[ARC5] Dear Vibrating Smart People....

[David Stinson]

I'm working on an RCA Light Aircraft set using the AVA-126A
Vibrator Pack power supply.  Here's a drawing of the input
of that supply, wired for 12 volt operation per the design:

http://home.netcom.com/~arc5/AVRAVT/AVA126Vib.JPG

All the old caps have been subbed.  
The OZ4 rectifier tests good.
The 10K resistor in the "snubber" is rated at 1/2 watt.
With or without load, it is drastically overheating.
I'm having this problem with two of these supplies.
A .02 temporarily touched across the resistor shows 
sparks, so it's "snubbing" some hefty spikes.

My current theory:  The vibrator is very old and the 
pressed-in, oxidized contacts are creating more "noisy,"
higher-voltage spikes than those for which the set 
was designed.  These are causing higher spikes 
in the secondary than the snubber can handle.
I got one of these vibrators working by soldering the
contact bases to the contact springs, but the top layer
of the bi-metal contacts themselves will not take solder.
The other worked without opening it.

I assume the RCA engineers knew what they were doing, 
so I don't want to "(dis)improve" their design.  But they 
could walk to the Supply Sargent and get a new vibrator.
I can't.  I don't want to fry them out with 115 VAC
as I've often heard mentioned, and I'd rather not replace them
with solid-state parts if I can avoid it.

Possible solution:  Modern MOVs (spike killers) are available
cheap for voltages from 2 to 500 volts that will handle amps.
Either low-voltage MOVs tacked-across the primary windings 
or a high-voltage MOV on the secondary, or both to eat the spikes.
That's what A.R.C. did on the MD-7 modulation transformer 
secondary; that large, conical-ended doodad under the chassis
is a big "varistor" (caveman MOV) that eats audio spikes 
to save the transformer.

Take a look at the diagram.  
Do you think MOVs might solve the problem?
If so, what values would you recommend?

TNX OM ES 73 DE Dave AB5S