Issues with Mini 3, left channel died, high current draw
Feb 6, 2010 at 3:24 AM Thread Starter Post #1 of 12

Teal

New Head-Fier
Joined
Jan 1, 2009
Posts
34
Likes
0
I recently built a Mini^3 amp and I tested it and it passed all the tests and worked wonderfully.

I then went to work on the case (A real pain) and noticed that the battery wouldn't fit in the case because of the contacts, so I desoldered them, cut away a portion, resoldered them back and connected the battery. When I went to see if it fits in the case, I didn't trim the leads at the bottom and the battery contacts shorted out on the case, and the blue LED lighted up.

Worried, I went through the initial check again. This battery is old and even though I recharged it, it gives out very little current. I measured the quiscent current at 15 mA. Then I switched it to a new 9 V that I did all my previous checks and listening tests with (it's too big to fit in the case). I measured the quiescent current at ~100 mA, and before the battery/case work it was at 25 mA (the proper amount).

I also found the DC offset on output ground and one of the channels to be ~0 and ~3 mV on the other. It was all 2-4 mV before.

When I tried a brief listening test (with a cheap pair of dollar store headphones), only the right channel was working. Also U5 started to noticeably heat up. Nothing heated up when I did the hour+ listening test the day before.

I have no idea what's wrong, but I feel I may have destroyed one of the op-amps. I don't however see how shorting out the battery terminals (all other leads were trimmed) could cause this. I checked every lead of the opamps for a short and there were none (a few had ~150 Ohms). When I tried to use wall power, I noticed U5 began to heat up (too hot to touch), so I didn't try any listening tests on it.

If anyone has experienced this problem and could shed some light, that would be greatly appreciated. Replacing one of these SMD opamps on a populated board is going to be nigh impossible, so I want to know if anything else could cause this before attempting that (in which case I'd lose the other opamp too).
 
Feb 6, 2010 at 4:09 AM Post #4 of 12
With great care.

SMT rework is tricky. Sometimes you have to see if you can feed some very thin wire under the legs on one side of the chip, so you have something to pull up on while you heat the pads. Adding extra solder can make it easier to heat all the pads at the same time.

The first side will be the hard one. Be careful not to burn or tear any traces off the board.
 
Feb 6, 2010 at 4:10 AM Post #5 of 12
Before you remove the damaged component read up about the various smt desoldering techniques. There are tricks to make it possible. There are many sites explaining this. It would probably be wise to use a magnifier also. Go slow. It is a learning experience rather than a disaster.

F
 
Feb 6, 2010 at 4:26 AM Post #6 of 12
I've had to replace U5's a couple of times. You can do it with desolder wick if you must. You'll need to remove the c5 caps and be careful with the input jack.

If you're going to be doing lots of electronic work, go ahead and buy a hot air pencil. It makes life easier when you make mistakes...and you will make mistakes. Those u5's are finicky little bastards.
 
Feb 17, 2010 at 3:08 AM Post #8 of 12
Replacement U5 came in today (have one extra), replaced the old one (by cutting off the leads of the old one with a sharp knife) and it all seems to be working fine now. Quiescent current reads at 25 mA. The only anamoly seems to be that the left channel DC offset is different from the right channel (0.9 mA for left and ground, and 2-3 mA for right). I trust that this won't be an issue, but just wondering why it's different.
 
Feb 17, 2010 at 4:45 AM Post #9 of 12
U5 has two opamps in it, they are completely independent except for the power supply pins. What you're seeing is normal tolerance.
 
Feb 22, 2010 at 3:53 PM Post #10 of 12
One last thing. When I plug in the walwart, sometimes the power turns on, even when the knob is in the off position. If I turn power on and off again, it stops. This doesn't happen all the time. Is there something up with the zener? Also, the walwart does make electrical contact with the case.
 

Users who are viewing this thread

Back
Top