Now that the dust has had a chance to settle on the "Improved AM Detector"
wars, here is an article I found that discusses the truth about diode
detector distortion. The article from no less of a source than Wireless
World magazine and author W.T.Cocking suggests that driving the diode with
a low impedance source like a cathode follower may be counter productive,
while stating that a cathode follower after the a diode detector is
helpful, but may not be needed.

http://users.rcn.com/jbyrns/stuff/171.jpg
http://users.rcn.com/jbyrns/stuff/172.jpg


Regards,

John Byrns


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

John Byrns wrote:
Now that the dust has had a chance to settle on the "Improved AM Detector"
wars, here is an article I found that discusses the truth about diode
detector distortion. The article from no less of a source than Wireless
World magazine and author W.T.Cocking suggests that driving the diode with
a low impedance source like a cathode follower may be counter productive,
while stating that a cathode follower after the a diode detector is
helpful, but may not be needed.

http://users.rcn.com/jbyrns/stuff/171.jpg
http://users.rcn.com/jbyrns/stuff/172.jpg



Lies, lies, all lies... :-)

The last paragraph says that the equations are not directly
applicable to the usual coupled pair of LC tuned circuits
of the bandpass type (I think found)in AA5s. But similar
equations do apply for a
similar effect.

Seriously, from my playing around it seems that the
most benefit can be had from splitting off the AVC
circuit (giving it its own diode), and the use of lowered
heater voltage on the audio's detector diode.
Careful selection of the IF filter caps (33p instead of
the more usual 120pF) avoids tangent distortion to 5KHz).
Less benefit is had from the use of a cathode follower
for the detected audio (direct coupling from the detector
avoids the AC/DC ratio issue).
The IF signal CF yields the least benefit and thus not
really needed. I though it might reduce the effects of the
loading by the AVC diode some, but other factors seem to
swamp that out (like nonlinearities in the audio diode).

Lately I've been trying the idea of reduced heater voltage
on the 12AV6 tube in AA5s. Only thing is that reduced
contact potential for the diodes has to be balanced off the
loss of contact potential in the triode. The sweet spot seems
to be at 10V for a tube meant for 12.6V. I suspect that the
diodes in the 12AV6 were designed for low contact potential
and the triode designed for more contact potential. Compare
visually the set of plates for the diodes in a 12AV6 vs a
6AL5. Looks like a lot less effective surface area in the 12AV6.
Seems that reduced heat helps it along some. As the triode
circuit has low plate current reduced heat shouldn't hurt
the tube's life.

John Byrns wrote:
Now that the dust has had a chance to settle on the "Improved AM Detector"
wars,

I just want to know if you guys could tell a blind test from one
detector to another to a ZN414 single chip receiver. Is there even a
station with a high enough quality broadcast to tell?


Adam

Behold, Adam Stouffer signalled from keyed 4-1000A filament:

John Byrns wrote:
Now that the dust has had a chance to settle on the "Improved AM
Detector" wars,

I just want to know if you guys could tell a blind test from one
detector to another to a ZN414 single chip receiver.

Yes.

Is there even a
station with a high enough quality broadcast to tell?

Probably not.

--
Gregg t3h g33k
"Ratings are for transistors....tubes have guidelines"
http://geek.scorpiorising.ca

John Byrns wrote:

Now that the dust has had a chance to settle on the "Improved AM Detector"
wars, here is an article I found that discusses the truth about diode
detector distortion.

You never let the dust from what you say are wars ever settle fully.

The article from no less of a source than Wireless
World magazine and author W.T.Cocking suggests that driving the diode with
a low impedance source like a cathode follower may be counter productive,
while stating that a cathode follower after the a diode detector is
helpful, but may not be needed.

http://users.rcn.com/jbyrns/stuff/171.jpg
http://users.rcn.com/jbyrns/stuff/172.jpg

Regards,

Mr Cocking says nothing to suggest I waste my time using my present methods
of using a CF after the last IFT LC, and to drive a diode + RC network where
a substantially constant current discharge of the peak and hold C is the
eminent feature.

When Mr Cocking takes the trouble of measuring his distortions,
and building my design of detector circuit, and measuring all those, I will
believe what he says,
but at present he offers me no better listening experiences than I have
offered myself.
He may go on and on with all those formulas and equations which convinces me
that
he's very good at that, but that he's useless with a soldering iron and
distortion meter.

Patrick Turner.





John Byrns

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

Adam Stouffer wrote:

John Byrns wrote:
Now that the dust has had a chance to settle on the "Improved AM Detector"
wars,

I just want to know if you guys could tell a blind test from one
detector to another to a ZN414 single chip receiver. Is there even a
station with a high enough quality broadcast to tell?

We have AM stations which transmit audio bandwidth out to 9 kHz,
and it isn't compressed and stuffed around with.
The AM set I use with 10 kHz of AF bandwidth and less than 1% of thd at 50% AM
is a pleasure to listen with, but nearly all SS sets are abominations to the
ear.


Patrick Turner.



Adam

In article , Robert Casey
wrote:

Lies, lies, all lies... :-)

The last paragraph says that the equations are not directly
applicable to the usual coupled pair of LC tuned circuits
of the bandpass type (I think found)in AA5s. But similar
equations do apply for a
similar effect.

Yes, the effects are much more complicated when a double tuned transformer
is used between the final IF amplifier tube and the detector diode.

Seriously, from my playing around it seems that the
most benefit can be had from splitting off the AVC
circuit (giving it its own diode), and the use of lowered
heater voltage on the audio's detector diode.

