notes on the Wenzel low power AM transmitter schematic

In many discussions of LPAM transmitter design you will hear or read references to "the Wenzel circuit" or "the Wenzel transmitter." These refer to the clever transmitter design downloadable at www.wenzel.com/pdffiles/amxmit.pdf

During the mid 1990s and early 2000s, this was the best available AM transmitter circuit if you wanted to build something yourself for around-the-house "yardcasting." The range is limited but the sound quality is good compared to other homebrew circuits that were published at the time.

An updated, illustrated web page about this circuit, including a modified version with a different output section, is published by Mr Wenzel himself at www.techlib.com/electronics/amxmit.htm

Below are some comments that were published in various newsgroups about the original version of the circuit. It's fascinating to see what a detailed analysis and spirited discussion this circuit received!

 
----------------------------------------------------------------

Subject: Re: Micro Transmitter Project
From: Syl 
Newsgroups: rec.antiques.radio+phono
Date: Sat, 30 Nov 2002 12:00 AM
Message-ID: <[email protected]>

Joe Bento a écrit :
> 
> Is that the little transmitter that Syl says is among the best he's
> tried?  If so, I am using that as a basis for my design.  If an LC
> works in place of the crystal, I'll certainly give it a try.
> 
> Joe

Yes it is. It's also on my website. The only really good design so far.
I use planar transistors (2N5551) to replace the 2n4401 because I had
them on hands by the dozen and I can drive them with 150V...(like I need it...)

I suspect any 2SCxxxx from the junk box might also work...

Now that Sal sent me a crystal, I'll replace the LC I replace it with...
I think Bill M. also did this with his...

Syl

----------------------------------------------------------------

Subject: Re: AM Transmitters.
From: Syl 
Newsgroups: rec.antiques.radio+phono
Date: Tue, 10 Dec 2002 10:27 PM
Message-ID: <[email protected]>

John Byrns a Žcrit :

> The "diff amp" type
> circuit favored by Syl, and BM, looks like it may have distortion problems
> related to antenna loading.

Tried that and it doesn't seems to affect the quality of the ouput.

Even on the scope. At least not within it's "useful" range with wrong
antenna loading. With the proper loading (which is somewhat important
for increased range), sound is excellent. If unproperly loaded, close 
to the transmitter the sound is excellent too but the range suffers a *lot.
It does distort outside the "useful" range but I think it is caused 
by the weak signal (some of my radios are better than other though)as the 
ouput is very clean close to the transmitter. I tune the antenna using
an adjustable coil (using a ferrite) I found in my junk box.

I could also use a fixed value for the L and use an adjustable C to 
tune the antenna I guess.

Syl

----------------------------------------------------------------

Subject: Re: AM Transmitters.
From: "Phil B" 
Newsgroups: rec.antiques.radio+phono
Date: Thu, 12 Dec 2002 1:53 AM
Message-ID: 

"John Byrns" wrote:
>  The "diff amp" type circuit favored by Syl, and BM, looks like it
>  may have distortion problems related to antenna loading.

If you look closely at the "diff amp" circuit at
http://www.wenzel.com/pdffiles/amxmit.pdf
you will notice a 470 ohm resistor paralleling the output tuned circuit.

This resistor is essentially the only load on the output, since the
short antenna doesn't even come close to loading the output.  Without
the resistor the tank just rings horribly and sort of loosely follows
the modulation envelope.  With the resistor the output is very clean and
linear and is quite unaffected by the antenna.  99% of the "100 mw"
power is absorbed in this resistor, but the waveform is linear.

The Vectronics doesn't have a load resistor.  It does have a better
output matching circuit.  It has a "reverse" pi network (low impedance
in and high impedance out).  This maximized the voltage presented to a
short antenna with its very low radiation resistance to improve the
loading (marginally).  The Vectronics circuit would benefit with a load
resistor of say 500 - 1000 ohms across C8 or a larger resistor of say
20k - 50k from the antenna terminal to ground.

This loading problem is common to all BC frequency transmitters with
short antennas.  You just can't get very much of the power to go to a
short antenna.  Another solution is to use an antenna of several hundred
feet or more and a good antenna matching network.

Phil B

----------------------------------------------------------------

Subject: Re: AM Transmitters.
From: Syl 
Newsgroups: rec.antiques.radio+phono
Date: Fri, 13 Dec 2002 11:58 PM
Message-ID: <[email protected]>

Bill M a Žcrit :

> Thats the good one even though John Byrns has reservations about its
> quality of signal :-)

Play a little with the osc. resistors to get a better sinewave. I changed
a few resistors until I got an almost perfectly uniform sinewave on
the scope. For some reasons, on some radios the signal got much better...

