallen to around -5C and nights are a bit chilly even with a heater in
the shed;
it merely generates a cool breeze.

I now have the SET audio amp to build, which I should be able to do while I am asleep.
I think I have a spare 6L6, which would be easier to drive than a 6CM5 in triode.

Patrick Turner.

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

I have seen no reference of your interpretive methodology in any
text books, and the text book methods to which I adhere to explain
it all nicely, and I don't have any intention of going right through
all that long and tortuous discussion again.

And I wouldn't ask you to, if you notice I am not disputing your method,
I am simply disputing your apparent claim that my method is invalid. I
would ask you for one favor though, could you cite some of the textbooks
that explain your method so nicely?

I would have a dozen on my shelf which explain radio theory sufficiently
well, including RDH4, and 11 others.

Ahh, the old weapons of mass destruction excuse, you have the RDH4 and 11
other text books, and yet you can't come up with a citation for your
method?

I won't have the same books as you have, but apart from RDH4,
I have Terman's Radio Engineering,
about 6 different dated copies of ARRL,
The british Communications Handbook, 5th Ed,
Phillips Radio Paractice, Essentials of Radio by Sluurzberb&Osterfield,
Applied Electronics by the staff of the Dept of Massacgusets insitute of
Technology,

Electrical and Electronic Engineering by John D Ryder,

Of those I have at least the RDH4, and Terman's Radio Engineering, plus
possibly one or two more, how about some page numbers where I can find an
explanation of your definition of the rate of increase of the attenuation
of a tank circuit around resonance?

and I am too lazy to copy out the titles of the other approximate 10 books
I have read on old fashioned electronics which all describe filters the same
way, but not the way you do.

Well I guess that about says it all, you are simply one of those old
fashioned blokes who can't change his ways to adopt newer and better
methods.

If you wanna uphold your methods, go write a book.

I suppose I could, but why, I am not a textbook author, and my methods are
not original with me, I am not nearly that clever. As I have said before
I took them straight out of the modern filter design textbooks, the books
on this subject have already been written by others, many times over, the
field is way too crowded. You need to expand your reading list beyond
those smelly old radio textbooks, the old blokes didn't know everything,
you might learn something new from some more up to date reading, if you
can even call it that.

All the books backing up what I am saying are on the shelves for you to read.

Page numbers please, if you can't cite page numbers it is nothing more
than BS! Don't worry, I'm not going to hold my breath waiting, or
anything like that.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

Patrick Turner wrote:
I finally got the revised Stromberg Carlson radio tuner section running about as
well as it ever will.

It includes the dual cathode follower stage for which I posted a schematic at abse and
abpr last week.

The only changes I made was to use IN914 diodes, and 200pF and 1M for the audio RC
peak&hold
network, as well as use a 12AT7 for the CF parts.

The idle DV at the cathode of V1 is at +48 v without a carrier, but
with a test signal carrier a little larger than the strongest station here,
The DV at the C after D1 was +103v, indicating a carrier of
55 pk volts, and the audio at 400 Hz at 90% modulation was 36 vrms, or
51v pk.

I measured the thd with a 1 kHz low distortion af signal,
and got 3% at this level of signal, which is about 2 dB short of total
overload if the IF amp.
At 30% modulation, thd was about 0.2%.

Don't forget to do a measurement at 5KHz, as various distrotion
products can go up for higher audio frequencies at the demodulator.
Namely tangent distortion, which may not show at 1KHz but may show
at 5KHz.


Then when I reduced the RF input by 30 dB, the audio output and carrier level
all fell by 10 dB, due to AGC action, and remeasured and got the same thd figures.

The distortion is so low in the receiver including the detector that its thd
cannot be measured because it is dominated by the thd in the RF test gene,
which measures similarly when I measured it alone.

The RF gene can achieve about 96% mod but the thd becomes quite high at about 7%,
because the pentode RF tube used does not cut off at a linear rate, and
I really should have used a pair of PP tubes with a NFB loop to make the RF modulated
signal
have far less thd in the AF envelope shape.

So the conclusion is that the radio I have just got running
does not produce the buckets of thd like so many other radios I have tested,
and anyone is welcome to use the design I had in my posted schematic.

It was of some importance to get the AGC application correct.
Too much directly applied AGC will virtually cut off the IF vari mu pentode IF amp,
so that with 28vrms of audio from as big a carrier which will support that,
you may only have 0.5 mA of anode current, and when you plot
the load line, it just isn't quite right.

