The diodes were my thought as well, but this may not give the necessary
control since voltage drop must be a multiple of the bias voltage. But this
will certainly get you closer and help determine if you have another problem
that is causing the filaments to glow so red hot.

Bob

"Thomas Tornblom" -to-reply wrote in message
...
Another possibility would be to connect two or four diodes in an
anti-parallel connection in series with the filament circuit, that
would subtract one or two forward drops, but would perhaps introduce
some unwanted harmonics into the circuit.

Thomas