My XPA-3 (Gen 1) blew up.

[ohiobuckeye]

Well my XPA-3 failed a few days ago. Turned it on, heard a crack, and then saw some smoke come out of it.

Did an inspection. Transistor Q015 on Channel #2’s PCB is toast. It literally exploded. (I found part of its plastic package on the bottom of the unit.) It is a 2SA1837 transistor.
And one of the big output transistors attached to the heatsink (Q014) shows evidence that it got really hot. It’s a NJW0281G transistor. Maybe it failed first, and then took out Q015?

At any rate, does anyone have a schematic for this thing?

Thanks

Pics from DB:

www.dropbox.com/s/fz2kaoedo75hnue/1.jpg?dl=0

www.dropbox.com/s/b5nuvwrlde9p4iq/2.jpg?dl=0

www.dropbox.com/s/w3h7j2960ap8v02/3.jpg?dl=0

www.dropbox.com/s/fz20lqz3odur3v5/4.jpg?dl=0

www.dropbox.com/s/v0ycl4h6csglpsg/5.jpg?dl=0

[Gary Cook]

Email Emotiva, the service technical guys are very helpful, Joe has saved my butt several times.

Cheers
Gary

[DavidR]

It may have been a counterfeit transistor. People who modify old Caver amps run into this problem from time to time.

[kraddy]

Me too. Same #2 PCB, different part(s). I've emailed Emotiva to question if I could purchase an entire new board, or if it makes sense to have local repair shop attempt a fix.

I will look back to see how you make out, so please update

[chops]

Nov 4, 2020 13:43:56 GMT -5 kraddy said:

Me too. Same #2 PCB, different part(s). I've emailed Emotiva to question if I could purchase an entire new board, or if it makes sense to have local repair shop attempt a fix.

I will look back to see how you make out, so please update

View AttachmentView Attachment

You're looks like it was just a couple of faulty caps that blew, but there could be something else in the chain that could have caused them to blow. What amp is this?

[ttocs]

You should check those two green resistors. They look a bit toasty. And look for any heat damage on everything while you've got the fun tools out.


For all of you with these explod'n amps, if you've got one of those blue ESR meters you can check the caps in-circuit. AFTER you discharge them.

[leonski]

Trouble with measuring ESR of a cap? What's GOOD? Unless cap is at Zero (shorted out) or 'infinite' which would indicate open....are either of those extremes posssible on that meter?
In between reading? Good or bad?

Toasty looking resistors ARE suspect, for sure. But is a dead transistor a CAUSE or an EFFECT? Replace bad transistor and it goes again?

C5 and C29 are of course, Toast. But Cause or Effect? Looks like something caused a SEVERE overvoltage to those parts.

I hope you get this resolved and maybe EMO ships you off a schematic.....than find a GOOD service guy.....

[ttocs]

Nov 12, 2020 0:25:51 GMT -5 leonski said:

Trouble with measuring ESR of a cap? What's GOOD? Unless cap is at Zero (shorted out) or 'infinite' which would indicate open....are either of those extremes posssible on that meter?
In between reading? Good or bad?

Uncle Doug on u-toob has a video using the Anatek ESR meter. He demonstrates checking caps with a multi-meter, and then with the Blue ESR meter, and explains the how and why. I suggest watching the entire video for complete context and how he determined what the value should be of the caps he is testing, but at 10:50 the video gets to the heart of the demo. He also explains how the Blue ESR meter can test in-circuit so accurately, so I won't paraphrase when he explains it so well. The title is: Using an ESR Meter to Determine the 'Health' of Electrolytic Capacitors in Guitar Amplifiers.

It's a great meter and makes short work of checking caps. I've used mine on old guitar amps mostly.

[leonski]

I've done the 'kick' test for years.....as a go / nogo indicator......

I don't know that the caps in at least ONE of the photos the OP posted 'need' such a test. They looked half-Vaporized!

[leonski]

Nov 2, 2020 21:11:04 GMT -5 DavidR said:

It may have been a counterfeit transistor. People who modify old Caver amps run into this problem from time to time.

Would you expect such a device in an unmodified EMO product?     

[audiobill]

If you replace those resistors, suggest higher wattage. They have them lifted off the board for a reason, heat.

[DavidR]

Nov 12, 2020 13:31:40 GMT -5 leonski said:

Nov 2, 2020 21:11:04 GMT -5 DavidR said:

It may have been a counterfeit transistor. People who modify old Caver amps run into this problem from time to time.

Would you expect such a device in an unmodified EMO product?     

Its possible

[KeithL]

I very much doubt that you had a counterfeit transistor.

The only reason to counterfeit a transistor is to be able to substitute a lower cost part for a more expensive one.

This happens most often with parts that are especially difficult to find - especially transistors that are no longer being made.

It is pretty uncommon with ordinary current parts that are still readily available...

And even less common when you source your parts from reputable vendors...

(Although, on rare occasions, even reputable parts vendors have been victimized.)

(People who modify old gear often encounter this problem because they require outdated parts that are no longer available.

And, with older parts, it's quite possible that no newer part exists that is an ideal and exact replacement.

As a result, "new old parts" must be purchased on the surplus market, or from someplace like eBay.

And, because they are difficult to obtain, and sell for a relatively high price, those parts then become worth counterfeiting.)

