For valve amplifiers, it is said, that applying anode voltage before the valve is heated reduces lifetime and can have other negative effects on circuit lifetime.

To avoid this, I want to add a thermistor to the HT voltage line, so the valve has time to heat, and the HT is applied in a more controlled manner.

Based on what parameters of my circuit should I choose a thermistor? Or what else should I use to "delay" the HT voltage?

Currently I am using a manual switch, which is good for standby mode, but I'd prefer to have something, that doesn't rely on the human factor.

Edit: In the past this was solved by using tube rectifiers, so they would need time to heat up too, but I want to use solid-state rectification, so that is not an option.

  • \$\begingroup\$ Traditionally, the rectifier warms up gradually... \$\endgroup\$ – Brian Drummond Jun 26 '15 at 14:21
  • \$\begingroup\$ @BrianDrummond Forgot to mention I am using solid-state rectification. Edited. \$\endgroup\$ – akaltar Jun 26 '15 at 14:27
  • \$\begingroup\$ If intention is to add delay till a certain temperature is attained, thermistor can be used with a comparator to trigger a gate at set temperature (thermistor résistance). Gate is like a switch which will supply power to the anode of the amplifier. \$\endgroup\$ – Umar Jun 26 '15 at 14:48
  • \$\begingroup\$ @Umar I heard there were/are thermistors, which heated themselves, thus acting like a delay, is there something like this nowdays? \$\endgroup\$ – akaltar Jun 26 '15 at 15:49
  • \$\begingroup\$ There are PTCs for delayed protection and NTCs for delayed response. Yes, I have been using it still. \$\endgroup\$ – Umar Jun 26 '15 at 15:55

The thermistor idea works well, I use it on all my amplifiers applied to the primary of the main power supply transformer. Connected this way it also limits the current inrush:

  • of the filaments when they are cold.
  • in the (big) power supply capacitors.
  • in the coupling capacitors which often discharge very slowly through high resistors.

This said, it will only limit the current for 1 or 2 seconds, you may wish to rise the voltage slowly on tubes anodes. This is a bad idea to wait for the filaments to be hot and apply the B+ suddenly because:

  • during the time the HT is not applied, the power supply has no load and then presents its maximum voltage
  • because it will create a current inrush specially in the power tubes that can -- for a short moment -- drive the tubes far above their max dissipation.

The idea is to apply the HT gradually. Using a vacuum rectifier is a way to achieve this. You can also set up a mosfet regulated power supply:

Gradual mosfet regulated power supply

Copyleft Yves Monmagnon. August 2009 from this link (in french)

Q1 and R1 create a CCS. This current will charge C1 through R2 and polarize Q2. When C1 is charged the voltage on the source of Q2 is roughly equivalent to Iccs * R

I use DN3545 for Q1 and IRF820 (up to 500V) for Q2. The zener is here to discharge C1 when the PSU is switched off and avoid Vgs > 20V (maximum specified by the DS).

  • \$\begingroup\$ Would it be off topic to ask what you do to limit inrush through the heaters to prevent premature burn out? \$\endgroup\$ – Paul Uszak Jan 27 '18 at 21:33
  • \$\begingroup\$ @PaulUszak use a NTC thermistor (33 ohms) in serie with the primary of your power transformer. Il will limit current inrush in both cold heaters and empty PSU capacitors. The thermistor impedance drops dramatically after few seconds. \$\endgroup\$ – greg Feb 6 '18 at 14:57

Since it takes the tubes some time to warm up, what you need is a time delay relay that starts to time out when mains power is applied to the amplifier and then closes a set of contacts between the B+ and its load(s) when the proper time has elapsed, perhaps 30 seconds to a minute, depending on which of the tubes takes the longest time to warm up.

Then, when mains power is cut off some time later, the relay opens immediately, allowing the tubes to cool down with no stress from the B+.

The relays are available commercially - quite pricey - and are called ON-DELAY relays, or you could easily put one together if you're so inclined.


I don't think your thermistor idea is going to work very well. The tubes will represent little load before the filaments are warm so a resistance in series won't drop much voltage. The current through the NTC has to be high enough to make it/keep it hot or it won't be low resistance.

You could automate the switch by replacing it with a time-delay relay- that would probably be the simplest solution.

If you really want to apply voltage gradually (rather than with a delay), and were willing to live with a few volts of drop, perhaps you could use a MOSFET voltage follower and an RC delay. To avoid the drop, create another supply about 10V higher than the B+ supply. You'll need a few parts to protect the MOSFET gate and to discharge the capacitor at a reasonable rate when power is removed.

