[Breck]

I'm guessing this has been discussed on numerous occasions already, but I wasn't able to find it anywhere on the forum even after extensive searching.  My question is this:
 
Why are there typically not standby switches on headphone amplifiers using tubes?  
 
A standby switch allows you to leave the plate voltage supply off while the tubes warm up, and then switch it "on" when you are ready to listen.  This can extend the lifespans of the tubes by as much as 50%.  I'm curious that I never see these on headphone amps.  Anybody have any input?

 

[Todd R]

Ray Samuels puts them on his amps. 

 

[Avro_Arrow]

My "Tube amp in the works" has one.
 
It uses a differential input with a diamond buffer output.
Without the feedback, the DC offset is not controlled.
The standby switch enables the tubes to warm up before
applying voltage...much safer for accidentally left connect
headphones.

 

[Breck]

I'm not familiar with diamond buffers.  A bit of googling suggests that its a fancy solid-state AB power buffer, yes?  I could see how DC swings during preamp warm-up could be ... detrimental to the life of one's headphones.
 
I almost wonder if the damage done by applying plate voltage before the tubes are properly warmed applies more to amps with (A) high plate voltages and/or (B) big power pentodes.  I know standby switches are popular on high-powered guitar amps, which tend to run their preamp triodes at around 200V at the plate as opposed to the 100v or so you tend to see on hifi amps.  They tend to run their power tubes pretty hard as well.  Could be at that high a voltage there's more damage with unwarmed tubes.
 
Then again, in the "Audio Cyclopedia", they talk about standby switches as well, and that's mostly talking about hi-fi amps.  Hmm...

 

[kevin gilmore]

Big power tubes are subject to cathode stripping if you apply high voltage before they are warm.
Tube rectifiers were designed to warm up much slower to prevent this from happening.
el34,kt88,6550 and especially 6c33 are some of the tubes that this happens too.
But it also depends on the plate voltage. Mcintosh mc275 is an example of a unit
with no delay, and solid state rectification and the tubes on that thing last a very long
time.
 
Not a problem with preamp tubes.

 

[Breck]

Thanks Kevin.  I figured it must be something like that.  Those dual triodes are nearly indestructible.  Well, if you run them overcurrent they tend to ... blacken ... somewhat, but you know what I mean.  Destroyed a nice pair of 12AU7s that way.  I'm guessing low-power pentodes like EL84s and the like probably don't have too much of an issue either.
 
I think I'll still plan on using one on my current build though.  Its kind of nice to have a standby switch, even beyond the tube-preservation angle.

 

[kevin gilmore]

If you have a low voltage around (say 12v) then a 555 timer and a relay makes
a good time delay. Easy if you use the tubes at 12v filament only.
 
But you really need two transformers this way, one for
the filaments, and another for the high voltage.  Switching the high voltage
live usually destroys the relay's very quickly. Or you need a pass fet with
an RC circuit to come up slow and go down quick.

 

[Breck]

I always wondered at the feasibility of using a single transformer and a boost converter for the high voltage.  I know the purists tend to frown on switching regulators for audio amps, but if you design them right they tend to have very little ripple.  Plus, if you switch fast enough - say 1MHz - you wouldn't be able to hear it anyway.  Then you could use a microcontroller to switch it, waiting a bit for the warm-up and then slowly bringing it up from a duty cycle of zero.  Plus, as long as you're running a Class-A amp, the current load will be basically constant, so you shouldn't have any overvoltage DCM issues.  Drawbacks would be the need for a pretty massive capacitor and good-sized inductor, but those might be cheaper than the conventional transformer power supplies.  I may try that for my NEXT next build.

 

[trimer1]

I have used boost and buck converters in my amp designs with really good results, as you said the switching frequency is high enough that our ears would never notice the ripple. To bring slowly the B+ you can use a mosfet with some zenners and a capacitor in the gate

 

[Breck]

Good to know.  I may try using a single high-voltage transformer for a mains supply and using a buck to produce the filament supply in a future build.  I'm not sure that it will really be any more space-, power-, or cost-efficient, but it would be an interesting exercise.