Perhaps I should add a cautionary risk-warning for the less experienced
who choose to energize the VFO during off-hours as described above:
During normal operation on mine, the current to the -10v zener set the
voltage at -10.1 volts. When I substituted an external source when the
transceiver was turned off, I selected a resistor to supply enough
current to read -9.9 volts at the zener....thus ensuring that the current
is not excessive. Accordingly, you should be sure to disable the external
current supply (and heater light) before you turn on the transceiver to
prevent BOTH currents energizing the zener. It would probably tolerate it
temporarily but if it were destroyed so as to "open up", the soaring
negative voltage might destroy the VFO transistors.
A wide variety of external power sources could be accommodated...AC or
DC, but if AC is used, it will require an external rectifier with the
Cathode toward the transformer, and Anode toward the internal zener. No
filtering required.
In my particular case, I had a unit labeled 13.5 vdc @ 1A.
On the auto brake/park bulb, I chose to solder to the two lead contacts
(ignoring the shell). This runs the Brake and Park filaments in series
and of course most of the illumination is the Park filament only. My
voltage across these measured 11.25 volts dc but of course it could
also be 11.25 ac. Although with my negative dc supply, I didn't need
the rectifier diode, but I used one anyway as a precaution against
accidentally reversing the dc supply. As noted above, this amount of
heat seems to simulate the normal operating heat quite well since I
experience virtually no warm-up time required for freqeuncy stability.
On accessory socket J6 on the Swan 500-C, chassis ground is pin #9
and the external zener current goes to pin #3 or #4 (They are jumpered
together in the octal plug) and of course you leave the plug inserted.
For other Swans that might have different wiring, the zener is a
25 watt, insulated, stud-mounted 1N2974-A.
----------------------------------------------------------
A couple of other tips that might be useful to some:
Although my receiver was functioning satisfactoriy, the S-meter
readings seemed stingy compared to my Icom. Removing a capacitor
is an easy fix for this:
The high-end of the Audio Gain pot (R-1201) has a capacitive-divider
of two .001uf's in series which divide down the audio voltage derived
in the Product Detector. This midpoint goes to AGC amplifier/detector
(pin#8 V-11, 6BN8)
to generate the AGC. Simply removing the lower (grounded) .001uf, C-1104
allows much more audio to generate AGC without affecting anything else.
In my case, this makes the S-meter readings more normal, and better
AGC control of signal levels.

Another change I did was to lengthen the AGC time-constant which seemed
much faster than the Icom (I use it excusively in SSB mode). They used
..01uf, C-901, from pin#6, V-11, 6BN8, to ground. I added an additional
..05uf across it for a longer release-time.

Although the Swan-500 is usually sufficient at its higher power levels,
I prefer to use mine to drive a Linear Power Amplifier. Even so, I prefer
to leave a 50 ohm dummy load paralled at its output to ensure stability
at light loading...the two 6LQ6's are still not working very hard and
should last a long time. The dummy load does not affect the Receive
sensitivity noticeably.