Soundcraft CPS-150 smoking capacitors

[HSLand]

I replaced the bridge rectifiers, huge caps and resistor as noted in this PDF: http://juggernautmusic.com/wp-content/uploads/2013/10/cps-150-mod.pdf

Here are the exact parts I used:
https://www.mouser.com/ProductDetai...ieJD0LNteWA==&countrycode=US&currencycode=USDhttps://www.mouser.com/ProductDetai...NUr6J7cWJWA==&countrycode=US&currencycode=USDhttps://www.mouser.com/ProductDetai...H9i0E7Pd/9w==&countrycode=US&currencycode=USD
Upon power up the 48v LED was off. There were also some PK lights on my mixer on, and before I could figure out why I saw smoke coming from the PSU. I shut it off, looked inside and noticed caps C33 and C34 bulging and smoking. I don't believe they were bulging beforehand.

What could have caused these two caps to do that?

 

[moamps]

Did you test the power supply alone before connecting it to the console? Most likely you did not replace the 2K7 resistor correctly or the phantom power signal LED is defective. In this case, the voltage on the electrolytic capacitors may rise above their nominal value and cause damage. Disconect the power supply from the console, take the C33 and 34 out, check the connection of the 2k7 resistor and the LED for 48V and try to turn on the power supply again without being connected to the console. Phantom power should also work without capacitors C33 and C34 for testing purposes.

 

[JohnRoberts]

Check polarity of capacitors and look for a maybe shorted rectifier diode.

JR

 

[HSLand]

I checked polarity of the rectifiers, caps, the solder joints of everything I soldered, and I don’t see an issue. The resistor tests ok.

https://ibb.co/wRJJrTMhttps://ibb.co/w78g03g
What isn’t testing ok is the 48v LED. I’m new to this stuff but I’m pretty sure that the 48v LED has gone kaput since when using the continuity/ohm setting on my meter, the good LEDs say about 5.3 ohms and the 48v gives me nothing (except for a couple times when it shot up to like 15).

Most likely you did not replace the 2K7 resistor correctly or the phantom power signal LED is defective. In this case, the voltage on the electrolytic capacitors may rise above their nominal value and cause damage.

I’m not sure if you mean either of these things can cause issues, but if just a little LED can cause caps to burn and smoke, that’s got to be what it is.

 

[HSLand]

Did you test the power supply alone before connecting it to the console? Most likely you did not replace the 2K7 resistor correctly or the phantom power signal LED is defective. In this case, the voltage on the electrolytic capacitors may rise above their nominal value and cause damage. Disconect the power supply from the console, take the C33 and 34 out, check the connection of the 2k7 resistor and the LED for 48V and try to turn on the power supply again without being connected to the console. Phantom power should also work without capacitors C33 and C34 for testing purposes.

Ok weird thing just happened. I took out C33 and C34, then I just jiggled the 48V LED a little and straightened it out. I turned it on right afterwards and now all LEDs are working. What I don't know is if the LEDs are now working due to my jiggling the 48v LED or if it was taking out C33 and C34.

What should I do at this point to get it working? Just replace C33 and C34? Every time I order from Mouser it's a $10 shipping fee, so I want to get all the right stuff if I make another order.

 

[moamps]

What should I do at this point to get it working? Just replace C33 and C34?

There are two ways you can proceed.
First, where you will replace the elcos with the same value but rated for higher voltage, 100V, and then replace the LED with a new one.
Second, which would be an improvement in design, because this original one from Soundcraft is suboptimal. It consists of replacing the C33 and 34 with 47uF/100V, by omitting the LED from the control loop so that the 2k7 resistor will be connected directly to ground, and the LED is connected using an additional resistor of, let say, 22k to 48V output (similar to how it was done for indication + 17V power supply). Picture attached. The 48V output voltage needs to be readjusted.

Every time I order from Mouser it's a $10 shipping fee, so I want to get all the right stuff if I make another order.

That's life.

 

[HSLand]

I've taken a day to think about it. Your first method seems much easier. The 2nd one -- I'm just not too good at reading schematics to figure out what goes where, but if I knew, I could solder pretty easily. I'd also have to look at the schematics in that pdf I posted and figure out where to put my multimeter probes and which pot to turn.

I don't think I should take much of someone's time and ask a million questions here on how to do the 2nd upgrade method, so I think I'll just have to stick with the first one. Changing out a couple caps and an LED shouldn't be a problem.

What I wonder though is that I've read forum posts by others that did exactly what I did and theirs worked flawlessly afterward. No burned caps or anything. I wonder why mine crapped out and theirs didn't.

 

[Matt Syson]

Just a warning that HSLand has been given plenty of advice but is then rude to people who know what the problems will be.
The 'modification to take the LED for 48 volts out of the voltage setting chain is a very valid point as the risk of failure of the LED due to mechanical damage is a lot higher than just the simple, but correctly rated 2k7 resistor alone. taking this to a 1 Watt part at least is very sensible.

Matt S

 

[JohnRoberts]

come on guys, chill out...

