need help troubleshooting an amp

[jaycee]

It might be worth checking what the voltage across them is. 25V capacitors would suggest 18V supplies are being used, but i wouldn't be surprised if the rail voltages are 25V - too close to the limit. Check this in the working speaker.

If it's using 25V supplies, replace the capacitors (in BOTH speakers) with some decent quality 35V units. Go for a good brand such as Nichicon or even Panasonic.

[emremeydan]

Well, i replaced that cap and the problem is gone

I couldn't find a 6800 uf cap the size of the original one (all the ones i found were a bit bigger and wouldn't fit), so i used a 4700 uf cap which had the same footprint (a little taller, but that wasn't an issue). This cap is just for power supply filtering, so it shouldn't make much difference whether it's a 4700 or 6800.

I didn't replace any other caps, cause de-soldering the old cap turned out not to be so easy - the copper "dot" on the PCB got removed as i was desoldering the legs of the cap, leaving behind only a tiny area for soldering the new cap. I managed to solder the new cap to that tiny trace, but didn't want to take any further risks trying to replace the other caps (which look fine for now).

And Jaycee, thanks for the advice. Turns out there are L7812's inside, which suggests that the amp runs on 12V, so i used a 25V cap again. Of course, it could be that it runs on +12/-12V, so 25 might not be high enough to be totally safe, but i'd have to go with an even smaller capacitance value if i were to pick a 35v cap (or it wouldn't fit). Besides, every single cap in the circuit (there is a total of 10 or so electrolytics) is 25V rated. It probably will work fine under normal conditions.

And thanks Bob, for the info. Good to know that they won't explode when not powered - there are many caps on my desk these days (as i'm building a headphone amp on my breadboard)

...
emre

[Bobken]

Hi,

I know the problem you mentioned, and it is not unusual, unfortunately.

What I do if this happens and the solder pad is damaged or lost, is to bend the cap's lead across the PCB underneath, cut it to a suitable length, and solder it to a part of the trace which isn't damaged.

You will probably need to remove any solder-mask which covers the trace with a bit of careful scraping so that the cap's lead can be soldered properly, but I have never been beaten by this, and I have done hundreds of these jobs. Often you can bend the lead and solder it to another solder-pad nearby, but make certain that this other pad is directly connected to the damaged pad, of course, and that it cannot short out on to anything else.

When removing caps, I find it needs more heat usually than when soldering a new joint, and plenty of flux, and even adding some more solder temporarily can help to get these cap leads out quickly. The main thing is not to apply heat for too long as this will tend to cause lifting of the pads.

If the lead cannot be 'moved' freely within a couple of seconds or so, it isn't good to keep the heat on this area. It is better to begin again, maybe with more heat (a hotter iron tip, if it is variable temperature) or maybe with a bigger soldering iron.

Far more damage is likely in my experience if the iron is too small (not powerful enough) or at a too low temperature than using more heat and completing the de-soldering job quickly.

Incidentally, it would be unusual in my experience if you do have a +/- PS and for any one cap like this to be across *both* supplies. With large electrolytics like you have there I would guess that one is across the +12v, and the other across the -12v rails, so you will have plenty of margin on voltage-rating if this is so.
Sometimes a cap is added across both PS rails, but I have never seen anything large in value used like this, and it is generally a very small cap (maybe 1uF or less, and not usually an electrolytic cap, either) which is installed to help just with HF decoupling.

I'm glad that you have solved the problem, anyway.

Regards,