Re: 'Twin Triode' AM Transmiter

[John Byrns]

In article >,
flipper > wrote:

> On Sun, 28 Aug 2011 20:09:47 -0500, John Byrns >
> wrote:
>
> >In article >,
> > John Byrns > wrote:
> >
> >>
> >> I will try to post a schematic of my proposed 12SN7 transmitter circuit
> >> later
> >> tomorrow.
> >>
> >> It also occurs to me that the "plate" winding of the oscillator
> >> transformer
> >> can
> >> probably be translated to the cathode circuit without changing the AC
> >> equivalent
> >> circuit. This would allow the use of a single winding coil with my
> >> partial
> >> "fix" idea. I will try to post a schematic of this too.
> >
> >I have posted a schematic of my "Twin Triode" AM Transmitter proposal here:
> >
> >http://fmamradios.com/TTAMXmtr.gif
> >
> >This version incorporates my "fix" for the grid bias modulation problem and
> >also
> >illustrates the use of a coil with a single tapped winding, the two winding
> >AES
> >(P-C70-OSC) coil is not required for this version.
>
> Oh, ok. Well, that's what I called the typical AA5 'cathode-grid'
> Hartley. Except, of course, there's not a modulator under it in the
> AA5.
>
> Well, we know circuitmaker isn't modeling the other one right so for
> what it is or not worth circuitmaker says the AM envelope is not
> symmetrical. The 'upper' envelope is a smaller amplitude than the
> lower.
>
> However, I think that's again due to the strange osc cathode voltage
> it comes up with and, so, 'not right'.

That is a disappointing result, I wonder if the problem is somehow inherent in
the morphed circuit, or if I morphed it incorrectly as I didn't draw out the
steps as I had planned to and instead morphed it in my head, or is it a problem
with the simulator? Did you use the same component values & 12AU7 as in your
original "sim", to eliminate double diddling?

That circuit wasn't the one I originally planed on posting first, I had intended
on posting my preferred "two winding" design first, as it has fewer changes from
your circuit. It would be interesting to see what your simulator does with my
preferred proposal. There are two steps to my preferred circuit. The first
step is to move the lead of C2 that is connected to L1 pin 3 to L1 pin 4
instead, and then scale the inductance of L1, adjust the core in the physical
implementation, to maintain the same carrier frequency so that the reactance of
C2 at the carrier frequency will remain the unchanged, before running a "sim".
The second step, which yields my target design, is to split the tapped coil into
two separate windings by disconnecting the grid part of the coil from the center
tap at pin 4, along with the end of C2 that had been moved to pin 4 in the first
step, and then reconnecting them both to the oscillator cathode, yielding my
proposed two winding circuit.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article >,
flipper > wrote:

