G&K 800RB keeps blowing up

[samgraysound]

Hi All,

Someone brought me a G&K 800rb to fix.  Completely trashed power amp. Tons of melted resistors, many failed transistors, two big scorched spots on the PCB, missing components from a previous repair attempt.

I thoroughly rebuilt the power amp, replaced all blown or missing components, tested all transistors. Tried to grind the scorching off the PCB. Replaced all electrolytics and diodes, reset bias.

They picked it up. It lasted through 1 practice then at a gig, starting popping and sputtering. I looked at again, it had blown 8 transistors. I replaced these and some other suspect components.

They picked it up gain. It lasted 30 minutes and popped a fuse. They replaced the fuse, lasted another 30 minutes, and it blew again, and would blow fuses on power up.

Took it apart again. It has blown transistors Q3, Q10, Q12, Q19, Q20, and Q21. 

What could be causing this? The transistors are all over the circuit. Q3 and Q10 are on the 100w cirucit which wasn't even in use.  Do I need to be matching these output pairs? Could it be an issue with the power supply?

Service manual attached. Schematic on page 16

 

[psych60s]

I'd probably try to sit right next to the amp when it starts to fall apart. Checking power supply voltages, components getting very hot.
Does bias tends to vary when the amp is sputtering?

I once have a Studer A68 on the bench. Changed blown transistors, everything was working fine, bias was ok, but after 10 min, as I was 5 meters away I heard popping, turned back and saw flames coming from the power amp board. Your amp didn't burn but you may have something that makes either the power supply or bias to be off specs.

 

[scott2000]

I have some working "legit" pulls from an old QSC amp of 15022 and 23 if you need.... just pm me if so.....

 

[mjrippe]

So it works fine on your bench, but every time "they picked it up" it blows up?  Could it be something in their setup?  Wrong impedance cabinet, guitar cable instead of speaker cable, running it from a generator at an outdoor gig, etc., etc.?

 

[Pip]

They must be run at exactly the impedance  hooked up to the exact output.  They hate impedance mismatch!

Are you using it in crossover mode?

https://medias.audiofanzine.com/files/800rb-service-manual-472053.pdf

 

[samgraysound]

So it works fine on your bench, but every time "they picked it up" it blows up?  Could it be something in their setup?  Wrong impedance cabinet, guitar cable instead of speaker cable, running it from a generator at an outdoor gig, etc., etc.?

I will double check the speaker impedance with them. Its a brand new cab supposedly. Will also ask about the cables. No generators involved to my knowledge.

Sam

 

[samgraysound]

They must be run at exactly the impedance  hooked up to the exact output.  They hate impedance mismatch!

Are you using it in crossover mode?

https://medias.audiofanzine.com/files/800rb-service-manual-472053.pdf

Maybe I should have them bring the cab in and double check they're whole setup.

 

[samgraysound]

I'd probably try to sit right next to the amp when it starts to fall apart. Checking power supply voltages, components getting very hot.
Does bias tends to vary when the amp is sputtering?

I once have a Studer A68 on the bench. Changed blown transistors, everything was working fine, bias was ok, but after 10 min, as I was 5 meters away I heard popping, turned back and saw flames coming from the power amp board. Your amp didn't burn but you may have something that makes either the power supply or bias to be off specs.

ok. I'll try that with a bench test today.

 

[pucho812]

I had an swr doing a similar thing.  In my case it was a heat issue.  The fan circuit failed.  With the lid and repairing had enough ventilation so the fan never kicked over.  But once the lid was on it never kicked over either.  Took some time to figure that one out due to the fan not always running.  So what was normal with the lid off was not when the lid add on.

 

[samgraysound]

I replaced all the blown transistors today. It will pop and spark once every 15 minutes or so. Haven't been able to see where yet, it always happens when I'm looking at the other side of the board. The 100w circuit is working but the 300w is not.

Sam

 

[CJ]

had a problem child like that, decided it was worth more by parting it out.  during dis-assembly we noticed that a ton of traces had bad solder around the fuse holders.

would never had found this if it were not for complete disassembly. it was too late to put humpty dumpty back together again but we did salvage  buckets of factory matched output devices which have been used to fix a multitude of other amps.  did this on a QSC and on a Crown.     

you can bypass the fan turn on circuit, just wire it to a plus rail with a dropping resistor. at this point i doubt the customer will complain about fan noise.

 

[mjrippe]

I replaced all the blown transistors today. It will pop and spark once every 15 minutes or so. Haven't been able to see where yet, it always happens when I'm looking at the other side of the board. The 100w circuit is working but the 300w is not.

I would agree with CJ here, check all the solder joints.  If this amp has been HOT before (you mentioned charred PCB) there may have been enough current/heat to cook the joints.

 

[JohnRoberts]

With amplifier output stage failures, broken output devices can cause subtle failures upstream in the driver transistors, so sometimes you have to replace the drivers, even if they appear OK.

A possible cause for repeat failures is unstable class A bias in output stage. From cold starts the amp can seem OK but after warming up, the increasing temperature, drops Vbe making bias hotter and sometimes running away... so confirm class A bias is correct, and stable.  Maybe set it colder until problem is sorted.

Good luck... Having to replace devices more than once sucks.

JR

 

[abbey road d enfer]

I replaced all the blown transistors today. It will pop and spark once every 15 minutes or so.

It is quite possible that the new 15022/15023 are faster than the originals. Do you see any sign of oscillation?
I would tentatively increase C11.
Indeed setting the idle current is crucial.

 

[samgraysound]

had a problem child like that, decided it was worth more by parting it out.  during dis-assembly we noticed that a ton of traces had bad solder around the fuse holders.

would never had found this if it were not for complete disassembly. it was too late to put humpty dumpty back together again but we did salvage  buckets of factory matched output devices which have been used to fix a multitude of other amps.  did this on a QSC and on a Crown.     

you can bypass the fan turn on circuit, just wire it to a plus rail with a dropping resistor. at this point i doubt the customer will complain about fan noise.

No fan on these models. Do I need to use matched output pairs on these?

 

[samgraysound]

I would agree with CJ here, check all the solder joints.  If this amp has been HOT before (you mentioned charred PCB) there may have been enough current/heat to cook the joints.

Yeah, I could go over the power supply board. The power amp board has basically been completely retouched at this point just from component replacement

 

[samgraysound]

With amplifier output stage failures, broken output devices can cause subtle failures upstream in the driver transistors, so sometimes you have to replace the drivers, even if they appear OK.

A possible cause for repeat failures is unstable class A bias in output stage. From cold starts the amp can seem OK but after warming up, the increasing temperature, drops Vbe making bias hotter and sometimes running away... so confirm class A bias is correct, and stable.  Maybe set it colder until problem is sorted.

Good luck... Having to replace devices more than once sucks.

JR

Thanks.

 

[samgraysound]

It is quite possible that the new 15022/15023 are faster than the originals. Do you see any sign of oscillation?
I would tentatively increase C11.
Indeed setting the idle current is crucial.

Haven't heard any oscillation during testing.

 

[JohnRoberts]

Haven't heard any oscillation during testing.

Oscillation is likely above human hearing, but it often burns up output stage RC traps (RC to ground) that aren't sized for full power oscillation.

It will probably be the last thing you check.

JR

 

[samgraysound]

Rebuilt it and load tested it yesterday biased to 5mv across the load resistors.

After 5 minutes driven, it had jumped to 125mv! After 10 minutes it blew a fuse and more transistors.

There's a circuit that supposed to be adjusting for heat on this right?