What factors affect amp warm-up time?

[Deke609]

A question for the experts.  My BH amps sound better to me after multiple hours of use. And I don't think it's brain adjustment, b/c I'll let them play music and walk away for hours without listening to them. And when I return and put the headphones on, they sound much  better - it's particularly noticeable with the pairing of BeePre and Kaiju - both of which are slightly modified. I'm wondering whether either of the following might contribute to long warm-up period: in the BP, I have 22mF cathode resistor bypass caps in place of the 10mF; the Kaiju currently has big parafeed Lundahl output transformers (LL1679) that I'm trying out -- although I noticed the same extended warm-up period phenomenon with the stock iron, but perhaps not as much.

Does it take smaller signal longer to "warm up" bigger iron designed for larger signal?  I don't think big caps should make a difference (difference in charge time between 10mF and 22mF would be tiny, right?), but what do I know - could they affect things?

MTIA for any insight/speculation you care to share, Derek

[Paul Joppa]

One theory is that the start-up transient will magnetize the parafeed output transformer core, and music will gradually demagnetize it. I'll speculate that the demagnetizing process is slower with small signal levels such as you might be using with headphones. (The shutdown transient is slower, so it will have less effect. Maybe. Sort of.)

A test might be to put in a switch to short the output transformer primary before shutdown and leave it shorted until after startup.

[Deke609]

Aha! That is an interesting conjecture and experiment.  Much appreciated PJ!

I will implement the switch -- and to that end, I have some practical follow-up questions:

(1) Do you have in mind a dead short (section of wire) in parallel with the primary - in which case the primary will still see the voltages of start-up and shut-down transients?  Or should the output transformer be completely removed from the path of the transients -- in which case the switch will "switch" between primary and dead short?

(2) I think you've told me before that the output tube doesn't care how it is AC loaded. If that's the case, would it be safe to accidentally send signal through the amp when the output tube has a roughly zero impedance path to ground for signal?  I wouldn't deliberately do this, but I might forget one day to flick the switch before hitting "play".

(I should add, further to your point about small signal use with headphones, that for the last while I have put a non-inductive resistor in parallel with each channel at the output, so that the output tube of the Kaiju sees a load close-ish  to design spec. And I do almost all my listening with the course attenuator of the BeeQuiet set to -18db, and occasionally -9db for quieter recordings or when really want some thunder. And the Kaiju PEC trimmer pots are fully open (zero attenuation). All of which is to say: by Kaiju standards, the amp produces a fairly large signal to do the demagnetizing. But I am still eager to try the switch experiment, since I would expect changes to signal size to change things by degree not in kind.)

Many thanks again, Derek

[Paul Joppa]

1) just a short across the primary

2) safe, but if there is a non-zero AC voltage at the instant of switching, it will produce a transient  which will magnetize the core again. That transient will be small unless the signal is really loud.

[Deke609]

Excellent! Many thanks PJ. A simple short across the primaries should be easy to implement.

cheers and thanks, Derek