Old computer/capacitor failure a regional phenomenon?

[A-ko]

I watch a lot of Youtube videos of people repairing old computers, the horrors of machines stored in less-than-ideal locations, humidity, moisture, etc killing machines. From this I'd be led to believe that these machines are incredibly fragile.

However, as somebody that's been collecting pre-1996 computers on and off since the early 2000's, had hundreds of systems over the years, been guilty of storing many, many machines in non-climate controlled storage units, dug many machines OUT of even worse situations, etc, I've only had maybe a 3% rate of non-functional machines, ever. Honestly I'm shocked if something DOESN'T just plain work when I get it home...with an exception being some of the early 90's Macs with the bad caps. And even then, they typically boot, but need help.

It has me wondering, is it luck, or a regional thing? I live in central Iowa, which has pretty drastic temperature changes and VERY humid summers. Youtube leads me to believe lots of these machines are just not gonna work when you get em home, but I've had the absolute opposite experience in my ~20 years collecting.

Any thoughts?

[Timo W.]

That's just biased awareness. If people have computers with failed caps, they tell the world about it. But no one does the same when the caps are still intact. Just like no one posts: "hey, woke up this morning and my TV set still works". But guess what they do when it fails.

[Dwight Elvey]

Electrolytics that have been sitting for a long time should be brought up to voltage slowly. It is because of the way they work. You power them up fast and they have to vent. Normally the seals are designed to vent very slowly but if you just hit them with full voltage, they have to vent faster and bust the seals. Even while running, they vent small amounts of gas. This is the way they work.
Tantalums are subject to sudden failure. These are not suppose to vent but I suspect these fail because of moisture getting into them. I've always thought one should use a deadman switch when bringing up a machine with tantalums. I had one N* with a 50 uf tantalum, on a RAM board fail. When it failed, it burst into flames and did damage to the board before I could switch it off and blow out the flames. Smaller caps tend to just blow up with less damage. I suspect a thermal imaging video would be the best way to find them before they can cause damage.
They even fail in use. I had a Sun machine with an external drive. It ran 24/7. After 4 or 5 years the cap blew. Suddenly smoke was blowing out of it, from the small fan. Luckily I was there to shut it off. I brought it to IT and they were just going to replace the drive. I had quite a bit of stuff I didn't want to lose. I asked them if they had any other broken drives laying around. I removed a tantalum from one of those and the drive ran for another few years before the work station was retired. I suspect for tantalums, it is just a matter of time before they fail, powered up or not. How it is stored has little effect.
Dwight

[A-ko]

That's just biased awareness. If people have computers with failed caps, they tell the world about it. But no one does the same when the caps are still intact. Just like no one posts: "hey, woke up this morning and my TV set still works". But guess what they do when it fails.

good point. Repairing a fully functional machine would make for a pretty short video.

[DeltaDon]

A novice buys a non-working computer and is told all he needs to do is replace all the capacitors and the computer will work. Then they use a 1000 watt soldering gun and acid core solder and then comes on this site asking what went wrong. Troubleshooting seems to be change the caps and punt.

Tantalum caps from the '80's seem to let the magic smoke out quite often. Easy to tell if they're defective. Chuck might be right in that moisture might get into them. I don't know, but I've had my share of them go up in smoke. I've read a few papers suggesting that a tantalum cap should be used with a much higher voltage rating than was the standard safety factor used back then. Say a 25v or 35 v cap on a 12 volt line vs. a more commonly used 16v cap. It is even more common for them to smoke when installed with the leads reversed. This "trick" saved me quite a bit of money some decades ago when a maker of S-100 boards managed to install a reversed cap on the -5v line of a 64K memory board using 4116's and I snagged a stack of them for little cost. Repair cost was a less than a dozen new cap's.

[Mad-Mike]

You know what, I hear about capacitor failures all day on vintage computers but so far I can only count on ONE hand the number of capacitor failures I've seen actually fail is almost nothing - this is having lived in Alabama, Washington, and Nevada. So I don't think it's regional. All of the failed caps were tantulums. The only time I've seen buldging or leaking electrolytics is on modern motherboards - mostly Dells circa 2005-2007 during that time. This is over a span of 20 years of working with "vintage" PCs (8088-Pentium).

The tantulums that blew were all power-filters on 286 systems - a Compaq Deskpro 286, and the other was my GEM 286, which has blown two, and it STILL Goes. Actually, that thing has one of those barrel batteries everyone insists on replacing - still does not leak either, it's not a Varta though.

