Bipolar Junction Transistor Testing Basics

A big score on the Dutch "marktplaats" last week!

I will have to refurbish it, switches are acting up since it is supposed to be not used for many years, but it is still working!
It is a Philips PM6507. guess much less heavy than the Tek above but still 20 kilograms...
I was looking for something like this for a long time,you do not see them often.

pm6507.jpg
 
How about Petticoat Junction ?

Kidding aside, I think it important to mention something here. I have had this several times and that is partly because I was the dog shooter. The tough dog troubleshooting, after nobody else could fix it. I am not bragging here just telling you how I got to see as many of this as I did.

Transistors that check good on the meter but simply do not transist. when you are in the circuit and see the waveforms it looks like an open collector because the EB is pulling current. So you get in there and with a meter check the EC junction and it tests fine, open in one way and a diode drop the other way. Seems good as hell. But it ain't.

I am at a loss to explain how - physically - this happens to a transistor, but it does happen. Not shorted, not open, just no hfe. I can say, over a dozen times I have run into this. With all the work I have done, yes it is rare, but we are in a time now when this stuff is getting really old and we are going to see different and perhaps strange modes of failure.
 
It was always thought that transistors either "work" or "don't work". But not anymore!
As transistors age, it seems like they may fail in ways never seen or experienced before.
Hope this sticky continues to develop and document these events as they unfold.
Thanks EW, this sticky has helped me fix a few amps.
Robert.
 
The idea with the heatsink grease is to eliminate the air gap between the transistor and the mica insulator, and between the mica insulator and the heatsink (air is a piss-poor conductor of heat). To accomplish this, only a thin coat of the grease is necessary, as the grease itself isn't such a great conductor either (but it beats an air gap). Too much grease is as bad, or worse, than no grease at all.

Sil-pad is short for silicone pad. They are available from many sources, and vary considerably in their ability to conduct heat. The only ones I use (when I use them, which is rarely because of cost...they are terribly expensive [$2.30 each]) are K-10 pads made by Bergquist: http://www.bergquistcompany.com/tm_sil_pad_detail.cfm?oid=104264



The other problem with the Sil-pads are that, once they are used (compressed), they must be replaced if the part is later removed. Mica insulators are cheap, reusable, and as long as they are nice and thin, adding a thin coat of heatsink grease makes for pretty decent setup (I buy them by the 100 lot, and pick the thin ones to use. For the ones I deem too thick, I either use an X-Acto knife to 'split' them into something usable, or just toss them out).
Where do you get them on line?
 
From my experience, the best way to test the transistors is to take the time to study the circuit, calculate all the DC voltage of all the the points on the schematic. Calculate the current through the transistors.........Then check the voltage on the real circuit and compare to the calculation.

Lately, I helped a few people fixing their amps on the forum, never have a chance to see the amp and touch the test instrument. All I did was obtaining the schematic, spent 30 minutes calculating and write all the voltages down. Asked what was the symptom and look at the likely problem area, then ask the person to measure and reported back the measurement. Then analyze the voltage and pin point to the bad component/s. It worked.

From reading a lot of posts here, I feel people are too hasty in taking out transistors, putting them on testers, replacing with new ones. A lot of time, I don't think people actually find the problem, but just blind shooting, hit and miss. Along the way, they might burn more stuffs, they beat up the circuit board and make the amp unreliable in the long run. I don't think that's the right way to approach this. Problem is there are a lot of people encouraging others to do this and become a general consensus here.

My experience is if it ain't broken, don't fix it. If there is a problem, take the time to study the circuit, test a lot and do a little. When you pin point the problem, then change ONLY the bad component and nothing more. You'll be surprised how many problem show up in just simple DC voltage measurement.

I have been working with transistors my whole life designing all sort of circuits including analog IC design. I never even use a curve tracer nor do I know how to use it yet. I understand what it test ( simple collector curve), but never use it nor find it necessary.

One very important thing, you dig into the heart of the amp, you better know electronics. If you take it on as a hobby of restoring, fixing or designing, take the time to study transistor basic, learn to do DC analysis. You cannot do hit and miss job. There is no magic, just logical analysis. Study, observe, think and calculate a lot, do a little. You might think you are good in soldering and rework, I can tell you I am very good in doing soldering and rework, it's my job. But I can tell you, the board I use to bring up my amp is NOT going to be in my amp. I have a second board that I accidentally shorted out something that I fixed, then I over heated and blew a power transistor. This board is not going to be in my amp. I since brought up two brand new boards after I rung out all the bugs, do the minimum verification to proof they work and these are the two that go into my amp. I am not taking a chance on the two that had been beaten up.

When you change transistors, you better know whether you need to match any characteristics. I have been experimenting in matching output power transistors, I yet to find an effective way to match Vbe better than 3mV between transistors. You play around and burn one, you are in deep dodo!!! Good luck in trying to find one to match to 3mV. When I pull out a transistor, I never put them back......I don't pull out transistor very often.

I am surprised people took their treasured amp apart, remove components to measure and put them back!!!! You just used up 8 of the 9 lives of the amp!!!! My experience is....unless the pcb is high quality ( made in USA), you can only remove and replace the component twice before the trace and pads start to be pull out on the pcb. I have good soldering and desoldering equipment, and this is my experience. If you don't have the desoldering station with electric solder sucker and use solder wick or hand solder sucker, good luck. Keep pulling components out and so you have a unreliable pcb and that spells the end of the amp.

