[Johnson] Viking I Modulation troubles

[k2cby]

       I'm new to the list, and I am having trouble getting an ancient
Viking I back on the air.
       I seem to have two problems:
       First: I am unable to get better than about 50% AM modulation.
       I rebuilt the speech amplifier as per the factory release titled
"Modification B for Viking II Transmitter" (courtesy of the BAMA website)
replacing all the paper and electrolytic capacitors in the process. 
       Feeding 1000 cycles into the microphone connector, my scope shows 7
volts peak-to-peak on the plate of V1, the 6AU6 speech amplifier.
       By the time I get to V2, the 6AU6 audio driver, I'm showing 1.1 V p-p
on grid 1 and 28 volts p-p of audio on the plate.
       The plates of the 807 modulators and the 4D32 PA carry 720 volts.
       The 807s show 55 V p-p audio to ground on each grid #1 with a bias of
-38V dc. The 807 screens show 260 Volts with R13 adjusted to produce a
no-signal idling current of 70 mA (as per the manual). The audio looks
pretty much undistorted on the 'scope, even though the 807s are obviously
being driven into the grid current region. I can run the audio drive up to
80 V p-p before the modulators start flat-topping.
       
       This is the second problem. 
       High voltage B+ is supposed to be 660 volts. I measure 770 volts in
the CW mode and 720 in phone, which is about 17% high. The ac line voltage
at the time these measurements were taken was 123 volts, which is only about
12% high. This, by the way, is with the original 5R4 rectifiers – not solid
state replacements. There is also more hum on the signal than I would like.
I originally thought that the high voltage choke L1 might be shorted
(converting this to a capacitor-input power supply) but its dc resistance
tallies with the manual.
       Oddly, enough, the medium voltage (+300 volts) and bias (-75 volts)
power supplies are pretty close to spot on.
       Questions:
       1. Do I have a problem in the HV supply, or is it just a question of
higher line voltage?
       2. Will bucking down the High Voltage B+ to nominal 660 improve the
modulation percentage?
       3. Will replacing the 807s do anything for me?
       4. Should I increase the negative bias on the 807s and push more
audio drive into them?
       5. Should I consider replacing the 4D32 with two 6146s and adding a
clamp tube a la the Viking II?
       What say some of you AM-ers who have had more experience than I?
       
       Miles B. Anderson, K2CBY
       16 Round Pond Lane
       Sag Harbor, NY 11963
       Tel.: (631) 725-4400
       Fax.: (631) 725-2223
       e-mail: k2cby at optonline.net