[Central-Electronics] 20A Sideband Rejection Problem Found

[Jim candela]

	Hi All,

	I know I've been boring you guys with my QRO escapades, but this last rig
has really been kicking my rear end. I don't want to admit how many hours
I've put into this one. I think I found the culprit tonight. The problem was
basically that as the RF power goes up using a single tone input (SSB Mode)
the carrier output from the tone starts showing modulation (distortion) that
gets progressively worse with increasing power. This 20A with a single EL-34
with 600 volts on the plate puts out 50+ watts carrier on 160 meters, and it
runs 12.5 watts AM 100% modulated, and the waveform is picture perfect.
Switch to SSB, null the carrier, and insert a clean 1 Khz sine wave into the
high level audio input, and advance the gain. The LSB carrier waveform is
pure DC up to 10 watts out, and then it goes south with increasing level.

	Looking at the audio going into the balanced modulator circuit does show
SOME distortion. Sure both the 6U8 triode, and the 12AT7 drivers all
contribute some distortion, but nothing like what I see at the RF output of
the rig (only on SSB). I went as far as adding a multi-pole negative
feedback circuit around the 6U8 triode to cut distortion, and Taylor the
frequency response some. Total waste of time. I started looking at the
balanced modulator diodes, and even swapped them with a known good set. No
change. Using a scope looking at the waveforms around the diodes (yes I know
the scope is detuning the circuits) showed some really ugly stuff that only
appears when the RF output shows distortion, and on AM it all looks good.

	Then it occurred to me. Many years ago a former owner of this rig replaced
the balanced modulator pots with 10 turn wire wound pots. They seem to work
great for nulling the carrier, so I never gave it a second thought until
today. The pots are bypassed end to end like stock with a .005uf ceramic
disc capacitor. As a test I bypassed one side of one pot end to wiper with a
,005 uf capacitor. Low and behold the distortion was reduced considerably,
but the null was upset a little. Still I could get to about 30 watts before
the distortion was nasty. It looks like it's the dam WW pots!

	So here is where I need help from the group. On the 20A the balanced
modulator runs at 9 Mhz. I suppose the inductance of a 1K wire wound pot is
going to behave differently than a 1K carbon pot. Can I just put two .005 uf
caps on each pot (wiper to each end), or is it going to require me going to
carbon pots? My head hurts trying to visualize the goings on with a phasing
balanced modulator circuit, but on AM only one side (2 diodes) gets audio,
and on SSB both sides get audio 90 degrees (or 270) to null the unwanted
sideband. If anybody knows the nuts & bolts theory of this circuit, please
tell me whether I can make the nice 10 turn wire wound pots work, or do I
need to go back to a single turn 1K carbon pot?

Regards,
Jim
WD5JKO

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.705 / Virus Database: 461 - Release Date: 6/12/2004