[DuMont]

        I had finished a complete recap of a Philco 46-480, as well as replacing original 1/4 watt resistors with 2 watt flameproofs where needed,
and a new loctal socket installed for broken one on the 7F8 converter and hand-cut a new insulator for one of the two electrolytic cans. (The original insulator, naturally, did not match the Richey Gold can I used as a replacement) 
Connected the chassis to the speaker, turned volume and tone off, and then a slow power up through the variac connected to the isolation transformer, and when I reached 117 VAC on the variac, everything seemed to be holding together, but the output tubes and the transformer were both running hot.
I rechecked everything, paying particular attention to making sure I had put in the correct resistors, then tried again with the same result.  Got the Fluke meter out and checked the B+, which should have been 195 DC according to Philco with a 117 VAC input.  It was up around 280 or so.
Checked everything AGAIN.   My house current from the wall receptacle is 120 VAC, then it goes through the isolation transformer, then to the variac - and then - WAIT A MINUTE.  The Variac?   I unplugged the Philco and plugged in my Fluke.  Where the variac dial and meter was reading 120, it was actually putting out over 130.  When I turned the dial down to about 96, it read an output of 117.   I left the dial there and plugged the Philco back in.  Bingo.  The voltage readings were spot on.
SO:  It may not always be some esoteric answer, but rather something as simple (and annoying) as a variac that's out of whack.
Sigh. 
DuMont

[morzh]

Strange.

130/120 is 1.08
117/96 is 1.22

If both your variac's dial and meter agree at both these points, it is very weird.

[DuMont]

No, the meter was telling me one thing, the variac another. With the variac dial set at about 96, I tested the variac outlet with the meter - and the variac was putting out 117 volts - even though its dial was set to 96. When Variac dial was set to 120, it was actually putting out 135.9.

[mikethedruid]

You may want to buy a project box and a cheap digital volt meter, and a 14 gauge extension cord, along with the needed fittings and build yourself a meter box to plug in line with the variac. You could also install an ammeter in it. Then you could monitor both the voltage being fed to a project and also the current being drawn.

[morzh]

I thought you said :

"Where the variac dial and meter was reading 120". I interpreted this that the Variac has its own meter, and both the dial (the graphics around the knob) and the meter (built-in) were showing the same voltage, and then you used the Fluke which showed 130V istead.
So the built-in meter is in agreement with the dial, which is strange...though it might be that the meter was wrong all along and then the dial was calibrated using that meter....but then these dials are not calibrated on case-by-case basis; I think they are mass-produced.
In any case, the dial is never too accurate. However the meter is supposed to be reasonably accurate.