1. I've been digging around in the receiver of my Yaesu FT-817ND
transceiver and simulating with LTspice the AGC and AM detectors to
investigate possible distortions due to non-linear loading of the IF
feed they share with the SSB/CW product detector when I chanced upon
an unexpected behaviour: the envelope of a CW IF signal at the input
of the product detector is amplitude-modulated by the beat frequency
output. The modulation is asymmetrical and distorted and is perhaps
only 5-10% deep but its fundamental component is clearly the beat
frequency. I presume that the output is therefore distorted to the
extent that the output amplitude at any time is somewhat proportional
to the amplitude of the input envelope. Is that a reasonable
supposition?

2. I suspect the cause is a dynamic load impedance of the product
detector being fed by a high impedance IF source - a 100pF coupling
capacitor (approx 3600Xc at 455kHz) preceded by an IF buffer amp with
at least 1000R source impedance. The product detector is NJM2594,
stated in the datasheet to have a 600R input resistance and schematics
for test setups showing 50R sources. Is that a reasonable explanation?
Are there any other probable causes?

3. If the hypothesis in 2 is correct, a probable solution would be to
lower the source impedance and that of the coupling capacitor to
something on the order of 50 Ohms. Is that a reasonable conclusion?
Are there any other possible solutions?

I've posted an article to the Files section of the FT817 Yahoo!Group:
http://groups.yahoo.com/group/FT817/...ulation_of_IF/

73, Tom VE3MEO

On Nov 15, 8:13 pm, Tom wrote:
1. I've been digging around in the receiver of my Yaesu FT-817ND
transceiver and simulating with LTspice the AGC and AM detectors to
investigate possible distortions due to non-linear loading of the IF
feed they share with the SSB/CW product detector when I chanced upon
an unexpected behaviour: the envelope of a CW IF signal at the input
of the product detector is amplitude-modulated by the beat frequency
output. The modulation is asymmetrical and distorted and is perhaps
only 5-10% deep but its fundamental component is clearly the beat
frequency. I presume that the output is therefore distorted to the
extent that the output amplitude at any time is somewhat proportional
to the amplitude of the input envelope. Is that a reasonable
supposition?

2. I suspect the cause is a dynamic load impedance of the product
detector being fed by a high impedance IF source - a 100pF coupling
capacitor (approx 3600Xc at 455kHz) preceded by an IF buffer amp with
at least 1000R source impedance. The product detector is NJM2594,
stated in the datasheet to have a 600R input resistance and schematics
for test setups showing 50R sources. Is that a reasonable explanation?
Are there any other probable causes?

3. If the hypothesis in 2 is correct, a probable solution would be to
lower the source impedance and that of the coupling capacitor to
something on the order of 50 Ohms. Is that a reasonable conclusion?
Are there any other possible solutions?

I've posted an article to the Files section of the FT817 Yahoo!Group:http://groups.yahoo.com/group/FT817/...rtion%20-%20Re...

73, Tom VE3MEO

Interesting that your posting should come just at this time. Perhaps
a week ago I was looking at a Spice simulation of a mixer to
baseband: really just a product detector. I noticed the same effect,
and concluded that it is indeed due to the change in effective load
impedance.

However, I do not believe that makes for distortion, directly. If the
source is linear, the loading that depends on the relative phase
between the LO and the IF signals does not cause distortion, just
modulation. On the other hand, if the source impedance looking back
into the IF is not constant with varying load, there will be
distortion.

If my analysis is correct, then there is not a need to prevent the
modulation of the IF output amplitude, at least not for the reason of
preventing distortion. A way to check this would be to put in two
signals and check the output for intermodulation products, or to just
check the sinusoidal output (appropriately filtered to remove high
frequency components) for harmonic distortion.

I'd be interested to know what more you learn about this. I may do
some playing with it myself, since the answer may matter quite a lot
to something I've just started working on.

Cheers,
Tom
@ K7ITM

On Nov 16, 12:28 am, K7ITM wrote:
On Nov 15, 8:13 pm, Tom wrote:





1. I've been digging around in the receiver of my Yaesu FT-817ND
transceiver and simulating with LTspice the AGC and AM detectors to
investigate possible distortions due to non-linear loading of the IF
feed they share with the SSB/CW product detector when I chanced upon
an unexpected behaviour: the envelope of a CW IF signal at the input
of the product detector is amplitude-modulated by the beat frequency
output. The modulation is asymmetrical and distorted and is perhaps
only 5-10% deep but its fundamental component is clearly the beat
frequency. I presume that the output is therefore distorted to the
extent that the output amplitude at any time is somewhat proportional
to the amplitude of the input envelope. Is that a reasonable
supposition?

