Are my caps leaking? (and did it toast something?)

[m5sime]

Hi! - Today while fiddling around with a new DC supply to my DAC with my Amp in standby.. I noticed two small plumes of smoke from the amplifier and the lovely (not) smell of toasting components.. I was surprised to say the least! (as it was in standby - relays not feeding power I assumed). Still not sure what fiddling with a 12vdc supply to my DAC would mean via the balanced XLR to the amp.. 

 

So I took the case off for a look and a sniff.. I am not sure whether the caps are potted/bonded and this has discoloured or have my caps started leaking some time ago?

 

Votes please.. Also looking on the rectifiers (where the smoke approx came from) I see white 'speckled dots' I wonder if the 3 ICs (mosfets?) were the things that cooked and boiled their thermo paste onto the rectifiers?

 

Amp still works as I pulled the power quickly. Still amazed at how HUGE the Toroid is in this amp.

 

Not happy tho

 

 

 

 

And now the 'spattered' bridge rectifier PCB?

 

 

 

[gmgraves]

 

This is pretty good indication of at least ONE shorted cap and perhaps both. Go on the reverse side of this circuit board with a voltmeter (a cheap VOM from Harbor Freight will work OK) set the meter to ohms/continuity and measure across + and - on those caps. If one or both are shorted, you will get full continuity on the lowest scale on the meter (a digital voltmeter will beep if there's a short). If you do get the indication of a short-circuit, and have any soldering/desoldering skills (or perhaps a friend who does) de-solder both caps, clean the discoloration off of the board with some solvent and check across the BOARD where the caps were soldered for continuity again. If the short is gone. Test each cap separately while it's in your hand. Perhaps only one is bad, If so, one will read shorted, the other will read open. I'd replace them both, either way. If the short's still there, check that third capacitor that shows some "brown" burn marks on it as well by removing it. If the short goes away when you remove the two blue and white caps or that smaller , order a pair on the internet. Everything you need to know about them is printed on them. Be sure to order the PC type of caps with both leads coming out of the same end, Observe polarity carefully when re-inserting. IOW, + to the + hole!. While your at it, check continuity between the output transistors' leads too. On any one transistor, there should not be continuity (a condition of no resistance) between any two of the three transistor leads For each transistor (or FET). There's no need to dismount these in order to test 'em. 

Look for burned or badly discolored resistors. Often shorts on circuitboards cause too much current to flow through resistors - much more that they are designed to carry, so they get hot and self destruct. Check small-signal resistors as well as larger power resistors. It might still "work" now, but if you leave it on it will begin to smoke again (that discoloration you see was caused by something getting hot!) and maybe catch fire. Not good. Those white specs are probably nothing and were mostly there from the factory. Perhaps some spattered adhesive or chemical used in construction. These are just suggestions. It is not possible to troubleshoot hardware by e-mail remote control. Good luck!

[m5sime]

Thank you @gmgraves!!  I'm pretty handy with soldering etc but this is one packed and cramped amp. I might just retire it. Thank you. I'll see how I feel about strip down tomorrow. 

 

Really appreciate your feedback 

[mansr]

1 hour ago, gmgraves said:

 

This is pretty good indication of at least ONE shorted cap and perhaps both.

I think that's more likely glue that's meant to be there. There's more of it elsewhere on the board.

[m5sime]

2 minutes ago, mansr said:

I think that's more likely glue that's meant to be there. There's more of it elsewhere on the board.

Thank you. That's my dilemma....  Is it glue gone brown or leakage..  I may well as suggested check cap resistance..  The amp is working and has been fine. 

 

The smoke was very odd 

[bgoc]

1 hour ago, m5sime said:

Thank you @gmgraves!!  I'm pretty handy with soldering etc but this is one packed and cramped amp. I might just retire it. Thank you. I'll see how I feel about strip down tomorrow. 

 

Really appreciate your feedback 

I just some cambridge audio 851s on Audiogon or US audio Mart from the Canadian distributor ,there refurbished but bran new with full warranty at 60% off, that might be the easy route.

[mansr]

I was recently goofing around with an old Cambridge Audio AVR when it started emitting smoke. I found this behind the speaker terminals:

Only visible after some disassembly.

[mansr]

10 minutes ago, m5sime said:

Thank you. That's my dilemma....  Is it glue gone brown or leakage..  I may well as suggested check cap resistance..  The amp is working and has been fine. 

 

The smoke was very odd 

Brown glue is not unusual. If you poke it with a screwdriver or similar, it should be obvious if that's what it is.

[m5sime]

More investigation needed but no obvious burns that I can see and I caught it quickly luckily 

[m5sime]

Just now, mansr said:

Brown glue is not unusual. If you poke it with a screwdriver or similar, it should be obvious if that's what it is.

None or the caps are bonded down that have the brown goo so hard to tell if it was supposed to be a bonding. They wobble a little 

[mansr]

3 minutes ago, m5sime said:

None or the caps are bonded down that have the brown goo so hard to tell if it was supposed to be a bonding. They wobble a little 

If you prod the brown substance, is it rubbery or brittle. Leakage from capacitors is usually nothing like glue in texture. Also, capacitors rarely smoke gently. They either leak slowly or blow up with a bang. Leaked electrolyte can cause shorts resulting in other components overheating, however.

