[R-390] AVC voltages?

[DJED1 at aol.com]

I've had a slight problem with my R-390A for awhile - the carrier meter reads 
about 20 dB too high.   I   finally pulled the radio out of the cabinet last 
night and did some diagnostics, but i need some sage advice from the group.   
first, I set the IF gain according to the Rippel method (measure AF noise 
level), which also gave me      -7V at the diode load with 3 microvolts in.   So 
that seems about right, but still the problem existed.   I measured the AVC at 
the back terminal strip: -12 V versus a nominal 9 V for 10,000 microvolts.   
This implies the AVC on the controlled stages is not high enough.    Measured 
with a 10 Meg voltmeter and a URM-25.
Then I measured the AVC voltage at all controlled tubes- With -12 V at the 
AVC terminals on the back, the IF stages had -12 to -10.5 V on their grids, but 
the mixers had -10.8 to -8.1 V , and the RF stage had -7.8V.   I tried 
substituting NOS tubes in the RF and worst mixer to see if the tubes were gassy, but 
no change.
I then measured the resistance to ground of the AVC bus, and it was about 
500K as expected.
Sooo- does anyone have data on AVC voltages at the grids on a known good 
radio?   If I had leaky caps pulling the AVC line down, I would expect a much 
lower resistance reading.   Any other suggestions?   
The radio has not been recapped, but I'm reluctant to pull the RF module 
unless I have some clear indication that that is the problem.
Ed   WB2LHI