2. I suspect the cause is a dynamic load impedance of the product
detector being fed by a high impedance IF source - a 100pF coupling
capacitor (approx 3600Xc at 455kHz) preceded by an IF buffer amp with
at least 1000R source impedance. The product detector is NJM2594,
stated in the datasheet to have a 600R input resistance and schematics
for test setups showing 50R sources. Is that a reasonable explanation?
Are there any other probable causes?

3. If the hypothesis in 2 is correct, a probable solution would be to
lower the source impedance and that of the coupling capacitor to
something on the order of 50 Ohms. Is that a reasonable conclusion?
Are there any other possible solutions?

I've posted an article to the Files section of the FT817 Yahoo!Group:http://groups.yahoo.com/group/FT817/...rtion%20-%20Re...

73, Tom VE3MEO

Interesting that your posting should come just at this time. Perhaps
a week ago I was looking at a Spice simulation of a mixer to
baseband: really just a product detector. I noticed the same effect,
and concluded that it is indeed due to the change in effective load
impedance.

However, I do not believe that makes for distortion, directly. If the
source is linear, the loading that depends on the relative phase
between the LO and the IF signals does not cause distortion, just
modulation. On the other hand, if the source impedance looking back
into the IF is not constant with varying load, there will be
distortion.

If my analysis is correct, then there is not a need to prevent the
modulation of the IF output amplitude, at least not for the reason of
preventing distortion. A way to check this would be to put in two
signals and check the output for intermodulation products, or to just
check the sinusoidal output (appropriately filtered to remove high
frequency components) for harmonic distortion.

I'd be interested to know what more you learn about this. I may do
some playing with it myself, since the answer may matter quite a lot
to something I've just started working on.

Cheers,
Tom
@ K7ITM- Hide quoted text -

- Show quoted text -

Interesting! I could find no Spice model for the NJM2594 and, as I was
primarily concerned with the always connected AM and AGC detector
diodes and their distortion of the IF envelope as seen by the product
detector, my LTspice simulation merely uses the special Modulate
function as an ideal p.d. It has effectively an infinite input
impedance so I do not see the beat modulation in the simulation.

I agree with your point about testing; sadly, I lack the equipment to
do so. I have one rather poor RF source - an old B+K function gennie -
and my notebook computer for audio analysis. My old Eico tube
oscilloscope does not readily show any distortion of the beat output -
it may not be easily discerned visually anyway without a reference
sine wave superposed or subtracted.

I'm not up-to-snuff on the math of this but I'm inclined to think that
amplitude modulation of the input envelope at the beat frequency,
however caused, should result in a distortion of the output beat
frequency signal that would not be present if the envelope was a
constant. Amplitude modulation at any other frequency would result in
a second beat note in the LPF output. I'll have to noodle about this
some more and see if I can come up with some test.

73, Tom VE3MEO

On Nov 16, 7:26 am, Tom wrote:
On Nov 16, 12:28 am, K7ITM wrote:



On Nov 15, 8:13 pm, Tom wrote:

1. I've been digging around in the receiver of my Yaesu FT-817ND
transceiver and simulating with LTspice the AGC and AM detectors to
investigate possible distortions due to non-linear loading of the IF
feed they share with the SSB/CW product detector when I chanced upon
an unexpected behaviour: the envelope of a CW IF signal at the input
of the product detector is amplitude-modulated by the beat frequency
output. The modulation is asymmetrical and distorted and is perhaps
only 5-10% deep but its fundamental component is clearly the beat
frequency. I presume that the output is therefore distorted to the
extent that the output amplitude at any time is somewhat proportional
to the amplitude of the input envelope. Is that a reasonable
supposition?

2. I suspect the cause is a dynamic load impedance of the product
detector being fed by a high impedance IF source - a 100pF coupling
capacitor (approx 3600Xc at 455kHz) preceded by an IF buffer amp with
at least 1000R source impedance. The product detector is NJM2594,
stated in the datasheet to have a 600R input resistance and schematics
for test setups showing 50R sources. Is that a reasonable explanation?
Are there any other probable causes?

