Odd Transistor Readings

[Cursitor Doom]

So I've had to pull 6 transistors one by one out of this amplifier board,
in the strong suspicion at least one of them was faulty. They all tested
fine - until the last one (typical!). Anyway, these are small signal PNP
BJTs in TO-8 tin cans. The last one checks out fine for base-emitter and
base-collector junctions giving about 650mV in one direction only on the
diode test setting - I very nearly didn't bother testing further at this
point, given it was looking increasingly futile. Anyway, for the sake of
completeness one last check across the C-E terminals and I got 295mV both
ways! Double checking on the resistance range confirmed 600 ohms between
C and E both ways. I've never known a BJT fail in *this* way. Has anyone
else?



--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.

 

[Phil Allison]

Cursitor Doom wrote:

-------------------

So I've had to pull 6 transistors one by one out of this amplifier board,
in the strong suspicion at least one of them was faulty. They all tested
fine - until the last one (typical!). Anyway, these are small signal PNP
BJTs in TO-8 tin cans. The last one checks out fine for base-emitter and
base-collector junctions giving about 650mV in one direction only on the
diode test setting - I very nearly didn't bother testing further at this
point, given it was looking increasingly futile. Anyway, for the sake of
completeness one last check across the C-E terminals and I got 295mV both
ways! Double checking on the resistance range confirmed 600 ohms between
C and E both ways.

** See, all you need was a simple ohm meter test.

Checking C-E is a standard practice, cos it's a common failure when a transistor is over voltaged or over heated under load.

Commonly known as " punch through" or "second breakdown".



....... Phil




I've never known a BJT fail in *this* way. Has anyone

else?



--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.

 

[Fred McKenzie]

In article <qov5uu$thh$1@dont-email.me>,
Cursitor Doom <curd@notformail.com> wrote:

So I've had to pull 6 transistors one by one out of this amplifier board,
in the strong suspicion at least one of them was faulty. They all tested
fine - until the last one (typical!). Anyway, these are small signal PNP
BJTs in TO-8 tin cans. The last one checks out fine for base-emitter and
base-collector junctions giving about 650mV in one direction only on the
diode test setting - I very nearly didn't bother testing further at this
point, given it was looking increasingly futile. Anyway, for the sake of
completeness one last check across the C-E terminals and I got 295mV both
ways! Double checking on the resistance range confirmed 600 ohms between
C and E both ways. I've never known a BJT fail in *this* way. Has anyone
else?

CD-

That sounds like leakage to me. Did you measure C, B and E voltages
in-circuit?

I wonder if a coupling capacitor from the previous stage is leaking, and
it might have damaged that transistor?

Fred

 

[Phil Allison]

Fred McKenzie wrote:

---------------------

That sounds like leakage to me. Did you measure C, B and E voltages
in-circuit?

I wonder if a coupling capacitor from the previous stage is leaking, and
it might have damaged that transistor?

** The OP is working on a scope horizontal sweep amplifier PCB.

It's direct coupled throughout.

Despite my advice, he failed to carry out basic ohm meter tests with the devices in circuit.

Folk pick their favourite replies here and pay the price for that.



..... Phil

 

[Cursitor Doom]

On Fri, 25 Oct 2019 16:35:34 -0400, Fred McKenzie wrote:

CD-

That sounds like leakage to me. Did you measure C, B and E voltages
in-circuit?

Yes, I did. They were all over place and not just for the transistor in
question:

Q10: Vc +1.1v, Vb +0.5v, Ve +1.2v
Q9: Vc -3.2v, Vb +1.1v, Ve -1.34v(this is the faulty one)
Q8: Vc -3.8v, Vb -3.2v, Ve -3.9v

Q15: Vc -2v, Vb +0.11v, Ve +0.78v
Q14: Vc -7.74v, Vb -2v, Ve -1.34v
Q13: Vc +3.5v, Vb -7.74v, Ve -7v

I'm not sure if just Q9 alone being faulty could account for three other
transistors showing 'impossible' bias voltages: Q8 fully saturated; Q10
inverse Vbc; Q13 inverse Vbe - but with them all being inter-dependent as
a consequence of direct coupling, who knows?

I wonder if a coupling capacitor from the previous stage is leaking, and
it might have damaged that transistor?

It's all *direct* coupled, which is 9/10 of the problem in trying to pin
down the fault. Here's the schematic:

https://yandex.com/collections/card/5db3651651aa90bd7e86a504/



--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.