Modulation Rise it's called in RDH4. If you have an RF amp stage, use
a variable mu tube there (signals are still small) and change the IF
tube from say a 6BA6 to a 6AU6 or such sharp cutoff tube. Another
solution is to use only a fraction of the AVC voltage on the IF tube.
Voltage divider is the usual method.

Better to make sure that with extreme levels of carrier, the anode current is
over 1.5 mA, and thd I expect is a little lower.

For lower levels of carrier, tube current will increase to a max of 5 mA
at no carrier at all.
There should be some method of applying at least about -1.5 v to IF and mixer tube grids
because
such tubes go a bit beserko when biased close to 0V.

The local oscillator (usually a 6BE6) develops a fair amount of negative
bias that you could tap. Connect a 10 or so megohm resistor physically
near the oscillator (to lessen added stray capacitence) G1 and connect
the other end to the AVC line.

Another method is to add more resistance to the cathode resistor on the
IF tube and bypass it to ground. This will make the G1 look to have
more negative bias on it as seen by the cathode. And reduce the gain
some. Additional shielding and careful lead dress might help tame that
tube.

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

I have seen no reference of your interpretive methodology in any
text books, and the text book methods to which I adhere to explain
it all nicely, and I don't have any intention of going right through
all that long and tortuous discussion again.

And I wouldn't ask you to, if you notice I am not disputing your method,
I am simply disputing your apparent claim that my method is invalid. I
would ask you for one favor though, could you cite some of the textbooks
that explain your method so nicely?

I would have a dozen on my shelf which explain radio theory sufficiently
well, including RDH4, and 11 others.

Ahh, the old weapons of mass destruction excuse, you have the RDH4 and 11
other text books, and yet you can't come up with a citation for your
method?

I won't have the same books as you have, but apart from RDH4,
I have Terman's Radio Engineering,
about 6 different dated copies of ARRL,
The british Communications Handbook, 5th Ed,
Phillips Radio Paractice, Essentials of Radio by Sluurzberb&Osterfield,
Applied Electronics by the staff of the Dept of Massacgusets insitute of
Technology,

Electrical and Electronic Engineering by John D Ryder,

Of those I have at least the RDH4, and Terman's Radio Engineering, plus
possibly one or two more, how about some page numbers where I can find an
explanation of your definition of the rate of increase of the attenuation
of a tank circuit around resonance?

Most of the radio books I have do have explicit graphs and explanations of the
response
of RF and IFTs, with varying amounts of mutual coupling.
You don't need the page numbers from me, the info is in there.
The attenuation rates are shown on the graphs
And also there is a statement in RDH4 about sideband cutting, with a narrow bw RF /
IF response,
which underlines the importance of requiring a wide RF bw to get a wide audio
response.



and I am too lazy to copy out the titles of the other approximate 10 books
I have read on old fashioned electronics which all describe filters the same
way, but not the way you do.

Well I guess that about says it all, you are simply one of those old
fashioned blokes who can't change his ways to adopt newer and better
methods.

Well in a later post I did take the trouble to name my sources.

And I am not a lazy old bugger who never gets off his arse to find out by looking
into things.

I shouldn't have to do all this for you; your library should be embellished with
enough old books about radio to make all of what I am saying perfectly clear.



If you wanna uphold your methods, go write a book.

I suppose I could, but why, I am not a textbook author, and my methods are
not original with me, I am not nearly that clever. As I have said before
I took them straight out of the modern filter design textbooks, the books
on this subject have already been written by others, many times over, the
field is way too crowded. You need to expand your reading list beyond
those smelly old radio textbooks, the old blokes didn't know everything,
you might learn something new from some more up to date reading, if you
can even call it that.

The smelly old textbooks say it all so well that there isn't any need to
re-invent the wheel.
The application of the theory contained didn't lead to great BCB AM radios very
often because the
radio industry was infested with bean counters and charlatans.
This fact don't detract from the wisdom of the old books.

The technology of tube radios is ancient history which will never again be
the important techno mainstream thing it was, like steam engines.
But the old technology is still fascinating, and great sound can be had with the
right circuits.



All the books backing up what I am saying are on the shelves for you to read.

Page numbers please, if you can't cite page numbers it is nothing more
than BS! Don't worry, I'm not going to hold my breath waiting, or
anything like that.

You have to do your own study; I can't and I won't do it for you.

I was frustrated when I started to study the subject 10 years ago, and nobody could
answer
1,001 questions I had, so I simply went to second hand bookstores and snapped up
whatever was there,
which seems impossible now because the sharks and collectors seem to have emptied
the stores,
and then I read and copied reams at the university libraries.