3. If the hypothesis in 2 is correct, a probable solution would be to
lower the source impedance and that of the coupling capacitor to
something on the order of 50 Ohms. Is that a reasonable conclusion?
Are there any other possible solutions?

I've posted an article to the Files section of the FT817 Yahoo!Group:http://groups.yahoo.com/group/FT817/...rtion%20-%20Re...

73, Tom VE3MEO

Interesting that your posting should come just at this time. Perhaps
a week ago I was looking at a Spice simulation of a mixer to
baseband: really just a product detector. I noticed the same effect,
and concluded that it is indeed due to the change in effective load
impedance.

However, I do not believe that makes for distortion, directly. If the
source is linear, the loading that depends on the relative phase
between the LO and the IF signals does not cause distortion, just
modulation. On the other hand, if the source impedance looking back
into the IF is not constant with varying load, there will be
distortion.

If my analysis is correct, then there is not a need to prevent the
modulation of the IF output amplitude, at least not for the reason of
preventing distortion. A way to check this would be to put in two
signals and check the output for intermodulation products, or to just
check the sinusoidal output (appropriately filtered to remove high
frequency components) for harmonic distortion.

I'd be interested to know what more you learn about this. I may do
some playing with it myself, since the answer may matter quite a lot
to something I've just started working on.

Cheers,
Tom
@ K7ITM- Hide quoted text -

- Show quoted text -

Interesting! I could find no Spice model for the NJM2594 and, as I was
primarily concerned with the always connected AM and AGC detector
diodes and their distortion of the IF envelope as seen by the product
detector, my LTspice simulation merely uses the special Modulate
function as an ideal p.d. It has effectively an infinite input
impedance so I do not see the beat modulation in the simulation.

I agree with your point about testing; sadly, I lack the equipment to
do so. I have one rather poor RF source - an old B+K function gennie -
and my notebook computer for audio analysis. My old Eico tube
oscilloscope does not readily show any distortion of the beat output -
it may not be easily discerned visually anyway without a reference
sine wave superposed or subtracted.

I'm not up-to-snuff on the math of this but I'm inclined to think that
amplitude modulation of the input envelope at the beat frequency,
however caused, should result in a distortion of the output beat
frequency signal that would not be present if the envelope was a
constant. Amplitude modulation at any other frequency would result in
a second beat note in the LPF output. I'll have to noodle about this
some more and see if I can come up with some test.

73, Tom VE3MEO

Hi Tom,

Well, you can simulate it and see if you get distortion in the
output. You can create a "product detector" using just Spice
switches, switched at the LO frequency. My model, which I haven't
gotten back to yet since seeing your orig. posting, is of a Tayloe-
type mixer, but I would not be at all surprised to see the same effect
with other configurations. In LTSpice you can do an FFT of the output
results to look for distortion products, but if you want to look at
really low distortion, you need to subtract out the fundamental(s);
I've seen a limit of around 80-100dB dynamic range (as I recall).

I do have equipment to test for distortion. I may be able to do some
of that this afternoon. I'll let you know what I find.

Cheers,
Tom

On Nov 16, 12:23 pm, K7ITM wrote:
On Nov 16, 7:26 am, Tom wrote:





On Nov 16, 12:28 am, K7ITM wrote:

On Nov 15, 8:13 pm, Tom wrote:

1. I've been digging around in the receiver of my Yaesu FT-817ND
transceiver and simulating with LTspice the AGC and AM detectors to
investigate possible distortions due to non-linear loading of the IF
feed they share with the SSB/CW product detector when I chanced upon
an unexpected behaviour: the envelope of a CW IF signal at the input
of the product detector is amplitude-modulated by the beat frequency
output. The modulation is asymmetrical and distorted and is perhaps
only 5-10% deep but its fundamental component is clearly the beat
frequency. I presume that the output is therefore distorted to the
extent that the output amplitude at any time is somewhat proportional
to the amplitude of the input envelope. Is that a reasonable
supposition?

2. I suspect the cause is a dynamic load impedance of the product
detector being fed by a high impedance IF source - a 100pF coupling
capacitor (approx 3600Xc at 455kHz) preceded by an IF buffer amp with
at least 1000R source impedance. The product detector is NJM2594,
stated in the datasheet to have a 600R input resistance and schematics
for test setups showing 50R sources. Is that a reasonable explanation?
Are there any other probable causes?

