Odd failure of dual matched transistors MAT02

[Audiorepair]

This is used on a microphone pre-amp on a mixing desk.

So each channel uses this device on a microphone balanced input circuit.
https://www.analog.com/media/en/technical-documentation/obsolete-data-sheets/MAT02.pdf

I have just pulled 12 of these from a desk, and all have failed the same way. (I suspect 48v phantom is involved)
One of the transistors, pins 1/2/3, is completely open circuit on all pins, the other transistor is intact.
Even the reverse biased diode on Base/Emitter has disappeared.

I've come across many failed transistors in my time, but usually there is a short, or no gain, or whatever, but never open circuit on all 3 pins.


Anyone able to shed any light?

[floobydust]

I find it's a common failure for mic preamp front ends, they get hit with a capacitive discharge from the console's phantom-power blocking caps i.e 47uF at -48VDC discharges into the front-end if you ever short an XLR signal pin to GND, like a shorting cable or connector. If so many channels are blown, the guy probably just kept trying a new channel with the offending cable/mic.

A decent mic-pre design has large clamp-diodes before the diff amp. that can survive high-current impulses. MAT02 internal diodes would not survive.

edit: Linear Systems has replacements for MAT series and LM394 like LS310.

[SilverSolder]

Is there a circuit diagram of the inputs?

Could it be ESD?

[Audiorepair]

The question is, I am curious how a transistor can go open circuit on all 3 pins.
And it is probably down to 48v phantom.

The original design used LM394, then MAT02 was substituted for repair when these became obsolete, now MAT12 is the very expensive option.

MAT12 specs, very speccy.
https://www.analog.com/media/en/technical-

Here is the circuit:

[SilverSolder]

So, which of the transistors is blowing - is it TR2-A or TR2-B, or do both go?

[Audiorepair]

TR2A blows, pins 1/2/3 all open circuit.

[SilverSolder]

It looks like @floobydust is probably spot on. If the + input is shorted to ground while phantom power is on, TR2-A will receive a beating.

TR2 is a sensitive, low current device - perhaps the base simply gets blown up so bad that both sides of it go open circuit.

Wonder why there is no current limiting resistor in the base circuit?

[floobydust]

I don't see much for high-current paths in the input circuit, not sure what is at MO1.2/MO1.4 though. The 2.2uF caps at 48V are around 2.5-5mJ enough to kill a semiconductor but they would have to discharge through an NE5534 pin 2 (-) or the substrate diodes on the LM394.

If you had phantom power on, then switch it off (all with no mic connected) the 2.2uF caps would hold a charge for a while, so the next plug-in can dump them into the transistors.

No series resistor is used because of the added noise. Big TVS also add a lot of capacitance.

I'm not sure where else enough energy to damage the transistors could come from.
Vacuum tube condensor mics use an output transformer to isolate the HV DC, unless he has a bad mic.

[SilverSolder]

What might be happening is that C2 is charged up to 48V when phantom power is On.  When the + input is then grounded with a bad cable or whatever, that would make the base voltage on the transistor instantaneously jump down to  -48V.   Since the collector is parked at +17V,  Vcb becomes 65V for a brief instant....   and the transistor is only rated for 40V Vcb...  so the collector - base junction likely suffers a breakdown...  and the current is only limited by the 10K resistor in the +17V line, so we could see a 6.5mA pulse flowing out via the base that way.  Possibly more than enough energy to completely vaporize the small and fragile transistor's base layer, leading to the transistor going open circuit on all its connections as @Audiorepair is seeing?

[Audiorepair]

I don't see much for high-current paths in the input circuit, not sure what is at MO1.2/MO1.4 though.

M01.x is the amps switched gain control.

[magic]

That thin, feeble base layer is just dopant atoms diffused into a solid chunk of silicon to a specific depth. It can't be completely vaporized without leaving a big crater in the silicon. Alternative possibility is fusing of bonding wires or metal traces on the surface of the die.

Any connection/diode/resistance between pins 1~3 and the metal can?

Maybe a teardown?

[Audiorepair]

....   and the transistor is only rated for 40V Vcb... 

Yes, the original LM394 is only rated 35v, both MAT's are 40v.

Not really enough to cope with the inevitable phantom kerfuffles they will meet I guess, but I suspect many would take a few punches before finally falling over.
These particular MAT02 are quite old now and have been in daily use for many years.

Musn't gwumble, eh.
 
But to lose so many in a short space of time does, I feel,  point to faulty equipment/cables or careless hot plugging.
Maybe DI's, where one might not be so concerned,  instead of expensive mics where you very much would turn the phantom off first.

[Audiorepair]

That thin, feeble base layer is just dopant atoms diffused into a solid chunk of silicon to a specific depth. It can't be completely vaporized without leaving a big crater in the silicon. Alternative possibility is fusing of bonding wires or metal traces on the surface of the die.

This from the MAT02 datasheet, dunno if that signifies an unusual construction or not.