But I also built and re-built a few radios.
Including AM/FM types.

Without having done anything in the workshop, I'd know SFA.

Patrick Turner.



Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

Robert Casey wrote:

Patrick Turner wrote:
I finally got the revised Stromberg Carlson radio tuner section running about as
well as it ever will.

It includes the dual cathode follower stage for which I posted a schematic at abse and
abpr last week.

The only changes I made was to use IN914 diodes, and 200pF and 1M for the audio RC
peak&hold
network, as well as use a 12AT7 for the CF parts.

The idle DV at the cathode of V1 is at +48 v without a carrier, but
with a test signal carrier a little larger than the strongest station here,
The DV at the C after D1 was +103v, indicating a carrier of
55 pk volts, and the audio at 400 Hz at 90% modulation was 36 vrms, or
51v pk.

I measured the thd with a 1 kHz low distortion af signal,
and got 3% at this level of signal, which is about 2 dB short of total
overload if the IF amp.
At 30% modulation, thd was about 0.2%.

Don't forget to do a measurement at 5KHz, as various distrotion
products can go up for higher audio frequencies at the demodulator.
Namely tangent distortion, which may not show at 1KHz but may show
at 5KHz.

At 35 vrms of undistorted output at 400 Hz, the detector seems fine, but at 2 khz the
tangential distortion starts.
But with 10vrms output of audio this distortion occurs first at a much higher F.
Reducing the value of the peak and hold cap from 200pF to 100 pF would improve the undistorted
bw.




Then when I reduced the RF input by 30 dB, the audio output and carrier level
all fell by 10 dB, due to AGC action, and remeasured and got the same thd figures.

The distortion is so low in the receiver including the detector that its thd
cannot be measured because it is dominated by the thd in the RF test gene,
which measures similarly when I measured it alone.

The RF gene can achieve about 96% mod but the thd becomes quite high at about 7%,
because the pentode RF tube used does not cut off at a linear rate, and
I really should have used a pair of PP tubes with a NFB loop to make the RF modulated
signal
have far less thd in the AF envelope shape.

So the conclusion is that the radio I have just got running
does not produce the buckets of thd like so many other radios I have tested,
and anyone is welcome to use the design I had in my posted schematic.

It was of some importance to get the AGC application correct.
Too much directly applied AGC will virtually cut off the IF vari mu pentode IF amp,
so that with 28vrms of audio from as big a carrier which will support that,
you may only have 0.5 mA of anode current, and when you plot
the load line, it just isn't quite right.

Modulation Rise it's called in RDH4. If you have an RF amp stage, use
a variable mu tube there (signals are still small) and change the IF
tube from say a 6BA6 to a 6AU6 or such sharp cutoff tube.

In this radio I wanted to have a vary U octal tube, and the 6G8 was chosen because I had it.
In another radio, I use a 6BX6, with fixed bias.

Another
solution is to use only a fraction of the AVC voltage on the IF tube.
Voltage divider is the usual method.

That is indeed what I am doing, but it takes awhile to get the divider values right.



Better to make sure that with extreme levels of carrier, the anode current is
over 1.5 mA, and thd I expect is a little lower.

For lower levels of carrier, tube current will increase to a max of 5 mA
at no carrier at all.
There should be some method of applying at least about -1.5 v to IF and mixer tube grids
because
such tubes go a bit beserko when biased close to 0V.

The local oscillator (usually a 6BE6) develops a fair amount of negative
bias that you could tap. Connect a 10 or so megohm resistor physically
near the oscillator (to lessen added stray capacitence) G1 and connect
the other end to the AVC line.

The ECC35 is a triode hexode, and the arrangements I have made for it are fine, and a bit
different to
6BE6, which I quite like.

Another method is to add more resistance to the cathode resistor on the
IF tube and bypass it to ground. This will make the G1 look to have
more negative bias on it as seen by the cathode. And reduce the gain
some. Additional shielding and careful lead dress might help tame that
tube.

In triode hexodes, the cathode should be grounded, lest the oscillator cathode current
be injected into the hexode cathode circuit; the grid of the triode oscillator
feeds into a grid of the hexode, and that's all that is wanted.

Patrick Turner.

Hi Patrick,

All your comments on text book references below are beside the point, what
I was asking for is a citation for a textbook that explains yours and
Phil's assertion that the rate of increase in the attenuation of a LC tank
circuit is greatest near the "nose", and decreases further from
resonance? Please note that I am not disputing yours and Phil's
viewpoint, I came to understand your perspective during the "Thick as a
Brick" thread back in January. You say this perspective is the one used
in old radio text books, but I have never seen it mentioned in an old
radio text book, hence I was hoping you could help me with a citation to a
text book that uses/explains your perspective? Just because I haven't
seen it doesn't mean it isn't there, since I never looked for it in the
past.