 

[Cursitor Doom]

On Fri, 25 Oct 2019 21:36:51 +0000, Cursitor Doom wrote:

Let's try and make that clearer:

Q10: Vc +1.1v, Vb +0.5v, Ve +1.2v

Q9: Vc -3.2v, Vb +1.1v, Ve -1.34v (this is the faulty one)

Q8: Vc -3.8v, Vb -3.2v, Ve -3.9v



Q15: Vc -2v, Vb +0.11v, Ve +0.78v

Q14: Vc -7.74v, Vb -2v, Ve -1.34v

Q13: Vc +3.5v, Vb -7.74v, Ve -7v

 

[Cursitor Doom]

On Fri, 25 Oct 2019 15:56:14 +0000, Cursitor Doom wrote:



Having checked for replies in google groups...

I did actually carry out in-circuit resistance checks with the board
unpowered using a meter capable of exceeding the 0.7v before removing
anything. It indicated discrepancies compared to an identical board in
the area around Q10, but not a specific enough area nor specific enough
discrepancies to be of use.




--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.

 

[Phil Allison]

Cursitor Doom wrote:
-------------------------

Having checked for replies in google groups...

I did actually carry out in-circuit resistance checks with the board
unpowered using a meter capable of exceeding the 0.7v before removing
anything.

** You are being very coy about what you REALLY did.

My post said to use an analogue ohm meter on the X1 range - nothing else works near as well, especially any DMM I know of.

The ohms reading found corresponds to millivolts and volts across the load - easily calibrated with a DVM and few resistors. I inked voltage numbers on the ohms scale of my meter 45 years ago.

Most importantly, such meters deliver up to 50mA into the load, making redundant any resistors that might be in parallel with device junctions.

Betcha you did nothing like that or else the near shorted C-E would have jumped up and bit you.




..... Phil

 

[Phil Allison]

Fred McKenzie wrote:

---------------------

https://yandex.com/collections/card/5db3651651aa90bd7e86a504/

CD-

Diagram appears to have a mix of PNP and NPN transistors.

** Hey - don't confuse the OP with facts !!



...... Phil

 

[Fred McKenzie]

In article <qovptj$thh$3@dont-email.me>,
Cursitor Doom <curd@notformail.com> wrote:

It's all *direct* coupled, which is 9/10 of the problem in trying to pin
down the fault. Here's the schematic:

https://yandex.com/collections/card/5db3651651aa90bd7e86a504/

CD-

Diagram appears to have a mix of PNP and NPN transistors.

Fred

 

[Trevor Wilson]

On 26/10/2019 2:56 am, Cursitor Doom wrote:

So I've had to pull 6 transistors one by one out of this amplifier board,
in the strong suspicion at least one of them was faulty. They all tested
fine - until the last one (typical!). Anyway, these are small signal PNP
BJTs in TO-8 tin cans. The last one checks out fine for base-emitter and
base-collector junctions giving about 650mV in one direction only on the
diode test setting - I very nearly didn't bother testing further at this
point, given it was looking increasingly futile. Anyway, for the sake of
completeness one last check across the C-E terminals and I got 295mV both
ways! Double checking on the resistance range confirmed 600 ohms between
C and E both ways. I've never known a BJT fail in *this* way. Has anyone
else?

**Frequently. Invest in a PEAK transistor tester. It will save time and
heartache.



--
Trevor Wilson
www.rageaudio.com.au

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

 

[Ralph Mowery]

In article <h1hsj2F7jmkU1@mid.individual.net>, trevor@rageaudio.com.au
says...

**Frequently. Invest in a PEAK transistor tester. It will save time and
heartache.

Instead of investing in the PEAK, go to ebay and get one of the
component testers for about $ 15 to $ 25.

About the same as the Peak and one box tests solid state and passive
components where Peak sells 2 boxes to do the same thing.

 

[Phil Allison]

Trevor Wilson wrote:

-------------------

** Invest in a PEAK transistor tester.

---------------------------------------------


** Never used one - do they inject at least 25mA testing BJT junctions?

Useless in-circuit if they don't.

25mA will override even a 50ohm B-E resistor




.... Phil

 

[Trevor Wilson]

On 26/10/2019 2:35 pm, Phil Allison wrote:

Trevor Wilson wrote:

-------------------


** Invest in a PEAK transistor tester.

---------------------------------------------


** Never used one - do they inject at least 25mA testing BJT junctions?

**Nope.

https://www.peakelec.co.uk/acatalog/dca75-dca-pro-semiconductor-analyser.html

12ma max.

Useless in-circuit if they don't.

**I don't trust in-circuit tests.

25mA will override even a 50ohm B-E resistor

**I'm sure it will. The PEAK DCA75 tester has not let me down yet (the
DCA55 did let me down once). In fact, it has been able to show very
slight faults in some semiconductors, that multimeter tests did not reveal.