JR

 

[swpaskett]

What I wonder though is that I've read forum posts by others that did exactly what I did and theirs worked flawlessly afterward.

Obviously, you did not do exactly what others have done or it would have worked flawlessly for you, too. I reviewed the mod schematic and instructions and everything is good there. I am not trying to rub your face in it, but sometimes to move on and solve a problem we have to admit to ourselves that we caused it. Short of that, there is nothing anyone on this forum can do to help you.

Before doing anything more grab your trusty voltmeter and measure the voltage across C31 and C34. You should see something around 84V across C31 and 48V across C34, and make sure the polarity is correct. If either is off by more than 20% you have problems.

While there is plenty to critique about the design, the PS worked as designed. Few things are ever made that cannot be improved upon and there are some good suggestions in this thread, but your first step needs to be getting it to a working state again, then you can deal with improvements.

I know of only one thing that causes electrolytics to smoke and bulge in a few seconds after applying power: reverse polarity. I have soldered caps in backwards and actually had the casing of one shoot off and hit my hand (it hurt), while others exploded leaving a serious mess, and I have a lot of experience as a builder so get used to the idea that you, too, can make mistakes. So go take another look at those diodes you replaced. I suspect one or more of them is backwards.

As for replacement parts, you obviously need to replace the two bulging caps. You should give serious consideration to replacing all other electrolytics; they too have been stressed and caps in this range are pretty cheap, so live it up. LEDs are cheap and if straightening it makes it work it may have an internal intermittent, so I would recommend replacing it, too. While you're at it, consider replacing the regulator unless you are confident it can survive reverse voltage. I don't know that part well enough to have an opinion but I grok the multiple order penalty when dealing with mailorder. Tayda is a foreign source that is much cheaper, ships quickly and sells reasonably high quality parts.

Inspect the board carefully around the LED and the diodes you replaced under a magnifier. Make sure there is no open track, especially around the LED as shifting the LED could have reconnected an open track around there. It would be worthwhile to check continuity -- open track can be very hard to see.

 

[HSLand]

Thanks for the reply. Before I proceed any further, here's what I mean:

IMG-0119
IMG-0115
IMG-0120

You can verify that the big caps are oriented correctly due to the marking on the board compared to the - symbol on the opposing sides, and you can verify that the bridge rectifiers are oriented correctly due to the + symbols matching up on the face of them as compared to the markings on the board. Only one rectifier and cap are shown because the other ones are too crowded to take a pic of, but believe me, they're oriented exactly the same as the other ones with all the +/- symbols matching up.

The only other thing I soldered in was the resistor and I know for certain the polarity is correct on that because I oriented/matched the color bands on it with the old one. Just kidding, just kidding!

I have a magnifying glass with an LED light on it and I inspected every solder joint 5-6 times over when it first happened, and just now as well. No visible bad/cold solder joints (all are shiny), no broken/lifted traces, and no solder bridging. That combined with checking the polarity on the caps and rectifiers about 10 times each is what is confusing the hell out of me, and you can see that it's right from the pics. So that's why I was pretty positive I did it pretty well.

Before I proceed with any more testing I wanted to post those pics and let you know that C33 and C34 are out of the PCB. Lemme know what you think, thanks.

 

[radardoug]

If the bridge has a shorted diode, you will put ac on the cap and it will smoke. Also, board makers do make mistakes, so check with a meter that the + on the diode bridge is going to the + on the cap.

 

[swpaskett]

Thanks for the pics. One caution: Nobody knows know how much you know, so somebody may suggest something that sounds like we think you're a total noob. We just don't know

I thought your problem was with the +48, so focused on that area. However, the pic of the back side of the board is rather telling. I don't know how much of this is your work vs. original, so a lot of this may be preexisting.

Lots of arrows here:
Green - good joints. You will see that they are clean, no voids, and have nice fillets (the solder flows into the track and wire with no bump or ridge)
Red - hard to tell what's under the black goo. Is that burned flux? You need to clean that off to see what's under it. Some look like they are cold soldered but a lot I just can't tell. There are a lot of these joints that may be black goo or may just be shadow. I didn't mark any where I couldn't be sure.
Blue - empty hole. I presume these are where the caps you removed go.
Purple - void or cold solder. Voids are caused by not enough solder and often do not hold. Cold solder joints are caused by a number of things: not enough heat, not enough solder, corroded wire or board or incompatible solder or flux or corroded/degraded solder/flux.
Brown - Solder bubble. These often hide cold solder joints. If you can't see a wire poking through it's really hard to tell if the joint is good or not.

Did you replace D33-D36? That is the bridge to focus on for C33-C34 problems -- 4 discrete diodes. If one of those is bad you could be getting AC on the caps -- not necessarily a bad thing so long as it stays above ground.