You seem to have forgotten that I changed the subject after you started
talking about the thickness of your skull, you acknowledged the change of
subject in a couple of posts, but now you have drifted back to an earlier
subject and assuming that is what I am asking for citations on.


Regards,

John Byrns


In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

I have seen no reference of your interpretive methodology in any
text books, and the text book methods to which I adhere to explain
it all nicely, and I don't have any intention of going right
through
all that long and tortuous discussion again.

And I wouldn't ask you to, if you notice I am not disputing
your method,
I am simply disputing your apparent claim that my method is
invalid. I
would ask you for one favor though, could you cite some of the
textbooks
that explain your method so nicely?

I would have a dozen on my shelf which explain radio theory
sufficiently
well, including RDH4, and 11 others.

Ahh, the old weapons of mass destruction excuse, you have the RDH4
and 11
other text books, and yet you can't come up with a citation for your
method?

I won't have the same books as you have, but apart from RDH4,
I have Terman's Radio Engineering,
about 6 different dated copies of ARRL,
The british Communications Handbook, 5th Ed,
Phillips Radio Paractice, Essentials of Radio by Sluurzberb&Osterfield,
Applied Electronics by the staff of the Dept of Massacgusets insitute of
Technology,

Electrical and Electronic Engineering by John D Ryder,

Of those I have at least the RDH4, and Terman's Radio Engineering, plus
possibly one or two more, how about some page numbers where I can find an
explanation of your definition of the rate of increase of the attenuation
of a tank circuit around resonance?

Most of the radio books I have do have explicit graphs and explanations of the
response
of RF and IFTs, with varying amounts of mutual coupling.
You don't need the page numbers from me, the info is in there.
The attenuation rates are shown on the graphs
And also there is a statement in RDH4 about sideband cutting, with a
narrow bw RF /
IF response,
which underlines the importance of requiring a wide RF bw to get a wide audio
response.



and I am too lazy to copy out the titles of the other approximate 10 books
I have read on old fashioned electronics which all describe filters
the same
way, but not the way you do.

Well I guess that about says it all, you are simply one of those old
fashioned blokes who can't change his ways to adopt newer and better
methods.

Well in a later post I did take the trouble to name my sources.

And I am not a lazy old bugger who never gets off his arse to find out
by looking
into things.

I shouldn't have to do all this for you; your library should be
embellished with
enough old books about radio to make all of what I am saying perfectly clear.



If you wanna uphold your methods, go write a book.

I suppose I could, but why, I am not a textbook author, and my methods are
not original with me, I am not nearly that clever. As I have said before
I took them straight out of the modern filter design textbooks, the books
on this subject have already been written by others, many times over, the
field is way too crowded. You need to expand your reading list beyond
those smelly old radio textbooks, the old blokes didn't know everything,
you might learn something new from some more up to date reading, if you
can even call it that.

The smelly old textbooks say it all so well that there isn't any need to
re-invent the wheel.
The application of the theory contained didn't lead to great BCB AM
radios very
often because the
radio industry was infested with bean counters and charlatans.
This fact don't detract from the wisdom of the old books.

The technology of tube radios is ancient history which will never again be
the important techno mainstream thing it was, like steam engines.
But the old technology is still fascinating, and great sound can be had
with the
right circuits.



All the books backing up what I am saying are on the shelves for you
to read.

Page numbers please, if you can't cite page numbers it is nothing more
than BS! Don't worry, I'm not going to hold my breath waiting, or
anything like that.

You have to do your own study; I can't and I won't do it for you.

I was frustrated when I started to study the subject 10 years ago, and
nobody could
answer
1,001 questions I had, so I simply went to second hand bookstores and
snapped up
whatever was there,
which seems impossible now because the sharks and collectors seem to
have emptied
the stores,
and then I read and copied reams at the university libraries.

But I also built and re-built a few radios.
Including AM/FM types.

Without having done anything in the workshop, I'd know SFA.

Patrick Turner.



Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/


Surf my web pages at, http://users.rcn.com/jbyrns/

John Byrns wrote:

Hi Patrick,

All your comments on text book references below are beside the point, what
I was asking for is a citation for a textbook that explains yours and
Phil's assertion that the rate of increase in the attenuation of a LC tank
circuit is greatest near the "nose", and decreases further from
resonance? Please note that I am not disputing yours and Phil's
viewpoint, I came to understand your perspective during the "Thick as a
Brick" thread back in January. You say this perspective is the one used
in old radio text books, but I have never seen it mentioned in an old
radio text book, hence I was hoping you could help me with a citation to a
text book that uses/explains your perspective? Just because I haven't
seen it doesn't mean it isn't there, since I never looked for it in the
past.

You seem to have forgotten that I changed the subject after you started
talking about the thickness of your skull, you acknowledged the change of
subject in a couple of posts, but now you have drifted back to an earlier
subject and assuming that is what I am asking for citations on.

I don't have the time to debate this any longer.
I don't want to repeat what I have already said.

I suggest yet again you satisfy your curiosity to inform yourself of the wonderments

we see with LC tuned circuits by reading whatever books exist on the subjects,
and I am sure there is a pile of material on the web.

My methods and perceptions have led to successfully building or modifying AM radios
to
a far better level of performance than the status quo, and I have thus prooved at
least to myself
the effectiveness of my education, which I promoted to be able to use it, and not
merely to be a
"knowledgeable do-little".

Patrick Turner.



Regards,

John Byrns

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

John Byrns wrote:

In article , Patrick Turner
wrote:

I have seen no reference of your interpretive methodology in any
text books, and the text book methods to which I adhere to explain
it all nicely, and I don't have any intention of going right
through
all that long and tortuous discussion again.

And I wouldn't ask you to, if you notice I am not disputing
your method,
I am simply disputing your apparent claim that my method is
invalid. I
would ask you for one favor though, could you cite some of the
textbooks
that explain your method so nicely?

I would have a dozen on my shelf which explain radio theory
sufficiently
well, including RDH4, and 11 others.

Ahh, the old weapons of mass destruction excuse, you have the RDH4
and 11
other text books, and yet you can't come up with a citation for your
method?

I won't have the same books as you have, but apart from RDH4,
I have Terman's Radio Engineering,
about 6 different dated copies of ARRL,
The british Communications Handbook, 5th Ed,
Phillips Radio Paractice, Essentials of Radio by Sluurzberb&Osterfield,
Applied Electronics by the staff of the Dept of Massacgusets insitute of
Technology,

Electrical and Electronic Engineering by John D Ryder,

Of those I have at least the RDH4, and Terman's Radio Engineering, plus
possibly one or two more, how about some page numbers where I can find an
explanation of your definition of the rate of increase of the attenuation
of a tank circuit around resonance?

Most of the radio books I have do have explicit graphs and explanations of the
response
of RF and IFTs, with varying amounts of mutual coupling.
You don't need the page numbers from me, the info is in there.
The attenuation rates are shown on the graphs
And also there is a statement in RDH4 about sideband cutting, with a
narrow bw RF /
IF response,
which underlines the importance of requiring a wide RF bw to get a wide audio
response.



and I am too lazy to copy out the titles of the other approximate 10 books
I have read on old fashioned electronics which all describe filters
the same
way, but not the way you do.

Well I guess that about says it all, you are simply one of those old
fashioned blokes who can't change his ways to adopt newer and better
methods.

Well in a later post I did take the trouble to name my sources.

And I am not a lazy old bugger who never gets off his arse to find out
by looking
into things.

I shouldn't have to do all this for you; your library should be
embellished with
enough old books about radio to make all of what I am saying perfectly clear.



If you wanna uphold your methods, go write a book.

I suppose I could, but why, I am not a textbook author, and my methods are
not original with me, I am not nearly that clever. As I have said before
I took them straight out of the modern filter design textbooks, the books
on this subject have already been written by others, many times over, the
field is way too crowded. You need to expand your reading list beyond
those smelly old radio textbooks, the old blokes didn't know everything,
you might learn something new from some more up to date reading, if you
can even call it that.

The smelly old textbooks say it all so well that there isn't any need to
re-invent the wheel.
The application of the theory contained didn't lead to great BCB AM
radios very
often because the
radio industry was infested with bean counters and charlatans.
This fact don't detract from the wisdom of the old books.

The technology of tube radios is ancient history which will never again be
the important techno mainstream thing it was, like steam engines.
But the old technology is still fascinating, and great sound can be had
with the
right circuits.



All the books backing up what I am saying are on the shelves for you
to read.

Page numbers please, if you can't cite page numbers it is nothing more
than BS! Don't worry, I'm not going to hold my breath waiting, or
anything like that.

You have to do your own study; I can't and I won't do it for you.

