EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Buriedcode on January 14, 2019, 11:57:04 am
-
Hi,
Bought some N-JFET's off ebay - I know I know, but the only other place is mouser. I messaged the seller beforehand asking if they were fake, and of course they replied that they were genuine, and he/she had used them in guitar pedals "for years".
Well, I still had my doubts so I plugged them into my generic "T4 transistor tester" (same one Dave tested in his blog)( and lo and behold - it shows a bipolar NPN with a Vbe of 6.59V and gain of 181. These are both pretty standard values for a generic NPN bipolar. Its pinout shows pin1 - emitter, pin2 - collector, and pin3 base. As that transistor tester can occasionally get things wrong, I did further testing myself.
Using the diode tester on my multimeter, I show a ~0.7V drop between pin 3-2 (base - collector), and pin 3-1 (base emitter), which I would expect for a bipolar NPN. Connecting pin 3 and 2 together - if its a J201 thats gate and source - and hooking a 9v battery between drain (+) and gate/source (-) via an ammeter I get ~ 0.4mA. If its an NPN, this would be collector and base connected together, and the Vbe reverse biased with 9V (!).
Am I correct in being certain these are just bipolar NPN's, re-stamped as J201 FETs? Wanted to ask if there are any further tests I can do before I tell the Ebay guy the bad news.
-
Not necessarily fake, JFets aren't an insulated gate device so you will get G-S and G-D readings on a diode test that look somewhat similar to a bipolar device
https://www.youtube.com/watch?v=F-o97JCQtes (https://www.youtube.com/watch?v=F-o97JCQtes)
-
Short the gate to the source. Then measure the resistance between the source and drain in series with a 1kOhm resistor using a basic multimeter.
You should read somewhat more than 1kOhm.
-
The simple diode test could be rather similar to an NPN. Even than one would already see a small difference: With a JFET the forward voltage for gate to drain/source is the same or very close for a symmetrical FET. With a NPN transistor the doping of emitter and collector is different and thus the voltage drop would be different.
The test with the battery is pointing towards a JFET and make an NPN very unlikely.
With a JFET the circuit is the simple constant current circuit. With an NPN transistor one would get no current with base connected to emitter and the positive side to the collector. In the reverse case with collector to the negative side one would get the collector to base brake down. This would act as a kind of zener diode - so still possible to get 0.4 mA with a breake down voltage just close to 9 V by chance. So if at all there would be small chance to mistake the gate for the base.
If it's a JFET it would act very much the same with pin 3 connected to pin 1 (source and drain are essentially interchangeable) . A NPN just block the current in this case.
With only 0.4 mA DS current the diode test might consider drain to source as high impedance, just because it can't drive the usual 1 mA test current. This might have cause the confusion.
Anyway it would be good to have an additional current limiting resistor (e.g. 1 K) in this test.
To be very sure, just build a simple circuit, that should work different with a JFET. A simple such test would be drain to +9V and gate and source via 10 K each towards ground. The voltages over the 10 K resistors should tell the difference.
-
Short the gate to the source. Then measure the resistance between the source and drain in series with a 1kOhm resistor using a basic multimeter.
You should read somewhat more than 1kOhm.
Reads ~2.05k with a 1k resistor (which itself read ~985ohm). With a 10k reads: 10.8k either polarity.
If it's a JFET it would act very much the same with pin 3 connected to pin 1 (source and drain are essentially interchangeable) . A NPN just block the current in this case.
With only 0.4 mA DS current the diode test might consider drain to source as high impedance, just because it can't drive the usual 1 mA test current. This might have cause the confusion.
Anyway it would be good to have an additional current limiting resistor (e.g. 1 K) in this test.
To be very sure, just build a simple circuit, that should work different with a JFET. A simple such test would be drain to +9V and gate and source via 10 K each towards ground. The voltages over the 10 K resistors should tell the difference.
.
Just did, and bearing in mind that I'm going under the assumption these are real J201's, so the pinout should be : drain, source, gate. Pins 2 and 3 each have a resistor to 0V, pin 1 to +9V.