>   I've never really haggled much over the output loading inductance
> because I have used 'illegal' antenna lengths and the coverage area has
> been sufficient for my needs.
> But I wonder if a 3 foot antenna would really present much difference at
> 1.6 versus 1.0?  Both present an extremely high impedance.  I think his
> measure of "half" the reading is based on reflection (VSWR) showing up
> on the meter.  I suppose ideally a tunable inductor that would go up to
> maybe 1.5mh

Exactly what I am using on mine. Dunno about the exact value. Right from the
junk parts bin. I tuned by ear for best signal. It does improve again the
signal...

Syl

----------------------------------------------------------------

Subject: Re: AM Transmitters.
From: [email protected] (John Byrns)
Newsgroups: rec.antiques.radio+phono
Date: Fri, 13 Dec 2002 10:07 PM
Message-ID: 

In article , Bill M
 wrote:

> 
> Thats the good one even though John Byrns has reservations about its 
> quality of signal :-)


Hi Bill,

Someone else pointed out in one of these transmitter threads that circuit
used in "amxmit.pdf" doesn't suffer the loading sensitivity problem, I
mentioned that type of diff amp transmitter has, because of the 470 Ohm
resistor that I failed to notice in the circuit, which provides the main
load for the transmitter, making the load from the antenna largely
irrelevant.  Think how much more power you could get out of that little
transmitter, if the 470 Ohm resistor were removed, and the antenna was
actually properly matched to the output of the diff amp.

As far as I am concerned it would be better to just use a simple class C
stage like the Vectronics, and the WPDJ, then you wouldn't have to worry
about proper loading, or dumping most of the power into a resistor.  A
similar transmitter with a class C output stage could also be done with 3
transistors.


Regards,

John Byrns

Surf my web pages at,  http://www.enteract.com/~jbyrns/index.html

----------------------------------------------------------------

Subject: Re: AM Transmitters.
From: Bill M 
Newsgroups: rec.antiques.radio+phono
Date: Fri, 13 Dec 2002 10:33 PM
Message-ID: 

John Byrns wrote:

> As far as I am concerned it would be better to just use a simple class C
> stage like the Vectronics, and the WPDJ, then you wouldn't have to worry
> about proper loading, or dumping most of the power into a resistor.  A
> similar transmitter with a class C output stage could also be done with 3
> transistors.

Most of the power is not "dumped into the resistor".  It only serves to 
flatten the Q to make the circuit more easily reproducible and flexible 
across the BCB range with equal results.

Granted that better Q could deliver more output (at some risk of 
affecting modulation peaks if one designs by the book) but I think the 
philosophy presented by the guy is more along the lines of making 
something that will work for the average homebrewer without poring over 
particular voltages, output transistor types, antenna lengths, etc.

This is a very important aspect for published designs that can actually 
work properly when the user may have nothing more than a junkbox and a 
broken VOM to work with.  To its credit, I deadbugged this circuit with 
junkbox equivalent parts into a ball of stray leads tacked together to 
see if it would work before I proceeded with a permanent mounting 
configuration.  Many 'idea' circuits won't pass such a test.

I'm powering it with a 9 volt battery and LC in place of the xtal.  I 
bypassed the 100 ohm emitter resistor with 4.7uf or so and I think it 
sounds superb compared to any of the tube-type versions of a 
"home-phono-oscillator" that I have tried.

Your improvements on the basic design would be appreciated.

-Bill

----------------------------------------------------------------

Subject: Re: AM Transmitters.
From: [email protected] (John Byrns)
Newsgroups: rec.antiques.radio+phono
Date: Sat, 14 Dec 2002 2:11 AM
Message-ID: 

In article , Bill M
 wrote:

> Most of the power is not "dumped into the resistor".  It only serves to 
> flatten the Q to make the circuit more easily reproducible and flexible 
> across the BCB range with equal results.

