Jfet pin ident?

[AndrewT]

Hi,
holding a To92 Jfet with label on flat side towards you and pins pointing downward. which pins are D, G, S?

DSG according to Fairchild 2n5457/8/9 & 2n5460/1/2
DGS according to Nat Semi BF244
DGS according to Toshiba 2sj/k 74, 108, 170, 369
GSD according to Fairchild BF245
GSD according to ONsemi BF245 (style 23)
SGD according to ONsemi BF244 (style 22)

I have tested Idss of my BF244 BF245 since these appear to have Drain on the right with S&G commoned for the test.
But are Nat Semi & Fairchild correct with Drain on the left.

How do I tell which pin order I have?

[KP11520]

Hello Mr T,

How are you?

I just recieved 100 Fairchild 2N5484s last week. I am checking the Idss (as you know) to group in matched pairs. The 2n5484/85/86 all have the same pinout.

Using the positioning as you describe (legs down and flat side facing me) they are D,S,G from left to right.

Funny, I built the device you shared with me on a stripboard to test the Idss with the 1K resistor and now my DMM won't range down to X.XXX volts DC. I hate when that happens!

I hope this helps take some of the mystery out of your question.

Regards//Keith

[jackinnj]

Originally posted by AndrewT

How do I tell which pin order I have?

Andrew:

If it smokes, you have the wrong pinout.

[KP11520]

Inhaling is optional!



Actually... We should be thankful for the Internet AND Datasheets!

[AndrewT]

I know what to expect feeding upto +25V into D with S+G commoned.
What kind of weird voltages will I get if +ve goes to S and D+G are commoned.
What happens if G goes to +ve and D+S are commoned?

Can this be done non destructively? i.e. feeding it through a CCS? 2mA or 5mA or more?

[Mooly]

Hello Andrew,
Weird things FET's. You can find the gate easily enough, it's identifying Drain and Source. Is that what you mean. In a lot of application it works either way round as well, but that does not mean it is correct of course.

[Mooly]

What about rigging up a resistor and LED as a load in the drain side, source to 0 volt. Connect the gate via a high value resistor, 1Meg say to a pot and have the pot so it can go both neg and pos on the wiper (2 PP3's maybe) and check the characteristics, pinch off volts etc. Would that work. When I have idley been playing with FET's I know you can fit them either way round and they still work in a lot of circuits.

[AndrewT]

Hi Mooly,
I agree, weird.
I thought I had tested the Idss of all my 244a&b & 245c. I assumed ONsemi pin outs and they seemed to measure OK.
So wired a cascode verson to see how it improved the dynamic impedance:- 2.3mA 244a as the CCS and a 7mA 244b as the cascode.
But it didn't pass any current.
Now I'm beginning to think I did all my Idss with the pins wrong.
Can't fathom out which way round to test.

Identify the gate?

[Mooly]

Hi Andrew,
This got me thinking, and I have not come up with much, just confirmed the "weirdness".
I rigged up a FET, the good old 2n3819 connected directly across 12 volts with a milliammeter in series with the drain. The gate was connected via a 6.8meg resistor to a pot that would swing the gate to + 12 volt through to - 12 volt relative to the source.
At any given drain current I could physically turn the FET 180 degrees so the source and drain were reversed and the current flow was absolutely identical.
I rather rashley said it's easy to identify the gate, well it is , sort of, but I don't think you would approve my method. I get an LED and 10K and a PP3 and a high value resistor. Just try combinations of connections so that the LED can be made to "flicker/ go on/off" by touching or just bringing a finger near the high value resistor which is connected to what will be the gate.
I googled "testing FET's" and that turned up a bit, but I checked the method given using an ohmeter and found it did not work. The troble is the FET can "absorb", is that the right word, charge from your meter and this gives misleading results. One of my FET's read 70mv (0.070) on diode test any way round, but it works perfectly, it's a good part.
Someone must have the answer to this.
Andrew want's to know.
I wan't to know.

[AndrewT]

Hi Mooly,
thanks for confirming why I'm completely lost.