## circuit for Transistor testing ?

[small_boy]

Hi guy´s,

has anyone a simple circuit with there i can test high power transistors as 2SC2922 and 2SA 1216.

For example a small circuit with LM317 or a L200 as power supply in that i can build in the 2922 or/and 1216 as current driver to bring them to there SOA (so 9 Ampere current) ?

Or a other simple Circuit with there i can test this Transistor up to the SOA border ?

greetings

Dirk

[AndrewT]

test at what temperature? heatsink or transistor case or junction temperature monitoring?
Beyond or within the voltage of the second breakdown knee?
Continuous (DC), or 1second, 100mS or 10mS single shot?

[small_boy]

complete DC without other Signals.

The transistor has on Emitter a resistor with 8,4 or 2 ohm with 100 to 200 watt. The transistor are passiv and activ cooled. See pic.

I have a power supply with 2*34 Volt with 9.5 Ampere and i will bring the Transistor near the SAO and will see that hold his data what stand in the datasheet.

Dirk

[cliffforrest]

Originally posted by small_boy
Hi guy´s,

has anyone a simple circuit with there i can test high power transistors as 2SC2922 and 2SA 1216.

For example a small circuit with LM317 or a L200 as power supply in that i can build in the 2922 or/and 1216 as current driver to bring them to there SOA (so 9 Ampere current) ?

Or a other simple Circuit with there i can test this Transistor up to the SOA border ?

greetings

Dirk

Errr! No!

That is very difficult and requires very fast short pulses with low duty-cycle (<1%) at very high power.

Otherwise the transistor will be zapped in milliseconds, even if decently heatsinked.

Industrial testers use an inductive load flyback circuit to move the DUT from a saturated collector current up to a secondary breakdown.

Once, and for only about 300uS.

SOA is a calculated curve from various parameters (thermal, Imax etc) and cannot be "measured" directly.

[Tarzan]

Once up on in time (1975?) there was an application note of RCA; I can't recall wich one.
It described a test setup for a transistor tester WITH the possibility to test the SOA of a transistor.
If my memory is correct it could go up to 300V and a huge amount of amps.
It checked the temp as well, allthough I'm not 100% sure.
I rather like to find that application note again.
It was a bit expensive for me at that time to build but I can now.
If only I had that AN...
Maybe one of the older members?
(look who's calling others "older"...)
Cheers,
Zilog

[small_boy]

@cliff

That is very difficult and requires very fast short pulses with low duty-cycle (<1%) at very high power.

Otherwise the transistor will be zapped in milliseconds, even if decently heatsinked.

Why is then in the datasheet from Sanken for SOA one line for DC and one for Pulse ?

@zilog

I´m old (39 year at begin of February ;-)) but on 1975 i was 6 and have don´t interest for Electronics

Dirk

[Tarzan]

Oh jee small_boy ,
Im 54 now, heading for 55.
If only I kept all my docs the time I as young.
You can never tell the way your life go's.
So if I may give some advice:
Never ever trow away your data books, magazines and other stuff.
One day they will come in handy.

So to the real "old" guys; look around if maybe you have that AN from RCA.
Cheers.
Zilog

[cliffforrest]

Originally posted by small_boy
@cliff


Why is then in the datasheet from Sanken for SOA one line for DC and one for Pulse ?

@zilog

I´m old (39 year at begin of February ;-)) but on 1975 i was 6 and have don´t interest for Electronics

Dirk

Yes, I lost 30 years worth of data books and notes in a fire

Confusing isn't it, small-boy ??

That curve says "natural cooling without heatsink", but the first break-point is 20A Ic at 10v VCE.

That is 200W.
Even with a fan cooled massive heatsink that is not real!

What do you want to TEST for SOA? If it doesn't meet SOA it dies. Period. No way back.

It is a design-aid curve.

[small_boy]

@Tarzan :

Oh jee small_boy ,
Im 54 now, heading for 55.

ok ok, then i´am a young boy


@cliff :

Yes, I lost 30 years worth of data books and notes in a fire

oh thats ****.

Even with a fan cooled massive heatsink that is not real!

What do you want to TEST for SOA? If it doesn't meet SOA it dies. Period. No way back.

Ok, i will test this transistors 2SC2922 with 8-9 ampere, over 1-2 hours and go near to the 200 watt Ptot to see that it don´t burn.

This is the only way to see that these transistors hold, what the datasheet says.

Because this are not from Sanken, there are from ISCSemi, and in the thread :

### Counterfeit transistors

you can read what problem i have.

A few told me, that this transistors are fakes, other that the trans from ISCSemi are ok. But now i will test it with power and look that there hold what the datasheet says.

What can i do to test this transistors and how ?

Dirk

[cliffforrest]

I do sympathise and you really do have a problem.

However, in my long experience, trying to test at DC is a complete waste of time. But I am not always right.

What are you going to do if they fail? The supplier will only accept THEIR test methods and not yours.

Have a look at:

SOA

particulary at the inductive test circuits.

As I said earlier, SOA curves are design aids. If you are having reliability issues and failures my first step would be to search for devices with 50% to 100% better SOA specs.

I don't think I can help more ....