So what do you do with this? I would scrub aggressively with 91% isopropanol and a stiff toothbrush. See if you can get all the black stuff off. If that is from your solder get rid of the stuff and get some from Kester or Alpha. While you're at it, buy some liquid flux and flux up everything before you touch an iron to it. You can run your iron at 800 degrees on this board. I am an old dude and stuck with 63/37 tin lead. It may not be your bag, but it sure is a lot easier to work with.

I have one other suggestion that comes from when I was a kid repairing crap stereos from the local KMart. If I needed to isolate part of a circuit I'd look for a series resistor and cut the lead on one side, leaving a bit of it poking up so I could solder the ends back together. If I couldn't do that I'd flip the board over, grab the xacto knife and slice the track. You might try that on the junction between C32 and the regulator. then you can save what's downstream if it isn't toasted already. Once you are sure the stuff on the left is working right (rectifiers and filter cap), solder the track back together and see what happens. If you aren't sure about your fix, solder in a 1/4A fuse where you cut the track. I used to buy these by the 100 with wire leads attached just for that purpose.

 

[jokeramik]

I replaced the bridge rectifiers, huge caps and resistor as noted in this PDF: http://juggernautmusic.com/wp-content/uploads/2013/10/cps-150-mod.pdf

Here are the exact parts I used:
https://www.mouser.com/ProductDetai...ieJD0LNteWA==&countrycode=US&currencycode=USDhttps://www.mouser.com/ProductDetai...NUr6J7cWJWA==&countrycode=US&currencycode=USDhttps://www.mouser.com/ProductDetai...H9i0E7Pd/9w==&countrycode=US&currencycode=USD
Upon power up the 48v LED was off. There were also some PK lights on my mixer on, and before I could figure out why I saw smoke coming from the PSU. I shut it off, looked inside and noticed caps C33 and C34 bulging and smoking. I don't believe they were bulging beforehand.

What could have caused these two caps to do that?

Hello,

when I look to your pics, I can see that C 34 is bloated. I think thats not OK.

C 34 is the ouput filtering electrolytic for phantom power. I recommend a change. Do you measure 48 volts on this cap?

Best regards!
jokeramik

 

[HSLand]

Ok, I'm back. I took a break from this because I was disappointed in the whole ordeal.

@jokeramik that picture was taken after C33 and C34 began smoking and bulging. I believe they were fine beforehand.

@swpaskett I believe what you're seeing are shadows. I looked at the board again and the joints are pretty clean and shiny. I also looked again at what I soldered with a magnifying glass and I'm not seeing any shorts at all.

I also need to reread your post and go through the board again, but right now I don't know if I should attempt any further because I feel like an idiot that's taking up everyone's time. I may be wrong but something is telling me that there's something stupidly simple going on with it. I really took my time and made sure to be careful. I wish I had a friend nearby that knew how to spot what's going on quickly.

 

[mjrippe]

Where is "nearby"? Your profile only says USA.

 

[HSLand]

I'm in Texas. There aren't many techs where I am, though.

 

[HSLand]

Alright, I'm just gonna send this to a tech. Just have to find one. Looking forward to figuring out what in the world went on. I'll update the thread to let you guys know.

 

[HSLand]

Guys here's an update. It's been over two months, and in that time I watched many youtube videos, consulted with an electronics guy, as well as bought a book on electronics. I didn't even understand most of the stuff posted in this thread at the time, but now I'm on my way.

The power supply is now working after some seriously rigorous testing and learning. While it's awesome that my supply is working, I replaced so many little things in it by this point that I actually don't know what caused the issue in the first place.

I won't go over the whole list of parts I've got in it now (it was a bunch of diodes and new potentiometers etc), but I still have the original C1 and C21 filter caps in there. I thought originally that was what caused the problem, but C1 and C21 are in the +17 and -17 sections, so caps C33 and C34 couldn't have smoked from that as far as I know.

Those old, big caps probably need to be replaced. I've got those 15,000uf/35v caps still but I also bought a new set of 4700uf/50v which is the original spec, and I'm debating which ones to put in. Someone actually told me that the 15,000uf/35v ones were 'bad advice' that I received. Funny how I saw so many others using them though. I've read that the stock C1 and C21 caps will eventually lead to catastrophic failure so I wanted to ask you guys which set of caps I should use.

In the end there must've been something in the 48v section that I messed up that I didn't notice. I was told that the problem was most likely somewhere around REG3, or perhaps the bridge rectifiers. I soldered in R32 which is right next to REG3 and I could've messed something up there. I also put in new bridge rectifiers with heat sinks. Perhaps what caused it was mounting the new rectifiers' metal side to little heat sinks with nothing but thermal paste instead of using an insulator. I've got those rectifiers in there without the heat sinks right now and I'm not sure if I should put them back on, by the way.

So that's the story. Thanks for the help. Still learning. Will be reading a lot in the future.

 

[swpaskett]

Glad you got it working. Doing electronics is sort of circular learning -- you try to understand a circuit before you build it, you build it, then you try to understand what you built. You chase your tail through that mess long enough you become some kind of expert. Then when you think you know it all, or a lot, anyway, something comes along that sucks your pride away.
Have fun with it!