Matching Tubes for Mutual Transconductance

[TD_Madden]

I've got a newly-calibrated B&K 707 tube-tester that supposedly tests "mutual transconductance".

I've usually used it to cull out dying/marginal tubes from my various stashes of tubes, but have been wondering if there's any value to using the meter's numbers to "match" output tubes; is gm a good thing to match?

Does this make any sense at all or just a dumb idea?

Thanks in advance for any help.

[Alex/Tubewonder]

Absolutely recommended. When you match for plate current at one specific bias point it's all you get: a match at that one point. Matching transconductance will tell how closely the tubes will track over their operating range. Very simplistically put of course.
More info HERE and you can find a lot more, just google "matching tube transconductance"

[Bruce / Mission Amps]

This is, of course my opinion:
As far as power tubes go... matching is a bit over rated.
Not considering Hi-Fi amps or high end audiophile equipment where sound reproduction is paramount, most guitar amps sound great with mildy mismatched power tubes.
Generally speaking, if the power tubes are within 5-10% of each other 99% of all players can't tell there is anything amiss and in many cases the darn things sound great being a little unmatched.
Adjust the bias voltages so the hotter running tube is not too hot-mushy when driven hard and the cooler one is not so cold as to be driven too easily into cutoff.
Amps with well balanced phase inverter preamp tubes seem to sound better with well balanced sections.
I have an rediculous number of vintage tube testers, including the billion dollar Hickoks... even when matching vintage NOS power tubes using their TC scales, the power tubes still don't idle at the same current... but I don't care if they fall into that 5%-10% idle current rule.

[TD_Madden]

I feel really dumb asking this, but have wondered about it for awhile....

I've never really thought about "hot" or "cold" bias....mainly because I've used B+ times current to ascertain output-wattage (trying to stay somewhat below max output).

For example, I usually bias my Mission Super (with TAD 6L6GCs) to about 17 watts each...using two meters and a calculator (402vdc x .042 gets me close)

That said, what does "hot"/"cold" really mean.....which is MORE negative?

[Amp Kat]

I think it's more of a slang than anything. But of course cold meaning so low of current draw that your nearing crossover distortion or even total cutoff where hot is at the point reaching total or 100% dissipation. Just like the term over and under biased can be confused and taken the other way very easily. I remember MKB gave me some info one time on transconductance testing that was really awesome stuff so maybe he'll chime in . I'm the same as Bruce with the tester as long as it's in the good area and there isn't any filament or cathode to grid shorts it's just a guessing game at that point as to how long they will last. It's the same as asking will my transconductance go up right before my tubes go because if you think about it don't they draw more current at that point so wouldn't the gm go up at that point ?

[TD_Madden]

that's what i thought, but decided to ask anyway....

thanks!

[TD_Madden]

actually it WAS sort of a dumb question since I don't even monitor the bias-voltage...I just watch the current-draw and the B+.....perhaps a better way to phrase it would be:

does increasing the negative voltage of the bias raise the current or lower the current draw?

Anyway, thanks for the heads-up.

[hasserl]

does increasing the negative voltage of the bias raise the current or lower the current draw?

More negative bias voltage (WRT ground) decreases current draw across the tube, less negative bias voltage increases current draw across the tube.

[stokes]

I'm with Bruce on the slight mis-match being desireable in guitar amps.We like more distortion in our outputs than Hi-Fi guys will tolerate.In fact what many of us consider "clean" will actually show some level of distortion on a 'scope.When you consider that most components in a typical guitar amp are +/- 5%-10% and more in a vintage amp,why would having perfectly matched power tubes be necessary?I have always been happy with a 5ma mis-match in output tubes.

[R.G.]

I do a monthly column for Premier Guitar magazine called "Tech Views".
My May 2007 column was on tube matching. The limits of a one-page column keep a lot of the more detailed stuff out, but it may serve as a quickie guide.

You NEED the DC bias currents to be matched for your output trannie to work properly.

Biasing the output tubes to just enough bias current keeps crossover distortion from making your amp sound harsh or sterile.

There is a wide range of "OK" bias points from the minimum to kill crossover up to the tubes running too hot or pulling too much power supply current.

DC bias point is a DC matching issue. Many cheaply matched tubes are matched for DC bias only. The more expensive matching probably matches them for tranconductance as well. You can think of transconductance as the AC gain of the tube. Matched transconductance is good for lowest distortion, but mismatched transconductance may sound good if not accurate.

But this is guitars, so accurate has to take a back seat to sounds good.

There is yet another layer of matching that could be done, matching the VARIATION of transconductance over the active range of the tube. I don't think anyone does this, but it is possible. I think...

[bob p]

... which reminds me ...

Geo could really use a project to build an automatic tube curve tracer. I know I'd build one.

[R.G.]

AAAAAAAAAAGGGGGGGGGGGGGGGGGHHHHHHHHHHHHHHHHHH!!!!!!!!!!!

(he screamed...)

The book on PCB layout is just about to ship. Can't you guys just read that for a while????

My irons in the fire have irons in the fire.

[Bruce / Mission Amps]

I feel really dumb asking this, but have wondered about it for awhile....
For example, I usually bias my Mission Super (with TAD 6L6GCs) to about 17 watts each...using two meters and a calculator (402vdc x .042 gets me close)

That said, what does "hot"/"cold" really mean.....which is MORE negative?

HOT = high idle current, low bias voltage
COLD = low idle current, higher bias voltage
Underbiased = HOT, less bias voltage... closer to zero
Overbiased = COLD, deeper negative bias voltage... further negative from zero

I run my MA 5F4's JJ6L6s at around 50ma to 54ma each but with an American made 5U4GB.
Another local runs his MA 5F4's NOS 6L6GCs at around 34ma-36ma with a GZ34. He likes this better then mine.
His sounds a little more modern to me and slightly brighter, snappier.
We both have WeberVST P10RTs (10A100T).

[TD_Madden]

may try that US-made 5U4 in the 5F4 as I have a few of them. The CC GZ34 I'be tried is a bit too stiff.

Got Sylvania, GE, EH, Sovtek, JJ, Shuguang, and the TAD 6L6s (gm-labled) to work thru.....and a handful of NOS 6L6GWBs as well.

It'll be fun trying them all!

[stokes]

... which reminds me ...

Geo could really use a project to build an automatic tube curve tracer. I know I'd build one.

Bob,I had a discussion with someone on a different board some time ago about using a transistor curve tracer (which go fairly cheap on ebay) as a tube curve tracer.He recomended the Tek 575 or 576,he has actually converted both.I had a passing interest in doing it,but got distracted and never went back to it.One day,when I have time.