when I tune to zero as instructed in the manual.
on 20m my power out appears to be about 200w less than when I peak it with power out.
I'm not surprised about that. However there is an alignment procedure in the manual (I think -- haven't looked at one lately) to actually "adjust" the "TUNE" (0 scale) meter so that it will agree with an output power peak. If you do that, they should both occur at the same point.
Would have been better if they just grounded the grids, fed positive bias to the cathodes, and elminated that silly feedback circuit they used. I guess whatever they did was to make it "easier for untrained operators in the field" as Phil says (sounds plausible), but the result was a silly design.
Collins also used excessive plate voltage for 811As, above their rating; I understand that was done with an agreement from RCA, who supplied all the tubes. Problem is, RCA stopped making 811As about 40 years ago and the Chinese tubes aren't as good.