Lytic capacitors always had and always will have a shelf life. Weather or not they see use does not matter for once their rubber seals give out they're gonna leak regardless. The only real way to keep on top of them is to either remove them prior to prolonged storage or replace them as per a schedule but do capacitors even have a listed MTBF? Even Tantalum becomes unstable with time but at least those only fail once power is applied

Storing machines upside down will not help as the electrolyte will wick its way to the board regardless.

As a rule of thumb I'd fully recap anything with 15+ years and 10+ of daily/constant service life.

Snayperskaya wrote:

As a rule of thumb I'd fully recap anything with 15+ years and 10+ of daily/constant service life.

This Tandy 1000TX is actually in mint condition; I doubt it's had as much as a week of daily use. Not so much as a smudge on the floppy drive eject or a spec of dust in the power supply. Case is perfect.

I'd like this to still be running in another 30 years, so I am considering recapping just to be certain. There's only a dozen or so visible capacitors, anyway.

Bad caps leak/rupture ... good ones don't.
When in doubt check the manufacturer and production dates - and then go on a wonderful google journey. On my sidejob I service old computers (dating back to 1976) ... except for the Amiga 600/1200 line no problems with leaking caps (leaking batteries are a problem tho).

Alternatively you can of course recap everything - but be aware this may not be as easy as it sounds (hello multilayer boards) and I've seen more boards die from "recapping" (tearing out vias/tracks while unsoldering) than being saved, when not required.

shock__ wrote:

Bad caps leak/rupture ... good ones don't.
When in doubt check the manufacturer and production dates - and then go on a wonderful google journey. On my sidejob I service old computers (dating back to 1976) ... except for the Amiga 600/1200 line no problems with leaking caps (leaking batteries are a problem tho).

Alternatively you can of course recap everything - but be aware this may not be as easy as it sounds (hello multilayer boards) and I've seen more boards die from "recapping" (tearing out vias/tracks while unsoldering) than being saved, when not required.

Thankfully this machine had no battery. I've worked with enough classic pinball machines to know how horrific a bad battery can damage a board...

I'm going to take a good look; I'm not even positive these are surface-mounted or through-the-board or what.

It is such a gorgeous, mint-condition machine that I feel almost compelled to make sure it lasts ANOTHER 30 years...

http://i.imgur.com/j1LlM6z.jpg <- motherboard for those interested

On one occasion that I recall, a board that had been sitting for a long time blew some electrolytic caps after I powered it up. The pop noise was audible, maybe about 30sec-1min after I turned the machine on. There were something like 2 or 3 caps visibly bulged afterward.
I've not personally ever seen *quality* electrolytic caps leak from just sitting. After powering up though, I did have some pop as mentioned above.

Defective capacitors are different though. Some capacitors have faulty chemistry and they will leak after a number of years, even if they've never been powered. Many people have found sealed NOS motherboards and video cards with leaking caps, but these were known defective caps from the so called "capacitor plague" era of the early 2000s. Their faulty chemistry causes them to build internal pressure and leak even without usage. I don't believe that quality caps have much likelihood of doing this in storage, but I don't want to say it's impossible.

The oldest devices I've looked in are some old game consoles from the 80s and an audio amplifier from the 70s. I did find bad caps (out of spec) in the game consoles, but they weren't leaking. In the audio amplifier, none of the caps were leaking but some of the wrappers were starting to pull away from their normal position, which implies the case might have been starting to deform. It wasn't visible other than the wrapper having crawled down the side of the cap.

Tantalums can be temperamental. They are stable in storage, but if they sit for several years without power, then when they see power again they are known for their potential to spark and explode.
I had that happen on a Pentium Pro board. Ever since, I've been afraid of it happening to a P2 Xeon board I have that has far more tantalums on it than any motherboard I've ever seen. If it sits for too long, the odds of just one tantalum deciding to explode seems pretty high. Therefore, I try to remember to turn it on every few months, even though I currently am not using it for anything anymore.
I think some of the axial capacitors in your photo might be tantalum, so I'd be worried about those sitting for years without power.
They are cheap parts to replace, but to do that you have to be able to figure out what the original specs were, and I'm not sure if they can fail short (which might cause further damage).

keenerb wrote:

Is it silly to consider storing older hardware upside-down so that any ruptured cap would not leak corrosive materials onto the motherboard?

No.

Whatever works, works! Right? 😁

I've had several components sitting in my storage for any number of years see develop leaking caps (I'm almost 100% certain the caps were looking fine when I stored them, I always check and label everything out of the ordinary), but usually the caps were either significantly more bulged and many had their tops cracked open and 1 or 2 had some brown crusty stuff coming from the top iirc.

Still, I've had zero spills of liquids on the insides of the bags I used to store them in and nothing on any of the PCBs, so the damage was always very local.

What struck me most is that almost all of the leaking caps I found this way were from graphics cards, even more so than from motherboards, which I thought was a bit odd, but then again it could just be random.

Tetrium wrote:

What struck me most is that almost all of the leaking caps I found this way were from graphics cards, even more so than from motherboards, which I thought was a bit odd, but then again it could just be random.

eVGA, by any chance?
They used a ton of Sacon FZ caps on their cards, which are some of the worst about this.

