Greetings All! I'm a long time lurker, but a first time poster.

I have been busy building a homebrew 20m SSB transceiver.

So far things are working fairly well, however, before I move on
and Finish the final PA stage, I would like to solve a couple of
mysteries with the driver stages leading up to the PA.

This is the project that I am copying: (Thanks to Ashar Farhan and to
W8DIZ, for all of their help and patience with my questions!)

http://farhan.net.co.nr/xcvr1.html

I have made some component substitutions, added varactor tuning to the
VFO, and I am using cascoded JFETs in the mixer instead of a single
JFET.

My problem at the moment lies with the tx linear amp chain.

So far I have built up the tx chain to the point up to the final PA
stage. I am using MRF517s instead of 2N3866s. I built the linear amp
driver for this rig on a separate board not wired to the rest of the
rig (so far). This consists of a chain of three nearly identical
bipolar feedback amp stages. For my tests, I attached a 50 ohm load to
the output of the third stage.

There are two strange things I have noticed about the transmit chain:

1) If the input to the first stage is shorted to ground through the
coupling capacitor (grounding at RF, but not DC), then the amplifier
will oscillate at about 45MHz. With a 50 ohm resistor grounding the
input to the amp, it does *not* oscillate. (The amp has a strong peak
in its gain around 45MHz, having twice the peak V as it does at 14MHz,
could this peak be related to the parasitic resonance making it
oscillate under the above conditions?) ((keep in mind that there are
no tuned circuits (by design anyway!) in the amp and its not attached to
any tuned outputs or filters yet.)) I've added 1K resistors across the
bifilar transformers to kill parasitic oscillations, which they do an
admirable job of with a normal impedance at the input. I've also used
all monolithic ceramic capacitors for bypass, coupling, and emitter
degeneration.

2) After I attached the amp chain to the rest of the rig and adjusted
the mic level so that no clipping or over driving of the output was
visible, I now have a residual carrier of about .5v peak when not
speaking. The signal level at the input to the amp is too low to measure
when not speaking. I substituted 1N5711s for the 1N4148 diodes in the
mixer, but I believe that is suppose to make it better because of the
lower forward voltage drop and the faster speed of the 1N5711. The
maximum output before clipping starts is 6V peak with 60mV peak input.
Higher inputs drive it into distortion and clipping. Seems normal to me,
but what do I know?

This amp is laid out very cleanly and constructed well. I have already
eliminated parasitics from the amp when disconnected with no signal
driving it.

With #1, I am concerned about the oscillation with the input shorted to
ground because it seems related to the peak in the gain around the same
frequency, even though I do not plan on operating it with its input
shorted to ground, nor operating it anywhere near 45MHz.

Should I be concerned? If yes, some suggestions of things to try
would be very welcome.

With #2, The residual carrier, I wish to eliminate or severely reduce it
before I add the final PA section. I will appreciate any suggestions of
things to try here too. BFO frequency adjustment? Bad mixer design? Bad
mixer diodes?

I am not sure where to start looking for trouble on #2

All of your help and your suggestions will be most appreciated. I can
supply photos of my layout and supply further details on where my
implementation deviates from the original schematic of Farhan's rig,
upon request.

Thanks in advance!

de Dan KC0LUY

Hi, welcome! :-)

Something you can try:

1) place a ferrite bead on the base(s)

2) Toss a low value resistor (10 ohms or less) on the collector as close
as possible to the transistor.

Good luck!

--
Gregg
*It's probably useful, even if it can't be SPICE'd*
http://geek.scorpiorising.ca

#1:

Yes, the 45MHz peak with a 50 ohm source is almost certainly related to the
45MHz oscillation with the input grounded. You can try to fix it, or you
can decide not to worry about it. Shunting the transformers will certainly
kill the peak, but at the cost of some gain. If you're _really sure_ that
you'll always present that amplifier with a 50 ohm impedance you're probably
OK. If it was me I'd probably try hard to chase it down, but I'd go on the
air even if I couldn't entirely fix it.