> On Mon, 29 Aug 2011 20:38:28 -0500, John Byrns >
> wrote:
>
> >In article >,
> > flipper > wrote:
> >
> >> On Sun, 28 Aug 2011 20:09:47 -0500, John Byrns >
> >> wrote:
> >>
> >> >In article >,
> >> > John Byrns > wrote:
> >> >
> >> >>
> >> >> I will try to post a schematic of my proposed 12SN7 transmitter circuit
> >> >> later
> >> >> tomorrow.
> >> >>
> >> >> It also occurs to me that the "plate" winding of the oscillator
> >> >> transformer
> >> >> can
> >> >> probably be translated to the cathode circuit without changing the AC
> >> >> equivalent
> >> >> circuit. This would allow the use of a single winding coil with my
> >> >> partial
> >> >> "fix" idea. I will try to post a schematic of this too.
> >> >
> >> >I have posted a schematic of my "Twin Triode" AM Transmitter proposal
> >> >here:
> >> >
> >> >http://fmamradios.com/TTAMXmtr.gif
> >> >
> >> >This version incorporates my "fix" for the grid bias modulation problem
> >> >and
> >> >also
> >> >illustrates the use of a coil with a single tapped winding, the two
> >> >winding
> >> >AES
> >> >(P-C70-OSC) coil is not required for this version.
> >>
> >> Oh, ok. Well, that's what I called the typical AA5 'cathode-grid'
> >> Hartley. Except, of course, there's not a modulator under it in the
> >> AA5.
> >>
> >> Well, we know circuitmaker isn't modeling the other one right so for
> >> what it is or not worth circuitmaker says the AM envelope is not
> >> symmetrical. The 'upper' envelope is a smaller amplitude than the
> >> lower.
> >>
> >> However, I think that's again due to the strange osc cathode voltage
> >> it comes up with and, so, 'not right'.
> >
> >That is a disappointing result, I wonder if the problem is somehow inherent
> >in
> >the morphed circuit, or if I morphed it incorrectly as I didn't draw out the
> >steps as I had planned to and instead morphed it in my head, or is it a
> >problem
> >with the simulator? Did you use the same component values & 12AU7 as in
> >your
> >original "sim", to eliminate double diddling?
> >
> >That circuit wasn't the one I originally planed on posting first, I had
> >intended
> >on posting my preferred "two winding" design first, as it has fewer changes
> >from
> >your circuit. It would be interesting to see what your simulator does with
> >my
> >preferred proposal. There are two steps to my preferred circuit. The first
> >step is to move the lead of C2 that is connected to L1 pin 3 to L1 pin 4
> >instead, and then scale the inductance of L1, adjust the core in the
> >physical
> >implementation, to maintain the same carrier frequency so that the reactance
> >of
> >C2 at the carrier frequency will remain the unchanged, before running a
> >"sim".
> >The second step, which yields my target design, is to split the tapped coil
> >into
> >two separate windings by disconnecting the grid part of the coil from the
> >center
> >tap at pin 4, along with the end of C2 that had been moved to pin 4 in the
> >first
> >step, and then reconnecting them both to the oscillator cathode, yielding my
> >proposed two winding circuit.
>
> Took me a few minutes to decipher those instructions but I figured it
> out.
>
> Good news is it 'works' and the plate signal looks right. Bad news is
> the plate signal is only 6.8Vrms so there's little RF power from
> there.
>
> More bad news, the 'big endian' coil we get 'mo power' from looks
> asymmetrical like the other one. Worse actually. It's not only
> asymmetrical but on deep mod the 'top side' is heavily distorted from
> what 'looks like' severe slew limiting. Which seems odd since that's
> what we're supposedly trying to fix.

No, the slew rate problem isn't the one I am trying to fix, your scheme of
reducing the time constant is the only way anyone has suggested to fix the slew
rate problem. The problem I am trying to fix is a slightly different but
related problem, which is also "fixed" by reducing the time constant as you have
done.

> The same 'tweaks' work with this one so, by using a 47k series grid
> limiter, I could get osc cathode to ride down around 122V, which
> eliminates the 'cathode crunch' up against the plate signal, so I have
> a sub fraction more semi confidence it might mean something.
>
> I left all values the same as mine except for increasing C2 to 180pF
> to get a similar frequency.

OK, that would reduce the capacitive reactance at the carrier frequency, that
may or may not be a big deal, it will probably increase the operating Q of the
coil considerably, resulting in greater high frequency rolloff.

> On mine I went back to 47k for R3 and it seems to have brightened
> things up a smidgen and it dawned on me,

Yes, that is part of the effect I noticed working with the "Zenith" circuit.

> when I was measuring
> frequency response and getting 'lower amplitude' at 10kHz I was using
> full mod. So maybe the small signal response does improve and it's
> slew limiting I was seeing limit the amplitude. Maybe just wishful
> thinking and I should take it back up to the scope but, for now, I'm
> listening to it.