I was frustrated when I started to study the subject 10 years ago, and
nobody could
answer
1,001 questions I had, so I simply went to second hand bookstores and
snapped up
whatever was there,
which seems impossible now because the sharks and collectors seem to
have emptied
the stores,
and then I read and copied reams at the university libraries.

But I also built and re-built a few radios.
Including AM/FM types.

Without having done anything in the workshop, I'd know SFA.

Patrick Turner.



Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

Surf my web pages at, http://users.rcn.com/jbyrns/

In article , Patrick Turner
wrote:

I don't have the time to debate this any longer.

Hi Patrick,

You seem to be the one that wants a debate where one is not necessary, or
asked for. There is nothing to debate here, I understand yours and Phil's
position in a qualitative sense, but I find yours and Phil's approach
somewhat counter intuitive, just as you find my approach counter
intuitive. You say that your approach to LC tuned circuits is described
in many text books, I have not found any, although admittedly I have not
been searching specifically for information on your approach to LC tuned
circuits. All I have asked you for is a simple citation to a textbook
where I can find out more about your approach, especially the quantitative
aspects of your approach, and how to make use of it in design problems.
While it is easy to understand your idea from a qualitative perspective, I
don't understand how to actually apply it. I would prefer to learn that
from a textbook rather than taking your time to have you explain it
further.

I don't want to repeat what I have already said.

I'm not asking you to repeat anything, I am only asking you to provide a
citation for one of the textbooks that gives a good description of your
approach to LC tuned circuit design, you have not done that before. Your
continued unwillingness to provide a reference to one of the many
textbooks that you have said describe your perspective on LC tuned circuit
design, while continuing to make the claim, has made it pretty obvious
that your claims relative to textbook explanations is nothing but BS.
That is not to say that your observations are BS, I unraveled that back in
January, it is just the claim that your method may be found in textbooks
that smacks of BS.

I suggest yet again you satisfy your curiosity to inform yourself of the
wonderments we see with LC tuned circuits by reading whatever books exist
on the subjects, and I am sure there is a pile of material on the web.

Yes, I have several books on my bookshelf that explain the "wonderments we
see with LC tuned circuits", that is not a problem, but they explain
methods that are quite different from yours, I am simply looking for a
reference that will more fully explain your methods than what I have been
able to figure out on my own, especially from a quantitative perspective.

Sadly the answer to my question has become crystal clear.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

John Byrns wrote:

In article , Patrick Turner
wrote:

I don't have the time to debate this any longer.

Hi Patrick,

You seem to be the one that wants a debate where one is not necessary, or
asked for. There is nothing to debate here, I understand yours and Phil's
position in a qualitative sense, but I find yours and Phil's approach
somewhat counter intuitive, just as you find my approach counter
intuitive. You say that your approach to LC tuned circuits is described
in many text books, I have not found any, although admittedly I have not
been searching specifically for information on your approach to LC tuned
circuits. All I have asked you for is a simple citation to a textbook
where I can find out more about your approach, especially the quantitative
aspects of your approach, and how to make use of it in design problems.
While it is easy to understand your idea from a qualitative perspective, I
don't understand how to actually apply it. I would prefer to learn that
from a textbook rather than taking your time to have you explain it
further.

I don't want to repeat what I have already said.

I'm not asking you to repeat anything, I am only asking you to provide a
citation for one of the textbooks that gives a good description of your
approach to LC tuned circuit design, you have not done that before. Your
continued unwillingness to provide a reference to one of the many
textbooks that you have said describe your perspective on LC tuned circuit
design, while continuing to make the claim, has made it pretty obvious
that your claims relative to textbook explanations is nothing but BS.
That is not to say that your observations are BS, I unraveled that back in
January, it is just the claim that your method may be found in textbooks
that smacks of BS.

I suggest yet again you satisfy your curiosity to inform yourself of the
wonderments we see with LC tuned circuits by reading whatever books exist
on the subjects, and I am sure there is a pile of material on the web.

Yes, I have several books on my bookshelf that explain the "wonderments we
see with LC tuned circuits", that is not a problem, but they explain
methods that are quite different from yours, I am simply looking for a
reference that will more fully explain your methods than what I have been
able to figure out on my own, especially from a quantitative perspective.

Sadly the answer to my question has become crystal clear.

Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

I hope I don't bore you with long winded repetitions
of what is contained in RDH4 and all the other reputable text books.
But whatever they say, I agree with, so I am not needed for citations and page
numbers.
Go find out for yourself like I did, and be confident.

