Hello all,
I've got a bit of a head scratcher regarding a dead lm3886. Just wanting to find out what caused it.
Many years ago I built the first version of Russ White's monobloc PCB of Mauro Penasa's My_ref Rev C amp (https://www.diyaudio.com/forums/chip-amps/54571-audiophile-lm3886-approach-38.html#post697610). The amp worked fine, but I had to put it away in storage for a couple of years whilst I was living abroad.
Recovered it recently and plugged it in again. The amp was working fine, but, the light dependant volume control in the amp was having a problem with balance (right channel was louder). As I didn't want to adjust it (annoying little thing to work with), I just removed it and did volume control through a standard 10k log pot. Rewired that, turned on the amp, measured a low DC offset at output (3mv) on the work bench.
Put the amp in the desired place, plugged an input (just an old mp3), and again measured DC offset voltage. 31V. 😕
At this point i suspect the LM3886 failed somehow. But I honestly can't find the reason for this (e.g., don't see any shortings, pcb not touching on chassi, fuses aren't blowing, ...). Now, I would imagine these things don't just fail out of the blue. Or do they? Anyone have any insight into whats going on here?
If ppl want, I can send some pics.
Thanks in advance!
I've got a bit of a head scratcher regarding a dead lm3886. Just wanting to find out what caused it.
Many years ago I built the first version of Russ White's monobloc PCB of Mauro Penasa's My_ref Rev C amp (https://www.diyaudio.com/forums/chip-amps/54571-audiophile-lm3886-approach-38.html#post697610). The amp worked fine, but I had to put it away in storage for a couple of years whilst I was living abroad.
Recovered it recently and plugged it in again. The amp was working fine, but, the light dependant volume control in the amp was having a problem with balance (right channel was louder). As I didn't want to adjust it (annoying little thing to work with), I just removed it and did volume control through a standard 10k log pot. Rewired that, turned on the amp, measured a low DC offset at output (3mv) on the work bench.
Put the amp in the desired place, plugged an input (just an old mp3), and again measured DC offset voltage. 31V. 😕
At this point i suspect the LM3886 failed somehow. But I honestly can't find the reason for this (e.g., don't see any shortings, pcb not touching on chassi, fuses aren't blowing, ...). Now, I would imagine these things don't just fail out of the blue. Or do they? Anyone have any insight into whats going on here?
If ppl want, I can send some pics.
Thanks in advance!
Sorry, that was the extent of my insight.. 🙂
If it's not source related the only other thing I could think of is bad solder joint.
Yes, also my guess. R11 is the 1ohm resistor between poeer and signal ground. Very probably You don't have a broken lm3886.
But. Why did it happen in the first place. I would not use again that source, it might have produced a high common mode output signal..
Best, George
But. Why did it happen in the first place. I would not use again that source, it might have produced a high common mode output signal..
Best, George
Some updates:
R11 is working just fine (even removed it from board to test it). So no luck there.
The signal source does not seem to oscillate or to input any strange voltage (measured only 1mv).
Will prob just try to sub the 3886 to see if I get a similar result. Unless someone has any other ideas.
Thanks!
R11 is working just fine (even removed it from board to test it). So no luck there.
The signal source does not seem to oscillate or to input any strange voltage (measured only 1mv).
Will prob just try to sub the 3886 to see if I get a similar result. Unless someone has any other ideas.
Thanks!
Next question: is it a metal tab LM3886 chip (LM3886T) or isolated package(TF).
If it's not isolated, metal tab -
Did You check conductivity from negative PS to heat sink?
Obviously if it worked before, than shouldn't be a problem, but maybe with time (and physically moving the case..)..
Anyway, just shooting, can be many things.
But these chips do not fail by themselves. Like oscillation, or short circuit, all this.. Too well protected.
One needs a very 'convincing' case.. 🙂
Ciao, George
If it's not isolated, metal tab -
Did You check conductivity from negative PS to heat sink?
Obviously if it worked before, than shouldn't be a problem, but maybe with time (and physically moving the case..)..
Anyway, just shooting, can be many things.
But these chips do not fail by themselves. Like oscillation, or short circuit, all this.. Too well protected.
One needs a very 'convincing' case.. 🙂
Ciao, George
It's a TF (insulated). So no problems with conductivity to HS.
I just really can't see how moving the case could result in this... 😕 I'm really just looking for any guidance here bc I'm quite at a loss.