Leaving aside for the moment exactly how much power is being dumped into
the resistor, the resistor clearly does more than simply "flatten the Q". 
The value of the resistor is an order of magnitude lower than is required
simply to "flatten the Q", in fact with the 470 Ohm resistor, the Q of the
circuit is only about 2 at 1 MHz.  Consider operation at 1 MHz, from the
chart in the transmitter documentation, the tank circuit L should be about
35 uH.  Lets say we want the audio response, without any antenna load at
all, to be down 0.26 dB at 5 kHz, and down 0.97 dB at 10 kHz, then
assuming no other losses in the circuit, the resistor would be 5,328
Ohms.  This is more than an order of magnitude higher than the 470 Ohm
resistor that is actually used in the circuit, plus there are other losses
in the real circuit, making the necessary resistor value even higher. 

Clearly the resistor is doing more than insuring that the Q is adequately
low for good audio response, even with no antenna loading.  I suspect that
purpose of the resistor is to swamp the output, to insure the modulation
of the diff amp is linear even when lightly loaded by the antenna.

My philosophy is to design so the adjustments are as simple as possible,
and I think a class C output stage is a better way to accomplish that,
than is the diff amp design, because a class C stage with high level
modulation will modulate properly even if it is not fully loaded.  The
design of the "amxmit" still requires the builder to futz with two
adjustments to the output stage, the tuning for resonance, and selecting
the proper loading coil.  The 470 Ohm resistor appears to have the
function of swamping the antenna load to keep the modulation linear, even
if the loading by the antenna is not correct.  The class C stage on the
other hand does not require this wasteful resistor for linear modulation,
even when it is lightly loaded.  With the class C stage, if one is not
concerned with squeezing out every last bit of range, you can ignore the
loading, and just peak the tuning.  To get the same simplicity of
operation with the diff amp circuit, the designer had to resort to the 470
Ohm resistor to insure that the load presented to the diff amp stage was
not too great for proper modulation.

One possible advantage that the diff amp may have is that it may be less
subject to the drive signal bleed through that limits the negative
modulation of a class C transistor amplifier, unless the drive signal is
also partially modulated.  I don't know if that is the case or not, it is
a question to be answered.  The diff amp may also be simpler in that it
doesn't require neutralization, while the class C amplifier may an extra
buffer stage to eliminate any possible need for neutralization, which
would otherwise violate the simplicity rule.  Both of these effects have
the same root cause.  That would bring the transistor count for the class
C design up to 5, or one more than the diff amp circuit uses.


Regards,

John Byrns


Surf my web pages at,  http://www.enteract.com/~jbyrns/index.html

----------------------------------------------------------------

Subject: Re: AM Transmitters.
From: "Phil B" 
Newsgroups: rec.antiques.radio+phono
Date: Sat, Dec 14, 2002 3:59 AM
Message-ID: 

"John Byrns"  wrote:

> design of the "amxmit" still requires the builder to futz with two
> adjustments to the output stage, the tuning for resonance, and
selecting
> the proper loading coil.  The 470 Ohm resistor appears to have the
> function of swamping the antenna load to keep the modulation linear,
even
> if the loading by the antenna is not correct.  The class C stage on
the
> other hand does not require this wasteful resistor for linear
modulation,
> even when it is lightly loaded.  With the class C stage, if one is not

> concerned with squeezing out every last bit of range, you can ignore
the
> loading, and just peak the tuning.  To get the same simplicity of
> operation with the diff amp circuit, the designer had to resort to the
470
> Ohm resistor to insure that the load presented to the diff amp stage
was

John,

You are correct in your analysis.  I posted spice simulation results of
this circuit on the binaries.  The effect of changing the 470 ohm
resistor is quite dramatic.  Here is what you will see on the binaries
post:

This post relates to the current discussion of the same title on RAR+P.

Here are spice simulation results for the "Personal Radio Station"
transmitter RF out and modulator sections showing the effect of changing
the 470 ohm resistor across the output tank.

The pictures show 3 values for this resistor.
1. 470 ohms (as specified in the schematic)
2. 1000 ohms
3. Resistor deleted.

The simulation was run with:
- 1610 khz RF, 3 khz modulation
- Antenna model is: 50 pf in series with .1 ohm radiation resistance in
series with 30 ohms to ground (typical ground loss)
- Waveforms taken across .1 ohm radiation resistance
- Tank L at resonance = 21 uH
- Tank C at resonance = 200 pF
- Modulation was limited to less than 100% due to a circuit limitation
that can easily be fixed with some resistor value changes. (I will
address this in another post).

Results clearly show that without proper resistor value, the tank
circuit rings and does not faithfully follow the modulation signal.