But to make an ideal AM tuner for local BCB to get at least 9 kHz of bw at low
thd and noise,
the most critical part of the exercize is to make the response of the IF
channel using two IFTs
to give around 8 kHz of bw of audio.

This means 16 kHz of IF bw.

To acheive that, its easier to utilise a pair of existing conventional
IFTs rather than wind one's own which necessitates the use of litz wire,
and special patterned widing techniques.

The IFTs chosen don't need to be a matched pair, but I prefer the types from
1950's radios with larger cans and simple solenoid windings with inductive
tuning.

A preliminary investigation should be made to ascertain the response
at 455 kHz by means of ideally placing the IFT as is into an existing IF amp,
with the detector with two CF as I have posted already established,
The input to the IF amp can be to the IFamp tube grid with about the normal
working bais level.
Only a small input signal is needed, to get about 20vrms of carrier signal at
the anode of the IFT tube
This can easily be measured using a simple peak reading volt meter using
a shielded low input capacitance probe, where you have a 500 pF cap
feeding an SS diode with its anode grounded, and the rectified peak +DV level
generated by the
simple detector is then divided by a 2.2M and 270k divider to approximately
give a 10 : 1 reduction of the DV level.
A schematic for such a simple detector is similar to the AGC diode detector in
the schemo I posted.
Many of the ARRL books carry such a simple RF detector schematic for
immediately
converting RF voltages into a DV which is so much more easily measured
remotely without losses or affecting signal level being measured in the high
impedance circuits concerned.
The peak DV can be read by a DVM across the 270k, and x 10 for the real value.
Its a primitive way to measure the anode RF level, but good enough for what we
want.
With such an RF voltage measurer, we can check the input and output levels of
the IFT,
to make sure the tube is not overloaded, and the insertion losses of the IFT
are not excessive.


The frequency adjustable test signal should be set at exactly 455 kHz, measured
with a ditital F meter,
and the carrier level adjusted for about 28 pk volts read from our network.
With the AGC voltage application shunted at the 0.05 uF first cap in the AGC
line

The IFT is tuned up to give the highest AGC voltage at the detector.

I have a modulated test carrier, so I use a low F of modulation of 100 Hz,
and 30% AM is sufficient.

A page is ruled up in an exercize book to record the response for 50 kHz each
side of 455 kHz,
with one line down the page representing -3 dB.
A CRO is used to monitor the level of recovered audio from the detector,
and perhaps you will have about 2vrms of audio to measure.

The 100 Hz sine wave is displayed on the CRO so it occupies the full height of
the screen.

Then you adjust the F of the sig gene and plot the -3,-6,-12,-18,-24 dB audio
levels and record the
carrier F at which these audio levels occur.
One does the graph for each side of the 455 kHz centre F.

The dots are joined, and the graph can be drawn of the IFT selectivity shape.

The 100 Hz audio will decline very nearly exactly with the decline in RF
response of the IFT,
because the 100 Hz modulation causes sideband frquencies only 100 Hz each side
of 455 kHz.

With an average garden variety IFT that an average radio maker of 1950 may have
made,
you will get -3 dB points at around 3 kHz each side of the 45 kHz, with steep
sides beyond this, then
some flattening out beyond the -12 dB points.
The -3 dB points are known as the F1 and F2 response poles, and F2 - F1 = the
bandwidth.
The overall response should look like a section through a bell,
and there should be an attenuation of -24 dB at about 50 kHz away from 455 kHz.

The aim is to make the IFT response wider, so the transformer is removed from
the chassis,
and the windings removed from the cans, and the tube on which the IF coils are
wound has a 5mm wide
peice of tube former cut out without wrecking the litz wire, cap leads, or
anything else.
Each of the 7 strands of fine wire in the litz wire must remain intact.

A plastic or carboard tube is found to tightly fit over the tube stubs of IF
coils, and the the coils moved closer together,
and reassembled into the can, and back onot the chassis.
The response measurement is repeated, and we should see a wider bandwidth
response,
but the slopes of the attenuation beyond the -3 dB will remain the same.
The reponse may even show a twin, but quite unevely peaked response, and that
because when the
IFTs are peaked up with coils closer the mistake was to tune the IFT so one
coil is centred on 455 kHz,
and the other centred on the side peak F which may be at 458 or 452 kHz, so a
if there are two peaks noticable when peaking up the IFT while aligning it, the
response peaks must be
symetrical each side of the 455 kHz, which might appear in the centre of a
trough
in the response.

Its all quite fiddly, and the novice will get trapped everytime.

The IF coil distance is repeatedly adjusted for a slightly troughed response
with two peaks,
and the bw should then be around 12 kHz.