Could be, I will have to take a few minutes to think about what you are saying
here.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article >,
flipper > wrote:

> On Tue, 30 Aug 2011 21:30:00 -0500, John Byrns >
> wrote:
>
> >In article >,
> > flipper > wrote:
> <snip>
>
> >
> >> when I was measuring
> >> frequency response and getting 'lower amplitude' at 10kHz I was using
> >> full mod. So maybe the small signal response does improve and it's
> >> slew limiting I was seeing limit the amplitude. Maybe just wishful
> >> thinking and I should take it back up to the scope but, for now, I'm
> >> listening to it.
> >
> >Could be, I will have to take a few minutes to think about what you are
> >saying
> >here.

I took 6 days off from thinking about exactly how I will rebuild my AES phono
oscillator with a high level series modulation circuit. In addition to my
wife's project to reduce what I call domestic entropy, which will probably
extend well into 2012, when you are "94", Patrick's number not mine, health
issues crop up which demand attention. The most time consuming though was
purchasing a new car for my wife, which involved scouring the country for one
that was built before, or shortly after, the tsunami in Japan ended the
availability of the optional $395.00 paint job she wanted. The tsunami seems to
have ended the production of a particular type of metallic flakes that were used
in the desired paint.

> Well, I did some more testing.
>
> I started with the intent to put my air variable in and see how it
> works over the whole AM band and, as suspected, I had to increase C1
> to 20pF for it to oscillate on the lower end.
>
> My cap is a typical AA5 dual gang with grounded frame so that meant I
> had to put it on the coil big end only, which meant I had to screw the
> slug full in.
>
> Still, I was able to get around 540 to 1700 and figured the top end
> would drop when I tacked the antenna back on but, to my surprise, not
> only did the top end come down but the lower end of the range came up
> to, say, 650 minimum.

That's interesting, do you have any tentative explanation for the low end
increasing from 540 kHz to 560 kHz when you connected the antenna, at first
glance it doesn't seem to compute?

> The BIG surprise though was the 10kHz amplitude response I talked
> about previously. With everything else remaining the same it changes
> with carrier frequency from rather lousy on the low end to looking
> pretty decent at the very top, and deep mod slew distortion follows
> hand in hand.

That shouldn't have been a big surprise, when the tuning is done with a variable
capacitor the circuit Q increases when tuned to lower frequencies, remember that
the grid resistance forms the main load on the oscillator, so the band width
becomes narrower at the low end, plus there is a double whammy because not only
does the Q increase, but the carrier frequency is lower further narrowing the
bandwidth.

> I don't understand why but it settles the tuning range question. Pick
> C2 so the range is pushed up to the top end.

The "top end" seems to be a better choice anyway because frequency availability
is better up there, and if you are going for a fancy matched antenna to maximize
range, the antenna is likely to be more efficient at the "top end".

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Don Pearce[_3_]]

On Tue, 06 Sep 2011 21:33:20 -0500, flipper > wrote:

>It's posted on the web page so see what you think.
>
>http://flipperhome.dyndns.org//Twin%20Triode%20AM%20Transmitter.htm

Only the photo of the breadboard is showing. And the text - pale blue
upon white - is near illegible.

d

[Don Pearce[_3_]]

On Wed, 07 Sep 2011 05:20:46 GMT, (Don Pearce) wrote:

>On Tue, 06 Sep 2011 21:33:20 -0500, flipper > wrote:
>
>>It's posted on the web page so see what you think.
>>
>>http://flipperhome.dyndns.org//Twin%20Triode%20AM%20Transmitter.htm
>
>Only the photo of the breadboard is showing. And the text - pale blue
>upon white - is near illegible.
>
>d

OK, a second try at loading the page sorted it.

d

[Don Pearce[_3_]]

On Wed, 07 Sep 2011 00:25:34 -0500, flipper > wrote:

>On Wed, 07 Sep 2011 05:23:39 GMT, (Don Pearce) wrote:
>
>>On Wed, 07 Sep 2011 05:20:46 GMT, (Don Pearce) wrote:
>>
>>>On Tue, 06 Sep 2011 21:33:20 -0500, flipper > wrote:
>>>
>>>>It's posted on the web page so see what you think.
>>>>
>>>>http://flipperhome.dyndns.org//Twin%20Triode%20AM%20Transmitter.htm
>>>
>>>Only the photo of the breadboard is showing. And the text - pale blue
>>>upon white - is near illegible.
>>>
>>>d
>>
>>OK, a second try at loading the page sorted it.
>>
>>d
>
>Ok, good.
>
>I've also noticed some odd Internet behavior but don't know if it's my
>end or the ISP (it's usually not my end).

I notice the site is directed through dyndns. Is it on your own home
PC or does the ISP host it?

d

[Don Pearce[_3_]]

On Wed, 07 Sep 2011 01:05:34 -0500, flipper > wrote:

>On Wed, 07 Sep 2011 05:40:19 GMT, (Don Pearce) wrote:
>
>>On Wed, 07 Sep 2011 00:25:34 -0500, flipper > wrote:
>>
>>>On Wed, 07 Sep 2011 05:23:39 GMT, (Don Pearce) wrote:
>>>
>>>>On Wed, 07 Sep 2011 05:20:46 GMT, (Don Pearce) wrote:
>>>>
>>>>>On Tue, 06 Sep 2011 21:33:20 -0500, flipper > wrote:
>>>>>
>>>>>>It's posted on the web page so see what you think.
>>>>>>
>>>>>>http://flipperhome.dyndns.org//Twin%20Triode%20AM%20Transmitter.htm
>>>>>
>>>>>Only the photo of the breadboard is showing. And the text - pale blue
>>>>>upon white - is near illegible.
>>>>>
>>>>>d
>>>>
>>>>OK, a second try at loading the page sorted it.
>>>>
>>>>d
>>>
>>>Ok, good.
>>>
>>>I've also noticed some odd Internet behavior but don't know if it's my
>>>end or the ISP (it's usually not my end).
>>
>>I notice the site is directed through dyndns. Is it on your own home
>>PC or does the ISP host it?
>>
>>d
>
>It's here at home on a old P133 notebook running Windows98SE. The web
>server is Savant. That's all it does.
>
>Local LAN access is zippity do da fine.

I think the problem is the dynamic dns server. It probably needs
reminding who you are from time to time. The web site would work much
better if you hosted it your ISP.

d

[John Byrns]

In article >,
(Don Pearce) wrote:

> On Wed, 07 Sep 2011 01:05:34 -0500, flipper > wrote:
>
> >On Wed, 07 Sep 2011 05:40:19 GMT, (Don Pearce) wrote:
> >
> >>On Wed, 07 Sep 2011 00:25:34 -0500, flipper > wrote:
> >>
> >>>On Wed, 07 Sep 2011 05:23:39 GMT, (Don Pearce) wrote:
> >>>
> >>>>On Wed, 07 Sep 2011 05:20:46 GMT, (Don Pearce) wrote:
> >>>>
> >>>>>On Tue, 06 Sep 2011 21:33:20 -0500, flipper > wrote:
> >>>>>
> >>>>>>It's posted on the web page so see what you think.
> >>>>>>
> >>>>>>http://flipperhome.dyndns.org//Twin%20Triode%20AM%20Transmitter.htm
> >>>>>
> >>>>>Only the photo of the breadboard is showing. And the text - pale blue
> >>>>>upon white - is near illegible.
> >>>>>
> >>>>>d
> >>>>
> >>>>OK, a second try at loading the page sorted it.
> >>>>
> >>>>d
> >>>
> >>>Ok, good.
> >>>
> >>>I've also noticed some odd Internet behavior but don't know if it's my
> >>>end or the ISP (it's usually not my end).
> >>
> >>I notice the site is directed through dyndns. Is it on your own home
> >>PC or does the ISP host it?
> >>
> >>d
> >
> >It's here at home on a old P133 notebook running Windows98SE. The web
> >server is Savant. That's all it does.
> >
> >Local LAN access is zippity do da fine.
>
> I think the problem is the dynamic dns server. It probably needs
> reminding who you are from time to time. The web site would work much
> better if you hosted it your ISP.

