Tube Tester Advice Needed

[Vince]

Hi,
I need to be able to match pairs of EL34 types and 7591's.
I had hoped that the individual bias adjustments of the "D" version
of the diytube ST-35 would be adequate. However on this thread:

http://diytube.com/phpBB2/viewtopic.php?t=1555

I was told that I still had to start with closely matched pairs.

Could someone please tell what are the parameters that must be "closely" matched?. Additional, what tester will measure these parameters? I do not need a "Bureau of Standards" one. I would just like one that is adequate?
I bought a B&K 650 with the 610 accessory panel and it will not do the basic tests on 7591's and I don't see any "gm"(?) capability for any tube.
Thanks, for your help, Vince

[EWBrown]

Some of the Hickok 539 and 600 / 6000 series, and the military TV-7 can do mutual conductance. perhaps the Eico "satan special" model 666?

I have a Hickok 600 and a 539 is currently located down at my place in NC. I only wish I had a TV-7 (I well remember when these were dirt cheap...)

A quick and dirty matcher that I use for 6Bq5s, EL84s and those bargain Russian made 6P14Ps, is a single hand-wired channel "borrowed" from an ST35, in place of the 400 ohm cathode resistors, I use LM317s as current regulators set at 35 mA (36 ohm, 1/4 watt resistor). I can test two tubes at a time, let them heat up and stabilize for a few minutes, then measure the cathode (pin 3) voltage with a DVM. WIth that, it's easy to grade and match up the tubes, with excellent accuracy. Perhaps some similar test rig can be done for the 7591s.

For larger power tubes, I use a Heathkit IP-17 tube regulated variable power supply, an octal relay socket (it has eight screw terminals) a couple of wire-wund resistors to simulate a plate load and cathode load, and can measure a tube's dynamics with that set up. It's a bit more complex procedure, but allows for tubes that my testers don't (or can't) test, like those weird but vwery nice Russian power triodes like 6C33C, 6C19P, 6C41C, etc...

Typically I set up for fixed bias, use a 225 ohm 10 watt plate load resistor, and a 10 ohm cathode resistor (just to meaure and calculate the current), and a 100K 1/2 watt resistor betwen the bias voltage output and the G1 grid.

The PSU has variable B+ and bias (negative) voltages, and 6.3VAC @ 4A, which is enough filament power for most tubes. If I need other filament voltages, then I drag out a solid state vaiable regulated supply.

The quad Tung-Sol 6AV5Gs I gave to Shannon were tested and matched up with this very setup.

/ed B in NH

[Vince]

Hi Ed,
I sincerely thank you for your time and advice! It is exactly what I was looking for.

So here is what I think I should do.
For grading EL84 types, I build an original type ST35 mono test amp(I don't have a donor) as shown in this link:
http://www.geocities.com/vintageaudio/st35b.jpg
and replace R8 & R9 with the LM317.

Not being the second coming of Tesla, is the LM317 you refer to this device:
http://www.national.com/pf/LM/LM317.html

How are they set for 35ma? Can I use a Fluke 8012A to set them?
How did you arrive at 35ma?

Does it have to have an OPT on it with speakers( or a dummy load)?
If so can a lower grade OPT be used?

Thanks again for you help!! Vince

[mesherm]

Here is good info on LM317s as constant current devices.

http://users.pandora.be/davshomepage/current-source.htm

R8 and R9 are grid stopper resistors. They need to stay.

An LM317/36ohm resistor combo is placed between the tube cathode and ground with the Vin terminal at the cathode. An electrolytic capacitor is also placed in parallel with the LM317/resistor for AC bypass. (use a 50 volt rating on the cap, with a value of 100 to 400uf)
Note that the original ST35 circuit had all four EL84s flowing through one cathode bias resistor/bypass cap. Its much better to give each tube its own bias resistor/bypass cap (or LM317/bypass cap). That way matched quads are not needed.

[EWBrown]

I just chose 35 mA as this is the normal cathode current for a 6BQ5 / EL84 / 6P14P, etc.

The LM317 can carry at least an amp, perhaps 1.5 amps, so 35 mA is just a sleepwalk for them, they won't need anyheatsinking.

The way an LM317 works, it likes to "see" 1.25VDC between the output and adjust pins, with "output" being the higher voltage level.

Using Ohm's Law, 1.25V / 35 mA comes out t 35.71428571428 ohms :o

Seeing that finding such a ridiculously precise resistor is totally unlikely, just use 36 ohms, which is a standard 5% resistor value. 1/4 watt is more than enough, and I've even used little bitty SMD resistors soldered directly to the LM's pins and it works. The "input" connects to the tube's cathode, and the "adjust pin" connects to ground / zero volts buss. Bypass from cathode to ground with 220-470 uF, 35VDC cap.

The LM317 has an upper input voltage limit, IIRC, it is around 35VDC, so this only works with tubes which operate with relatively low cathode bias voltages.

An alternate approach, if your junkbox has lots of 7805s but no LM317s:

Cathode to input, center pin to ground, load the output with resistance somewhere between 140 and 150 ohms @ 1/2 watt. (142.857 ohms being the theoretical perfect value, but I digress once again, Mister Spock). Bypass cap from cathode to ground. in the usual practice. . If you want to make the bias current "tunable", then use a 100 ohm resistor and a 100 ohm pot in series, as the LM7805 load. Then you can tweek anywhere between 25 and 50 mA.

There are higher current and voltage CCD devices available, I posted about them a few weeks ago... These can take up to 350 and 450VDC across them, so they should make nice CCD plate loads, as well.

I wouldn't run a nice vintage NOS Amperex or Mullard 6BQ5 / EL84 at 50 mA, but it's no "great crime" to treat one of the $2-$3 Russian 6P14Ps in such a "hostile" manner, for truly "hard" Class A SE or PP operation.

They're cheap and plentiful, so long tube life isn't such an economic issue under these conditions.
Oh, yeah, you'll need a HEFTY power trannie to get away with this...

/ed B in NH
.