bench matching rig?

[Andy Le Blanc]

A bud of mine mushed a 12BH7A in his old Macintosh and asked if I had a
replacement, I have a few, but its a stereo and I have to find the closest
match with the 12bh7 from the other side.

So I need to scab together a test circuit on the bench to find;

amplification factor

u = -d ep/ -d eg (-d= -delta or change... ip is constant)

plate resistance

rp = d ep/ d ip (eg constant)

transconductance

gm = d ip / d eg (ep is constant)

I'll want to use gm = u/rp

So..... I think I can do it with a variable 200v B+, a variable battery grid bias,
a signal generator in series with the grid bias, and a meter in the plate circuit.
I should be able to get static and dynamic characteristic.

I've seen forum posts that mentioned this and wondered how others have
approached this?

It would be a hoot to get some form of practical tube matching from bench scraps.

[FunkyE9th]

Go to http://www.pmillett.com/technical_books_online.htm and download

Basic Audio, Norman Crowhurst, 1959, Vol II

He shows you how to measure transconductance, Dynamic gain etc...

[Andy Le Blanc]

I all ready have it, thanks for the reminder.
I've got some mystery trannies in the bone yard, be good for a low draw power supply.
Maybe a fairly compact rig with two 120/ 12ct v PT, should give ac for the fil.
around 24 v dc for a variable grid, and around 200 to 250v B+.

[FunkyE9th]

With a VVR, you can make a nice variable power supply for testing.

[Andy Le Blanc]

Another solid state wonder, the features come and go, but never really help your playing. ........... , I dont have one ...

But I got a mess of bench scraps....

[Sonny ReVerb]

Andy, You could probably modify Steve Bench's RAT tester to meet your needs. His website has lots of cool info and projects: http://greygum.net/sbench/sbench101/

Look for the RAT gm/Mu Tube Tester Project about halfway down the page.

There's also a copy at Triode: http://www.triodeel.com/tester.htm

I've been meaning to build one - I actually bought the trannies and an enclosure, but unemployment put a kink in that plan...

[Andy Le Blanc]

I've got most of a chassis dressed today, providing the right static voltages etc....
If its like testing a opt it wont be bad at all.

I'm working so hard at being un-employed I don't have the time to look for a job.

If you can pull together the tools, you can have a job, the next two nights
are sound gig's, a few repair's and the cell phone's paid.

The only thing keeping the electric on is the insulin in the refrigerator.

[Andy Le Blanc]

So.... two 120/12 trannie's back to back worked great with no load.

as soon as there was a load it became very obvious that it was not going to fly.
so far I've got a good variable heater supply, it'll go from 14 to 8 vac with a
50 ohm large amp+ rheostat from an old speaker crossover and 1 120/12v.
I'm still going to try back to back trannies but as a low volt supply for the grid
voltage source. I had to run a doubler which reduces the already limited current.
I should get +/- 120 or there abouts with out it.

which leaves a HT supply, and as cheap as it can get.

Ive got an old PT for two 6v6, should be plenty of ma. I was surprised at just
how much draw this basic test requires. The problem is how to make a
variable supply with as few parts as can be had. I found an old audio express
with an article that uses a constant current source instead of a choke,
a MOSFET, the circuit is a direct analog to a tube CCS.
This set up is supposed to have a very high power supply ripple rejection.
It is also very close to a typical tube series regulator, but I think can be
tried with minimal parts. Max plate on a 12BH7 is 3.5w per plate.
Total dissipation for a 12AU7 is 5.5w. It looks like I can get away with a
12au7 as the CCS and be with in max ratings for tube type and application.
Most testers used a multi tap PT to accommodate variation in wall volts
and to provide the various voltages to the test circuit, if you've ever seen
the spaghetti dress inside of one, its a chore to dress.
This CCS uses one pot. to vary as a current sense resistor, or one muti-tap
switch and a fine adjust, simple and appealing, dial in the conditions for
the test, take you readings, do the math.

