Audio Amplifier; another Capacitor "upgrade" thread > am new to this!

[tl01magic]

Heya, a tinkerer of audio amplifiers who's almost at par for played with and broke compared to repaired.

Super fun hobby, omg if I could turn back time

So the subject

Teac A-X75mkII DC integrated stereo amp; after cleaning up / maintenance type work I see what appears to be capacitor juice stains from the "main caps" (do they have a specific name?) as imaged below.



Caps were elna 50v 10,000uf  (two of course).  I only have a cheap components tester and the cap measures proper, and they're both close in "measure" as well. 

In another forum more audio specific the reply from seeing the juice stain pic was yes to replace the caps.   

closest I had were 64v 15,000uf. (originals were elna 50v 10,000uf )

Unfortunately I had swapped before actually testing the amp with the suspect original caps at a "demanding" volume. The original caps seem robust and have no bulging / evidence of swelling / overheating ect. BUT I did find traces of the juice stains on the leg of one of the caps, so think that is conclusive. I do not have the experience to look at the stains and know with confidence what am looking at...could be coincidental location of a spilled drink.

I've not played the amp past a few watts with these 50% increase capacitance and x% increase in voltage (does voltage rating even matter in this case of being higher than?) 

So I guess my question is

from your experience / understanding is it worthwhile for me to put the old ones back in and see if the audio is poor when played at volume

remove those too high spec'd caps and find some closure to spec

am fine with the replacement caps in there and the old ones are NOT worth checking out / using even if they sound good.....could fail in short time / cause stress / upset engineering Gods / ect

Thanks so much for your consideration! it is much appreciated

Oh and neat at least to me, I think the rectifier is four "independent" diodes, you can see the legs of them bottom left of the pic...they're big diodes though not sure the spec

total side question, I see like four different colors of electrolytic caps in this amp which I've not seen before; is that uncommon enough to suspect some have been replaced? (nothing obvious from solder)

[TimFox]

The difference between 15,000 uF replacement and 10,000 uF original is probably not important.
Different colors of sleeving around aluminum electrolytic capacitors is normal, depending on vendor and part number.
Hard to tell from the picture if the "juice" stains are from electrolyte (bad) or some kind of fixing glue (ok).
If you have already replaced the capacitors with units that match the PCB hole pattern, it is safer to leave the new ones in than to risk damage to the PCB.

[tl01magic]

Thanks for the reply and sounds good; the pcb at this area is really good so switching back wouldn't be too much risk from my perspective.

The glue was very distinctly different; here's a pick where we see the color of it clearly; can even see some of the dark brown / reddish suspect juice stain on the leg of this cap.

Oh and the replacements were exact same dimensions except just taller. even the legs were same. They're prob about 25 years apart in age.

[TimFox]

With a larger capacitance and higher voltage rating, the "new" capacitors should have more volume than the "old" capacitors, absent technological change between manufacture dates.

[bob91343]

Are you sure that 'leakage' isn't just glue to support the weight of the parts?

[tl01magic]

With a larger capacitance and higher voltage rating, the "new" capacitors should have more volume than the "old" capacitors, absent technological change between manufacture dates.

? less the very thing that is my point; they're 25 years apart!

anyways they're not much bigger, i'd say maybe 20% bigger

[tl01magic]

I've posted the pics and with respect to what I see in person they are representative.  imo it is clear the liquid was very viscous as it had flowed and was dammed by surface tension......imo that is clearly seen

not sure the degree of certainty you're pressing me for...but yes am sure it is NOT glue

here is a pic with the originals in before cleaning dust


online pic upload

and a pic after cleaning dust


bbcode link image

another pic of the original caps bottoms; clearly see the glue. it was actually still nicely plyable; the dark reddish brown stuff was flaky and stained.

[tl01magic]

here's a pic that shows the four diode rectifier I find neat

looking at the bottom two diodes looks like corrosion on the legs perhaps supporting leaked cap.

So is it an unknown whether a capacitor can leak juice and still measure proper capacitance? would it effect other things drastically that my little Polly Pocket tester doesn't "measure" if first statement true?

can tell am biased towards putting the originals back in. (i don't want higher "peaks" anywhere in these 36yr old components if I can avoid it, but do want to be able to use it at volume occasionally)

[andy3055]

I don't think the caps were leaking. If they were, you should be able to see the traces of the chemicals coming out at the bottom. They look clean enough. That glue most of the manufacturers of that era used is known to cause corrosion on metals. Those diodes may have had them exposed to it. Just clean the leads and re-tin them and they will be fine. Make sure to put them back the way they are. As for the caps, leave them as they are. Just clean all that glue from everywhere you can see it.

[tl01magic]

I don't think the caps were leaking. If they were, you should be able to see the traces of the chemicals coming out at the bottom. They look clean enough. That glue most of the manufacturers of that era used is known to cause corrosion on metals. Those diodes may have had them exposed to it. Just clean the leads and re-tin them and they will be fine. Make sure to put them back the way they are. As for the caps, leave them as they are. Just clean all that glue from everywhere you can see it.

Awesome! Thanks Andy; am going to put the originals back in and see if it seems to perform proper.