I bought a Chinese DAC which has two NE5532 op-amps for I/V and two 470uf 25V capacitors are driving these op-amps as far as I could guess (by tracing the lines on PCB from the caps leading to op-amps). These two capacitors started to bulge in a matter of days. I thought I was unlucky - these must have been defective. So I bought the exact same DAC board again and installed it - and voila, the same two capacitors started to bulge again in a matter of days!
Maybe the voltage is not sufficient ? Out of curiosity I de-soldered the two problematic capacitors and soldered more expensive high quality Panasonic caps, this time with 470uf, 50V (double the voltage). After about a few days, sure enough, they started to bulge again. This is just so weird. It is always these two that are next to the Op-amps. I don't think op-amps are responsible because each time I replaced the DAC board or capacitors, I used different op-amps.
Can someone offer any possible explanations and solutions for this problem? When I touch the caps, they are barely lukewarm (not hot at all). So this can't be due to overheating or being overdriven. It is especially weird because it takes only a few days for this to happen. Would it help to try with 1000uf, 50V caps this time? (Just stabbing in the dark out of desperation)
p.s. I circled the two cap locations, one spot is de-soldered. You can see the bulge on the cap although it is not that obvious from this viewpoint. Please disregard opposite polarity of the cap placement - I just placed the desoldered cap on that spot in order take the pic
Maybe the voltage is not sufficient ? Out of curiosity I de-soldered the two problematic capacitors and soldered more expensive high quality Panasonic caps, this time with 470uf, 50V (double the voltage). After about a few days, sure enough, they started to bulge again. This is just so weird. It is always these two that are next to the Op-amps. I don't think op-amps are responsible because each time I replaced the DAC board or capacitors, I used different op-amps.
Can someone offer any possible explanations and solutions for this problem? When I touch the caps, they are barely lukewarm (not hot at all). So this can't be due to overheating or being overdriven. It is especially weird because it takes only a few days for this to happen. Would it help to try with 1000uf, 50V caps this time? (Just stabbing in the dark out of desperation)
p.s. I circled the two cap locations, one spot is de-soldered. You can see the bulge on the cap although it is not that obvious from this viewpoint. Please disregard opposite polarity of the cap placement - I just placed the desoldered cap on that spot in order take the pic
Last edited:
Have you measured the DC voltage across each cap and compared it to the marked value. That is the first check to make... then assuming it is to high we look for why that might be. Also check the polarity of measured voltage matches the markings on the cap.
Also... you need to measure the voltage across the cap with it coupled up to whatever you are using it all with. Could the next stage have a problem ??? and be reverse biasing these caps which look to be output coupling caps,
Also... you need to measure the voltage across the cap with it coupled up to whatever you are using it all with. Could the next stage have a problem ??? and be reverse biasing these caps which look to be output coupling caps,
Mooly's right, the first thing to do is to measure the voltage across the capacitor(s) with everything connected together, including the equipment which drives the input and the equipment which receives the output. It's possible that the PCB silkscreen is wrong, and it's also possible that the downstream gear is reverse biasing the output capacitor(s).
A quick-and-dirty diagnostic arrangement, AFTER measuring the DC voltage as above, would be to solder a 1N4148 diode (!!) across each capacitor on the bottom side of the PCB. Cathode {band end} to capacitor_plus, Anode to capacitor_minus. This limits reverse bias across the capacitor to 0.7 volts, whereas it takes about 1.5 volts or more of reverse bias to damage an electrolytic capacitor. So the diodes are bulge prevention contrivances. Listen to the entire musical chain with the diodes in place and see whether the sound has changed. If yes, then there's a problem either with this board or with the interaction with other boards. What I would do in this case if I were you, is replace the polarized electrolytic capacitor with a NONpolarized electrolytic capacitor. What seems to be in stock everywhere, is a 220 microfarad, 25 volt, nonpolarized cap. Not quite as many microfarads, so your bass response may be rolled off a little. [link to price and availability] . If it were me, I'd purchase six of those nonpolar caps, just to have enough in hand, to connect two in parallel later (giving 440uF) if the experiment is a success: if nonpolar caps actually do make good sound with no bulges.