Phil B

----------------------------------------------------------------

Subject: Re: AM TRANSMITTERS
From: "Phil B" 
Newsgroups: alt.binaries.pictures.radio
Date: Sat, 14 Dec 2002 11:28 PM
Message-ID: 

Since the "Personal Radio Station" transmitter web page seems to
emphasize using a loading coil, I re-ran the spice with a loading coil.
I neglected to mention in the last post that those spice results were
WITHOUT a loading coil. I got to thinking that a loading coil could make
a big difference in the amount of power extracted from the tank circuit.

I spent an inordinate amount of time playing with the numbers until I
came up with a loading coil value and changes to the L/C tank values to
get a peak.  The loading coil affects the resonance of the tank.  A
truly practical version of this transmitter would require a tunable L
inductor for the tank and a tunable loading coil to get an optimum peak.
This is a fact glossed over in the author's description.

Anyway, here are the same three pictures WITH a loading coil.  Things to
notice:  The output voltage across the radiation resistance is almost 10
times higher (shows the value using a loading coil). The waveform is not
distorted as greatly with load resistors higher than 470 ohms.  But the
waveform is still significantly distorted, so the resistor is still
necessary.

As before:
The pictures show 3 values for this resistor.
1. 470 ohms (as specified in the schematic)
2. 1000 ohms
3. Resistor deleted.

- 1610 khz RF, 3 khz modulation
- Waveforms taken across .1 ohm radiation resistance
- Modulation was limited to less than 100% due to a circuit limitation
that can easily be fixed with some resistor value changes. (I still need
to address this in another post).

Things I changed:
- Antenna model is: 40 pf in series with .1 ohm radiation resistance in
series with 30 ohms to ground (typical ground loss).  Changed from 50 pf
to be possibly more accurate representation of a 3 meter antenna in
proximity to objects.
- Tank L at resonance = 9.5 uH
- Tank C at resonance = 200 pF
- Loading coil = 223 uH

Phil B


begin 666 PAMT_Orig_LoadCoil_470.gif

{the illustration was not preserved in our archives, sorry}

----------------------------------------------------------------

Subject: Re: AM TRANSMITTERS
From: "Phil B" 
Newsgroups: alt.binaries.pictures.radio
Date: Sun, 15 Dec 2002 2:08 AM
Message-ID: 

"Joe Bento"  wrote:

> I don't claim to understand spice models - something I should probably
> learn.  But I know what an AM envelope should look like, and your
> models with graphs open a lot of areas to investigate with this
> transmitter.  It appears capable of achieving near 100% modulation,
> and some of you models seem to show it can do so cleanly.

Joe,

Yes indeed, this transmitter can do 100% modulation, but not without a
simple bias change on the differential output transistors.  This is
something I discovered in doing the simulations.  My posted simulations
showed something like 60% or 70% modulation.  I found that the
differential output transistors overloaded above this level.  Since I
was posting on the effect of the 470 ohm tank resistor, I didn't want to
complicate matters by introducing distortion from an unrelated source.
I am working on the optimum bias resistor change and will post when I am
satisfied.

BTW, when I was a kid I had a battery operated Allied 20:1 electronics
kit that included a simple AM transmitter circuit. It used a single
metal-diaphragm headphone as the mic.  I asked my father to talk while I
listened on the car radio about 30 ft. away.  I was astonished at the
clarity.  The antenna was about 2 ft. of wire hanging off the board and
there was no real ground because he was holding it in his hands.  Go
figure.

Phil B

----------------------------------------------------------------

Subject: Changes to Wentzel "Personal Radio Station" At 100% Mod
From: "Phil B" 
Newsgroups: alt.binaries.pictures.radio
Date: Mon, 16 Dec 2002 1:30 AM
Message-ID: <[email protected]>

If you choose to run this transmitter into a short antenna without a
loading coil, it will distort at 100% modulation even with precisely
tuned L/C tank.  If you use a precisely tuned loading coil and you
precisely tune the L/C tank, then it works fine at 100% modulation.

I have found with my spice modeling that it is very important to
precisely tune the L/C tank and the loading coil to get a good waveform
(at any modulation level).  I would think that builders may choose to
exclude the loading coil to reduce complexity.  Without the loading
coil, the RF output stage exceeds its bias levels at 100% modulation.

You may want to make the following bias resistor modifications to
eliminate the problem.  These changes simply improve the margins and
will have no effect when a loading coil is used, but will eliminate
distortion when a loading coil is not used.

