Audison eats output N mosfet..why?

[pergo]

Hi everyone!

i buoght a semi-new Audison LRx1.400 with broken power supply and output mosfets.

I substitute original 60N06 with new IRF3205 in the power supply (with 10ohm gate resistor) and the PS works fine.

The output stage is made by 8 IRF9640 + 6 IRF640N.

I have a problem that i cannot resolve:

At idle (1.4A current draw..correct) the amplifier works fine. Non heat, no ringing, everything's ok.

When i put a test signal (sinewave 50hz for example) 1Vrms and gain at minimum, with 4ohm resistive load (or speaker), the amplifier works fine for some seconds...than it goes on protection because there's an output N mosfet broken.

Only N mosfet (IRF640N) get broken..and it's random which mosfet fails.

I check all gate resistor (47ohm) and other resistors beside the output stage. All value are ok (all measure with desoldered component, obviously).

I replace output stage 10 times!!! Every time one, only one, N mosfet fail.

I try to reduce biasing..but same problem.
I try to increase biasing (3A idle) but same problem.

Always 3-4 seconds play......then, without blow, heat or fire.....random one fails!


At idle, gate N voltage is 4.5V. P gate voltage is -4V. All clear.

Rail voltage is +-52V regulated.

Emitter resistor are 0.15ohm. All desolder have similar 0.15ohm value. seems correct.

Driver stage is made by 2SD2400A and 2SB1569...could be them? They seems fine..

With o-scope, before the failure, everything seems ok. No distortion, no clipping..nothing. Only clear signal.

The failure happens with easy load (4ohm) and low power (20Wrms).
The amp should be capable of 900Wrms @1ohm!!!

With no-load, the output stage don't fail, even with clipping.


How could it be??

I'm going crazy!
I repaired lots of amplifier in my country, but i cannot fix it!

Please help.

thank you guys!

[1moreamp]

"Driver stage is made by 2SD2400A and 2SB1569...could be them? They seems fine.."

Possibly it could have bad driver transistors.

I work on Audisons here in the US and I have repaired a few of these but nothing like what you described. Perhaps we are missing something like the drivers and integrator circuitry, I have seen bad ICs, and leaky drivers cause similar issues. After all this is class D either the fet is full on or full off. Have you checked dead time for cross conduction issues ? This might explain repeated blown single outputs device at any power level.


Audison is local to you, I am amazed they have not been of assistance to you already. Can they supply you a PDF of the output and driver stage of this amp ? I have never needed one so far. And my only reference is the LRx5.1k..But I will try to help if possible....Ciao

[pergo]

Hi,

thank you for reply.

This is NOT a class D

This is a class AB with drivers in full class A (heat a lot!).

I've repaired other Audison like VR/LR series and Thesis Series, but anyone had this type of schematic.

The schematic is simple, like all class A/B.
Differential amp, driver, output mosfet.
Differential amp is in SMD, but works because there's not distortion.
For that 3-4-5 seconds, the amps work.

I try connect a ReAudio subwoofer and play music for 2 minutes. Amplifier seems work but then......failure..same random N channel mosfet

I never see that problem before in class AB amplifier

I know that Audison/Elettromedia is in my country but they always answer "if you have a problem, send the amplifier to our tech assistance....." (i add....""and pay so much!!"")


I post some measure.

- Gate resistor are 47 ohm.

- There is a 22ohm SMD resistor with one leg on emitter, and other leg together with other 22ohm resistor from other output N mosfets.
If i measure every resistor desoldered, i measure 22ohm correct.
But, if i mount all of them, i measure 3.8ohm...
Maybe 22ohm paralleled for 6 times, makes 3ohm....mmmm..maybe to little.
This afternoon i try to re-check those resistor, but in the P-channel section, all 22ohm paralled for 8 times, makes 3ohm

- For emitter resistor measure i use a volt-ampetometric measure. I put 1V and measure the current. Then i calculated the resistance.
Not professional, but efficent. all value are 0.16, 0.17,... 0.15 +- 0.02. Seems ok for indicated 0.15 value.



Do you have and output mosfet (IRF640/IRF9640) or bjt??
If you have mosfet, could you measure the gate idle voltage??
Mine is 4.5V for n-channel and -4V for p-channel. Seems right


Thank you!
Ciao!


ps: driver pin voltage are:
- external (1 and 3) to rail voltage (both positive or both negative)
- central (pin2) has the gate idle voltage: 4.5V for gates N channel mosfet, -4V for gates P-channel mosfet

[Perry Babin]

Were the original parts IRF640s or IRF640Ns?

[pergo]

Original original i don't know because it arrived with only 2 P-channel and no N-channel.

Original P's are IRF9640 from IR.

Then, i buy fist IRF640N from Ebay (not fake), than IRF640NPbF from RSComponent.

Same problem.

Yesterday i replace all output, both P's & N's with all Pb Free parts, IRF9640PbF and IRF640NPbF.....but same problem.

Random N-channel mosfet fail without blown or something.

Failure consiste in both Gate-Drain and Drain-Emitter 0 ohm reading.

[1moreamp]

OK so we are dealing with a linear amp stage not a digital one, opps my bad .
I see so many of the D class amps. I have not seen a linear mosfet output stage Audison product. All the LRx 400's I have seen here were D class not linear AB.
I currently have 8 LRx5.1k's and four LRx500.2 and three LRx400.2's here for repair. None of which have a linear mosfet output stage in any of them.

The random fet failure is puzzling, I have seen insulator failure in these amps. With holes being blown into the sink through the insulator that has failed, and of course output failures. Audison loves to run high voltage rails inside their amp products, everything I have here has +&- 80 volt rails inside of them.

Are you just replacing the single bad fet upon failure ?

I don't have a sample amp to check the 4 and 4.5 volt gate drive voltage reference your mentioning but something has to be causing these fets to be pushed into failure like they are. And if its not bad outputs then it has to be the driver circuitry breaking down somewhere and possibly causing the outputs to be pushed into failure mode. The speed at which it is happening is just crazy fast.

I have seen similar issues in Phoenix Gold mosfet amps and it was always a problem with the diff amp and driver stage and the bias circuitry running outside of spec. How is this 4 and 4.5 volt signal derived in the circuitry, and what is it using for a temp sensor for thermal tracking ? Please can you post a close up picture of the amp driver section and output stage for me, or PM me with them.
I need to be on the same page with you, so I do not speak redundantly about this amps troubles.

Ciao

[Perry Babin]

I've never had a problem with IRF640Ns but I know that the IRF540Ns cause problems where IRF540s and IRF540As do not. I'd suggest that you try IRF640s (no N) or the IRF640A (if you can find them).

Whatever you find as the solution, please post it here.

[pergo]

Now i go lunch

Then i'll try to replace mica insulator, and i'll post some pics

thank you!

[1moreamp]

Originally Posted by pergo

Now i go lunch

Then i'll try to replace mica insulator, and i'll post some pics

thank you!

Its 4:30 AM here I will nap a bit as i had a late poker game last night and have not rested yet...More later....

[richie00boy]

The reason the N suffix devices are not suitable is that they are the next generation die, which is smaller and actually not recommended by IR themselves for any linear applications.