3. If the hypothesis in 2 is correct, a probable solution would be to
lower the source impedance and that of the coupling capacitor to
something on the order of 50 Ohms. Is that a reasonable conclusion?
Are there any other possible solutions?

I've posted an article to the Files section of the FT817 Yahoo!Group:http://groups.yahoo.com/group/FT817/...rtion%20-%20Re...

73, Tom VE3MEO

Interesting that your posting should come just at this time. Perhaps
a week ago I was looking at a Spice simulation of a mixer to
baseband: really just a product detector. I noticed the same effect,
and concluded that it is indeed due to the change in effective load
impedance.

However, I do not believe that makes for distortion, directly. If the
source is linear, the loading that depends on the relative phase
between the LO and the IF signals does not cause distortion, just
modulation. On the other hand, if the source impedance looking back
into the IF is not constant with varying load, there will be
distortion.

If my analysis is correct, then there is not a need to prevent the
modulation of the IF output amplitude, at least not for the reason of
preventing distortion. A way to check this would be to put in two
signals and check the output for intermodulation products, or to just
check the sinusoidal output (appropriately filtered to remove high
frequency components) for harmonic distortion.

I'd be interested to know what more you learn about this. I may do
some playing with it myself, since the answer may matter quite a lot
to something I've just started working on.

Cheers,
Tom
@ K7ITM- Hide quoted text -

- Show quoted text -

Interesting! I could find no Spice model for the NJM2594 and, as I was
primarily concerned with the always connected AM and AGC detector
diodes and their distortion of the IF envelope as seen by the product
detector, my LTspice simulation merely uses the special Modulate
function as an ideal p.d. It has effectively an infinite input
impedance so I do not see the beat modulation in the simulation.

I agree with your point about testing; sadly, I lack the equipment to
do so. I have one rather poor RF source - an old B+K function gennie -
and my notebook computer for audio analysis. My old Eico tube
oscilloscope does not readily show any distortion of the beat output -
it may not be easily discerned visually anyway without a reference
sine wave superposed or subtracted.

I'm not up-to-snuff on the math of this but I'm inclined to think that
amplitude modulation of the input envelope at the beat frequency,
however caused, should result in a distortion of the output beat
frequency signal that would not be present if the envelope was a
constant. Amplitude modulation at any other frequency would result in
a second beat note in the LPF output. I'll have to noodle about this
some more and see if I can come up with some test.

73, Tom VE3MEO

Hi Tom,

Well, you can simulate it and see if you get distortion in the
output. You can create a "product detector" using just Spice
switches, switched at the LO frequency. My model, which I haven't
gotten back to yet since seeing your orig. posting, is of a Tayloe-
type mixer, but I would not be at all surprised to see the same effect
with other configurations. In LTSpice you can do an FFT of the output
results to look for distortion products, but if you want to look at
really low distortion, you need to subtract out the fundamental(s);
I've seen a limit of around 80-100dB dynamic range (as I recall).

I do have equipment to test for distortion. I may be able to do some
of that this afternoon. I'll let you know what I find.

Cheers,
Tom- Hide quoted text -

- Show quoted text -

With a LTspice simulation of a MC1496, I've confirmed that the
impedance between the CW source and the product detector input does
have an effect on the distortion of the audio beat note output as well
as distorting the envelope of the input signal. Now I cannot vouch for
the accuracy of the model - it's one that I found in a message on a
forum but the envelope modulation of the CW carrier can look very
similar to what I see in the FT-817 with the appropriate choice of
amplitudes and series resistance. I stepped the source amplitude
through the distortion minimum of the audio output for three levels of
source resistance: 1K, 51 and 1 Ohm and obtained THD of 7.7%, 5.3% and
5% respectively. All these were done with the BFO source resistance at
51 Ohm; at 1 Ohm for both the signal and BFO, the output THD dropped
to about 3.5%; distortion of the BFO envelope by dynamic loading is
visible at 51 Ohms.

Now these levels of THD are not especially impressive which is why I'm
questioning the model. The circuit is that for a product detector
straight off the ON-semi datasheet. I think the NJM2594 bears some
resemblance to the MC1496 - both switching differential amplifiers
rather than diodes but maybe offers a higher input impedance.