When getting a replacement cap, how do I know what is good quality?

Avoid these:
http://www.badcaps.net/forum/showpost.php?s=f … 685&postcount=3

Bad caps leak/rupture ... good ones don't.

My laserdisc player begs to differ.

Again, all lytic caps have a finite life.

One thing I'm not sure about when trying to get a replacement for my caps that are starting to burst, what is the difference between all the sizes?

I'm searching for 560uF 6.3V and I keep finding so many shapes and sizes. Some super tall, some really small.

http://www.ebay.com/itm/560uF-6-3V-8x23-Rubyc … R0AAOSwEeFU4YDF
http://www.ebay.com/itm/10pcs-560uf-6-3v-Radi … 1kAAOSwJQdXAIwF

Here is my bad cap

ESR, temperature rating, and ripple current will both factor into package size.

CelGen wrote:

My laserdisc player begs to differ. […]
Show full quote

Bad caps leak/rupture ... good ones don't.

My laserdisc player begs to differ.

Again, all lytic caps have a finite life.

That is most probably Chemi-Con LXF series, which was a defective series, so that's an exception. Capacitors that used quaternary ammonium salts in their electrolyte are prone to that kind of failure. Nichicon PL and PF are also an example of this. They may do well if kept away from heat, but that particular one is sitting right next to a heatsinked TO-220 device.

Thraka wrote:

One thing I'm not sure about when trying to get a replacement for my caps that are starting to burst, what is the difference bet […]
Show full quote

One thing I'm not sure about when trying to get a replacement for my caps that are starting to burst, what is the difference between all the sizes?

I'm searching for 560uF 6.3V and I keep finding so many shapes and sizes. Some super tall, some really small.

http://www.ebay.com/itm/560uF-6-3V-8x23-Rubyc … R0AAOSwEeFU4YDF
http://www.ebay.com/itm/10pcs-560uf-6-3v-Radi … 1kAAOSwJQdXAIwF

Here is my bad cap

That is a Nichicon HN series 8x20 capacitor. It has bulged because Nichicon HM and HN were defective in certain years (all parts with datecodes before 2006 - though there are some exceptions).

It should be replaced with an identical Nichicon HN/HZ 8x20 capacitor. The value isn't important. You can use 6.3V/10V 560uF and above, or even a 16V 1000uF 8x20 part.

This will work:
http://ca.mouser.com/ProductDetail/Nichicon/UHN1A102MPD9/

Don't buy capacitors off eBay. They're counterfeit.

mockingbird wrote:

It should be replaced with an identical Nichicon HN/HZ 8x20 capacitor. The value isn't important. You can use 6.3V/10V 560uF a […]
Show full quote

It should be replaced with an identical Nichicon HN/HZ 8x20 capacitor. The value isn't important. You can use 6.3V/10V 560uF and above, or even a 16V 1000uF 8x20 part.

This will work:
http://ca.mouser.com/ProductDetail/Nichicon/UHN1A102MPD9/

Don't buy capacitors off eBay. They're counterfeit.

Hey thanks a bunch! You saved me a lot of searching and future problems.

😎

Got some more bad caps to replace.

Can I replace KZG6.3VB222M10X20LL (united chemi-con 2200uf, 6.3v) with EEU-FR0J222 (panisonic 2200uf, 6.3v) ?

Trying to figure out the best replacement, using the mouser website. I just want to double check 😀 The ripple rating seems slightly different on them

CelGen wrote:

but do capacitors even have a listed MTBF?.

Yes any sub dissent company will rate the mtbf on their caps.
Cheaper and general use caps are around 1k-10k Hours. Better caps are almost always over 10k hours.

shamino wrote:

Tantalums can be temperamental. They are stable in storage, but if they sit for several years without power, then when they see […]
Show full quote

Tantalums can be temperamental. They are stable in storage, but if they sit for several years without power, then when they see power again they are known for their potential to spark and explode.
I had that happen on a Pentium Pro board. Ever since, I've been afraid of it happening to a P2 Xeon board I have that has far more tantalums on it than any motherboard I've ever seen. If it sits for too long, the odds of just one tantalum deciding to explode seems pretty high. Therefore, I try to remember to turn it on every few months, even though I currently am not using it for anything anymore.
I think some of the axial capacitors in your photo might be tantalum, so I'd be worried about those sitting for years without power.
They are cheap parts to replace, but to do that you have to be able to figure out what the original specs were, and I'm not sure if they can fail short (which might cause further damage).

My Western Digital MFM HDD from 1988 won't agree with you. When I got it, it had 2 tantalum capacitors that were shorting, and this prevented the whole computer to even start up. Then when I disconnected them, another one exploded. In order to know which one was shorting prior the explosion, I checked the resistance of each one. Almost all of them were either Ohm or infinity Ohms so I think that they were all dead

Deksor wrote:

Almost all of them were either Ohm or infinity Ohms so I think that they were all dead

Sure, if it says near zero ohms, it's shorted dead.
But you can't judge a capacitor is dead just because ohmmeter shows infinity.

Even if the cap is just fine, it will show infinity ohms after the cap has fully charged to the ohmmeter testing voltage so current does not flow any more.