JMHO
 
Last edited:
Well, in speaking of opinion, I have a slightly different one, in my world (of precision RF power electronics) a curve tracer, in particular the Tektronix, is a must have in our design Lab, yes we have two of them in well kept condition. As a DIY I am totally salivating to get my hands on one of these extinct units. The purpose and usefulness of such instrument may be unknown to some people. DIY individuals who own a curve tracer know very well for what purpose this instrument was made keeping it around and alive for that. While professionals are struggling to find a suitable replacement unit with equal or better performance.

And yes as mentioned, semiconductors can partially fail and may still function to a degree but crippled somehow. One needs a lot of extra persistence and a good pinch of luck when troubleshooting such failure modes of crippled semiconductors when armed only with a voltmeter and a calculator. In cases with totally dead transistors a voltmeter surely suffices and offending parts are quickly and easily located no calculator needed.
It may be an extinct skill these days, some people don't need a calculator much, they can compute such simple basic math in they're head.

I believe high quality PCB's are made outside the USA also.
 
I've been in this field for 30 years designing all sort of transistor circuit, RF, low noise and high voltage circuits and even spent two years designing bipolar analog IC, never have I use a curve tracer nor find it necessary. More importantly, people here have a bad habit of pull parts out to test and put it back it. This is a very bad habit in my book. You study, learn the circuit, check out all the voltage to make sure it makes sense. Unless there is a problem, you don't pull things out just for the sake of pulling them out and check. If you suspect a problem, show me the symptom first, not pull the transistor and run curve tracing.

I don't expect a lot of the cheap commercial amp use top quality pcb.

I don't pull transistors out for no reason, when I pull any out, I intend to replace them with new ones no matter what.


EDIT:

Again, the best weapon of testing transistor is knowledge. Study how transistor works, spend the time to get the schematic of the amp you are using, study the schematic. Write out all the DC voltages in the circuit, check it against the real circuit and compare.

If you cannot write out the DC voltages, go back to the transistor basic books and study how to. You waste time pulling out transistors, wasting time to fix if something gone south along the way during the pulling, your time is well spent to work and write out all the DC voltages.

You can spot the problem 90% the time just by a simple volt meter and the knowledge. then if all else failed, then consider more drastic approaches like pulling, not from the get go, particular not if the amp is working and you pull it all just for the sake of pulling.
 
Last edited:
I've been in this field for 30 years designing all sort of transistor circuit, RF, low noise and high voltage circuits and even spent two years designing bipolar analog IC, never have I use a curve tracer nor find it necessary. More importantly, people here have a bad habit of pull parts out to test and put it back it. This is a very bad habit in my book. You study, learn the circuit, check out all the voltage to make sure it makes sense. Unless there is a problem, you don't pull things out just for the sake of pulling them out and check. If you suspect a problem, show me the symptom first, not pull the transistor and run curve tracing.

I don't expect a lot of the cheap commercial amp use top quality pcb.

I don't pull transistors out for no reason, when I pull any out, I intend to replace them with new ones no matter what.


EDIT:

Again, the best weapon of testing transistor is knowledge. Study how transistor works, spend the time to get the schematic of the amp you are using, study the schematic. Write out all the DC voltages in the circuit, check it against the real circuit and compare.

If you cannot write out the DC voltages, go back to the transistor basic books and study how to. You waste time pulling out transistors, wasting time to fix if something gone south along the way during the pulling, your time is well spent to work and write out all the DC voltages.

You can spot the problem 90% the time just by a simple volt meter and the knowledge. then if all else failed, then consider more drastic approaches like pulling, not from the get go, particular not if the amp is working and you pull it all just for the sake of pulling.


I too have been working professionally on electronics for many years (40+ but who's counting) and concur: pulling working parts for specious testing is poor practice. Almost nobody has a car tire removed from the rim to "inspect the inside for abnormal wear". Heating, removing and handling functioning components, without solid foundation, is just adding more reasons for premature failure. Measure voltages first and compare with manufacture's data. When schematic is not available, draw out suspect stage and do the math. Voltage drop is absolute best starting point to asses current conditions (pun intended). Only times I ever used curve tracing were when I was pushing components in my custom designs. Getting out to the SOA limits occasionally gave merit to using curve tracer and matching everything. For repairs in general, especially systems designed over 30 years ago, basic troubleshooting starts with DC voltage measurements and common sense applications of circuit theory. Before looking into any curve tracing options, it is much more beneficial to invest in a good quality 5 amp VARIAC. DBT will not help one trying to reform 'ltyic capacitors.
 
hello folks , i need a opinion on my problem.......denon pma 1520 , big baby , power transistors are 2 x 2sa1492+2 x 2sc3856 per channel.........one channel broken ........unfortunatelly i am in an isolated area now-heavy snowfall- and i do not have a multimeter to test those 4 transistors so i did this.........desoldered all 4 trans. and i got relay coupling on , other channel working ok , so those 4 trans are my problem.......since no multimeter, i resoldered one by one , after each one i powered amp and wait relay click on.......in one case , after resoldering a 2sa 1492 relay didnt click on .......now 3 are soldered on motherboard 2 x2sc3856 and 1x2sa1492...........it is correct to think that 2sa1492 - in which case relay didnt click- is my broken one? it is possible 1 of 4 to break down?
 
I can't very well read all this entire thread to find an answer to my question, so I'll start fresh. I have a 1966-69 Concord 776 stereo tape recorder with fold-out speakers. There is a hiss in the sound and after many hours of satisfying myself that the problem originates in the first 2 transistors of each channel. They are pnp germaniums. I do not know case styles, but they're your average small round metal can style. I have a parts donor, but question the worth of it. Germaniums are germaniums. I see no point in swapping and hoping. Is there an at-home test to detect hissy transistors? Thank you.
 
Back
Top Bottom