Hot Transistor while testing HFE

[chromesphere]

Hey everyone!

I decided tonite to hook up the small bear circuit for testing HFE of PNP transistors, in preparation for an order of GE trans that should be here soon.

I used a 2n3906 to test, and it gave me a varying Ic, starts at 300ma and goes down slowly, which with a 9ma Ib should equal, according to the small bear calculation: 33.33 gain (ic / ib = 300 / 9 = 33.33 hfe).  I soon realised, the transistor was getting ridiculously hot.   It melted 3 of the sockets on my breadboard (and burnt my finger).  Am i supposed to do the test quickly?  Is that normal?  Does this circuit not work for silicon transistor?

If i flip the transistor around (what im sure is the wrong way around) i get 0ma.

Any ideas? Thanks in advance for any help!
Paul

Small bear circuit (bottom of the page):
https://ww.smallbearelec.com/HowTos/FuzzFaceFAQ/FFFAQ.htm

2n3906 Datasheet:
http://www.fairchildsemi.com/ds/2N/2N3906.pdf

[smallbearelec]

the transistor was getting ridiculously hot...Is that normal?

No. Some mistake in your wiring is causing grossly too much current flow through the transistor. It may have been damaged at this point. Compare your wiring Very carefully to the schem and try to figure out what is not connected correctly. Also: Is the resistor 1 Megohm? Too low a resistance would cause trouble.

The circuit will work fine with silicon devices, but the leakage current will be zero. Germanium will leak anywhere from 10 to 1000 microamps or more.

SD

[chromesphere]

The resistor is definitely 1meg.  There must be a wiring issue, maybe one of my alligator clips is busted.  Thanks for the confirmation ill check it over tonite!
Paul

[Earthscum]

Try a battery... I cooked a couple germs. Basically, my power supply was pushing as much current through the transistor as it could. The Ge I used had quite a bit of leakage and low gain, I already knew from testing before using R.G.'s rig. It warmed up, gain went up... by the time I realized it wasn't going to settle on a number I pulled the plug... my PSU was cooking hot. The germ is toast. Did a second one right after that and put everything away. Those were questionable ones I tried,,, up until that point I had no issues, and ran through about half a dozen.

[chromesphere]

I can build an maestro FSH but this circuit is killing me lol

Thanks for the idea earthscum, you know i actually thought about that (too much current from the PSU).  I'm sure i've just stuffed up the wiring somewhere but ill use a battery instead this time.  I'm going make a small vero board, something permanent for testing this stuff.  For some reason whenever i use a breadboard i end up with small component burn marks on my finger tips 

Paul

[Earthscum]

I think the way it's intended to work (correct me if I'm wrong, steve) is that the battery's internal resistance will limit the current, where R.G.'s uses current comparison across a resistor, which limits the current to the transistor, making it safe to use with a PSU.

[chromesphere]

Well it makes sense to me.  Ill give it shot tonite or tomorrow nite.  Thanks for the explanation and help!
Paul

[PRR]

The circuit works by applying 8.5 micro-Amps (uA) of current to the Base. This gets multiplied by hFE. hFE of all small-signal transistors (old radio to snazzy preamp) is say 40 to 800. The Collector current should be 340uA to 6800uA, which is 0.3mA to 6.8mA. Transistor dissipation is current times the 9V supply, 3mW to 61mW.

So the worst-case could be 61mW. 2N3906 is rated 300mW. It should not cook. It should hardly be warm.

A Darlington transistor could get in trouble, but 2N3906 is not a Darlington.

Something is not wired like you think.

Yes, if the transistor has been so hot it burned socket, it's probably permanently toasted.

It would be more mistake-proof with some resistance in the Collector leg. 100 ohms will have negligible effect on hFE measurements, but will limit worst-case transistor dissipation to 200mW. This may be similar to the idea of using a battery: a not-fresh 9V battery has many dozen Ohms of internal resistance.



If you are testing 1950s transistors, some of which were rated 50mW, use 390 ohms. This will also work with lower-hFE 300mW devices but will add some measurement error on the very highest hFE devices.

[chromesphere]

Awesome PRR!  I'm going to play around with this circuit tonite and will put in the 100 ohm resistor.  The germaniums i will (eventually when they deicde to show up) test will be russian 90's GT108V's:

10V| 50mA| 75mW| Hfe 60-130| Kn 6dB| Fe 1MHz

I've heard these have "bugger all" leakage anyway, but i would obviously still like to test them for HFE (without them igniting).

Thanks again, i'll give the 100 ohm resistor a try!
Paul

[chromesphere]

Ok, wired this up on vero tonite, left off the 100 ohm for now as i'm thinking i just stuffed something up.  All looks good.  Plug in a new 2n3906.  Its not getting hot, but im just getting zero microamps on the MM.  checking polarity of everything etc.  All looks ok.

