Sound clips of resistor types

[Lauri]

I think one of the reasons why null testing can't be done with tube amps is because of how much tube parameters change depending on how long the amp has been on. I've done some tube matching for studio equipment and there can be a huge difference in currents, gain and transconductance compared to if the tubes have been powered for 5 minutes or 20 minutes.

[Stevem]

Agreed with the above.

Tubes have been around for so long that we forget that originally they where called thermonic valves.

[maxkracht]

I would still assume null testing in a DAW should give some sort of a pattern, assuming there is enough difference between component types. If the natural variance in the amp over time is more significant a change than swapping a component, that's also a result.

[R.G.]

I think one of the reasons why null testing can't be done with tube amps is because of how much tube parameters change depending on how long the amp has been on. I've done some tube matching for studio equipment and there can be a huge difference in currents, gain and transconductance compared to if the tubes have been powered for 5 minutes or 20 minutes.

Null testing has been done, with success on tube amps. The instance that comes to mind is the Bob Carver Challenge; for some fun reading on this, see the conclusion at https://www.stereophile.com/content/car ... nge-page-4 or just read the whole article. Stereophile mag admits they're not an impartial observer, yet also admits that Carver made his amp sound indistinguishable from their La-De-Da (tm) fancy high end tube amp.

On the thermal drift thing: it's common in formal testing to let things come to thermal equilibrium first. Logically, a tube's parameters simply can't keep changing forever as it heats up. Eventually it gets as hot as it's going to get, and should stop changing. Out of curiosity, did you happen to measure the current/gain/transconductance every five minutes or so for an hour? If you did, I suspect that the rate of change diminished steadily until it leveled out.

For the simple one-tube test I thought-proposed, we could just leave the triode on for days, perhaps under a thermal cover (i.e. a cardboard box inverted over it) so it gets a good, long chance to equalize. Small triode tubes have shown lifetimes of tens of thousands of hours when operated within their specified limits. 10,000 hours is 416 days, so there would be plenty of time to stabilize thermally, and through initial lifetime shift and work for many months before wearing out if you just left it on. Maybe you could let it warm up for a day or two for some tests, then turn it off for several months until you tested again. One tube, if initially good, ought to last a good long life as a tester.

It's also possible that the tube drift doesn't matter much. Using a single duotriode for the tester you would expect the two halves to drift similarly, not differentially. As long as they're matched, the effect on the resistors being tested should not matter much. Maybe - maybe - the tube warmup could be for some minimal time.

[Lauri]

Out of curiosity, did you happen to measure the current/gain/transconductance every five minutes or so for an hour? If you did, I suspect that the rate of change diminished steadily until it leveled out.

I've only measured them every 5 minutes for about 20 minutes and then measured again after it's been on for an hour or so. After about 15 minutes all the parameters have pretty much settled and there's only very small changes.

[martin manning]

I found similar stabilization time using my uTracer. After about 40 minutes the parameter readings (Ia, Ig2, gm, ra) flattened out.

[R.G.]

It's good to see that real data matches my guesses. So one instruction for use for a duotriode based resistor-difference tester probably ought to be "for best results, let it warm up for an hour before running tests."

Hmmm. given that triodes change gain/mu/bias point/etc. for fifteen minutes to an hour, that probably means that tone evaluation testing should be done only after a long warmup. I hope that the tone of the amp doesn't depend on the power transformer and output transformers. Those things have thermal time constants measured in hours. Secondarily I wonder if interviewing resistor types, capacitor types, and so on might have hidden misdirections lurking under triode warmup time.

Real data is so much more satisfying.

[martin manning]

t's good to see that real data matches my guesses...

Here is the test using uTracer. Note the temperature rise is from the heater only (DC powered) since the tracer measures tube parameters with 1.2 ms pulses. As you can see after the 5 minute mark the current drifts down 2% on this 6L6GC, and the second run using resistors shows that a third of that was due to the instrument itself settling in.

https://ampgarage.com/forum/viewtopic.p ... 38#p327038

[R.G.]

Good work!

That also brings up another thought. I used a duotriode as the supposed test element primarily to forestall "but you used silicon!!!" style comments. The two halves of a 12A?7 will have some differences that need nulled out, and will need some support to stay at a constant idle point. It's possible to wash this out by doing a servo on the grids of the 12A?7 driver to make its cathode voltage park at a constant value. I'm guessing that, plus the inherent rejection of taking the difference at the plates should be good enough.

The two-resistor test really comes down to constructing a current-fed bridge.