best way to match tubes?

Someone wrote, current draw in different conditions, not just matching trasnconductance/gm
is the best way to match tubes.

Is this a good idea? How is this done?
My usual priority, at least with push pull amps & power output tubes, is to focus first on current, then Gm. I prefer to measure the current across a low value wire-wound resistor I've placed in the cathode path, as that will capture current that is going to the screen as well as to the plate. This is more important if you're using them in ultralinear or triode mode, where both currents will be going through the transformer primary. I've seen tubes that were matched on plate current, but turned out NOT to be matched in use because the screen currents weren't the same (sucks!).

Of course, if your amp is a McIntosh, you can't use the resistor in the cathode circuit because they take part of their power off the cathode. But anyway...

The goal, of course, is to minimize the imbalance across the transformer primary caused by tubes that pass different current. The imbalance shows up as hum or distortion (one half of the waveform is amplified more than the other).

Of course, if your amp allows you to individually bias the output tubes, then it's probably better to focus FIRST on the transconductance because you can bias each tube to match on current.

So the answer is... it depends!

I keep an old Dynaco ST-70 amp around for matching tubes in an amplifier (under use). It has a modified bias circuit, adjustable bias for each tube, and 10-ohm resistors on the cathodes. You can measure 6L6's or EL34's or KT88's etc. in this setup, under actual operating conditions. Of course, you can also measure current across the screen resistor as well, so you can figure out the plate current and screen current. Anyway, with plate voltage around 400volts, if the tube can draw around 50ma or so with a nice high negative bias of -45 or more, you know it's a good strong tube. You can also look at how well matched each tube is, based on it's ability to pass current at the same level of bias. That is, keep the bias voltage constant and measure the plate current on each one. I've gone thru my whole collection of octal tubes and matched them up this way. Sounds great when I get them in amps.

Hope this helps!

A vacutrace (sp?) from looks to me to be about the ideal. Have not used one.