I kind of cringe whenever I hear modern "retro PC enthusiasts" talk about "I got a new 486, what's the first thing I need to do?" "REPLACE THE CAPS AND THAT VARTA BATTERY!" It's like my experiences in the world of Guitar whenever anyone buys a Fender Jag-Stang they are told to put a bloody Seymour Duncan JB in it and lock down the vibrato - it's like "let the guy do what he wants with his thing".

Just because a humbucker guitar pickup is 7.16K ohms does not make it a bad pickup, nor does the fact a computer is nearly 20-30 years old mean every single electrolytic and battery is leaking. I'd say inspect and test first. My approach probably matches Adrian's Digital Basement the most, although I'm not even half the tech he is (at least, not yet).

[TheDrip]

I think there is also some misinformation or misinterpretation regarding the types of caps.

Should RIFA caps be replaced at 30 years old (or other power filter caps)? probably.

Should every single capacitor of any type be replaced on every board that is over 20 years old? I believe you're more like to do more damage than good by just blindly replacing that many components.

That said, replacing all the capacitors is a trouble shooting step akin to rebooting a computer. Its easy to advise, and it seems to be pretty effective. If nothing else, it generally means every board is removed, cleaned, card edge connectors disconnected and put back. All of that handling seems to fix as many bad connections as replacing the caps fixes. Taking everything apart to facilitate the cap replacement also makes for a much better inspection than shining a flash light on a dusty board buried in the back of a case.

[Chuck(G)]

I suspect that more electrolytic capacitors fail as a natural consequence of operation than by storage conditions. The usual rated life of a wet aluminum electrolytic cap in continuous operation is about 10 years. That assumes proper circuit design and cooling.

The other determining factor is the use of a capacitor in a circuit. Those exposed to high ripple currents will experience more self-heating due to ESR effects. Simple decoupling or timing applications less so.

I also suspect that 10 years is what systems designers use as a guideline. To be sure, there are better alternatives to the wet aluminum electrolytic, but the wet ones are far cheaper.

[A-ko]

I'm no electronics repair expert but the thing that makes me cringe about the "recap everything" mentality is I'm sure a lot of those people are taking something with perfectly good, very high quality Japanese electrolytics that will probably remain fine for years to come and replace them all with probably close to the cheapest replacements they could find to keep costs down, without testing a single one. Aaand it probably ends up being a failed tantalum or chip, or whatever, that actually took down the system. I don't even touch a circuit board if the diagnosing is outside of my knowledge base. I learned that lesson when I was younger and thought I could be a "motorcycle customizer" and just ended up cutting up complete bikes and turning them into worthless piles of scrap because I didn't know the next step.

I HAVE had capactitor failure for sure but the only ones I've ever had on a pre-90s system were the RIFA filter caps on older TRS-80 models...and they still "work" after that, so I don't really count them. Same with leaky caps on 90's Macs. It's kinda to be expected.

Suprisingly for the amount of XT class systems I've had, I have yet to have a blown tantalum (knock on wood) that I know of. I just feel like if you weren't into the hobby, you'd be under the impression that nothing old and electronic works anymore and they all need lots of repair, when that seems to be the complete opposite of the case.

[DeltaDon]

Last night I was working on an S-100 z80 CPU card that had failed to boot when I purchased it some time ago. I had tested it and put it aside for a future "round-2-it" day. Well, yesterday was the day. As I applied power I heard a pop and quickly powered the computer down. No bad fried tantalum smell and when I pulled the board I found a 4.7micro farad@25V Tantalum on a -12V supply with a side blown off. No smoke or sign of flames. It was replaced with a 4.7@35v which was the closest I have in stock. It has to have been the 5th or 6th that I've replaced in the past year or so on boards from the mid 80's. Most were on power up on boards that saw some use recently - not first time in decades.

Regarding electrolytic cap's, I often replaced most if not all of them a few years ago when I was repairing a dead LCD monitor. All were one-hung-lo brand cap's, replaced with, decent, but not the best cap's. Quite often more than one had been bulged and needed to be replaced anyways. If I offered a warranty on my work I figured it was cheaper to so up front rather than to have to eat a second repair cost/time later on. Meanwhile, old American made test equipment from the '60's & 70's is often still going strong with the original electrolytic caps. I worked for a electronics dis't back before going off to the Vietnam War and we sold Cornell Dubilier cap's, among other brands, and most I used seem to still be holding up in some of my early projects.