On the other hand, if diodes-across-polarized-electrolytics do not change the sound: rejoice. Leave them in for a couple weeks and check whether new bulges have emerged. If not you can now flip a coin to decide whether to go with polarized electrolytics and diodes {i.e. do nothing more} , or to purchase and install nonpolarized electrolytics.
A quick-and-dirty diagnostic arrangement, AFTER measuring the DC voltage as above, would be to solder a 1N4148 diode (!!) across each capacitor on the bottom side of the PCB. Cathode {band end} to capacitor_plus, Anode to capacitor_minus. This limits reverse bias across the capacitor to 0.7 volts, whereas it takes about 1.5 volts or more of reverse bias to damage an electrolytic capacitor. So the diodes are bulge prevention contrivances. Listen to the entire musical chain with the diodes in place and see whether the sound has changed. If yes, then there's a problem either with this board or with the interaction with other boards. What I would do in this case if I were you, is replace the polarized electrolytic capacitor with a NONpolarized electrolytic capacitor. What seems to be in stock everywhere, is a 220 microfarad, 25 volt, nonpolarized cap. Not quite as many microfarads, so your bass response may be rolled off a little. [link to price and availability] . If it were me, I'd purchase six of those nonpolar caps, just to have enough in hand, to connect two in parallel later (giving 440uF) if the experiment is a success: if nonpolar caps actually do make good sound with no bulges.
On the other hand, if diodes-across-polarized-electrolytics do not change the sound: rejoice. Leave them in for a couple weeks and check whether new bulges have emerged. If not you can now flip a coin to decide whether to go with polarized electrolytics and diodes {i.e. do nothing more} , or to purchase and install nonpolarized electrolytics.
Hey, Mooly and Mark
The silk of the board that is slightly visible under the negative side of the EC18 doesn't seem to have the diagonal line that should be on the negative side...
The silk of the board that is slightly visible under the negative side of the EC18 doesn't seem to have the diagonal line that should be on the negative side...
Sorry mason_f18 to confuse you, I just put that (de-soldered) cap on that spot to take pic, and the polarity happened to be reversed. 🙂
Thank you Mark and Mooly. I am a novice when it comes to circuits, but it sounds like the easiest way for me is to just go with bipolar capacitors. (non-polar caps look weird, with a lead coming out of opposite ends). Bipolar one has two leads come out at bottom just like the polarized ones. I'd assume this is OK ? I did a quick search and I found 470uf 35V bipolar cap from Nichicon. I am hoping that whatever circuit design flaw or reverse bias there, will be graciously resolved by the bipolar caps. If this fails (sound distorts etc), then I will try the diodes way.
But either way, I'll try measuring the voltage of caps first as you guys suggested...
But either way, I'll try measuring the voltage of caps first as you guys suggested...
[Here is the manufacturer's datasheet] for Nichicon ES "Bipolar" (a/k/a nonpolar) electrolytic capacitors with the green sleeve. You'll notice that 470uF / 35V capacitors in this product line, have a lead spacing of 7.5mm and a body diameter of 16mm.
However, your PCB appears to expect capacitors with a lead spacing of 5mm and a body diameter of 10mm. You can of course jiggy-bendy the leads to fit the holes, but I worry that the bottom capacitor "EC21" might not have enough clearance to accept a cylindrical capacitor with 16mm diameter. It's a $17.52 roll of the dice.
However, your PCB appears to expect capacitors with a lead spacing of 5mm and a body diameter of 10mm. You can of course jiggy-bendy the leads to fit the holes, but I worry that the bottom capacitor "EC21" might not have enough clearance to accept a cylindrical capacitor with 16mm diameter. It's a $17.52 roll of the dice.