[m5sime]

14 minutes ago, mansr said:

If you prod the brown substance, is it rubbery or brittle. Leakage from capacitors is usually nothing like glue in texture. Also, capacitors rarely smoke gently. They either leak slowly or blow up with a bang. Leaked electrolyte can cause shorts resulting in other components overheating, however.

The dismantling has begun and will continue tomorrow am (UK time) bit early signs look like it may be baked old adhesive bonding...  Thank you 

[m5sime]

I got to the bottom of the cap issue.. They are fine.. They measure ok with a multimeter and I found on stripping down the amp a little that other caps have this. So it looks like old glue.

 

[m5sime]

I did get to the bottom of my smoke though.. two teeny resistors gave their life for a noble cause.. Now just to find out from the manufacturer the values as I cannot see from the damage.. Same each side. Thank you @gmgravesand @mansr

 

[mansr]

3 hours ago, m5sime said:

I did get to the bottom of my smoke though.. two teeny resistors gave their life for a noble cause.. Now just to find out from the manufacturer the values as I cannot see from the damage.. Same each side. Thank you @gmgravesand @mansr

 

This looks similar to what happened with mine. Is the CN3 connector by any chance the speaker output? You'll probably find that the resistor is connected in series with the little capacitor (C29) next to it across the output. It's likely that only the paint on the resistor was smoked and it is still close to its proper value which if I'm right is around 10 ohms.

[m5sime]

1 minute ago, mansr said:

This looks similar to what happened with mine. Is the CN3 connector by any chance the speaker output? You'll probably find that the resistor is connected in series with the little capacitor (C29) next to it across the output. It's likely that only the paint on the resistor was smoked and it is still close to its proper value which if I'm right is around 10 ohms.

Hi Mansr, No, the connector is the input I believe.. This amp board is pretty complex being all discrete components.. I think the resistor is part of the on-amp board PSU. Both smoked and both measure 19k8 so I am hopeful they survived. I have run the amp without inputs for 30 mins and checked temps and all ok. 

 

I am getting the service manual from Cambridge Audio to check the resistor values.. So bizarre that when tinkering 'upstream' with the DAC with the amp in standby that these could both smoke.

 

I note that Cambridge Audio reduced the complexity with the 851 replacement.

[m5sime]

An internet image showing how much going on in there. Still chuckle at how deep that toroid is..

[mansr]

43 minutes ago, m5sime said:

Hi Mansr, No, the connector is the input I believe..

In that case it's a different problem than I had.

43 minutes ago, m5sime said:

This amp board is pretty complex being all discrete components.. I think the resistor is part of the on-amp board PSU. Both smoked and both measure 19k8 so I am hopeful they survived. I have run the amp without inputs for 30 mins and checked temps and all ok. 

Strange. Those look like ⅛ W resistors. To reach smoking temperatures they'd need to be dissipating quite a bit above their rated power which at that resistance means a considerable voltage. 140 V would give about 1 W. Burnt resistors generally increase in resistance as the conductive material is destroyed, so it's possible the original value was lower.

[m5sime]

48 minutes ago, mansr said:

In that case it's a different problem than I had.

Strange. Those look like ⅛ W resistors. To reach smoking temperatures they'd need to be dissipating quite a bit above their rated power which at that resistance means a considerable voltage. 140 V would give about 1 W. Burnt resistors generally increase in resistance as the conductive material is destroyed, so it's possible the original value was lower.

Plot thickens.. I have the repair/service manual. Both resistors are not the right value now.. But they seem to be some kind of protection circuit.. There is the full PCB schematic for the amplifier board. and then this on the page.

 

 

would you have any idea what this is for? Is this a PCB ground circuit to limit current?

[mansr]

19 minutes ago, m5sime said:

Plot thickens.. I have the repair/service manual. Both resistors are not the right value now.. But they seem to be some kind of protection circuit.. There is the full PCB schematic for the amplifier board. and then this on the page.

 

 

would you have any idea what this is for? Is this a PCB ground circuit to limit current?

It probably has something to do with the chassis grounding assuming that's where the screw goes.

[m5sime]

2 hours ago, mansr said:

It probably has something to do with the chassis grounding assuming that's where the screw goes.

 Think I solved my mystery.. My new DIY starquad Dc cable managed to short +12v from my external dac PSU. To the shield of the connector when I bent it as installing it. The chassis of the external linear PSU is grounded so I assume I sent +12vdc into mains earth. This in turn created a voltage across the amp resistors that isolate the chassis earth from the PCB signal ground. 

 

I found the cable short when investigating...  Cable being re made. I hope nothing else suffered! 

 

Thank you for helping 

[mansr]

2 minutes ago, m5sime said:

 Think I solved my mystery.. My new DIY starquad Dc cable managed to short +12v from my external dac PSU. To the shield of the connector when I bent it as installing it. The chassis of the external linear PSU is grounded so I assume I sent +12vdc into mains earth. This in turn created a voltage across the amp resistors that isolate the chassis earth from the PCB signal ground.

That makes sense. 12 V across 47 ohms generates 3 W of heat, far more than that resistor is designed to handle.