• Please review our updated Terms and Rules here

PSU recapping - what's the general consensus on when to do it?

pkhoury

Experienced Member
Joined
Dec 6, 2007
Messages
274
Location
Bandera, TX
Like the title says - when should we recap? I picked up a SPARCstation IPX a few days ago on ebay with a known 12v power issue. I've read A LOT about failed IPC/IPX power supplies and figure I'll recap it to see if it solves things, and as a precaution.

That being said - if I have another supply that works just fine, like in my SPARCstation 2, is there a problem with recapping it just to recap it, or should I be waiting for a problem to crop up first?
 
Well I agree with Chuck, Only when sober. I also tend to avoid re-capping unless there is an issue. The quality of capacitors seems to vary at random some go on for years, others seem to fail at random. I guess temperature makes a difference as caps have a temperature rating. The fault doesn't need to be the classic PSU only starts or the first Tuesday in the month, but intermittent resets and other odd results. It can also be obvious signs of stress such as fluid leakage or bulging. So when these happen I get the de-solder gun and iron out and fix...
... but if all is well I tend to stay chilled.
 
I think you should do it when indoors.

Also do it with proper equipment.

And much like blaming the flyback for everything CRT related, maybe dont rush to the conclusion of a bad psu. I recommend you do everything else first as its probably not the psu. Can you just test the output voltages on it? If they are correct... leave it be.
 
It is very easy to rush to the wrong conclusion, even if you're 99% aware of the most common issues. I had an IBM 5160 PSU of unknown working condition, and I used it to test a clone motherboard. No power. Thinking that the motherboard must have had a short, I unhooked it and still no power. So then I assumed that it was the PSU and not the motherboard, so I connected a brand new XT/AT clone PSU of much higher wattage. Immediately there were the sparks of a shorted tantalum. I had briefly forgotten that the 5160 PSU needs a load to start up. After the beefy PSU cleared the short, I went back to the 5160 PSU and it worked fine.
 
The SPARCstation IPC/IPX PSUs (same model made by Sony in both) are known for capacitor faults. I have both, and the PSU caps had already leaked in each of them, so I recapped. I think PSUs in other SPARCstations are not as bad as in the IPC/IPX. Small tip: after recapping, you may have to readjust the potentiometer for the 5V rail. After repair, my PSUs would shut down again immediately after power on. The 5V setting was a tad too high and tripped the over voltage protection. Took me a while to figure out. After adjustment, they have been working very reliably.
 
The SPARCstation IPC/IPX PSUs (same model made by Sony in both) are known for capacitor faults. I have both, and the PSU caps had already leaked in each of them, so I recapped. I think PSUs in other SPARCstations are not as bad as in the IPC/IPX. Small tip: after recapping, you may have to readjust the potentiometer for the 5V rail. After repair, my PSUs would shut down again immediately after power on. The 5V setting was a tad too high and tripped the over voltage protection. Took me a while to figure out. After adjustment, they have been working very reliably.
Thanks for this tip. I get to open mine back up and check this out. Brought the IPX with me for our club's GTG and it wouldn't power up. Thankfully, I also brought my SS20 with me.
 
A decade ago I would of left it alone until it actually caused a problem. That is not the case anymore, in part because I now own several really nice and somewhat rare machines where I ignored preventative maintenance and now have power supplies with major damage resulting from electrolyte leakage or cascading MSOFET failures. Machines with complicated switch-mode power supplies and no known schematics. Machines where at that point, they are rendered static display pieces.

If it's 20 years old, regardless of build quality it MUST be serviced. I've been curb stomping this into people's heads now for several years and they still get hung up on "well they aren't all bulging and leaking so I don't see an issue."
Recapping is my professional $JOB. I'm not skimming Google or watching youtube videos. I get a [REDACTED] on my desk and it's not even ten years old I can bet you $50 I know what is making it not power on.

In the past I would of done it simply because they were known awful brands like Teapo or CapXon which were known to fail. When it became clear about 8 years ago that surface-mount capacitors (and early miniaturized radial capacitors) from the late 80's and early 90's were not so much failing like a normally bad capacitor but simply failing and drifting out of spec from time-based chemical breakdown it was obvious that the old visual check and brand assurance was no longer going to work. Panasonic, United Nippon Chemicon, Elna and even Rubycon. Brands that we considered "good" during the first wave 20 years ago when it was still Athlons and P4's I'm now finding have again failed purely due to age. Anything beyond that point while build quality did lend a hand in the longetivity of the PSU the capacitors are now out of their expected lifespan and no longer on borrowed time.
Most importantly is they are leaking out their electrolyte, not venting. The exceptionally juicy ones will leave a mark but I've found caps now where the electrolyte eats and eventually causes the electrode leads to fail. No visible leaking, no bulging and not signs of failure but you pull the cap and either a leg falls off or it reads completely open.

The lunchbox power supplies are especially bad. When they leak they drip down and ruin the logic board. I have two machines I ahve yet to get to but I store them upside-down. That means the PSU gets a soaking but at least a PSU replacement is easier to make than a logic board.
 
The caps that were at this location looked great:



And so did these:



The supply was working fine, machine was running without issues. Had to take it apart for dust removal cleaning and decided to recap because I caught the faintest whiff of leakage with the top off.
 
So my 70-year old Western Electric power supply needs servicing? Wonder where I'd find Pyranol can-type replacements. Seems to be working fine, however--regulation is spot on +250V with less than 10 mV of ripple.
My point is that linear power supplies are generally easier (i.e. less self-heating) on filter caps than when caps are used in a SMPSU. However, my 50 year old Lambda rack-mount PSU did dry out its "Long life" Mallory cap did need that one replaced.
 
