As for high frequencies making across the dropping capacitor into the heater element, some testing will be required; maybe a simple filter is needed, such as a small series resistor and a small shunting capacitor across the heaters.
   This reminds me of an issue I do not believe I have mentioned here before. If you hook up your scope to just the heater winding of a tube amplifier, often you will see an odd distorted waveform rather than the clean 50 or 60 cycle waveform. Why? The small aperture of charge conduction time caused by large power supply reservoir capacitors is the culprit. As the charging window narrows, the current spike increases. Because all windings share the same transformer core, the windings interact. The solution I have found to work very well is to use separate transformers for the B+ and the heaters. An added advantage of two transformers is an easier staggered turn-on arrangement: heaters first, then B+. I  have used a 6 position rotary switch that powered on the heaters and provided a small trickle voltage for the B+ capacitors at the first click, at the final click the full B+ came on. As fast as I could turn the knob, the scope revealed a gently blossoming B+ voltage.

Subject: heater supplies
   Hi there,  once again an excellent issue,  I was really impressed by your ideas on voltage regulators etc. for heaters. One way of getting a cheap and cheerful voltage regulator is to use a 5v device with a couple of diodes in series from  the GND pin to the actual ground.  this lifts the output of the regulator to 6.4 volts (nearly 6.3, this is made up from two times 0.7v forward bias on the diodes, plus the 5v on the device itself).  whilst this does not limit the current as your model with the resistor in the output,  it satisfies my engineering desire not to waste power ! I am not keen on heating up the room unnecessarily. just a thought for you, thanks for your excellent articles.

bill

Subject: Low voltage heater operation
   I was interested to see the lifetime curves for lower heater voltage. It is sometimes claimed that reducing heater voltage (hence temperature) will cause interface poisoning, i.e. reduce the Gm of the tube. Some references say  that there is a critical temperature (around 900 degrees K) below which this occurs. Heater temperature is a function of the 2/5th power of voltage, so a 20% reduction in voltage gives a 9% reduction in temperature, still comfortably above this level.
    Another advantage, theoretically at least, of lower heater temperature is that the non-linearity due to the virtual cathode is reduced. The potential of the virtual cathode has an exponential relationship to the current, which at low currents will result in higher high-harmonic distortion.
   I tried this with a 6922-based line stage. Reducing heater voltage to 5v did reduce harmonic distortion by about 10%. However I did not measure whether this was directly due to the lower temperature, or whether it was a second-order effect due to the changed Gm and
hence operating point - it is on the list of things to do one day!

John

   I remember reading an article that compared many pentodes as low distortion audio amplifiers that stated that the clear winner in the first round of testing failed the next day when retested. The variable had been isolated and it was the change in heater voltage. With less voltage on the heater, the tube was a champ; with more, a loser. I have a few theories about what is going on and I hope to do some testing of my own soon.

Subject: shock hazard
   I have read all of your web based magazine articles, and I thoroughly enjoyed them. They provoked many new experiments, as I tend to build a circuit and test it rather than analyze it to