However it would be lacking in cool factor if hosted at the "ISP".

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article >,
flipper > wrote:

> On Sun, 04 Sep 2011 20:25:25 -0500, flipper > wrote:
>
> <snip>
>
> >
> >Some sort of modulation meter would be more useful but neither easy
> >nor particularly inexpensive, compared to the rest of it.
> >
>
> Well, I kludged up a simple 'modulation indicator' in spice, but I
> haven't actually built it.
>
> It's posted on the web page so see what you think.
>
> http://flipperhome.dyndns.org//Twin%20Triode%20AM%20Transmitter.htm
>
> The commonly obvious thing would be to make a 'LED VU' meter with
> something like the LM3915 but the modulator drive has an interesting
> problem in that trying to sense 'max' positive mod may not be so
> simple due to positive grid drive potentially clipping the signal. It
> shouldn't if everything is 'right', but it might not be, and then
> there's the matter of 'calibrating' the thing even if all else is
> okay.
>
> Positive mod isn't the 'real' (or biggest) problem anyway, it's
> carrier cutoff on negative mod, and 0 modulator current has the
> advantage of being 0 no matter what 'idle' or peak positive mod is. So
> I look for modulator cathode approaching 0, rather than worry about
> 'calibrating' to positive mod, then peak hold/decay it for visual.
> (Note, the LM3915 doesn't peak hold so you could easily miss seeing
> carrier cutoff events even if you got it calibrated right)
>
> The yellow 'mod' LEDs are for visual effect. The R15/C19 filter senses
> 'idle current' and U1B, then, generates (rectified half sine) pulses
> proportional to positive mod to drive the LEDs. We don't care if it's
> terribly accurate as it's sort of like those 'musical lights' things
> where light intensity dances around to the music.
>
> The idea is a sort of 'glow indicator'. Like, maybe, aiming the LEDs
> into a diffuser, or not. I haven't worked out the details. Probably
> need to buy some 'wide angle' flat LEDs and see just how well they
> disperse and how it looks.
>
> In theory it would be nice to know our peaks are -3dB, or whatever,
> but for something this simple I think people would slam it as high as
> they can anyway, so a red 'too much' indicator is probably sufficient,
> and calibration, as mentioned, would be non trivial so it's unlikely
> -3dB would have much meaning to the actual problem. At least that was
> my thinking: a modulation 'idiot light', to use a car analogy.
>
> A simple tripler off the heater supply should be enough to power it.

It's an interesting coincidence that you are working on this idea, as I have
been thinking along similar lines the past couple of days.

As you allude to above, I think the important thing would be to monitor the
onset of oscillator dropout and flash warning indicator at that point.

The idea that struck me was to use a neon bulb like some 1950s era tape
recorders used as recording level indicators. To a first approximation a neon
bulb and two resistors would do the job. One connection of the neon bulb would
be go to the modulator plate, with the second connection to the junction of two
resistors forming a voltage divider between B+ and ground. The potential at the
junction of the two resistors would set the firing voltage for the neon bulb,
and the fact that the firing voltage is higher than the conduction voltage would
provide a sort of faux peak hold effect. Calibration of this circuit should be
easy as we only have to identify the point where the oscillator drops out, not
the precise point of 100% modulation.

There is at least one obvious problem with this elegantly simple scheme, which
is that when the neon bulb fires it would drag the modulator plate down to a
lower voltage, creating serious distortion. This problem could be overcome at
the cost of an additional triode operating as a cathode follower to drive the
neon bulb, buffering the modulator plate. Unfortunately besides requiring the
extra triode, an additional higher voltage power supply would be required for
the cathode follower plate, to prevent it from saturating before the point where
the oscillator drops out.

