Amp troubleshooting... what kind of sound is this?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Trying to troubleshoot a Swan M200MKII which has a LM4766. Below is a video of ARTA measuring the amp's right channel via soundcard:

2018 07 13 22 28 24 - YouTube

The video goes: Off, on, minimum volume, back to original volume, off

So I can understand the "eeooon" part (lightsaber sound?) caused by clipping from insufficient Vcc at powerup and powerdown - the amp continued running for a while after power down so the bulk caps look ok.

But two things I don't understand:

1) Upon power up, why does the amp spend a while at a lower volume while having lots of harmonics of 200Hz?

2) The unstable bass noisefloor is due to a repeating "ts-ts-ts-ts-ts-ts-ts-ts-ts" sound which sometimes last a bit longer like "tssss-brszks" and is only audible on the right channel. (Although no guarantee what's on the left channel until I measure the insides)

I'm thinking the amp chip is a goner.
 
I would replace the local bypass caps and check the main supply caps. It's a shot in the dark, but a pretty cheap thing to try.

Tom

It's a good shot, tho I had to wait a few days for the parts. And $7.50.

So after some reading I learnt that I have motorboating in one channel and after a while oscillation in both channels. LM4766 datasheet does mention the possibility of both occurring without sufficient bypassing.

So I went to check the local bypass caps... and...

ipomKJk.png


What local bypass caps?

So, swapped the old JEC main caps for some low ESR ones. That improved the motorboating to ignorable levels (but not eliminated) although it also changed the channel the motorboating is on. Meanwhile the oscillation came back harder so at least now I'm sure it is power supply related.

Now the question is, for HF passing I only have two 0.22uF MKS capacitors, meanwhile LM4766 each amp has its own Vcc although sharing Vee. Do you think I can make do with both channels sharing the caps or must it be one pair of caps for each channel?
 
I remembered what I forgot to ask...

The datasheet suggests a 10uF for LF bypass and 0.1uF ceramic for HF bypass. But parallel-ing different caps just seems weird to me... even if it would be as close to the amp as possible. Do people do that in practice?
 
Ok, I managed to solve motorboating with 0.1uF to eat up the rectifier noise

And I solved the whine by undoing a previous repair job - electrolytics for the 7815/7915 have been moved too far from the regulator outputs.

But seems that with the above I got the amp back to the original failing state: After a while, I start losing gain and the slew rate gets limited - higher volumes become a triangle wave.

Now, a hint is that this only happens when the mains 50Hz is on. After I switch off the mains, when the amp is powered by only the 10,000uF caps, I get back the gain and a sine wave. Mmm?
 
Oscillation is usually due to poor decoupling, bad pcb layout and long tracks in feedback path.

I had a valve amp motorboat and it turned out to be rectifier diodes having switching spikes. I changed to schottky diodes and that fixed it.

I had a tda7294 amp oscillate like mad. Turned out there is a minimum gain for the tda7294 and I was well below it causing it to oscillate. I fixed that but still got oscillation, that turned out to be too long a feedback resistor paths.
A better pcb layout fixed that problem.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.