Aaagh - I have to revise this. While writing this, I discovered that
the product detector schematic in the Application Note AN531/D has a
couple of extra components that make a huge difference in distortion
and output level - now getting as low as 0.1% for that last example!
That's where it should be. I'll re-run these sims and revise the
results (and maybe the conclusion!).

Tom VE3MEO

On Nov 22, 11:57 pm, Tom wrote:
On Nov 16, 12:23 pm, K7ITM wrote:





On Nov 16, 7:26 am, Tom wrote:

On Nov 16, 12:28 am, K7ITM wrote:

On Nov 15, 8:13 pm, Tom wrote:

1. I've been digging around in the receiver of my Yaesu FT-817ND
transceiver and simulating with LTspice the AGC and AM detectors to
investigate possible distortions due to non-linear loading of the IF
feed they share with the SSB/CW product detector when I chanced upon
an unexpected behaviour: the envelope of a CW IF signal at the input
of the product detector is amplitude-modulated by the beat frequency
output. The modulation is asymmetrical and distorted and is perhaps
only 5-10% deep but its fundamental component is clearly the beat
frequency. I presume that the output is therefore distorted to the
extent that the output amplitude at any time is somewhat proportional
to the amplitude of the input envelope. Is that a reasonable
supposition?

2. I suspect the cause is a dynamic load impedance of the product
detector being fed by a high impedance IF source - a 100pF coupling
capacitor (approx 3600Xc at 455kHz) preceded by an IF buffer amp with
at least 1000R source impedance. The product detector is NJM2594,
stated in the datasheet to have a 600R input resistance and schematics
for test setups showing 50R sources. Is that a reasonable explanation?
Are there any other probable causes?

3. If the hypothesis in 2 is correct, a probable solution would be to
lower the source impedance and that of the coupling capacitor to
something on the order of 50 Ohms. Is that a reasonable conclusion?
Are there any other possible solutions?

I've posted an article to the Files section of the FT817 Yahoo!Group:http://groups.yahoo.com/group/FT817/...rtion%20-%20Re...

73, Tom VE3MEO

Interesting that your posting should come just at this time. Perhaps
a week ago I was looking at a Spice simulation of a mixer to
baseband: really just a product detector. I noticed the same effect,
and concluded that it is indeed due to the change in effective load
impedance.

However, I do not believe that makes for distortion, directly. If the
source is linear, the loading that depends on the relative phase
between the LO and the IF signals does not cause distortion, just
modulation. On the other hand, if the source impedance looking back
into the IF is not constant with varying load, there will be
distortion.

If my analysis is correct, then there is not a need to prevent the
modulation of the IF output amplitude, at least not for the reason of
preventing distortion. A way to check this would be to put in two
signals and check the output for intermodulation products, or to just
check the sinusoidal output (appropriately filtered to remove high
frequency components) for harmonic distortion.

I'd be interested to know what more you learn about this. I may do
some playing with it myself, since the answer may matter quite a lot
to something I've just started working on.

Cheers,
Tom
@ K7ITM- Hide quoted text -

- Show quoted text -

Interesting! I could find no Spice model for the NJM2594 and, as I was
primarily concerned with the always connected AM and AGC detector
diodes and their distortion of the IF envelope as seen by the product
detector, my LTspice simulation merely uses the special Modulate
function as an ideal p.d. It has effectively an infinite input
impedance so I do not see the beat modulation in the simulation.

I agree with your point about testing; sadly, I lack the equipment to
do so. I have one rather poor RF source - an old B+K function gennie -
and my notebook computer for audio analysis. My old Eico tube
oscilloscope does not readily show any distortion of the beat output -
it may not be easily discerned visually anyway without a reference
sine wave superposed or subtracted.

I'm not up-to-snuff on the math of this but I'm inclined to think that
amplitude modulation of the input envelope at the beat frequency,
however caused, should result in a distortion of the output beat
frequency signal that would not be present if the envelope was a
constant. Amplitude modulation at any other frequency would result in
a second beat note in the LPF output. I'll have to noodle about this
some more and see if I can come up with some test.