That's what I fear. If the IC is so resilient, then it suggests the flaw is somewhere else. I think I will just try substituting the IC to see if I get similar results...
The feedback is closed across the input stage..
So also the lm318 could be a suspect. I mean that part of the circuit, because also that chip is not that
easy to fry..
I would start to check all the internal 'service' voltages, like +/- 35V (or zone), +/- 12V for the opamp, all connections are ok from opamp output to lm3886 topology?
Input resistance stays connected to ground? ( 100k)
Ciao, George
So also the lm318 could be a suspect. I mean that part of the circuit, because also that chip is not that
easy to fry..
I would start to check all the internal 'service' voltages, like +/- 35V (or zone), +/- 12V for the opamp, all connections are ok from opamp output to lm3886 topology?
Input resistance stays connected to ground? ( 100k)
Ciao, George
Once I developed an open on a My_Ref board in the feedback loop. The lead to a resistor was a cold solder joint. After locating with a voltmeter it was visible from the top that the solder had not flowed to the lead. Just touched with an iron to fix this problem.
It occurred after much use. It also dumped rail voltage to output, tripping the relay.
Try working back from the output to see if this is the issue. The circuit is easy to follow.
It occurred after much use. It also dumped rail voltage to output, tripping the relay.
Try working back from the output to see if this is the issue. The circuit is easy to follow.
Managed to sub the 3886 today and the DC offset is back to normal (3-4mv). I'm letting it run now with a dummy load to check for temp and to check whether the output voltage remains low. So, I guess I can safely assume the 3886 was the problem (although I really am quite confused as to why this is the case).
Checked all the voltages around the 318 and couldn't find anything wrong (taking the datasheet as reference).
Yes 🙂
I agree that measuring beats imagination 😀 But measuring every pin sounds like a nightmare. I'll do so as a last resort. 😉
Quickly re-soldered all resistors in feedback loop (just in case).
I'll see how things progress and will post updated here.
Checked all the voltages around the 318 and couldn't find anything wrong (taking the datasheet as reference).
Yes 🙂
I agree that measuring beats imagination 😀 But measuring every pin sounds like a nightmare. I'll do so as a last resort. 😉
Quickly re-soldered all resistors in feedback loop (just in case).
I'll see how things progress and will post updated here.
After running the amp with input shorted for 45 min, temp was good and so was the DC on output. However, when I soldered the input to the pot and turned the shaft all the way clockwise, I measured 2V on output (turning the pot as if lowering volume brings voltage back to 3mv).
I'm not quite sure what's going on here. Can someone give me some guidance?
Btw, I've wired the pot as a typical passive volume control (pin 1 is ground, 2 to amp, 3 to source).
I'm not quite sure what's going on here. Can someone give me some guidance?
Btw, I've wired the pot as a typical passive volume control (pin 1 is ground, 2 to amp, 3 to source).
Try C9 for short circuit.. It had seen a constant high level DC.. (the 220uF feedback cap)
If not that, then unfortunately also the LM318 could be damaged.
But Your description fits to a C9 cap shorted out..
And.. Now I realize that something similar happened also to me: my trusted and too much abused test modul had just failed from one moment to the other, and it was the same high output, and LM3886 gone.
Then I realized that with all these years, my non-existent mecchanical support >> and the too much flexing had just worn out some of the pins..
But that was a typical floating prototype 'build'..
Ciao, George
If not that, then unfortunately also the LM318 could be damaged.
But Your description fits to a C9 cap shorted out..
And.. Now I realize that something similar happened also to me: my trusted and too much abused test modul had just failed from one moment to the other, and it was the same high output, and LM3886 gone.
Then I realized that with all these years, my non-existent mecchanical support >> and the too much flexing had just worn out some of the pins..
But that was a typical floating prototype 'build'..
Ciao, George
I am surprised it lasted his long
George,
When I saw your amplifier two years ago I was surprised it still worked. From opamp swaps, feedback mods, and other testing the PCB had lived a hard life.
Hopefully the PCB is still fine and a new LM3886 is all that is needed.
George,
When I saw your amplifier two years ago I was surprised it still worked. From opamp swaps, feedback mods, and other testing the PCB had lived a hard life.
Hopefully the PCB is still fine and a new LM3886 is all that is needed.
Thanks for idea, Geroge. But, alas, C9 is fine... I subbed the 318N and DC offset was still around 2V when input was connected to the pot.
I'll just go every single component to try to find something wrong. Let's see...