Proper way of testing for jfet IDSS?

[Pojo]

So I have a batch of Vishay 2N5457's that I suspect are out of spec. I used the Fetzer Valve method of testing and they show Idss between .2mA and .6mA. I contacted the vendor who said they tested a bunch from their stock on my behalf and gave a reading of 1.03mA...which is the lower boundary of the spec...but in spec. They also shot me over a pdf with the circuit they used. Here's a screenshot:



Here's a datasheet to see the specs: http://www.onsemi.com/pub_link/Collateral/2N5457-D.PDF

Looks like they're using +15Vds and -20Vgs. But the spec in the datasheet shows it at 15Vds and 0Vgs...could the -20Vgs be skewing their results? I don't have a supply capable of providing either 15VDC or 20VDC, just the common Boss style adapters which give me 9.6VDC so I can't reproduce the same test. I've always been using the method outlined here:

http://www.runoffgroove.com/fetzervalve.html#11

This seems to work for me but I have no way of actually verifying it. I have a bunch of J201s and some 2n5457, 2n5458, MPF102 from previous orders and all of them seem to be in spec when I measure with the ROG method...the Vishay's from the batch in question are the only ones that show out of spec results.

Thoughts?

[R.G.]

"Idss" literally means "the current through the drain with the gate shorted to the source".

The current in a JFET channel is relatively fixed by the Vgs once you're over a few volts and significantly below the channel breakdown voltage. So if you short the gate and source, apply a voltage of more than a few and less than the breakdown, the current you read is Idss, by definition.

If you put a resistor in series with the drain or source, the only way this affects the reading is if the voltage dropped across the resistor forces the JFET voltage to be less than "more than a few volts". With that in mind:

So I have a batch of Vishay 2N5457's that I suspect are out of spec.

What leads you to suspect that they're out of spec?

I used the Fetzer Valve method of testing and they show Idss between .2mA and .6mA. I contacted the vendor who said they tested a bunch from their stock on my behalf and gave a reading of 1.03mA...which is the lower boundary of the spec...but in spec. They also shot me over a pdf with the circuit they used. Here's a screenshot:
...
Looks like they're using +15Vds and -20Vgs. But the spec in the datasheet shows it at 15Vds and 0Vgs...could the -20Vgs be skewing their results?

If they're using -20Vgs, all of their JFETs would show zero current on that test; a Vgs of -20 would shut all normal JFETs off entirely. It's not a subtle error. So (1) yes, if they are using -20V for Vgs, they're doing it wrong and (2) it's highly unlikely that they were actually using -20Vgs, because none of the JFETs would pass as the drain currents would all be zero. So this does not compute. Either you misunderstood them, they told you wrong, or some other disconnect happened.

I don't have a supply capable of providing either 15VDC or 20VDC, just the common Boss style adapters which give me 9.6VDC so I can't reproduce the same test.

Doesn't matter. The exact voltage on that test is not critical, and almost does not matter. You could use two 9V batteries in series, as an example.

I've always been using the method outlined here:
...
This seems to work for me but I have no way of actually verifying it.

That's an OK test. It's better for Idss than Vp, but still OK for what it's used for. The voltage across that 100 ohm resistor is an OK measurement of Idss. Notice that the tolerance of the resistor will skew the result, as will your meter's accuracy at low voltages, so the accuracy is into the lower ranges, but then high accuracy is not needed for this measurement.

I have a bunch of J201s and some 2n5457, 2n5458, MPF102 from previous orders and all of them seem to be in spec when I measure with the ROG method...the Vishay's from the batch in question are the only ones that show out of spec results.

Trust the ROG test first. There is something grossly wrong with the "test circuit" from the vendor, or how it's annotated.

[Pojo]

You're the man RG! Thank you for your response 

What leads you to suspect that they're out of spec?

I'm measuring Idss from .2mA to .6mA on every transistor from a bag of 20. And actually, looking again through my notebook, I see the highest measurement was only .47mA. The datasheet shows 2n5457 as having a minimum Idss of 1.0mA and a maximum Idss of 5.0. The readings I'm getting seem to be more in line with a J201 (min .2mA, max 1.0mA) than a 2n5457.

If they're using -20Vgs, all of their JFETs would show zero current on that test; a Vgs of -20 would shut all normal JFETs off entirely. It's not a subtle error. So (1) yes, if they are using -20V for Vgs, they're doing it wrong and (2) it's highly unlikely that they were actually using -20Vgs, because none of the JFETs would pass as the drain currents would all be zero. So this does not compute. Either you misunderstood them, they told you wrong, or some other disconnect happened.

I'll pass on this info to them. They're an honest vendor that many here use and it shows a lot that they were willing to have their techs take the time to test and report back with their results and pictures. BTW, here's the pictures they sent along showing their test setup. It's too low res for me to see what's going on and too much is off screen. Maybe you might notice how they're getting more than '0' with a Vgs of -20V?



The left pic is them testing a Vishay 2n5457 and the right pic is a Fairchild 2n5457.