1
\$\begingroup\$

As I understand vacuum tubes (like triode / pentode) usually are designed to be heated with certain current through the filament.

And usually it is about low-voltage and comparatively high-current. E.g. 6.3 V at 300mA.

Schematics I usually find around use transformer to get this heating voltage. I wonder, whether I can instead connect heater filament directly to 220V AC via suitable capacitor to limit current. E.g. I think 4.7 uF should provide about the said 330mA. And it can feed several lamps in chain.

What could be the drawbacks? I can suppose that temperature will be pulsating slightly (like incandescent bulbs do about 5% light pulsation) and this can affect the anode current... But vacuum tubes have slower temperature time-constant...

UPD: some clarification - this scheme of course should only be considered for isolated filament (how do we call it... indirect cathode heating?).

Also "simplified" calculation I=Uac*(w*C) which I mean is correct only while total voltage on filament (or chain of filaments) is small in comparison to full Uac.

\$\endgroup\$
10
  • 3
    \$\begingroup\$ Think about it for a moment. You are using the capacitor as a voltage divider, with the heater filament as the second part of the divider. If you add more tubes in series, you change the divider ratio. Also, heater filaments aren't simple resistors. Like light bulb filaments, their resistance is low when cold and higher when hot. \$\endgroup\$
    – JRE
    Aug 27, 2019 at 10:48
  • \$\begingroup\$ Interesting approach. Be careful with placing un-isolated 220 volts between filament and cathode. \$\endgroup\$
    – analogsystemsrf
    Aug 27, 2019 at 10:50
  • 1
    \$\begingroup\$ The RC highpass filter disagrees. \$\endgroup\$
    – JRE
    Aug 27, 2019 at 11:05
  • 1
    \$\begingroup\$ And, yes voltage divider. \$\endgroup\$
    – JRE
    Aug 27, 2019 at 11:07
  • 2
    \$\begingroup\$ Long ago, there was a 5-tube radio whose filaments were wired in series, driven by 120V AC: 50V + 35V + 12V + 12V +12V (all indirectly heated cathodes). I recall that one or more flared alarmingly at turn-on until all thermally stabilized. Have also seen a series resistor to take up the slack...never a capacitor. \$\endgroup\$
    – glen_geek
    Aug 27, 2019 at 13:31

2 Answers 2

Reset to default
2
\$\begingroup\$

AC voltage on the heater will drive the filament to a positive voltage every half cycle. Any voltage difference with respect to the cathode will draw electron current away from the cathode, thus modulating the voltage gain of the tube at AC frequency. Likely highly undesirable.

\$\endgroup\$
6
  • 5
    \$\begingroup\$ Many tube circuits use AC for the heater filament. They are designed for that. \$\endgroup\$
    – JRE
    Aug 27, 2019 at 11:04
  • 1
    \$\begingroup\$ Ah, yes. I clearly missed this. 300V swing to both positive and negative - comparable or more than anode voltage.Thanks a lot! \$\endgroup\$
    – Rodion Gorkovenko
    Aug 27, 2019 at 11:04
  • \$\begingroup\$ @JRE they are designed for roughly 6V AC, not 220AC... Interesting to check how AA5 was designed in this respect, I think they have higher-voltage filaments. \$\endgroup\$
    – Rodion Gorkovenko
    Aug 27, 2019 at 11:05
  • 1
    \$\begingroup\$ Filaments are usually driven from a separate winding on the transformer so they're isolated from the anode and cathode and won't cause a current to flow. \$\endgroup\$
    – Finbarr
    Aug 27, 2019 at 12:21
  • 1
    \$\begingroup\$ simply tie one end of the filament to GND, and other end to the voltage-dropping capacitor. The cap handles most of the 220 volts. \$\endgroup\$
    – analogsystemsrf
    Aug 27, 2019 at 13:51
2
\$\begingroup\$

I wonder, whether I can instead connect heater filament directly to 220V AC via suitable capacitor to limit current. [...] What could be the drawbacks?

Well, the primary drawback would be the dramatic increase in deadliness of whatever device you're trying to avoid putting a mains transformer in. As others have mentioned, in the Before Time(tm), companies were somehow allowed to get away with exactly what you're describing (the AA5's, too many guitar amps, etc. etc.). This is thankfully no longer allowed in general (and certainly nowhere a vacuum tube would be used today, i.e. some type of audio amplifier).

Even if you were to limit the discussion to heaters only, and were otherwise properly using a mains transformer for the HT+, the safety hazard still exists: e.g., heater-to-cathode shorts do happen, and when they do, you would now have your circuit ground and all the exposed metalwork connected to it now shorted to the AC mains through a ~1.5K resistor and a cap (neither rated to handle 220VAC BTW). So now you have a live chassis. Not good.

For much more information, dive into the regulatory requirements covering the different electrical safety classes, e.g. Class I and Class II would be the most applicable here.

This truly is one of those "never, ever, consider doing this" sort of things we occasionally run into in our field. Always put in a proper mains transformer between the mains and any vacuum tube (or other) device you're designing.

\$\endgroup\$
1
  • \$\begingroup\$ I agree that it’s not the best idea from electrical safety perspective, but a DIN-rail mounted GFCI mounted inside the unit would go a long way towards catching heater-to-cathode shorts. If the unit is a self-contained radio with no external exposed antenna connections, and all the controls are insulated, as well as the case, to provide double insulation - then it should be about as safe as anything else with double insulation. An external antenna connection would need a safety-rated coupling transformer. Two coils on opposite sides of a thin glass sheet would be fancy way to do it. \$\endgroup\$
    – Kuba hasn't forgotten Monica
    Aug 12, 2022 at 2:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.