[marozie]

So here's what happened: I had a perfectly functioning mini^3 which I put together a couple weeks ago, but I didn't have a recharcheable 9v so I didn't bother putting in U2 - I was using it with a non-rechargeable that I hacked to fit inside the case. Meanwhile, I ordered a CTA 325mAH 9v, which came yesterday. So I open it up, put it U2, and plug in the battery. But my battery contact posts were a bit high and it wouldn't fit in the case, so I took them out to grind them down. In the processs of desoldering them I made a funky contortion and while my iron was touching a battery pad (the battery was not plugged in of course) it also hit another pin and there was a small but bright spark - I'm guessing it was from one of the C3s. Now I have no left channel. So, assuming something has been destroyed, how do I diagnose the part(s) that need to be replaced?

 

[CodeToad]

I did something similar once. I inserted the board, with the newly ground battery contacts installed with a battery connected, into the case without clipping the now much longer battery contact leads. Zap.

I had to replace U5.

 

[marozie]

Yes, I'm guessing that's where the problem is, as the L/R counterpart voltages at the U5 are not the same. Damn, how'd you replace it with everything populated?

 

[CodeToad]

Hehe...don't worry too much, you can do it.

First though, what are the dc offsets on the output channels. Actually run through the whole testing procedure again and report back the things that aren't in spec.

I doubt anyone can top this dumbass mistake. I decided the battery contacts needed to be ground down, removed them using desolder wick (this was pre 808), ground them down and them installed them backwards, blew the d4 (thank God it was there). Had to take them out again.

 

[marozie]

I should have written all the values down as I went along. I'll run through it again and report back tomorrow night.

 

[marozie]

I haven't had time to go through any more diagnostics. I'm just going to work under the assumption that I fried U5 since that seems to be a pretty common result of shorting a battery connection. So, I ordered a new chip. I'll report back once I've replaced it.

 

[CodeToad]

If you read through the threads you find that 95% of the problems with the Mini3 are related to fried U5's. It's a very unforgiving little chip. I don't think that most DIYer's are prepared for the precision necessary in assembling one, thus, it's a learning experience. I know it was for me. I think I've made every mistake you could make with the Mini3. Even small mistakes like the meter's probes slipping can cause much heartache.

The upside is clearly evident in the oscilloscope pics on the Mini3 web site.

_Personally_, I've found it much easier to work on U5 by removing C5's. YMMV.

 

[Teal]

Quote:

Originally Posted by CodeToad /img/forum/go_quote.gif I did something similar once. I inserted the board, with the newly ground battery contacts installed with a battery connected, into the case without clipping the now much longer battery contact leads. Zap. I had to replace U5.

That is EXACTLY how I fried my U5. Same symptoms, left channel died, quiescent current at 200 mA. Replaced U5 2 days ago and everything is working fine now.

The way I replaced it is I took an X-acto knife and cut off the leads of U5 (I didn't remove any parts and you don't have to, heck you might destroy something). I was careful not to score the board and leads cut off pretty easily (do it right at the top where it meets the component and move the blade back and forth like a saw). Then you remove the component and use the soldering iron to individually remove the pins.

Make sure to brush away any small bits of metal. Be careful not to touch the metal pad in the middle with your soldering iron as it will cause bumps and your part will not sit properly on the metal plane (and thus won't dissipate heat as effectively). Then use a soldering wick if necessary to flatten out the pads and it will look just like you never soldered on U5 in the first place.

 

[marozie]

Cool. Thanks for the advice guys. Yeah, U5 seems to be the consensus "I shorted something and now I don't have a left channel" culprit. I should have read AMB's initial checks section more closely though, he says specifically that if you're going to do any work you need to turn the power on after you remove the battery to discharge the rail caps. One thing's for sure, I've certainly never work with such a delicate chip!

 

[Spacehead]

I once inserted AD8397 wrong way to dip8 socket (on poorman's browndog dip8adapter).
It smoked a bit and a little zap. I turned power off very quickly and just removed and inserted it back right way.

It worked

Some distortion added and little less punch, but fine anyway.

More common op amps survive without smoke because they have output short circuit protections. They shut down if there is too much current drawn.

With AD8397 you need to be extra careful. I wouldn't plug the headphones while it is on, that can short circuit it too.

Good luck fixing your amplifier.

 

[marozie]

Just to follow up, replaced U5 and we're good to go...

 

[ericj]

Quote:

Originally Posted by CodeToad /img/forum/go_quote.gif If you read through the threads you find that 95% of the problems with the Mini3 are related to fried U5's. It's a very unforgiving little chip. I don't think that most DIYer's are prepared for the precision necessary in assembling one, thus, it's a learning experience. I know it was for me. I think I've made every mistake you could make with the Mini3. Even small mistakes like the meter's probes slipping can cause much heartache. The upside is clearly evident in the oscilloscope pics on the Mini3 web site. _Personally_, I've found it much easier to work on U5 by removing C5's. YMMV.

I'd built several amps before i built my mini3, and i fried my U5.

Didn't have any trouble replacing it. Even installed an opa2132u in there just to make sure the rest of the board worked until the replacement turned up.