I have an old Heath SG-6 signal generator. It uses two triodes.
I'd like to convert it to solid state.

Bill-

Others have supplied more good suggestions than I thought were possible, but I
couldn't resist adding to it.

With lower voltage solid state circuits, you probably will need additional
amplification to provide the same output you now have. Keeping the output
level constant over the entire range could be a challenge.

I seem to recall that there are high voltage FETs available that might be
candidates for direct substitution for the tubes. Keeping in mind that a
reactive load could stress the FET at double the supply voltage, you might have
to reduce B+ slightly.

The suggestion to regulate B+ seems a good idea, but perhaps you only need to
do it for the oscillator triode(s). I once replaced a 105 volt VR tube in a
Heath HW-100, with a ten watt zener. It was a reverse-polarity diode, so the
stud mounted directly to the chassis. That approach might be easier than
regulating higher current for the entire circuit.

If you try the approach of keeping the tubes with a regulated DC filament, I
wonder if a vacuum tube would be susceptible to the same "filament notching"
that panel lamps experience? According to a report I read many years ago, it
seems that lamps operating from DC fail sooner on average than lamps operating
from AC. Failure analysis showed a notch in the filament develops only on the
DC lamps, with failure occuring at that point.

73, Fred, K4DII