A guitar string attack has a sharp rising edge. This is when the PT core saturation is needed the most. Having half the input winding will not restore the core saturation as quickly for the next attack depending how close that next attack is to the preceding one. So, what you hear is going to be related to the signal source you provide. I always used to use a signal generator and a dummy load to monitor the amp and find where the distortion begins. The percussive attack of a guitar string just doesn't have what it takes to verify a clean amplified signal.
If you don't care about signal throughput purity at full output power, then it is easy to understand why you think the PT now works as well when half of it is disabled. I don't believe it was over designed at all, provided you test to the full original specifications. Clean 120 watt amps were not that easy to make in the 60s and 70s. (think PA use where you do NOT want to color the vocal sound texture.)
Guitarists, on the other hand WANTED that amplifier and speaker coloration added to the sound source signal from their guitar. It gave the guitar sound new life. I don't recall any of the Ampegs being popular with guitarists because they were so clean. (except maybe Jazz or country guitarists.) But, that was the very reason why bassists liked it. Clean power at full output.
The pre CBS/Pre 68 fender twins actually colored the guitar signal. When the CBS engineers took over, they came from an audiophile background where amplifiers were NEVER intended to color the signal source sound. Many believe that CBS ruined the fender amps, by making them SOO clean that they had no "tone" or added no life to the sound throughput at all, unless you pushed them to 90% output power, which is damned loud. In the those days pre-effects weren't quite as prevalent, and those that were available didn't have much dynamic range for throughput, in other words, they changed to guitar signal, even when bypassed. When engaged, the guitar harmonics were corrupted or lost, and the tone that the amp normally added was changed.
My CBS twin is great for vocals, if you use it just for that. O any other clean signal you wish to reproduce at a louder level. Nowadays, if you have pre-effects, the sound from those is accurately reproduced with a nice clean throughput amp.
BTW, my Fender twin has tilt back legs for that in-your-face sound projection. But, I've learned to use earplugs with that amp.
Anyway, while the plate voltage may meet spec with no signal input or throughput, the loss of pt core saturation will cause it to sag at some point below the peak rated output of the amp. Those tubes need the most power when pushed hard and that power comes from the PT core. I also wonder how you are measuring the B+. With a Oscope? If a meter, what is the frequency response of the meter? Alternately, over what period does it sample the average voltage?
In those day when this amp was available, audiophiles would use very powerful amps, and then only use 10% of the output power capability. The reasoning was that the frequency response curve was not linear through the entire power available range, but it would remain linear when using a very small part of that total range, giving clean, quiet, accurate reproduction of all frequencies throughput. I will freely admit not ever encountering this particular VT-22 amp. Maybe they put that triode section in to get the guitarist desired "tone". But, I would expect they didn't want the final amp section to add "tone", and that's where the power hog lives in that amp, and why the PT is designed as it is. You push the power amp and deplete the PT core saturation, and the B+ will sag changing the bias stability on the all the tubes. But, as I said, you might just like the tone changes that this adds. I remember when Fuzz boxes were popular, changing sine waves to square waves, and then letting the amp color the harmonics produced.
There were many guitarists who wanted the amps to color the sound. Off Biasing the amps was one such method. You can get the sound to color in the various stages of the amp, preamp on through power amp sections.
Do you have access to a THD HotPlate Power Attenuator? Or, some other output attenuator that allows driving the power amps to full spec independent of speaker load?
See:
http://amptone.com/Just to brag a bit, I also own a '55 Fender 5E8 twin amp. It's 50 Watt and plenty loud for clubs.
Cheers,