[Central-Electronics] 20A Sideband Rejection Problem FIXED

[George Maier]

Hi Jim:

The objective is to do exactly what you did, get into a QSO with the SSB
Hi-Fi gang, but with some CE gear that's been tweaked and stabilized to the
point of being transparent.  If you look at the latest high end Japanese
radios, they essentially use phasing SSB generation with a DSP to control
the filter bandwidth and phase relationships.  I'm convinced that DSP based
phase shift and audio filter adapters could be built for the 100/200V, but
have not pursued it to any extent.  I tracked the VFO in my 200V against a
high resolution, oven controlled freq counter, and it was stable nearly from
turn on, but the mixers bounce around a bit, and need some attention.  

 

A more fun challenge would be to get a 20A to that same stage, so your
comments are right on target.

 

Vy 73,  

George - K1GXT

 

  _____  

From: Jim candela [mailto:jcandela at prodigy.net] 
Sent: Wednesday, June 23, 2004 8:43 AM
To: George at MaierGroup.com
Cc: central-electronics at mailman.qth.net
Subject: RE: [Central-Electronics] 20A Sideband Rejection Problem FIXED

 

resend

 

 

George,

 

            I'm not sure if you refer to the SSB Hi-Fi crowd, or to some of
the AM'ers who run broadcast audio. The CE equipment is capable of doing
both. I once broke into a Hi-Fi SSB group on 40meters with my 20a. I bout
got ran off the band! I drifted up about 100 hz! I got NO slack from them
boys. To them 20 hz drift is out of the band. I'm currently gathering parts
to build a WA1FFL DDS VFO for my 20A. Time to retire the BC-458 (especially
on 40m where the output is tripled).

 

            As for AM HI-FI, I been playing with that a little too with 20A
QRO #3. I had to re-do the circuit driving the 6E5 tuning eye tube (because
the switch from 12BH7 to 6S4 deleted this feature), and while at it I
thought why not make the carrier rectifier linear, and use that as a
negative feedback source when running AM? Doing so turned out to be dirt
simple, and early tests show it to work extremely well. The 6BA7 mixer plate
tuning is critical (L-C tank Q too dam high on 80 meters), and tuning alters
the linearity of that stage. The L-C tank was originally shunted with either
a 10K or 22K resistor to lower the Q. This resistor also lowers the gain,
and for QRO I need more gain to drive the final. So clip the resistor, get
the gain, get some tendency towards instability, and distortion. I found
that the top of the coil is bypassed to ground with a .005 uf capacitor.
This bypass is critical. Use a silver mica (early 20a's) versus a ceramic
(later 20a's) and the gain is much higher on 80 meters. Bypass instead with
say a .047 polypropylene, and gain is even higher on 80 m, but severe
distortion sets in that is non linear with signal level (seems to be
regeneration). I found that using a .01 uf poly in parallel with a series
R-C network consisting of a .047 poly, and three 2 ohm .5 watt resistors in
parallel does the trick. I get the gain, tuning is broadened, and distortion
is down. Still the Mixer-driver-final combo does show some distortion. Sine
wave drive shows a slight widening of the waveform, and slight compression
at the peaks (10 watts carrier output, or 40 watts PEP). Adding the negative
feedback from the 6E5 tuning eye drive circuit to the AM modulation
transformer output (secondary winding feeding the balanced modulator) boosts
the compressed peaks, and sharpens the waveform. If I overlay the 12At7
modulator tube grid waveform with the RF modulated envelope (dual trace
scope) I see the distortion. Close the feedback loop, and the waveforms are
identical! This is for AM only however. The feedback amount is less than 6
db. Still plenty of gain left when I use a D-104 Mic (sorry Bill). Listening
to the difference with my monitor reveals the difference, and the feedback
definitely cleans up what already sounded pretty good. With a sine wave
drive it sounds clean up to about 80%, and then I hear harmonic energy that
worsens with increasing modulation level. Close the feedback loop, and it
goes away (can't hear it) up to 100% modulation. I wish I had a distortion
meter. The negative FB can be switched in for AM only. There is a unused set
of contacts on the mode switch.

 

            Well off to the salt mines.then off to Boston on business for
two weeks starting this Sunday.

 

Regards,

Jim

 

-----Original Message-----
From: George Maier [mailto:George at MaierGroup.com]
Sent: Tuesday, June 22, 2004 11:37 AM
To: 'Jim Candela'
Subject: RE: [Central-Electronics] 20A Sideband Rejection Problem FIXED

 

Thanks for the words of wisdom, Jim.  I'm always fascinated by the amount of
detail in your work.  

It's pretty doubtful that I'll be going QRO with my 10A/B or 20A's, although
it does touch on an area of interest.  

A friend and I have been thinking about clean, undistorted SSB with carrier,
just to confound and confuse the HI-FI crowd, without telling them what
we're running !!!!    That in itself may be an issue from what I've read
about your experiments.  The good news is that I also have a 200V and  600L
that gets pretty good reports on DSB with carrier, when I'm really careful.
Some good EQ might do the trick, but we've been thinking about some modern
technology phase shift networks as the real way to do it.   

 

My real (slow moving) AM project is another Leo WRL radio, a Globe King
500B, but that's a horse of another color !!!

 

73,

George

 

.  

 

  _____  

From: Jim Candela [mailto:jcandela at prodigy.net] 
Sent: Tuesday, June 22, 2004 9:06 AM
To: George at MaierGroup.com
Subject: RE: [Central-Electronics] 20A Sideband Rejection Problem FIXED

 

George,

 

    I don't know too much about the EL-38 except it is rare, and was another
one of Leo Myerson's tube picks for WRL. The AX9909, 5514, V70D are a few
others he chose. Nice tubes, but try to find them today.

  If you attempt to QRO the 10A consider a few things I've learned the hard
way. The 20A as designed offers 10 volts peak drive to the 6AG7's with
little more in reserve. The Cgp of the 6AG7 is a puny (~ 0.06 pf) amount so
the miller effect on the driver stage is small. The EL-34 has about 1.1 pf
Cgp per tube! Most power tubes need 30 to 50 volts peak drive (Class Ab1),
and the 20A cannot do much above 15 without serious modifications to the
driver. A pair of 7591's only need 19 volts peak drive for 50 watts output,
and they only have 0.25pg Cgp per tube. To use the higher Cgp tubes, you
MUST neutralize the final. A balanced circuit helps, but the input
capacitance as seen by the driver is still high, so the driver must be
beefed up considerably. That is why I have went to a 6S4 driver instead of
1/2 12BH7. The 6DQ6 needed nearly 40 volts peak drive, so that got axed. The
EL-34 can run with about 25-30 volts peak drive, and only costs about 10
bucks apiece (Sovtek). 

    If you go down the QRO road with your 10A, please keep us posted on your
progress.

Thanks for the comments,

Regards,

Jim

WD5JKO