The schematic to reference is at:
 http://www.wenzel.com/pdffiles/amxmit.pdf

Changes:
1. Change the resistors from base to ground on the two differential RF
output transistors from 1.8k to 1.0k.
2. Change the audio amp emitter-to-ground resistor from 100 ohms to 82
ohms.
3. Change the audio amp base-to-ground resistor from 2.2k to 1.8k.

As a side effect there is a minor added benefit: the audio gain is
increased slightly.  You will now reach 100% modulation at 1.6 v p-p
input instead of 2.0 v p-p.

----------------------------------------------------------------

Subject: Re: Changes to Wentzel "Personal Radio Station" At 100% Mod
From: "Phil B" 
Newsgroups: rec.antiques.radio+phono
Date: Tue, Dec 17, 2002 12:20 AM
Message-ID: 

"John Byrns" wrote:
> Am I correct in assuming that the 470 Ohm resistor was still in the
> circuit for the simulations of the modified circuit?  If it is, the
real
> trick would be getting clean modulation without the 470 Ohm resistor,
and
> no antenna connected.
>
> A question whose answer should be available from your simulation, with
the
> loading coil in place, and everything properly tuned up, how much RF
power
> is going into the 470 Ohm resistor, and how much into the combination
of
> antenna system radiation resistance and ground resistance.  In other
> words, what is the RF Voltage at these two points?

Yes, this still includes the 470 ohm resistor.  Without the resistor, I
can't get an undistorted waveform under any tuning conditions, with or
without the loading coil, and with or without my bias changes.  I only
suggest the bias changes for anyone that chooses to run without the
loading coil.  The loading coil increases the p-p voltage across the
antenna radiation resistance by more than 10 times, but I have found
that it must be precisely tuned to the antenna capacitance or the
waveform becomes distorted.  The need to tune the loading coil precisely
is a PTIA as far as I'm concerned.  If the range is adequate without it,
I would be in favor of eliminating it and possibly going with a somewhat
longer antenna instead. Tuning the L/C tank is much less critical for
undistorted output because the Q is lower than the Q of the loading
coil.

For the simulations I just ran for Bill M (using the 2SC1815
transistors) I have the following readings for the modified version:

 WITHOUT the loading coil:
- Collector of output transistor:  20v p-p, centered at 15 v (this is
the rf across the 470 ohm resistor
- At antenna terminal after .1 uf blocking cap: 20v p-p centered at 0 v.
- At top of series connected .1 ohm radiation resistance and 30 ohm
ground resistance: .320 v p-p. (this point is after the 40pf antenna
capacitance).
- Across the .1 ohm radiation resistance: 530 microvolts p-p

And WITH the loading coil:
- Collector of output transistor:  17v p-p, centered at 15 v (this is
the rf across the 470 ohm resistor
- At antenna terminal after .1 uf blocking cap: 17v p-p centered at 0 v.
- At top of series connected .1 ohm radiation resistance and 30 ohm
ground resistance: 1.9 v p-p. (this point is after the 40pf antenna
capacitance).
- Across the .1 ohm radiation resistance: 6200 microvolts p-p.

Phil B

----------------------------------------------------------------

Subject: Re: Changes to Wentzel "Personal Radio Station" At 100% Mod
From: [email protected] (John Byrns)
Newsgroups: rec.antiques.radio+phono
Date: Tue, Dec 17, 2002 12:49 PM
Message-ID: 

In article , "Phil B"
 wrote:

> And WITH the loading coil:
> - Collector of output transistor:  17v p-p, centered at 15 v (this is
> the rf across the 470 ohm resistor
> - At antenna terminal after .1 uf blocking cap: 17v p-p centered at 0 v.
> - At top of series connected .1 ohm radiation resistance and 30 ohm
> ground resistance: 1.9 v p-p. (this point is after the 40pf antenna
> capacitance).
> - Across the .1 ohm radiation resistance: 6200 microvolts p-p.


With the loading coil, if I am operating my calculator correctly, this
means 76.9 mW of RF power is dissipated in the 470 Ohm swamping resistor,
and 14.6 mW makes it to the actual load composed of the antenna radiation
resistance, and the ground system resistance.  This seems to confirm my
initial suspicion that most of the actual RF power is dissipated in the
470 Ohm resistor.  With a class C output stage, of the same power output,
we would have more than six times as much RF power going to the
antenna/ground system.


Regards,

John Byrns


Surf my web pages at,  http://www.enteract.com/~jbyrns/index.html

Hosted by www.Geocities.ws

1