Or something like this 20 second Amperite delay relay might be more in the spirit of your 'bear skins and stone knives' technology:


  • \$\begingroup\$ There is already some RC there for smoothing, couldn't they be the load so the series resistance will drop a lot? \$\endgroup\$ – akaltar Jun 26 '15 at 16:15
  • \$\begingroup\$ @akaltar No, I don't think so. \$\endgroup\$ – Spehro Pefhany Jun 26 '15 at 16:19

"For valve amplifiers, it is said, that applying anode voltage before the valve is heated reduces lifetime and can have other negative effects on circuit lifetime."

This is not, actually the case. Standby switches were the invention of Fender. I would highly recommend you read the work of Merlin Blencowe as well as this piece from Peavey's site.

Keep in mind that tube TVs, tube radios, tube PAs, and other tube devices never had standby switches. Only guitar amps have them. With the exception of gain structure and features, there's little difference between a guitar amplifier and a PA.

  • \$\begingroup\$ It should also be noted that valves actually last longer if they have some cathode current flowing while they are heated, which again is contrary to the theory of the standby switch. Source: Radio Designer's Handbook, 4th edition. \$\endgroup\$ – user207421 Feb 14 '18 at 1:46

Use a sensitive bi-metallic switch whose contact closes when warmed by the tube. Old tech solution to match the amplifier technology.

I guess a related question might be, do you leave anode voltage on until cooled down OR, should anode voltage be collapsed before deactivating the heater supply. You need to think about this.

  • \$\begingroup\$ While this might be a good solution, I would prefer something, thats not coupled to the tube, as that would be hard to do mechanically because the tube is made of glass, also I have multiple tubes, so a solution regardless of tubes, and with passive elements would be the best. \$\endgroup\$ – akaltar Jun 26 '15 at 15:51
  • \$\begingroup\$ Try feeling around the socket the tube is in - there's still quite a bit of heat to be had without contacting the tube. \$\endgroup\$ – Andy aka Jun 26 '15 at 16:17

As others have pointed out and have also provided links to other sites to add citation, a simple standby switch can actually shorten the life of valves and can seriously damage a valve rectifier if it is placed before the reservoir capacitor. And as has also been pointed out, theoretically at least, there is no evidence that the phenomenon of cathode stripping due to HT being applied before the valves are up to operating temperature actually occurs at least not to any significant amount that needs to be worried about.

There is, however subjective evidence that when both a mains soft start circuit (to manage the inrush for the whole thing lasting only for a few mains cycles) and a means of slowly building up the HT as the valves are warming (either using a valve rectifier or a pair of mosfets after solid state rectification), rather than suddenly applying full HT after a delay with a human operated switch or a timed relay, that amplifiers with a reputation for repeatedly blowing valves for no apparent reason, do then settle and stop blowing valves.

  • 1
    \$\begingroup\$ Well, that's your opinion. What I have written is based on years of experience as an electronics engineer. If you can't read a couple of paragraphs like that without understanding and feel you have to be so rude, then perhaps it is you that will get down voted. I'm all for constructive criticism, but there is no need to be so offensive. There's too much information to include all the points on one or two lines. Learn to read and comprehend before being so rude. If you have to comment be polite like people are on most such forums, if you feel you are so superior. \$\endgroup\$ – David Warsop May 9 '17 at 13:29
  • \$\begingroup\$ Just wanted to say, after having a warm welcome and congratulations for a good answer from the moderator, it was really lovely to know from your comments that I have joined such a friendly and open discussion site where people can share their skills and experience in electronics out of a genuine desire to help others and share experiences without abuse and rudeness about perceived lack of written English skills. \$\endgroup\$ – David Warsop May 10 '17 at 7:40
  • \$\begingroup\$ That was yet another rude attack on English grammer. How wonderful to be so perfect and self satisfied, having to correct others in order to boost your own ego. This is an electronics site, not an English test! \$\endgroup\$ – David Warsop May 15 '17 at 10:59
  • \$\begingroup\$ If you want to set up your own discussion site then do so. You can then show how good you are at English and continue to put people down to boost your own ego there. This is not the place. If someone's comments or answers passes moderation, you are only entitled to comment on the technical content, you are not entitled to appoint yourself judge and jury in a rude and counterproductive way. Grow up and stop being so superior and self important. Leave people alone who just want to help each other and share their experiences I a civil way. If you need to criticise, do so constructively to help on \$\endgroup\$ – David Warsop May 15 '17 at 11:42
  • \$\begingroup\$ I may have a lot to learn about English Grammar, however I am not a proof reader or author and do not pretend otherwise. You however have far more to learn about basic politeness, the difference between abuse and constructive criticism and chat room etiquette. \$\endgroup\$ – David Warsop May 15 '17 at 15:21

The time delay timer seems like a good idea, I found a delay on timer on Ebay for 99p up to 10 sec from a 12V supply, I supply my valves with a volt reg (7806), so have 12V on power up, hope this gives others some design ideas


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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