Source 10k resistor: 400mV, Gate 10k resistor, ~0V. Just for kicks I measured the current going through a 10k from +9V to pin 3 (gate, or base if its NPN) and it was 0.3uA.
I'm confused now. It sure doesn't look like a bipolar, but that 400mV is 40uA from drain to source with the gate grounded. To top it off, I did basic measurements with a multimeter:
ohms (infinite = >10M):
positive negative resistance
1 2 90k climbing from 60k
1 3 infinite
2 1 90k climbing from 60k
2 3 infinite
3 1 infinite
3 2 infinite
diode test (inf = out of range, about 1V on this meter)
positive negative result
1 2 inf
1 3 inf
2 1 inf
2 3 inf
3 1 0.709
3 2 0.704
I also have plenty of 2N5457's and 2N3819's to test, all of which show up as JFETs in the transistor tester. The 2N5457 with the 1k in series with gate/source tied measured ~1.2k.
The same JFET with 10k resistors from gate and source to 0V measured ~1V across the source resistor, and 0V across the gate. It also reads ~0.7V from gate to source or drain with a diode tester.
So.. it seems it probably is a JFET but with different specs to the 2N5457. I expected Vgs of >2V and Idss of mA but its 10x less than this at <100uA.. I know JFET's specs vary wildly but this is ridiculous.
The other thing that made me suspicious is the markings. The face of the TO-92 is rough, like its been sanded, and the markings are way off centre. I've seen this before with fake (as in mislabeled) MOSFETS. It could well be another JFET that has been re-marked, apparently this is quite common and is mentioned in DIY guitar pedal forums.
I'll knock up a half decent test jig at some point so I can characterise every part I have. Thanks for the help lads!
Update: Found out I have a J201 model in LTspice. Just simulated several JFETs, and... they give pretty much the same results as the breadboard. It really does seem the J201 has a very low Idss (~0.3mA) and very low Vgs (about 0.8V).
-
JFET current source
-
Might be rejected parts. J201 is rated at 2mA and 0.3mA seems suspiciously low even for a JFET with wide spread. I've previously had some seriously hooky JFETs off ebay. J310 and MPF102s are the worst.
I've taken to sticking to a bag of 2n4416 and 2n3819 devices I managed to snag with a 1971 and 1973 date codes respectively. Bag is large enough for me to drop dead before I run out :D
-
Why do people often scream some component is fake, if they just don't understand how it is supposed to work?
-
The Fairchild Datasheet specifies IDSS at 0.2 to 1 mA for the J201. So 0.3 or 0.4 mA is perfectly OK. These are rather small JFETs with a low threshold and this can confuse the transistor tester.
Low threshold and low Idss usually come together. So the values are scattering, but it's mainly one degree of freedom. There are often graphs on how IDSS, threshold and R_on correlate.
-
Why do people often scream some component is fake, if they just don't understand how it is supposed to work?
Well I didn't "scream it" its why I posted, because I wasn't sure. And I've had plenty of fake (as in re-labeled) parts from fleabay - it's not completely unheard of. Perhaps I should have titled the topic "possible" fake J201's.
-
Short the gate to the source. Then measure the resistance between the source and drain in series with a 1kOhm resistor using a basic multimeter.
You should read somewhat more than 1kOhm.
Reads ~2.05k with a 1k resistor (which itself read ~985ohm). With a 10k reads: 10.8k either polarity.
Looking at the datasheet from Fairchild, the value of about 1kOhm you get for the drain to source impedance when Vgs=0 is within the specs for the JFet.
-
Why do people often scream some component is fake, if they just don't understand how it is supposed to work?
Well I didn't "scream it" its why I posted, because I wasn't sure. And I've had plenty of fake (as in re-labeled) parts from fleabay - it's not completely unheard of. Perhaps I should have titled the topic "possible" fake J201's.
A question mark would have done the job and if this were in the beginners section people would've been more lenient too.
Have you tried any real J-FET circuits with it yet?