Hi All...
I had a LM3875 chip let the magic smoke out yesterday, not unheard of, but it was first time it's ever happened to one of my builds. But here's the kicker, it wasn't being used when it happened, it was just sitting there with the power on and no music playing. I've been using that amp for about 9 years without any problems, never abused it at all, and it never showed any signs of impending failure. I checked it out and couldn't find anything else wrong, power supply voltages all OK, no swollen caps, nothing shorted anywhere, and none of the other components are damaged or out of spec. The only question I have at this point is; can leaving a system powered-up 24/7/365 lead to this kind of failure?
Mike
I had a LM3875 chip let the magic smoke out yesterday, not unheard of, but it was first time it's ever happened to one of my builds. But here's the kicker, it wasn't being used when it happened, it was just sitting there with the power on and no music playing. I've been using that amp for about 9 years without any problems, never abused it at all, and it never showed any signs of impending failure. I checked it out and couldn't find anything else wrong, power supply voltages all OK, no swollen caps, nothing shorted anywhere, and none of the other components are damaged or out of spec. The only question I have at this point is; can leaving a system powered-up 24/7/365 lead to this kind of failure?
Mike
I have heard other wiser (than me) guys here in another thread talking about oscillation caused with a volume pot at one of its extremes on travel (I cant honestly remember whether its was fully open or closed)
Was the volume set at minimum?
Perhaps some of those wiser guys can help you further.
As far as 'should they fail after time has elapsed?'
YES.
Depending on the IC, dissipation, quality of manufacture etc all contribute to the ultimate lifespan. From the little I know, a substrate temperature above 55*C lowers life span.
Or is it possible the amp oscillated with the volume down fully?
Could.it.be possible you were running fairly warm but not overly so, and yet amassed many many hours listening (perhaps with a QC bare pass device?)
This is where my (questionable) knowledge ends 🙂
Good luck
Was the volume set at minimum?
Perhaps some of those wiser guys can help you further.
As far as 'should they fail after time has elapsed?'
YES.
Depending on the IC, dissipation, quality of manufacture etc all contribute to the ultimate lifespan. From the little I know, a substrate temperature above 55*C lowers life span.
Or is it possible the amp oscillated with the volume down fully?
Could.it.be possible you were running fairly warm but not overly so, and yet amassed many many hours listening (perhaps with a QC bare pass device?)
This is where my (questionable) knowledge ends 🙂
Good luck
Last edited:
That's what I thought. I guess I'll just chalk it up to a random failure. I'll just replace the 3875 sometime later today and see how it goes.
Mike
Thanks for all the replies. The rail voltage is +/- 26v at idle, and there are no mains issues as far as voltage or noise (I had access to a mains monitor in one of my recent jobs, borrowed it out of curiosity to check our power at home, it was exceptionally clean and steady at that time and I have no reason to think anything has changed.) I don't ever listen to music at excessive volume levels (I like my ears to be in working condition.) At idle the heat sinks have never gotten beyond slightly warm, whether at idle or playing music.
So I guess I'll just fix it and move on. Maybe I'll turn the power off when not in use from now on. 😉
Mike
So I guess I'll just fix it and move on. Maybe I'll turn the power off when not in use from now on. 😉
Mike
Or maybe I did something (Or failed to do something) to pi$$-off the Gods of Electrons...I don't know and willingly accept that I may never know exactly what or how it happened. It is a little weird that just this one chip failed and none of the other five in the system did, in fact, none of the other components failed, no burned resistors or damaged caps at all.
Mike
It's naff I agree. Id blame a bad device, you seem to have covered the likely suspects and come up answerless. I guess a new IC is the only way you'll know. If there is still a failure you'll know you missed something 😉
Maybe I should find a Witch Doctor, that dark cloud seems to drop by for a visit far too often lately.
Mike
Maybe fitting more RF attenuation.
On ALL the cables entering the amplifier chassis.
Have you acquired a new "wireless device"?
On ALL the cables entering the amplifier chassis.
Have you acquired a new "wireless device"?
Just a follow up...
I replaced the chip, put it on the 'scope and everything is A-OK, so I'm back in business. I guess I'll just right it off as one of those mysterious things that can happen without warning. Thanks for the replies.
Oh, look at the time...I'm gonna go listen to some tunes. 🙂
Mike
I replaced the chip, put it on the 'scope and everything is A-OK, so I'm back in business. I guess I'll just right it off as one of those mysterious things that can happen without warning. Thanks for the replies.
Oh, look at the time...I'm gonna go listen to some tunes. 🙂
Mike
Just a follow up - follow up...
I have an idea what may have caused that chip to fail. When I was assembling the amp I soldered the chips to the PCB's first and then attached them to the heat sinks. On one channel the assembly went fine, on the other I remember the chip was at an angle and wouldn't lay flat on the heat sink, so I just bent the chip over to force it, so I'm thinking that's where it got damaged and caused it to eventually fail. So in the future I'll mount the PCB to the heat sink first, then the chip, and lastly, solder the chip pins to the board.
Mike
I have an idea what may have caused that chip to fail. When I was assembling the amp I soldered the chips to the PCB's first and then attached them to the heat sinks. On one channel the assembly went fine, on the other I remember the chip was at an angle and wouldn't lay flat on the heat sink, so I just bent the chip over to force it, so I'm thinking that's where it got damaged and caused it to eventually fail. So in the future I'll mount the PCB to the heat sink first, then the chip, and lastly, solder the chip pins to the board.
Mike
Poor contact with the heat sink would mean running hotter *all the time*. Bending the leads or tab to force it may compromise the seal and allow moisture ingress. Either one alone will accelerate failure and both certainly would.
If you had been cranking it up on a regular basis, it probably would have failed MUCH sooner.
If you had been cranking it up on a regular basis, it probably would have failed MUCH sooner.
Definitely hydrogen embrittlement of the leads. Alpha particles wouldn't ever get past the epoxy. 😀
(This post is in jest)
Glad the OP is back running!
The reason for forcing it was to ensure that it WAS laying flat for good contact with the heat sink. I believe it probably was damaged from physical strain on the leads, causing damage to the chip where the leads enter the case. I never crank it up enough to cause running hot, even after several hours of use the heat sinks are only slightly warmer than at idle. And remember this occurred after many years of use, so maybe moisture intrusion could have happened from physical damage to the chip. Anyhoo... Lesson learned!
Mike
- Status
- Not open for further replies.