73, Tom VE3MEO

Hi Tom,

Well, you can simulate it and see if you get distortion in the
output. You can create a "product detector" using just Spice
switches, switched at the LO frequency. My model, which I haven't
gotten back to yet since seeing your orig. posting, is of a Tayloe-
type mixer, but I would not be at all surprised to see the same effect
with other configurations. In LTSpice you can do an FFT of the output
results to look for distortion products, but if you want to look at
really low distortion, you need to subtract out the fundamental(s);
I've seen a limit of around 80-100dB dynamic range (as I recall).

I do have equipment to test for distortion. I may be able to do some
of that this afternoon. I'll let you know what I find.

Cheers,
Tom- Hide quoted text -

- Show quoted text -

With a LTspice simulation of a MC1496, I've confirmed that the
impedance between the CW source and the product detector input does
have an effect on the distortion of the audio beat note output as well
as distorting the envelope of the input signal. Now I cannot vouch for
the accuracy of the model - it's one that I found in a message on a
forum but the envelope modulation of the CW carrier can look very
similar to what I see in the FT-817 with the appropriate choice of
amplitudes and series resistance. I stepped the source amplitude
through the distortion minimum of the audio output for three levels of
source resistance: 1K, 51 and 1 Ohm and obtained THD of 7.7%, 5.3% and
5% respectively. All these were done with the BFO source resistance at
51 Ohm; at 1 Ohm for both the signal and BFO, the output THD dropped
to about 3.5%; distortion of the BFO envelope by dynamic loading is
visible at 51 Ohms.

Now these levels of THD are not especially impressive which is why I'm
questioning the model. The circuit is that for a product detector
straight off the ON-semi datasheet. I think the NJM2594 bears some
resemblance to the MC1496 - both switching differential amplifiers
rather than diodes but maybe offers a higher input impedance.

Aaagh - I have to revise this. While writing this, I discovered that
the product detector schematic in the Application Note AN531/D has a
couple of extra components that make a huge difference in distortion
and output level - now getting as low as 0.1% for that last example!
That's where it should be. I'll re-run these sims and revise the
results (and maybe the conclusion!).

Tom VE3MEO- Hide quoted text -

- Show quoted text -

Major retraction! The appnote has one extra component and I missed one
on the datasheet when capturing to LTspice. With those in place, I see
no modulation of the input CW envelope by the product beat note and
the output distortion is consistently much lower than what I saw with
the erroneous schematic. Now that leaves me wondering if the circuit
around the NJM2594 in the FT-817ND is optimum or am I seeing a
simulation that represents an ideal that is unattainable in practice.

Tom VE3MEO

On Nov 15 2007, 11:13*pm, Tom wrote:
1. I've been digging around in the receiver of my Yaesu FT-817ND
transceiver and simulating with LTspice the AGC and AM detectors to
investigate possible distortions due to non-linear loading of the IF
feed they share with the SSB/CW product detector when I chanced upon
an unexpected behaviour: the envelope of a CW IF signal at the input
of the product detector is amplitude-modulated by the beat frequency
output. The modulation is asymmetrical and distorted and is perhaps
only 5-10% deep but its fundamental component is clearly the beat
frequency. I presume that the output is therefore distorted to the
extent that the output amplitude at any time is somewhat proportional
to the amplitude of the input envelope. Is that a reasonable
supposition?

2. I suspect the cause is a dynamic load impedance of the product
detector being fed by a high impedance IF source - a 100pF coupling
capacitor (approx 3600Xc at 455kHz) preceded by an IF buffer amp with
at least 1000R source impedance. The product detector is NJM2594,
stated in the datasheet to have a 600R input resistance and schematics
for test setups showing 50R sources. Is that a reasonable explanation?
Are there any other probable causes?

3. If the hypothesis in 2 is correct, a probable solution would be to
lower the source impedance and that of the coupling capacitor to
something on the order of 50 Ohms. Is that a reasonable conclusion?
Are there any other possible solutions?

I've posted an article to the Files section of the FT817 Yahoo!Group:http://groups.yahoo.com/group/FT817/...rtion%20-%20Re...

73, TomVE3MEO

I revised my FT-817ND so that the product detector is now fed off the
IF buffer's 220R emitter resistor (not the 1k collector) via a 0.1u
coupling capacitor. This lowers the source impedance seen by the
product detector and now I see no modulation of a CW IF input signal
by the pd's beat frequency output. I do see a trace of it on the input
from the BFO. Unfortunately, no distortion measurements on the audio
output but I believe, based on what I could simulate with LTspice,
that there is some improvement.

73, Tom VE3MEO