An AGC rectifier feed from the same IF amplifier clearly complicates the
situation, aside from obviously improving the AC/DC load ratio of the
audio detector, it will also affect Cocking's analysis, but it isn't
immediately clear to me if the overall impact on the effect Cocking
illustrates is positive, negative, or neutral.

Careful selection of the IF filter caps (33p instead of
the more usual 120pF) avoids tangent distortion to 5KHz).

I believe "tangent distortion" is also less serious than often assumed
when the RF/IF filters are used provide the de-emphasis necessary to
complement the modified 75 usec. pre-emphasis used in the US. For example
if the necessary de-emphasis is accomplished in the RF/IF filters, the
maximum modulation at 5 kHz would be about 40%.

Less benefit is had from the use of a cathode follower
for the detected audio (direct coupling from the detector
avoids the AC/DC ratio issue).

"tangent distortion" and "AC/DC ratio" are separate issues with different
causes and solutions.

The IF signal CF yields the least benefit and thus not
really needed. I though it might reduce the effects of the
loading by the AVC diode some, but other factors seem to
swamp that out (like nonlinearities in the audio diode).

If you believe what Cocking says, an IF signal CF is counter productive
because a high source impedance has desirable effects.

Lately I've been trying the idea of reduced heater voltage
on the 12AV6 tube in AA5s. Only thing is that reduced
contact potential for the diodes has to be balanced off the
loss of contact potential in the triode. The sweet spot seems
to be at 10V for a tube meant for 12.6V.

Your 10 volt figure is interesting because that is the heater voltage at
which H.H. Scott, and some other Hi-Fi manufacturers, ran their 12AX7, and
similar, small signal triodes.


Regards,

John Byrns


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

Seriously, from my playing around it seems that the
most benefit can be had from splitting off the AVC
circuit (giving it its own diode), and the use of lowered
heater voltage on the audio's detector diode.


An AGC rectifier feed from the same IF amplifier clearly complicates the
situation, aside from obviously improving the AC/DC load ratio of the
audio detector, it will also affect Cocking's analysis, but it isn't
immediately clear to me if the overall impact on the effect Cocking
illustrates is positive, negative, or neutral.


Careful selection of the IF filter caps (33p instead of
the more usual 120pF) avoids tangent distortion to 5KHz).


I believe "tangent distortion" is also less serious than often assumed
when the RF/IF filters are used provide the de-emphasis necessary to
complement the modified 75 usec. pre-emphasis used in the US. For example
if the necessary de-emphasis is accomplished in the RF/IF filters, the
maximum modulation at 5 kHz would be about 40%.

Then I should be in good shape when working with a signal and an IF
that will pass a full 20KHz AM radio signal to detect a full 10Khz
worth of audio. I chose to look at 5Khz a distortion products
above that would fall above 10Khz and could be low pass filtered out
in the audio stages if necessary. So far it doesn't seem to be
necessary.

Something else I would want to look at closer is the result
of an asymmetrical IF response on envelope detector distortion.
IIRC an envelope detector doesn't do well on a single side
band plus carrier signal. It may need both sidebands to get
low distortion. An asymmetrical IF response might mean that
there would be mismatched amplitudes for upper and lower sideband
frequencies and thus distortion.



Less benefit is had from the use of a cathode follower
for the detected audio (direct coupling from the detector
avoids the AC/DC ratio issue).


"tangent distortion" and "AC/DC ratio" are separate issues with different
causes and solutions.

Agreed. I may not have been clear in my writing in that that my
writing might have merged them by mistake.


The IF signal CF yields the least benefit and thus not
really needed. I though it might reduce the effects of the
loading by the AVC diode some, but other factors seem to
swamp that out (like nonlinearities in the audio diode).


If you believe what Cocking says, an IF signal CF is counter productive
because a high source impedance has desirable effects.


Lately I've been trying the idea of reduced heater voltage
on the 12AV6 tube in AA5s. Only thing is that reduced
contact potential for the diodes has to be balanced off the
loss of contact potential in the triode. The sweet spot seems
to be at 10V for a tube meant for 12.6V.


Your 10 volt figure is interesting because that is the heater voltage at
which H.H. Scott, and some other Hi-Fi manufacturers, ran their 12AX7, and
similar, small signal triodes.



That would indicate that the 12AV6 won't be hurt by low heater
voltage.

In article , Robert Casey
wrote:

Something else I would want to look at closer is the result
of an asymmetrical IF response on envelope detector distortion.
IIRC an envelope detector doesn't do well on a single side
band plus carrier signal. It may need both sidebands to get
low distortion. An asymmetrical IF response might mean that
there would be mismatched amplitudes for upper and lower sideband
frequencies and thus distortion.

Yes, envelope detectors need symmetrical sidebands for distortion free
demodulation. Asymmetrical IF response is bad, it creates the same sort
of distortion in an envelope detector as SSB-FC, but not quite as severe.

Your 10 volt figure is interesting because that is the heater voltage at
which H.H. Scott, and some other Hi-Fi manufacturers, ran their 12AX7, and
similar, small signal triodes.

That would indicate that the 12AV6 won't be hurt by low heater
voltage.

The vast majority of the H.H. Scott 12AX7 stages used cathode bias so it
may not be relevant to the 12AV6 with contact potential bias. Scott did
use contact potential bias in a few applications, but the one that comes
immediately to mind probably had a full 6.3/12.6 volts on the heater.


Regards,

John Byrns


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