#2:

Where are you measuring things? Is this the input and output of the mixer,
or somewhere else?

Did you adjust the carrier balance (that 100 ohm pot that the diodes connect
to after T5)? It should be diddled until your carrier is at a minimum. I
would be very tempted to put some resistance in to reference the audio to
ground, but I'm not sure if it would improve anything.

You may also want to match the forward drop of the diodes, but with the
carrier balance control such measures are probably superfluous.

Even what you have probably meets minimum standards (whatever they are), but
you probably want to fix it before you fly.

"Dan" wrote in message
om...
Greetings All! I'm a long time lurker, but a first time poster.

I have been busy building a homebrew 20m SSB transceiver.

So far things are working fairly well, however, before I move on
and Finish the final PA stage, I would like to solve a couple of
mysteries with the driver stages leading up to the PA.

This is the project that I am copying: (Thanks to Ashar Farhan and to
W8DIZ, for all of their help and patience with my questions!)

http://farhan.net.co.nr/xcvr1.html

I have made some component substitutions, added varactor tuning to the
VFO, and I am using cascoded JFETs in the mixer instead of a single
JFET.

My problem at the moment lies with the tx linear amp chain.

So far I have built up the tx chain to the point up to the final PA
stage. I am using MRF517s instead of 2N3866s. I built the linear amp
driver for this rig on a separate board not wired to the rest of the
rig (so far). This consists of a chain of three nearly identical
bipolar feedback amp stages. For my tests, I attached a 50 ohm load to
the output of the third stage.

There are two strange things I have noticed about the transmit chain:

1) If the input to the first stage is shorted to ground through the
coupling capacitor (grounding at RF, but not DC), then the amplifier
will oscillate at about 45MHz. With a 50 ohm resistor grounding the
input to the amp, it does *not* oscillate. (The amp has a strong peak
in its gain around 45MHz, having twice the peak V as it does at 14MHz,
could this peak be related to the parasitic resonance making it
oscillate under the above conditions?) ((keep in mind that there are
no tuned circuits (by design anyway!) in the amp and its not attached to
any tuned outputs or filters yet.)) I've added 1K resistors across the
bifilar transformers to kill parasitic oscillations, which they do an
admirable job of with a normal impedance at the input. I've also used
all monolithic ceramic capacitors for bypass, coupling, and emitter
degeneration.

2) After I attached the amp chain to the rest of the rig and adjusted
the mic level so that no clipping or over driving of the output was
visible, I now have a residual carrier of about .5v peak when not
speaking. The signal level at the input to the amp is too low to measure
when not speaking. I substituted 1N5711s for the 1N4148 diodes in the
mixer, but I believe that is suppose to make it better because of the
lower forward voltage drop and the faster speed of the 1N5711. The
maximum output before clipping starts is 6V peak with 60mV peak input.
Higher inputs drive it into distortion and clipping. Seems normal to me,
but what do I know?

This amp is laid out very cleanly and constructed well. I have already
eliminated parasitics from the amp when disconnected with no signal
driving it.

With #1, I am concerned about the oscillation with the input shorted to
ground because it seems related to the peak in the gain around the same
frequency, even though I do not plan on operating it with its input
shorted to ground, nor operating it anywhere near 45MHz.

Should I be concerned? If yes, some suggestions of things to try
would be very welcome.

With #2, The residual carrier, I wish to eliminate or severely reduce it
before I add the final PA section. I will appreciate any suggestions of
things to try here too. BFO frequency adjustment? Bad mixer design? Bad
mixer diodes?

I am not sure where to start looking for trouble on #2

All of your help and your suggestions will be most appreciated. I can
supply photos of my layout and supply further details on where my
implementation deviates from the original schematic of Farhan's rig,
upon request.

Thanks in advance!

de Dan KC0LUY