[R-1051] FIXED: Low output from 500cps synth

[David Wise]

The output of the R-1051 plain's 500cps synthesizer is a CW signal at approximately 7.1MHz .
The exact frequency represents the error in the 1-and-10KHz crystals, plus the
000/500/Vernier setting.  A mixer combines the 1-and-10 module's ~7.089MHz
error output with a locally-generated signal in the 11-12kHz range which is
the 500cps module's VCO divided by 10.)  The result goes through a narrow
7.1MHz crystal filter, and is then amplified and sent on its way to the 100kHz
module where it becomes part of the second RF Translator injection signal.

This injection signal is weak in my radio, so the 500cps module is up for scrutiny.

I found a "13W3" hybrid-d-sub cable that works just fine to extend the module
out of the six-pack.  It's the video monitor cable that Sun Microsystems used with
their popular line of workstations back in the 1980's, and used ones are
still abundant and fairly cheap.  I got mine for $12, which is about the price
of two individual coax contacts new.  (I wonder if I could harvest the contacts
from several cables to make other ones.)

The 2N Repair Manual is the only source I have for alignment procedures.
Does anybody have anything else?

I'm one of those techs who likes to use alignment as a diagnostic aid.
The manual says to peak A1T1 for maximum 7.1MHz output, and then adjust a
gain trim, A1R16, for 8mV output.  I couldn't get a peak, and the output was only
5mV even though some previous tech had already turned R16 to max.

Having previously noticed in the Spectrum Generator that its input coil A1L2 wouldn't,
and, as far as I could tell, couldn't peak except perhaps when used with the earliest
type of Frequency Standard, I wondered if the same was true here.  The schematic
clearly shows a cap across the winding, however, which I figured was there
in order to resonate.

Then I wondered if the low-impedance load at the destination was swamping the
resonance, so I tried again with the module powered up alone on the bench.
I backfed the output with my tracking generator, through a 1K resistor to raise
the Q.  I watched for a peak at the output using my spectrum analyzer and
active probe.  Nothing.

Disappointed, I decided I needed a break.  I started powering things off.  As luck
would have it, I powered down the module first.  And Voila!  A clean, sharp peak
appeared on the spectrum analyzer screen.  I turned the supply back on.
Flat line.  Just for fun, I slowly turned the voltage down.  The peak reappeared
at about 0.3V .  The germanium diode drop.

I powered back up and probed output transistor A1Q3 with a DMM.  It's wired
as a common-emitter amplifier, with bias stabilization via 1.2K emitter resistor R13
and 10K/15K base voltage divider R14 and R15.  Its base and emitter are supposed
to be around 11V, about half of the 20V Vcc.  (It's a PNP - a 2N1142 - "hanging from
the ceiling", so the collector is 0V.)

I didn't see 11V; the emitter was 0.5V, the base only 0.1V .  A B-C short.
I replaced it with a 2N3906, and the output sprang to life with plenty of amplitude
and a tunable peak.

I could have called it good at this point, but the last guy did that, and I wanted
to do better.  I was still worried about the amount the signal varied as I turned
the vernier dial - 8dB.  Since the 7.089 was constant, it had to be the 11kHz.
It was constant coming out of the divider, at A1TP6.  From there, it is amplified
(Q5), filtered (T2/C17) to removed harmonics from the divide-by-ten square wave,
buffered (Q4), and applied to mixer Q1.  At the buffer, TP3 showed a clear
variation.  I wondered if the filter Q was too high.  Sure enough, swamping
resistor R22 was out of tolerance.

But only a little.  I used it as an excuse for experimentation.  I wondered how
much swamping was acceptable before the filter became ineffective.  I kept
the spectrum analyzer on TP3, and saw third harmonic about 40dB down,
fifth about 60, and seventh just barely measurable.  Watching the third
harmonic, I started loading the T2 primary.  At 4.7K, the harmonic had
only risen 1dbc, and the amplitude now varied only 4dB instead of 8.
I decided to go with this.  I also noticed that the tuning was a bit high due to
tolerances, so I strung a small cap across my new resistor to pull it down.
This might be overkill, but I won't have to worry about my 7.1 .  If there's still
trouble, it's somewhere else.

Dave Wise