Shorted transistors are pretty easy to find. But leaky one's can be a pain

Ray Gianelli

Super Member
I had a Yamaha RX-900U on my bench recently. Belongs to a friend. Pretty easy diagnosis; burned components in the right channel. Shorted outputs, drivers and open fusible resistors. Checked back from the outputs until everything else checked good. Replaced shorted and burned components, and fired it up on the DBT. Major negative offset, and since it's direct coupled like most modern amps, it's bad everywhere.

I checked and rechecked everything, and couldn't find a problem. I finally pulled a transistor that showed some voltage between collector and emitter on diode test; usually I find that when they test like that in circuit they're fine when you pull them. Not this one. Replacement of this transistor fixed the amp.

So this transistor E/B and B/C junctions both tested good. It was only when testing the E/C that it showed continuity, and only on one orientation of the meter leads.

Not the usual failure, but transistors can and do fail in other manners besides short or open. Thankfully I rarely find one that only fails under higher voltage, but leaky can be just as hair pulling.

So, does anyone have any advice on how to find these? The only reason I did was circuit analysis, helping zero in on the area I thought was causing the trouble.
 
Sometimes you can find them with freeze mist or heat. Freeze mist is expensive so I'll often use a Q-tip moistened with alcohol. It cools as it evaporates, though not near as much as mist. Still enough to find some defective devices.
 
Ha! Those would be the ones in my Crown power amp where the curve tracer says they can't support the full supply voltage at high currents. Work great until you stress the amp. Now I only install sets that I've verified myself as being up to the task. Actually new parts are great, and only the old OEM ones are risky. The common transistors testers won't give you a clue in a case like that.
 
If using a diode test on a multimeter, you MAY have to remove it from the circuit first (P-I-T-A I know!).

Mark T.:music:
 
I've said this before, but ALL RETS (ring emitter transistors) and other diffused emitter types should be tested on a curve tracer or high voltage transistor tester at various voltages up to their ratings to determine if they fail at those high voltages. I've lost count of how many dodgy Sanken/Fujitsu 1970s RETS test fine on basic testers, but go awol at power amp rail voltages.
 
It has been my experience that if my tester says it's bad, it's bad. The problem is when it says it's good. Most times, it is good, but sometimes......it ain't.
Like Tom mentioned, that Huntron Tracker, or an Octopus, can be a huge time saver.
 
If using a diode test on a multimeter, you MAY have to remove it from the circuit first (P-I-T-A I know!).

Mark T.:music:
Mark, you're right. That's the only way I found it. Fortunately, it was the first one I pulled after studying the schematic.

Considering how long it took to find it I may have to look at a Huntron or equivalent.
 
Back
Top Bottom