I even tried my only current germanium aprehensively.  Just getting zero microamps.  I'm now thinking that this multimeter has issues reading microamps.  I've got it set to the 2000 microamp scale (the smallest). 

Theres a link below to the multimeter. You can zoom in and see it.  Tried swapping the cables around on the MM incase that was the problem. Just dont know why its not working...I see no reason why this MM couldnt read microamps but i'm stumped....

http://dicksmith.com.au/product/Q1459/digital-palm-multimeter

[chromesphere]

Ok, all this for a blown fuse in the MM!!!  Cant believe it! 

I very rarely use current settings and then only for low current pedal stuff...well, there you go. 

im now getting 2.2ma for collector current = 2200ua / 9ua = hfe = 244.   According to the data sheet the 2n3906 has a gain of 100 - 300 with an IC of 10ua. 

So why am i posting this if i got it going? because im excited that i just learnt something

Thanks for everyones help!

Paul

[Earthscum]

Ok, all this for a blown fuse in the MM!!!  Cant believe it! 

I very rarely use current settings and then only for low current pedal stuff...well, there you go. 

I feel for you! I did this exact same thing when I got the results I mentioned above. I even went to HF and got a new DMM (the transistor test socket went to pot on my other cheapy, anyways) after blowing the fuse in both the Centech and my Craftsman. I got home and popped the fuse right away, so I went back to using a resistive load.

im now getting 2.2ma for collector current = 2200ua / 9ua = hfe = 244.   According to the data sheet the 2n3906 has a gain of 100 - 300 with an IC of 10ua. 

So why am i posting this if i got it going? because im excited that i just learnt something

Thanks for everyones help!

Paul

Good to read through... someone else may come across it and remember to check the fuse. I'm curious, though, why we were cooking with transistors. Like you, I have everything working fine now, but really not sure what went wrong.

I'm wondering if there isn't some kind of rig a person could come up with to use an Op Amp, like the Jfet tester. Really, your base is all that changes between readings, and if you could get a circuit to self balance leakage and gain at a voltage reading relative to the true gain, that would be slick. One reading, just wait for it to stabilize. I know... K.I.S.S. and all, but I do like R.G.'s jfet matcher. It comes in handy for way more than just matching for phasers.

[chromesphere]

Well, that sort of thing is way beyond my current abilities.  Im high on success for just getting this going

Yeah i have no idea why it cooked...I checked over the circuit 10 times.  Only thing would be if the 1m was removed from the circuit (somehow), then you'd just be getting full current through the transistor?  I seem to recall my first readings where in the 300ma range, enough to burn sockets in my breadboard
Paul

[PRR]

> Just getting zero microamps.
> a blown fuse in the MM!!!  Cant believe it!

Believe it. This happens a lot. Usually by (mistakenly) reading the current of a wall-outlet or other potentially HIGH-current source.

> im excited that i just learnt something

Here's another learning. Always test the meter. DC V can be tested against the nearest 1.5V or 9V battery. AC V can be tested against an AC wall-wart or (carefully!!) a wall-outlet. Ohms test against any handy resistor or damp skin.

But how to test Amps? The most handy way is a power source and a resistor. Take a 9V battery and a 1K resistor, wired in series. Test the battery, test the resistor. Now try the amps of the meter. This should read 9mA. (9V and 9K would read 1mA; 9V and a common 10K would read 0.9mA.) If the battery and the resistor and the connections are good (and the meter switch and jacks), but it insists on "zero", the meter is sick.

[chromesphere]

But how to test Amps? The most handy way is a power source and a resistor. Take a 9V battery and a 1K resistor, wired in series. Test the battery, test the resistor. Now try the amps of the meter. This should read 9mA. (9V and 9K would read 1mA; 9V and a common 10K would read 0.9mA.) If the battery and the resistor and the connections are good (and the meter switch and jacks), but it insists on "zero", the meter is sick.

Funny you say that Paul, thats exactly how i came to the conclusion it was the meter!  I tried 100k with 9v batt "still nothing".  Maybe the range is still too high...10k, nothing, 5k, nothing, 1k nothing, 500 ohms nothing.  Ta-da!  The MM's got issues   I placed some wire across the fuse to get it going last night, obvioulsy going to have to fix that soon or next time i will need a new MM

Thanks again for your help!

Paul

[Jdansti]

Glad you got it going. I too have had the blown fuse issue with a meter. Here's something that an EE taught me that I do religiously:  If your meter has an on/off switch separate from the function switch, always set the function switch to the highest AC volts setting when you turn it off. That way, when you turn it on, you won't be on a setting that could damage the meter if you connect the probes to a voltage source without checking the function switch first.