Last edited:
So my 70-year old Western Electric power supply needs servicing? Wonder where I'd find Pyranol can-type replacements. Seems to be working fine, however--regulation is spot on +250V.
Right, being obtuse about it helps keep machines running.
 
Not being obtuse, just pointing out that there are huge differences in power supplies as a group.

Those big can-type capacitors are rated at 40 µF at 350VDC. They've got a lot of life left in them. Particularly on really vintage equipment, I hate to see "shotgun" replacement of components just because someone "says". In general, new capacitors are smaller and more prone to failure than their old great-grandfathers.
 
I really haven't brought this issue up with collections management about all of the 90s era equipment
in CHM's collection yet. They did take care of battery mitigation, but there has never been a discussion
about how to deal with electrolyte leakage.
 
Those big can-type capacitors are rated at 40 µF at 350VDC. They've got a lot of life left in them. Particularly on really vintage equipment, I hate to see "shotgun" replacement of components just because someone "says". In general, new capacitors are smaller and more prone to failure than their old great-grandfathers.
Western Electric, mil-spec, etc. is by and far the exception to "doesn't need replaced" when it comes to truly old stuff and capacitors, and I suspect you *know* that. How many times have you seen where someone stripped the rectifier plates, or worse, smoked a B+ transformer, because "it's working with those old dry 'lytics, I don't see why I should replace them?" I've got to be half your age and have seen it enough times, often in pretty nice stuff, to know better.
 
If it's average consumer-grade stuff, yes, I'd get out the meter and check the electrolytics. But I'm not going to junk a bunch of oil capacitors because they're "too old". And there's a whole bunch of fairly recent stuff that's been destroyed by "shotgun replacement" of caps. Any time that soldering iron comes out, one must take into account the possibility of damage.

It's the "shotgun" advocates that I have the biggest problem with.
 
If it's average consumer-grade stuff, yes, I'd get out the meter and check the electrolytics. But I'm not going to junk a bunch of oil capacitors because they're "too old". And there's a whole bunch of fairly recent stuff that's been destroyed by "shotgun replacement" of caps. Any time that soldering iron comes out, one must take into account the possibility of damage.

It's the "shotgun" advocates that I have the biggest problem with.
So you do know old electrolytics blow stuff up, then? OP is asking about electrolytics; specifically, low ESR stuff in semi-modern power supplies. Things that are known to have problems, due to a number of issues ranging from growing pains to counterfeits. Where doing nothing gets you boards like I provided pictures of.

I have seven 90s pick-and-place machines that have required full recaps in their AC servo drives. Nothing else is wrong with the drives, but if you run them like that, resetting or locking out error conditions, eventually you blow up power electronics in them. You can still get them from machine rebuilders for around $1000/ea.

The controllers for those servo drives have 90s surface mount capacitors on them, under conformal coating. It helps to hold the leaking electrolyte in. Every single one we opened had at least one leaking surface mount electrolytic.

If I open a power supply with any real age on it for basically any reason, it now gets a recap. If I get some vintage test equipment in, it gets a recap before I bother to calibrate it. Tube equipment gets a recap not only for electrolytics (which all go, no matter what) but also wax paper capacitors and anything that looks like a wax paper capacitor. I don't care if it *might* be a di-film or if some guy's tester with a 9V battery says it's OK, or even if the HV Heathkit cap popper says it's OK. It's not worth blowing up something nice for $10, $20, or even $50 in capacitors. Does that make me a "shotgun advocate?"

W.R.T. stuff being destroyed by bulk recapping, I would argue that the kind of folks who butcher a cap job would've probably wrecked it anyway.
 
I really haven't brought this issue up with collections management about all of the 90s era equipment
in CHM's collection yet. They did take care of battery mitigation, but there has never been a discussion
about how to deal with electrolyte leakage.
I toyed with this a few times for items I knew I was not going to get back to for any length of time and quickly found out that "just pull the caps" is considerably harder than it sounds.
Other than the obvious manifest of where each capacitor was and what their value was you also have to include their composition, special characteristics like it is bipolar, size, lead spacing and after all that you need to safely store that manifest with the machine somewhere where it will not be lost because you lose it, have fun. Adding to that you are removing KNOWN bad capacitors. We could very well disregard others because TODAY we know they are stable but next week it could be different. Those vertical standing green dip film capacitors are starting to fail in radios and we usually associate film capacitors as pretty darn reliable.
 
So you do know old electrolytics blow stuff up, then? OP is asking about electrolytics; specifically, low ESR stuff in semi-modern power supplies. Things that are known to have problems, due to a number of issues ranging from growing pains to counterfeits. Where doing nothing gets you boards like I provided pictures of.
Oh dear, my Sony AM FM Radio is going to dissolve in a pile of steaming goo! Funny that I listen to it almost daily.

The receiver in my office is a Sansui 2000 that I bought sometime in the late 1960s (do Chicagoans remember Toad Hall); other than cleaning the pots and replacing burned-out dial lamps, I've done noting. It's fed by a Sony CDP-102 CD player--other than lubricating the CD slide rails, I've done nothing to it. It works fine, driving those Dyna A-25 speakers.

One aspect of electrolytic capacitors is that the show went to hell when PCBs were banned. The capacitor business has yet to find an equivalent substitute.

To be sure, I've replaced caps in lots of stuff (LCD monitors are notorious for their bad designs), when it was apparent that they were going to fail. But replacing things "just because", in my opinion, isn't warranted.

Let just agree to disagree.
 
Last edited:
Back
Top