Output transistors and gain

[Axtman]

I recently pulled the output transistors from a Sunn Beta Lead amp that I am servicing. One side had 2N3055 transistors. The gain (hFE) tested 8, 8, and 35. According to the datasheet, those 2N3055 transistors should test 25 to 70.

Do transistors that age or are stressed get lower in gain? Should I consider transistors lower than spec to be bad?

[Helmholtz]

hFE varies with collector current.
If you tested at low mAs, hFE will be low.
Datasheet hFE is at 4A.

[R.G.]

The answer is full of asterisks and footnotes. The current gain of bipolar devices is not constant; it varies with collector current, temperature, and phase of the moon. Testing at different currents will definitely produce different numbers on the same device.

Bipolars do age, especially with high temperatures. A gain of 8 is not by itself suspect, but a triplet of 8, 8 and 35 will have problems. I would guess that the two 8's are the paralleled output devices and the 35 is the driver.

That being said, If I had the outputs out, I would put in a new set of MJ15022 devices, and if it's a PNP bottom side, I'd put in a set of the complementary MJ15023s. The whole output stage will be more reliable.

[JP-Stereo]

Hi, i've enconter a few instances when transistors were passing the basic diode check but were still deffective.
So i acquired an HeatKit transistor tester in order to verify more parameters and it really speak a lot more.
Bias current, Leaks, Gain, etc.
With this tester verifying a minimum of 5 other parameters, i am catching a few more, but some are still getting through.
The last addition to my test sequence, suggested by fellow members, is the measurment of the Emiter to Base voltage in circuit.
That catched the most concelled ones as of now.
Time will tell if some defective transistors still get through...

[R.G.]

Be aware that transistor testers have difficulty testing power devices. They're great in many cases for small signal stuff, but they just can't pour enough voltage and current into a power transistor to be anything more than a quick check. They're great for "is this transistor completely broken?" on power devices, but they will pass devices that may not be able to function at high power. That is - they can tell you it's definitely broken, but may give some false positives.

[JP-Stereo]

Yes i know, i've read a little about it before making a move.
I've found an HeatKit IM-36 that have a very big internal battery bank and even accomodation for external power...


Bias to 5v, Collector voltage to 50v and Leak voltage to 150v.
I don't tell you i mastered it yet, but it's the device to make them work.

Next will be a curve tracer, if some transistor still make it through.

[Axtman]

I had a BK Precision 520B Transistor Tester but sold it. Perhaps I shouldn't have gotten rid of it. I use my Peak Atlas DCA55 Semiconductor Component Analyser.

I assume, like tubes, the best way to test a transistor is in the actual circuit.