Jfet matching setup

[lloop9]

Hiya fellows, I'm going to do some jfet matching for some 2SJ103, 2SK170, etc. I would like to know if the below schematic is appropriate for my purpose?

[davada]

Originally Posted by lloop9

Hiya fellows, I'm going to do some jfet matching for some 2SJ103, 2SK170, etc. I would like to know if the below schematic is appropriate for my purpose?

I can't comment on this matcher because I've never used it. But Anatech, one of the moderators here, designed a pretty good transistor matching jig that also matches jfets. I have used it with great satisfaction. I believe Anatech posted it in a thread somewhere on DIYAudio. You can try searching for it or just PM Anatech.

The only advice I can give you is to match your transistors at the current and temperature you will be using them at and allow sufficient time for the transistors to reach a settled temperature. Do a rough match first and catalog the results then a second accurate match.

This work for me.

Cheers,

David.

[lloop9]

Thank you for your respond David, I'll try looking up for Anand's circuit.

cheers

[thaumaturge]

Good suggestions. I'd add one of the National temp sense IC's to your setup. A touch of Heatsink compound between the sensor and the test device. Take your readings at the same temp on each device you test. I had to teach that trick to the Racal Dana engineers who were hand selecting Transistor arrays for best offset drift on an exotic function generator. It works.
Doc

[davada]

Hi loop9.
It's Anatech. If you search for anand's you find much.

Good suggestion thaumaturge. I like the idea of the thermal sensor.

Anatech's matcher is a differential circuit. Two transistor are being matched at once.

I keep the transistors against one another so they are thermally coupled and let them self heat at room temperature.

Another aspect of matching is that each transistor may have a different current-temperature curve. Therefore it maybe beneficial to test them over a range of currents and temperatures rather than just one point. This insures the transistors will stay matched dynamically.

I have done point matching to 0.1% with anatech's jig.

One suggestion from Anatech is to take one of those foam packing peanuts, hollow it out a bit and place it over the transistors under test. This jig is so sensitive that even a small room temperature change upsets the differential balance.

Cheers,

David.

[Zero Cool]

Please post a link if you find it. I didn't find it.

[davada]

Hi Zero Cool,


I think it was in Fred's link for the Adcom power amp repair. Everybody was in on that one.

I'll see if I can find it. Better yet how about I just post the JPG of the jig.

I built this on a single perf board with the circuits back to back and used dip sockets and header sockets to plug the transistor into.
For the different polarities I just flip the board around.

The current set resistors are an example. Other values can be chosen for other currents. I used a header to connect the resistors and made the connection using one of more of those shunt jumpers you'd find on a hard drive.


Cheers,


David.

[davada]

Hi All,

Here are a couple of pics of my matcher.

The 10K and 100 ohm resistors need to be hand matched.
The nominal value is not as important

Cheers,

David.

[Conrad Hoffman]

Long ago National Semiconductor published an app note for building an FET curve tracer to work with a 'scope. It you have a scope it's a worthwhile thing to build, though it only does small signal FETs. It gives a single curve, not a family like the Tek and other tracers do. I don't think the circuit is on-line anywhere, at least I didn't find it, but I can probably dig it out and scan it if anybody's interested.

[Zero Cool]

Mega Thanks!