Unmatched, different brand power tubes?

[Silvio55]

Hi, a couple of days ago I came across with 3 NOS EL84 tubes, one is an Ei, one Hitachi, and the other is an International C (USA rebrander) from yugoslavia. I tested them on my 18 watt lite build and they all worked fine, the amp sounded really good, but I don´t know if it's bad for the amp, or OT to run unmatched power tubes. Anyone?
Thanks!

[J M Fahey]

How do you know whether they are matched or not?
Measure idle current for each of them and post it.
Also measure the ones you had there earlier.

[Silvio55]

Si, tenes razón Juan Manuel! I don´t know for sure if they are matched or not, I thought that maybe cause they´re different brands they could be unmatched, and wanted to know what could possibly happens if they are.
Well, I'll see if I can measure them and post the results

[Enzo]

It doesn't matter if they are unmatched. This is a guitar amp, not a precision NASA space probe. Amp makers at the factory don't install matched tubes. The amp won't care. The only real advantage of matching power tubes is hum reduction and potentially lower distortion. Note that the hum reduction would ONLY be the hum coming from the power tube stage itself.

If a set of completely different tubes sounds OK to you, then it IS OK.

[R.G.]

Dead straight, Enzo.

For the techies, matching comes in at least three flavors. One is DC bias point matching. Another is AC gain matching. Yet another is gain-curve distortion matching. Shoot, now that I think about it, there's drift and others.

Class A and Class AB OTs want the DC bias point to be matched so the DC through both halves of the transformer is as much the same as possible at zero signal. This puts the magnetic field as close to zero as possible and lets you have the biggest undistorted signal out of it. DC matching is what we do with individual pots per output tube, and that is a huge advance over using "matched-some-unspecified-way" tubes. It's the first thing you need to get right. I suspect almost all "matched" tubes are matched for plate current at some grid-cathode voltage.

AC gain matching is the next refinement step. The ideal of two absolutely identical tubes would also have each tube have the same AC gain at every point in its operation. Two same-type, same brand (and preferably same manufacturing lot and same DAY!) tubes will have similar AC gains. But unmatched ones may have, for instance, the "top side" tube with an AC gain of 37 and the bottom-side tube a gain of 41. So the top and bottom halves of the waveform in Class AB get different gains. I don't know if any tube matchers even look at AC gain matching. You can match modestly UN matched AC gain tubes by the expedient of changing the amount of AC signal it gets. This is like a post PI master volume, but with two NON-dual pots to change the amount of PI signal each one gets.

One could argue that this is unneeded, as many people do prefer the sound of slightly unmatched tubes. But it could and has been done (by that I mean I've done this and checked it out on a live amp). Also, amps with feedback will make AC gain mismatching much less prominent. I was worried about it on the Workhorse amps because we used no feedback.

Beyond that, the AC gain of each tube is actually different from low to high currents, and everywhere along the way. This is probably a quite-small effect. I haven't measured this myself on tubes (I have on solid state devices) but it exists. I am pretty sure that no tube matchers do this; it's even more esoteric than matched AC gains.

And then there's drift. Drift never sleeps. Every tube drifts, and with my luck, my tubes do NOT drift in parallel, but head opposite directions, even for "matched" ones. This is why we periodically rebias. It would take a true mad scientist to match drifts after matching all that other stuff.

But I'm blathering. As Enzo said, if you like the sound, it's good; go for it. Tubes, and for that matter amps themselves are just expendable supplies to be used up in the name of making good music. Don't waste the good unnecessarily, but don't be afraid to use it up in the pursuit of good music, either.

[Gregg]

Next comes the question "At what bias and what anode voltage the tubes were matched?". You can see eventually some current values and gm on the boxes but you don't know how did they get them. So when you put the tubes in YOUR amp which has different anode voltage and bias you can get surprised.

[kleuck]

Unmatched (current-wise) tubes can lead to issues you know, they can for example in a cathode biased amp oscillate, i saw this in one of my amps, it was a very low frequency, but perhaps can it happen if HF ?
A tube can draw almost all the current too, giving a weak sound a putting the OT in magnetic saturation territory.
Thats said, it's very unlikely to happen in a hot cathode biased amp, but that's another story in a cold and fixed bias (think Mesa)
There's no one size fits all response i think.
I must say that i personally use and prefer unmatched tubes, but my amp has separate cathodes resistors (done that after the LFO episode)
As for matching, i don't know exactly how manufacturer are doing, but i know that most of the time current and transconductance is measured at a pretty low voltage (250 volts) not necessarily matching the real life in an amp.
But i know a tech who uses an Amplitrex, up to 500 volts, and he matches the curves.

[R.G.]

But i know a tech who uses an Amplitrex, up to 500 volts, and he matches the curves.

That's pretty cool! That takes dedication, and it's about as good as you can do for matching.

[Gregg]

...but i know that most of the time current and transconductance is measured at a pretty low voltage (250 volts)...

That's some kind of joke, right? Everybody "matching" tubes at 250V should look for another occupation. That's why for new builds I always put two bias trimpots (for a pair) and sometimes 4 for a quad. This way you can always DC "match" the tubes for yourself.

[kleuck]

Well, in most tube testers manuals, that's the reference voltage, so...and most can't go any higher anyway.
And, that's too the reference provided by manufacturers in datasheets very often.
Look at the 6L6 datasheet : http://www.mif.pg.gda.pl/homepages/f.../021/6/6L6.pdf
Nominal characteristics are given for 250 volts, though we use them up to 450 irl.

[jpfamps]

That's some kind of joke, right? Everybody "matching" tubes at 250V should look for another occupation. That's why for new builds I always put two bias trimpots (for a pair) and sometimes 4 for a quad. This way you can always DC "match" the tubes for yourself.

Many of the UK valve vendours use 250V triode strapped as their standard testing conditions.

Yes, there are several modern testers on the market that use high voltages (eg the excellent Amplitrex AT1000), but they don't test many valves at the same time, limiting throughput.

If you are getting a custom rig built then keeping the plate voltage down will save you a lot of money.

[jpfamps]

I think main motivators for the AC parameter measurements are the valve "hi-fi" enthusiasts, who are obviously using stereo amps, and thus "need" to have both amps matched for performance.

You can imagine the catastrophic consequences to the stereo image if unmatched valves were employed.........

[nashvillebill]

I think main motivators for the AC parameter measurements are the valve "hi-fi" enthusiasts, who are obviously using stereo amps, and thus "need" to have both amps matched for performance.

You can imagine the catastrophic consequences to the stereo image if unmatched valves were employed.........

Almost as catastrophic as using common zip cord for speaker wire, of slightly unequal length (oh the horror!) or plugging their amp into a regular wall receptacle instead of the audiophool-grade $800 receptacle (gasp! oh the inhumanity!)

[R.G.]

I think I'll ask our local electric cooperative company to give me a quotation for installing solid silver transmission cables back to the primary generators, then I'll get started rewiring the house with silver.

[J M Fahey]

Don't forget to ask them to use a silver wire wound *generator* too; otherwise all your effort will be in vain.
Amd if they use a Nuclear Power plant, refrigeration water should not be cheap tap water, but bottled-in-France Perrier.