[Amps] Why does power output increase?

[Paul Christensen]

>"I still wonder if the filament voltage should be measured while transmitting or on standby..."

Vic,

This may be unrelated to your amp issue, but I nearly ruined a FU728F tetrode tube because of filament voltage starvation.  That tube uses a 9.0 VAC filament.  The amp is an OM Power 2000A+.  The installation section of the manual recommends setting the transformer tap to match the AC lines voltage - in my case that's 250VAC.  

Unlike your problem, output power would drop to 50-70% of the initial keyed power.  Filament voltage dropped to under 8V when trying to make 1500W.  That's a big problem with indirectly-heated cathodes.  Changing to the 240V transformer tap fixed the problem.  Filament voltage at full power is 9.0V and 9.4V during key-up.  The moral of that story is to ignore the installation instruction and choose the tap setting that results in 9.0V filament voltage at 1500W.  

So, to answer your question, I believe filament voltage should normally be set at full output power.  OTOH, if there's AC line sag due to a long cable run, or a long undersized power cord, then that should be corrected in order to minimize the filament voltage differential between key up and key down at full output power.  Otherwise, it's possible to end up with dangerously high key-up filament voltage when filament voltage is set for full output power.  

By the way, your 20m sigs have been very loud into N FL for the last two weeks -- sometimes the strongest DX on the band.  

Paul, W9AC