Jfet Vgs Tester

[nickbungus]

Hi

I am looking for the best circuit to get a vgs reading for matching Jfets for phasers.   For six stages, this ebay China jobby has worked fine.

But for 10 stages, any mismatch can become audible to my ear so I've been looking more accuracy (more decimal places!).  I've tried building a few matching circuits from the www but the problems I seem to be having are:
1.  The vgs reading never remains stable (slowly decreasing in value), so I just have to switch on the circuit and take a reading after say 2 seconds.
2.  If I re-measure a Jfet I will get a slightly different reading using the same technique

So, theres the background, the question is, is there a recommended circuit or tool that will overcome these issues?

[antonis]

1.  The vgs reading never remains stable (slowly decreasing in value), so I just have to switch on the circuit and take a reading after say 2 seconds.
2.  If I re-measure a Jfet I will get a slightly different reading using the same technique

1. It's called "measurement stabilization"..
2.  //    //    "relative measurement error"..

[merlinb]

It's really not hard to measure Vgs(off), all you need is a 9V battery and your DMM. Presumably your DMM can measure at least 2 decimal places, which should be plenty. I get perfect repeatability and no drift with this method.


You can of course change the supply voltage to better reflect your intended circuit conditions, e.g. you could use 4.5V instead of 9V. Also, if you switch your DMM to measure current then the same setup will measure Idss.

[anotherjim]

I don't think you only need to match -Vgs, but also select for a lowest maximum. 9v powered circuits are not too happy if the -Vgs is large.

[nickbungus]

Excellent - thanks guys.

Merlin, I was tinkering with your method.  I'm getting much more repeatable results and a very little drift working at 3 decimal places.

I'll bash up a circuit with an adjustable regulator (to reflect the intended circuit) and some smoothing caps and have a play.

Thanks again

[merlinb]

I don't think you only need to match -Vgs, but also select for a lowest maximum. 9v powered circuits are not too happy if the -Vgs is large.

+1.
Or it may be that he needs to match Vgs and Idss, in order to get matched gm.

[amptramp]

There is the tester on the geofex site on the 04/16/02 entry.  (I am respecting R. G.'s request to not copy anything from his site) but it is located at:

www.geofex.com

[Tobyk]

It's really not hard to measure Vgs(off), all you need is a 9V battery and your DMM. Presumably your DMM can measure at least 2 decimal places, which should be plenty. I get perfect repeatability and no drift with this method.


You can of course change the supply voltage to better reflect your intended circuit conditions, e.g. you could use 4.5V instead of 9V. Also, if you switch your DMM to measure current then the same setup will measure Idss.

I've seen a lot of different versions of Vgs(off) testers, some with shunt resistors, some with higher voltage, switches and buttons. Maybe we should collect them all in one place for evaluation. But I tend to trust Merlin, and this tester is indeed very simple. It would be great to have Merlin's version of an Ids tester as well!
The agreed on one seems to be gate and source connected to a 100 or 10 ohm resistor to ground. Check voltage across the resistor and apply Ohm's law to calculate the current, which will be Ids.
Curious if Merlin agrees with this.

[Mark Hammer]

Over time, my view on matching JFETs for phasing has changed, or rather "relaxed", a bit.  If one is aiming for ultrawide sweeps that can go from very low to very high, then matching IS important, since you don't want one or more JFETs to crap out on you during the intended sweep and stop changing D-S resistance.

If, on the other hand, one's typical use of phasing is a pleasant medium speed, or even faster speed, "wobble", that neither sweeps very high or very low, a great many JFETs will be up to that task, straight out of the parts bag, especially if part of the same batch you ordered.  What one is aiming for is not that the D-S resistance be identical in each JFET, but that they are all responsive and change D-S resistance with the range of current applied to their gate by the LFO.  A bias trimmer will set the range, and if the JFETs can't manage the sweep intended, just insert a bit more resistance between the LFO output and gates.

Some folks like phasers to provide drama, and some use them to provide texture.  Drama needs matching,but texture often doesn't.

[Tobyk]

Over time, my view on matching JFETs for phasing has changed, or rather "relaxed", a bit.  If one is aiming for ultrawide sweeps that can go from very low to very high, then matching IS important, since you don't want one or more JFETs to crap out on you during the intended sweep and stop changing D-S resistance.

If, on the other hand, one's typical use of phasing is a pleasant medium speed, or even faster speed, "wobble", that neither sweeps very high or very low, a great many JFETs will be up to that task, straight out of the parts bag, especially if part of the same batch you ordered.  What one is aiming for is not that the D-S resistance be identical in each JFET, but that they are all responsive and change D-S resistance with the range of current applied to their gate by the LFO.  A bias trimmer will set the range, and if the JFETs can't manage the sweep intended, just insert a bit more resistance between the LFO output and gates.

Some folks like phasers to provide drama, and some use them to provide texture.  Drama needs matching,but texture often doesn't.

I don't use them for phasers but as gain stages. My main goal with checking them is to be sure they are within specs, as Chinese copies are almost the only available ones nowadays. I have less interest in matching them.

[merlinb]

It would be great to have Merlin's version of an Ids tester as well!

Just switch the meter to measure current (milliamps), no other changes required!

[Mark Hammer]

Well that's different.  Please ignore me, then.

[Tobyk]

It would be great to have Merlin's version of an Ids tester as well!

Just switch the meter to measure current (milliamps), no other changes required!

Ha, that's about as simple as it gets, thanks Merlin. I wonder why everyone else feel the need for shunt resistors here.

[Rob Strand]

Ha, that's about as simple as it gets, thanks Merlin. I wonder why everyone else feel the need for shunt resistors here.

Plenty of reasons

Anytime someone uses a DMM on current on a regular basis they will end up blowing up the meter fuse by measuring voltage in current mode.   

Also, for a JFET tester, a shunt resistor can be used to limit the current if the user happens to insert the JFET the wrong way around - which *is* going to happen.