Most old vacuum tube era professional AM broadcast modulation monitors required
calibration vs. an external source such as Patrick's trapezoidal CRO display.
There was one old modulation monitor though which was self calibrating, that was
the Gates M5693 Modulation Monitor. The ideas from this monitor should be easy
to adapt to the modern solid state opamp era. A year ago I worked out a scheme
to implement this self calibrating idea in a single quad opamp to drive an
analog peak modulation meter. A little additional circuitry, as in the actual
MS5693, could provide for driving a red peak flasher LED, and the analog meter
could be replaced with an orange LED "VU" meter.

An AM demodulator would be required to drive this circuit, the oscillator plate,
which is a lower impedance point than the grid circuit where the antenna is
connected, might provide enough RF drive voltage for a demodulator diode without
unduly loading the oscillator circuit.

If you want to puzzle out the self-calibrating principle used in the M5693, the
operating instruction manual and schematic can found here.

http://louise.hallikainen.org/~harold/BroadcastHistory/uploads/GatesM5693.pdf

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Don Pearce[_3_]]

On Wed, 07 Sep 2011 15:47:48 -0500, flipper > wrote:

>On Wed, 07 Sep 2011 15:45:35 GMT, (Don Pearce) wrote:
>
>>On Wed, 07 Sep 2011 01:05:34 -0500, flipper > wrote:
>>
>>>On Wed, 07 Sep 2011 05:40:19 GMT, (Don Pearce) wrote:
>>>
>>>>On Wed, 07 Sep 2011 00:25:34 -0500, flipper > wrote:
>>>>
>>>>>On Wed, 07 Sep 2011 05:23:39 GMT, (Don Pearce) wrote:
>>>>>
>>>>>>On Wed, 07 Sep 2011 05:20:46 GMT, (Don Pearce) wrote:
>>>>>>
>>>>>>>On Tue, 06 Sep 2011 21:33:20 -0500, flipper > wrote:
>>>>>>>
>>>>>>>>It's posted on the web page so see what you think.
>>>>>>>>
>>>>>>>>http://flipperhome.dyndns.org//Twin%20Triode%20AM%20Transmitter.htm
>>>>>>>
>>>>>>>Only the photo of the breadboard is showing. And the text - pale blue
>>>>>>>upon white - is near illegible.
>>>>>>>
>>>>>>>d
>>>>>>
>>>>>>OK, a second try at loading the page sorted it.
>>>>>>
>>>>>>d
>>>>>
>>>>>Ok, good.
>>>>>
>>>>>I've also noticed some odd Internet behavior but don't know if it's my
>>>>>end or the ISP (it's usually not my end).
>>>>
>>>>I notice the site is directed through dyndns. Is it on your own home
>>>>PC or does the ISP host it?
>>>>
>>>>d
>>>
>>>It's here at home on a old P133 notebook running Windows98SE. The web
>>>server is Savant. That's all it does.
>>>
>>>Local LAN access is zippity do da fine.
>>
>>I think the problem is the dynamic dns server.
>
>That wouldn't explain the recent 'in general' odd Internet behavior
>I'm seeing since none of that is dyndns related.
>
>> It probably needs
>>reminding who you are from time to time.
>
>I have a semi static IP. I haven't ever seen it change 'on it's own',
>it's just not 'guaranteed'.
>

It will have an assigned lifetime and will expire if you don't log in
for some time. That time is probably upwards of a week.

>> The web site would work much
>>better if you hosted it your ISP.
>>
>>d
>
>I prefer the control and it being ISP 'independent'.

Hardly ISP-independent. Your IP address still comes from the block
assigned to your current ISP.

d

[Don[_4_]]

In article >,
says...
Where one *finds* 6DR4s, though, I don't know
>as ABCvacuumtubes doesn't list it and they seem to be rather rare.

http://www.tubedepot.com/tcp8.html $4.95