Hi Mark, you are right and I was kind of worried about that as well. What I will do is to remove the terminal block (blue) in order to make more room. I hope this is OK as it has no connection to anything.....
It would still be good to find the reason why because you might be masking a real issue somewhere. Measuring the voltage takes seconds to do, you could even try it with the cap removed and measure across the pads to see what is there.
Hi Mooly, I've put in an order for multimeter. I will measure capacitors in action. I guess the root cause lies in the circuit somewhere else (not the caps themselves) because I tried 3 sets of capacitors as I described in my original post above and they all bulged within a week. I wouldn't be surprised even if it turns out the polarity markings were reversed on the PCB or something crazy like that ....
The only reasonable explanation is that the capacitors are installed in the wrong polarity.
Do not proceed further until you have a DVM so you can verify the correct polarity on the board,
regardless of the marking on the pcb, which can be wrong. A schematic is essential to a novice
for building and testing a pcb, so you should ask for one and post it here.
Do not ever expose either new op amps or new polarized capacitors to an unknown voltage polarity, even briefly.
Also, do not use capacitors that do not properly fit the pcb.
Do not proceed further until you have a DVM so you can verify the correct polarity on the board,
regardless of the marking on the pcb, which can be wrong. A schematic is essential to a novice
for building and testing a pcb, so you should ask for one and post it here.
Do not ever expose either new op amps or new polarized capacitors to an unknown voltage polarity, even briefly.
Also, do not use capacitors that do not properly fit the pcb.
Last edited:
Yes I am dying to get DVM so I can confirm or rule it out... I'll report back on this later for sure
"They are bulging because of the build up of internal pressure. If it builds up enough, the cap may well explode."
So either caps wrong way around or voltage across them is too much.
So either caps wrong way around or voltage across them is too much.
Member
Joined 2009
Paid Member
Nasty. My vote goes for wrong way around, less likely but possible if fake that caps are defective and unable to achieve their rated tolerance
I remember in 1980 doing a city and guilds electronics course.
One project was to build a LM386 audio amplifier.
There were about 18 of us in the class.
We all set off building up the amp.
Then throughout the afternoon now and then there would a loud crack and a capacitor flying about the room.
Clearly some students were putting in decoupling cap wrong way around.
Being a smug git I managed to get mine right way around.
Forty years on and I have never had a cap blow up on me from being wrong way around.
Until recently when I put a 470uf cap wrong way around.
Rather than explode it just sizzled and hissed a bit.
Not so smug now......
One project was to build a LM386 audio amplifier.
There were about 18 of us in the class.
We all set off building up the amp.
Then throughout the afternoon now and then there would a loud crack and a capacitor flying about the room.
Clearly some students were putting in decoupling cap wrong way around.
Being a smug git I managed to get mine right way around.
Forty years on and I have never had a cap blow up on me from being wrong way around.
Until recently when I put a 470uf cap wrong way around.
Rather than explode it just sizzled and hissed a bit.
Not so smug now......
The board could easily be marked with the wrong polarity. The caps would not necessarily fail initial board testing and be shipped with reversed caps without the manufacturer or designer discovering the error until the complaints roll in. Then the easiest fix would be to add an assembly instruction to install the caps contrary to the marking. That works for a while till someone on the manufacturing line is new or forgets.
"That works for a while till someone on the manufacturing line is new or forgets."
Or someone fixes the pcb markings but doesnt tell production !
Or someone fixes the pcb markings but doesnt tell production !
You'd be doing these guys a favor by telling them about this problem.
At least they could rework the finished boards, or include an addendum
to the assembly instructions on how to properly install the capacitors.
Apparently only two of them are reversed on the board.
At least they could rework the finished boards, or include an addendum
to the assembly instructions on how to properly install the capacitors.
Apparently only two of them are reversed on the board.