Odds are that, in this case, a part simply failed.

While we use good quality parts, nothing is perfect, and failures do happen.

(It's one reason we have things like warranties and repair technicians.)

It is very difficult to successfully test capacitors with a meter.

And, in many cases, you cannot completely test a capacitor in-circuit.

So you would be talking about un-soldering it, testing it, then replacing it in the circuit.

And, once you've gone to that effort, it usually makes more sense to simply replace it with a new tested part.

(Individual small capacitors don't actually cost very much individually.)

I should also point out that, except in certain specific situations, simply "replacing the burnt parts" is often not an effective repair strategy.

Many parts, especially semiconductors, can suffer a complete failure without showing any outward signs.

Therefore, in addition to the obviously burned transistors, there may well be others that have failed, but don't show it.

And, if that's the case, when you replace the burned ones, your replacements will simply fail again.

So, when a failure occurs, it's often obvious to someone who understands the schematic that other parts should also be changed.

And, in some cases, you replace parts that seem likely to have been "overstressed", even if they are still functional.

(For example, in some circuits, if an output device has failed, you also replace the associated driver and pre-driver...)

 

[kraddy]

Mine was also 1st Gen XPA-3.

I've had Emotiva mail me a proper box, and it's getting shipped back to them to fix. They would not sell parts.

[leonski]

Nov 12, 2020 18:32:32 GMT -5 KeithL said:

I very much doubt that you had a counterfeit transistor.

The only reason to counterfeit a transistor is to be able to substitute a lower cost part for a more expensive one.

This happens most often with parts that are especially difficult to find - especially transistors that are no longer being made.

It is pretty uncommon with ordinary current parts that are still readily available...

And even less common when you source your parts from reputable vendors...

(Although, on rare occasions, even reputable parts vendors have been victimized.)

(People who modify old gear often encounter this problem because they require outdated parts that are no longer available.

And, with older parts, it's quite possible that no newer part exists that is an ideal and exact replacement.

As a result, "new old parts" must be purchased on the surplus market, or from someplace like eBay.

And, because they are difficult to obtain, and sell for a relatively high price, those parts then become worth counterfeiting.)

Odds are that, in this case, a part simply failed.

While we use good quality parts, nothing is perfect, and failures do happen.

(It's one reason we have things like warranties and repair technicians.)

It is very difficult to successfully test capacitors with a meter.

And, in many cases, you cannot completely test a capacitor in-circuit.

So you would be talking about un-soldering it, testing it, then replacing it in the circuit.

And, once you've gone to that effort, it usually makes more sense to simply replace it with a new tested part.

(Individual small capacitors don't actually cost very much individually.)

I should also point out that, except in certain specific situations, simply "replacing the burnt parts" is often not an effective repair strategy.

Many parts, especially semiconductors, can suffer a complete failure without showing any outward signs.

Therefore, in addition to the obviously burned transistors, there may well be others that have failed, but don't show it.

And, if that's the case, when you replace the burned ones, your replacements will simply fail again.

So, when a failure occurs, it's often obvious to someone who understands the schematic that other parts should also be changed.

And, in some cases, you replace parts that seem likely to have been "overstressed", even if they are still functional.

(For example, in some circuits, if an output device has failed, you also replace the associated driver and pre-driver...)

 

I worked for 15 years or so for a company making HEXFET and FRED and IGBT discrete devices.   COST was king.   Huge production levels kept end-unit costs low.   And they were Made In USA.    

Keith's right about troubleshooting.    Sure...a BlownUp transistor needs to be replaced.   Now?   Upstream cause?   Downstream effects?  Even a transistor which tests OK might have sustained enough stress to eat into eventual lifetime.
We would test new devices from a Test Lot (a lot of 50 or 100 wafers) for 1000 hours.    That's the burn-in level for new parts.    Failures were anazlyed for pattern, cause and potential end-user problems.   The REL (reliability) group got MY vote every time.     Such a lot of NEW devices must be a 'clean' run thru the fab....no reworks beyond a certain type allowed.    ANY furnace abort, for example?    Scrap it.     

And while Keith refers to caps as Cheap......inexpensive may be more like it...you can get a premium part for not much MORE than 'budget', the real commodity parts are stuff like we made in our wafer fab.    As a side note?    A name familiar to some around here.....Bob Carver....Used devices from my company in some of his Sunfire line.    The measure is MTBF.   Mean Time Between Failures.     You pay MORE for parts with higher ratings......and pass the cost on to the consumer.    Bryston will do a burn in on EVERY amplifier they sell.....which is one reason they can offer a 20 year warranty on those components.   

Keith gets a 100% on his answer.   His ONLY omission?    Look at the BOARD, too.   Sometimes a burned part will 'charcoal' a board.  THAT is conductive, too, and created problems later.    In emergency situations, I'd bridge burned sections with insulated wire just to get a machine going again.   Some of that old equipment was ready for the junk heap, anyway, and no boards were available, though 'parts is parts' ruled.   We had the semiconductor replacement handbook and since we couldn't get some parts, like Delco Transistors, made do with the nearest spec'd part.    An adventurous soul among us could fabricate simple boards using stuff available at the local Radio Shack...But NO multi-level or complex geometries.