Same transistor curves, but one is broken?

[BlownUpCapacitor]

I am repairing an SC502 TM500 series oscilloscope. One channel was faulty, having a lot of DC offset where you can't even center the trace. This was channel one, the faulty one. I narrowed down the problem to Q125, a dual N channel J-FET.

I "knew" Q125 was faulty because since both channels are identical, I swapped channel 2's J-FET and put it in the socket for channel 1's. Channel 1 worked just fine now, and when I put the "faulty" J-fet into the socket for channel 2, channel two was now faulty.

But when I put the FET's into my 5CT1N, both transistors were IDENTICAL. I made sure to check both transistors in each 6-legged package. Sure, there are some minor differences like one has slightly more gain than the other, but I do not think this would make such a huge effect that the channel with the "faulty" transistor gets a massive DC offset.

I just wanted to ask before spending $30 on ebay for a single replacement transistor, is the J-fet really faulty or is it something else that's wrong in the scope?

[TimFox]

Can you compare the gate currents on the two dual JFETs?

[BlownUpCapacitor]

Not too sure what you mean, but I can provide some images of the curves of each transistor. The curve tracer I'm using isn't calibrated, and it's actually a bit out of cal, so just take this as a qualitative measurement.

Image 1: Good FET: side A
Image 2: Good FET: side B
Image 3: "Bad" FET: side A
Image 4: "Bad" FET: side B

Also yes, I made sure to use the FET function setting on the curve tracer.

[TimFox]

Your curves show the drain current as a function of gate-source and drain-source voltages.
The gate current (not measured here) should be low, but perhaps the defective part’s gate current is abnormally high:  that could induce a voltage offset.

[BlownUpCapacitor]

The current flowing through the gate of the good FET is 3.5 mA AC RMS, and the current flowing through the "bad" FET is 2.2 mA AC RMS. Would this minute difference cause a massive voltage offset?

[TimFox]

Those current values are huge for a JFET.
How did you measure them?
I was referring to DC current into the gate itself;  the gate-source and gate-drain diodes should be reverse biased in a normal amplifier circuit.

[BlownUpCapacitor]

I measured those values with a DMM in the mA range.

The curve tracer was using three steps, so the DC voltage would vary. This is why I used the AC current.

Also the gate current varies with the Drain to source voltage. A higher drain to source voltage on the curve tracer gives me a lower gate current.

[TimFox]

How did you connect the mA meter?
2 mA gate current is much too high for a properly operating small-signal JFET.

[BlownUpCapacitor]

transistor Drain and source go directly to the curve tracer's drain and source, then I connect the gate of the transistor to one side of the ammeter, then the other side of the ammeter connects to the gate receptacle of the curve tracer.

[TimFox]

Then, 2 mA gate current is far too large.
If you fix the gate-source voltage at, say, -2 V, what value and direction is the gate current?
By the way, what range of Vgs did you use for the curves?
For that application, you want Vgs negative-definite.
Your high AC current might be due to forward biasing something inappropriately.

[MathWizard]

Do you have a pic of the scope's schematic/circuit ?

ok it's an older Tektronix,  no doubt with a great manual.
http://manuals.chudov.com/Tektronix/Tektronix-SC502-Oscilloscope-1986-WW.pdf

[magic]

TimFox is talking about gate leakage current under normal operation, when the gate is reverse biased with respect to source and drain and should be conducting no current in principle.

Not sure what happened in your test, perhaps your curve tracer forward biased the gate and you measured that? At some unknown duty cycle? Quite meaningless...


I would crudely test it as follows:
Measure or guesstimate the reverse bias Vgs under operating conditions in the oscilloscope. Say it's 2V, as suggested by TimFox.
Apply 2V positive (N-ch) or negative (P-ch) to the source and drain together.
Connect the gate to ground through the 10MΩ input impedance of a DMM in volts range.
Measure the voltage across the DMM with the DMM itself.
Use Ohm's law to calculate gate leakage current.

[BlownUpCapacitor]

After doing every single test possible to determine if the transistor was defective or not, non of them revealed ANYTHING out of the ordinary.

But I figured out why.

The transistor was inserted in the wrong direction

Originally, the transistor was inserted incorrectly, so I assumed that to be the correct orientation. Turns out not. 

One day I had had enough BS from that FET, so I decided to try many possibilities including using BJTs and inserting the FET in a 180-degree orientation. And voila, the stupid thing works perfectly now with two fully functional channels.