[Sonny ReVerb]

Sounds like you're making progress, Andy. I had to rebuild a Sencore MU-140 tester I got from eBay. I had to fix some broken traces on the PCB, replace a bunch of E-caps, and a cracked power resistor. I believe they used a multi-tap PT.

I don't think most, if any, traditional testers test the tube under actual operating conditions. Many don't test at full plate voltage.

There's always the Maximatcher if you get desperate. Alas, no 12BH7...

Here's one that handles the 12BH7.

These might give you some ideas, anyway. Keep us posted.

[Andy Le Blanc]

Ive looked at both, with envy, but I'm cheap.
It would be be great to make a go of it with bench scrapping.
Hopefully it be an easy project when all the rig-a-ma-role is finished.

That transcendent tester looks good don't it.

[Andy Le Blanc]

Its funny how PS choices can make or break a project. Going to a more traditional
supply cured all of the projects ill's. Im going to have to yank one of the 12v
tranny's to make room for the PT but thats minor. The CCS will need more
tweaking but that will be fun. Its basically dressing a dual triode as one big
diode and using the grids to play with the current the whole thing passes.
I'm still going to have to pay close attention to the load.....
Its very close to a series regulator, anyway, its all easy common parts and tubes.
No sense reinventing the wheel with NOS.

So.... after the rectifier I've used a 12au7 as an extra diode. There's a 100 ohm
on the grid to prevent oscillation and a variable rheostat on the cathode.
Put a ma. meter between these and the load to set the ma..
There's another spot to do a very simple regulator on the center tap.
Again, its playing with the tube as a diode, this time between the ct. and ground.
The grid is given a potential from a divider between B+ and ground.
With this project the max draw of the tube under test is known, so tube
choice for this again is common tubes, providing regulation.
You could dress a large pentode to do the same for a whole amp, lots of
fun. I'll post scheme after a bit more. The goal is tube testing after all.

[Phil_S]

Nice job, Andy. You aren't the first to build your own rig and you won't be the last. It takes a special kind of person to even want to do it! Looking forward to seeing whatever documentation you post.

[Andy Le Blanc]

here's the general idea for the PS. Its just a big bias supply in the end.
Ive planned to switch between two plate voltages, there will be a 1 ohm
in the plate circuit to get a reading over. I tried to over complicate the PS
from the get go, except for trying out some simple tube regulation ides the
whole thing will be as simple as the scheme implies. A 12-25w PT should be
good. I can get B+ from 85v to 350v and it seems pretty stable.
The one thing that bothers me is that I can watch the line ac drift.
so.... I'll be noodling around..... the goal isn't perfection, more like usable numbers.....

[Andy Le Blanc]

Here's the component board for the power supply. Every thing except the
rolled eyelet's is bench scrap and has been recycled, at least once.
The board is old clipboard material. It is a natural hard fiber.

Physical properties that improve with age
High strength-to-weight ratio
High dielectric strength/excellent electrical insulator
Maximum temperature: 425 F

Its easy to machine with hand tools and old clip boards are every where for free.

[Andy Le Blanc]

It's beginning to progress....

Things that didn't.....

A simple attempt at a regulator on the center tap of the PT didn't.
Bypassed, but left in the chassis to noodle with later.

The voltage divider for controlling voltages didn't either.
Its not "just like a bias supply" at all. The current needed for the test is
5w worth of draw, it was nice to think it was a bias supply.

Things that worked out....

A 12AU7 dressed as a constant current source.
This works surprisingly well. What happens is that the tube under test becomes
the load for the PS. You use the CCS to adjust the current to get the
voltage over the load. This becomes your B+ adjustment.
I had planned to switch between a couple set voltages, switches , knobs....
I'm thinking that the CCS allows a great simplification of the PS and test bed.
It easily ranges from 150v to 300v dc.. I still need to see if the CCS will interfere
with the test.