Then we add some damping resistance of say 150 k to each of the LC windings.

This will usually reduce the sightly twin peaked reposne to a single one,
but which has a broad response of 10 kHz.

Sometimes 100 k dmaping R should be added, but the general idea is to
close the distance between IF coils, and use the least R dampers to produce the
widest
single peaked bandwidth with a nicely symetrical shape, which indicates each LC
is exactly
tuned to 455 kHz.

We repeast the whole process again with the second IFT.

The IFTs can then have their coils carefully glued so their distance cannot
vary, and finally
reassembled and mounted in the set with the mixer tube added.

With the RF input tuning gang tuned to the lowest possible RF frequency,
the IF gene signal can be reduced, and fed into the RF input, and enough 455
kHz will
get through to the mixer anode to test the IF response again,
only this time we can have the AGC allowed to be operational, lest we overload
the
mixer and IF amp with too much signal.

The IFs should be realigned for the highest AGC negative voltage, and
symetrical
response shape, which now should show the -3 dB points at say 7 kHz each side
of 455 kHz,
and at least twice the rates of attenuation recorded with one IFT.

The input test F should then be changed to an RF input signal of say 550 kHz,
and the set tuned to this F for maximum AGC, and quick check of the IF
frequency should reveal
455 kHz, if not, the set slightly tuned so that is the case.
The tuning gang should nearly be fully closed.
If not, the oscilator coil slug is adjusted to where 500 kHz should be.
The set should tune up to the high end of the band, so that RF of 1,650 kHz can
be tuned,
if not, the oscilator gang trim cap adjusted to allow 1,650 kHz.
With the set tuned to 550 kHz, the RF input coil slug should be adjusted for
max AGC,
which shows the RF LC is now tuned to the low RF while producing 455 kHz IF.
The set is then tuned to around 1,400 kHz, and the trim cap on the RF tuning
gang adjusted for
max AGC.
This indicates the RF input is tuned to the high RF whilst producing 455
kHz.IF.

With a constant level of RF input, the AGC level across tha band should vary
by not more than +/- 3 dB.
Sometimes its best to align the best tracking at 650 kHz and 1,300 kHz,
if the wanted stations are all in this bandwidth.

Notice how we have not simulated or calculated a single thing during
this whole tedious process, which should take the unititiated a couple of days
to get right,
using the right type of gear, and making only a few mistakes which are
discovered
along the way by those with a skeptical distrust of their abilities, and an
attitude that they
must proove beyond any doubt that everything they measure all adds up to the
design aim.

The same tedious methods for alignment exist in the text books I have
described.

Little attention in the text books has been given to achieving a pass band as
wide as
8 kHz at least, because the textbooks were often written for the mass market
makers
by manufacturers of tubes, who wanted to make it easy to use their products
with as much ease as possible.
The books give little account of using an audio step filter to slightly
emphasize the
recovered HF audio so that we can boost the initial audio roll off caused by
the
IF response so that the response is stretched a bit further from 7 kHz to 9
kHz.
Trying for more than this is high impossible because an RC compensation network

can only have a slight boost, before the roll off due to the IF shape causes a
massive
roll off rate which is impossible to compensate with any RC network.
And we don't want to over compensate, and end up with a hump in the audio
response
at say 2 kHz.

I am sure anything I left out can be found in the books, so don't bother being
lazy
and asking me questions they can answer.

Patrick Turner.

In article ,
Patrick Turner wrote:

I hope I don't bore you with long winded repetitions of what is
contained in RDH4 and all the other reputable text books. But
whatever they say, I agree with, so I am not needed for citations and
page numbers.

Why don't you print all your posts in a book. You certainly have written
enough of them.

So far in June - July you have cross posted 70 threads and hundreds of
messages to RRS.

Stop being a cross posting moron.

--
Telamon
Ventura, California

Telamon wrote:

In article ,
Patrick Turner wrote:

I hope I don't bore you with long winded repetitions of what is
contained in RDH4 and all the other reputable text books. But
whatever they say, I agree with, so I am not needed for citations and
page numbers.

Why don't you print all your posts in a book. You certainly have written
enough of them.

So far in June - July you have cross posted 70 threads and hundreds of
messages to RRS.

Stop being a cross posting moron.

I continue to write as much as I feel like to whoever I feel like.

You will always fail 100% to limit my rights to free speach.

Get your book reading skills up to scratch.

You don't know how to contribute to the discussion, so why not stay out of
it?

Patrick Turner.



--
Telamon
Ventura, California