--
Trevor Wilson
www.rageaudio.com.au

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

 

[Trevor Wilson]

On 26/10/2019 2:29 pm, Ralph Mowery wrote:

In article <h1hsj2F7jmkU1@mid.individual.net>, trevor@rageaudio.com.au
says...

**Frequently. Invest in a PEAK transistor tester. It will save time and
heartache.





Instead of investing in the PEAK, go to ebay and get one of the
component testers for about $ 15 to $ 25.

**I've used one. Impressive value for money.

About the same as the Peak and one box tests solid state and passive
components where Peak sells 2 boxes to do the same thing.

**I've not seen one that can do what the PEAK DCA75 can do:

https://www.peakelec.co.uk/acatalog/dca75-dca-pro-semiconductor-analyser.html

Note the ability to display curves on a computer.

Please point me to one that can do what the DCA75 does.



--
Trevor Wilson
www.rageaudio.com.au

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

 

[Phil Allison]

Trevor Wilson wrote:

--------------------

** Invest in a PEAK transistor tester.

---------------------------------------------


** Never used one - do they inject at least 25mA testing BJT junctions?

**Nope.

https://www.peakelec.co.uk/acatalog/dca75-dca-pro-semiconductor-analyser.html

12ma max.

** Marginal then.

Useless in-circuit if they don't.


**I don't trust in-circuit tests.

** Not even Bop Parker's magic ESR meter ?

Or if a junction shows open or short ?

Or a C-E test shows low ohms or near short ??

Shirley you jest ?




...... Phil

 

[Cursitor Doom]

On Sat, 26 Oct 2019 14:24:20 +1100, Trevor Wilson wrote:

> **Frequently.

Well I don't do repair work for a living so defer to your wider knowledge
of the subject. I just find it baffling how such a transistor can give
readings like those I described, with a Rc-e lower than each of its Rc-b
and Rb-e readings!

Invest in a PEAK transistor tester. It will save time and
heartache.

It wouldn't be worth if for the number of times a 'mass extraction' like
this has been necessary, which before this happened was zero.

It would be nice to have *all* the meters in the Peake range, but on an
occasional need basis I can't justify it.





--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.

 

[Phil Allison]

Cursitor Doom wrote:

-------------------

Trevor Wilson wrote:

**Frequently.


Well I don't do repair work for a living so defer to your wider knowledge
of the subject. I just find it baffling how such a transistor can give
readings like those I described, with a Rc-e lower than each of its Rc-b
and Rb-e readings!

** Never seen a shorted power transistor ?

The two junctions test OK, but there is a near dead short from C to E.

If you open one, the chip usually has a black spot on it where the silicon has melted.

Your example involves a low resistance, likely the result of an over voltage or over current spike.

Mistreat a BJT and a short from C to E is the MOST likely outcome.

Same goes for mosfets, the short is then from D to S.

Seen many hundreds of both.

Diodes fail short too, plus zeners, triacs and SCRs.

Silicon dies are fragile things.




..... Phil

 

[Cursitor Doom]

On Sat, 26 Oct 2019 15:52:30 +0000, Cursitor Doom wrote:

Not the same exact model as couldn't find one on Ebay to show, but same
make made about 60 years ago.

Which is a lot more 'modern' than the one in the Ebay advert, plus mine
goes up to 2,500VDC.




--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.

 

[Cursitor Doom]

PITA having to cut and paste from Google Groups but anyway...


"** You are being very coy about what you REALLY did.

My post said to use an analogue ohm meter on the X1 range - nothing else
works near as well, especially any DMM I know of.

The ohms reading found corresponds to millivolts and volts across the
load - easily calibrated with a DVM and few resistors. I inked voltage
numbers on the ohms scale of my meter 45 years ago.

Most importantly, such meters deliver up to 50mA into the load, making
redundant any resistors that might be in parallel with device junctions.

Betcha you did nothing like that or else the near shorted C-E would have
jumped up and bit you. "

Nope. I actually used one of these:

https://www.ebay.co.uk/itm/Selectest-Testmeter-Dlll-Old-Multi-
Meter/333272931128?hash=item4d989d5b38:g:jD8AAOSwxGBdNGvF

Not the same exact model as couldn't find one on Ebay to show, but same
make made about 60 years ago. I also keep an AVO 8 but the batteries are
dead in that one. And yes, I *did* check it on the 1X range. YOU could
probably have made sense of the readings I was getting; they were
definitely lower on the damaged board in that region and *now* I know
why, but I'm no repair technician and never pretended to be so the
significance on me was a bit lost.

Are we clear now?




--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.