My tube matcher circuit.

[mac]

I hate re tubing my amp. I have to find the better combination and placement of a quartet of el84's.
24 possible combinations if you are a perfectionist  , 3 if not. A annoying and slow process. 

So I came up with this idea, a diy circuit to test those sometimes unmatched "new matched tubes".
It is intended to work as close as the real thing, a laney lc30, but trying to be simple as well:



The general idea is to read plate, screen grid and cathode current of a pair of tubes. If I'm not wrong, unmatched cathode currents can be the cause of a tube going bad.
I just number the tubes and check them in pairs, and write down the readings. The currents across the 3.9 resistors should be within 5%, at least for my standars
It works for me. Last time I ordered from tubedepot the circuit detected one combination not so good, and it sounded really muddy.
The other pairs were ok, one really great.

I guess that it can be used to burn in tubes as well.
Ideas are welcome.

Question: to test 12ax7 for phase splitters, any simple circuit to check if the two halves are matched?

mac

[Jordan A.]

Question: to test 12ax7 for phase splitters, any simple circuit to check if the two halves are matched?

mac

Input a test signal and use a dual channel scope to monitor both outputs of the PI?

- General thoughts from a non expert -

Since cathode current is the sum of plate and screen current, you could skip one measurement in your tester.  Measure cathode current, then measure screen current and subtract it from your first reading, whatever is left (which should be the vast majority) will be plate current.  Either way seems ok i guess, you could loose two resistors if you tie the plates high.

If you wanted to get fancy you could build an adjustable negative supply for the grids instead of the cathode bias setup, it seems a bit more controllable to me that way.

How come not 1 ohm or 10 ohm resistors on the cathodes and 100 ohm for the screens?  It makes the math really easy that way.

You might check how much power the 1k pot is dissipating at various settings.

I kind of feel like the best tube tester is the amp itself though, you can do all these same measurements in the amp.  If you are not running your B+ too close to the maximum for your tubes you can just stick two in at a time to match up pairs.  1 ohm resistors between the cathodes and ground or Rk make it super easy.

If your amp is a pain to take apart and/or you don't have bias test ports then it does make sense to check the tubes outside of the amp i guess.

Fun with tubes!

jordan

[mac]

If your amp is a pain to take apart and/or you don't have bias test ports then it does make sense to check the tubes outside of the amp i guess.

Exactly.

How come not 1 ohm or 10 ohm resistors on the cathodes and 100 ohm for the screens?  It makes the math really easy that way.

Yes, I thought about 10 ohm, but if the voltages are similar then the currents are, so I used what I had in my box. Too lazy to walk 8 blocks to the electronic shop 

You might check how much power the 1k pot is dissipating at various settings.

The pot was a bad idea. It went very hot. I removed it and put 3 // 330 ohm = 110 ohm which makes the tube draw about 30ma.
My amp has a 56 ohm @ 9.5v, 42ma per tube. I guess the difference is B+, lower in my tester.

Since cathode current is the sum of plate and screen current, you could skip one measurement in your tester.  Measure cathode current, then measure screen current and subtract it from your first reading, whatever is left (which should be the vast majority) will be plate current.  Either way seems ok i guess, you could loose two resistors if you tie the plates high.

Yes.
But I thought that the sum of screen and plate currents could be equal for a pair of tubes, being the screens and plates currents different.
For example: for the first tube you read plate: 29ma, screen: 1ma so cathode: 30ma, ie, 29ma + 1ma = 30ma, and for the second 25ma + 5ma = 30ma. I don't know if in this case there is a sonic difference in the OT. That's why I read them all.

mac

[Gus]

I like this circuit.  I sometime build a fragment or the complete circuit to match and/or select devices at the voltages and current they operate at.

You can measure the voltage drops across all the resistors and nodes to ground and use Ohms law to calculate the screen currents and other currents.

One thing to keep in mind is that if the tubes have big difference in transconductance you will want to start at the lowest currents so not to stress one tube more than it's ratings.

You also might want to do a two or more point match maybe switch out one or two of the 330 ohms for less current and check the results kind of like a curve tracer.  You might match at one point in the tubes curves but not at other operating points.

[iccaros]

Just some thoughts

the load on your plates should match what you expect from your OT, IMO.. So if you have a 2K OT well the plates should match, you can place a 1R resitor inline and read voltage over that, it will drop almost nothing but a 1:1 current for voltage
your screens should match your amp, as the screen resistor is to protect your screen.

AX84 has a easy way to do ajustable bias
http://www.ax84.com/static/hioctane/AX84_Hi-Octane_Options_101004.pdf


since you are running two tubes together so its readings may be off, I would test one tube at a time

[Ronan]

Very good idea. Good for running in the tubes too, let them settle in.

No reason to run one tube at a time Steve, and the screen current will be at idle current level with no AC input, 82 ohms should be fine.

[mac]

You also might want to do a two or more point match maybe switch out one or two of the 330 ohms for less current and check the results kind of like a curve tracer.  You might match at one point in the tubes curves but not at other operating points.

Nice idea!

the load on your plates should match what you expect from your OT, IMO.. So if you have a 2K OT well the plates should match, you can place a 1R resitor inline and read voltage over that, it will drop almost nothing but a 1:1 current for voltage
your screens should match your amp, as the screen resistor is to protect your screen.

I am testing DC so I used the OT wire resistance (which is in the order of some hundreds of ohms) instead of its impedance.

No reason to run one tube at a time Steve, and the screen current will be at idle current level with no AC input, 82 ohms should be fine.

My amp has 82 ohm screen grid resistors, and 56 ohm cathode resistor for 4 tubes, that's why I used 110 ohm for 2 tubes.
This circuit is intended for a Laney LC30, but could easily be adapted to any amp I guess.

I'll draw Gus mods.

mac