Storing Vintage Macs & Capacitors

[trag]

In my (seemingly never ending) project to clean all my junk out of the living room, I was about to put two SE/30s in the attic. I decided I should pull the batteries, if any, first. When I opened up the first one, I found a battery with an August 1989 date code. Thankfully it had not leaked.

However, upon examining the board, I found little beads of liquid on the vias (holes through the board) near the surface mount electrolytic capacitors. Apparently the stuff was leaking out and flowing along the capacitor traces until stopped by a via where it accumulated. This is the first time I've seen actualy beads of the stuff. Usually it's just a discoloration on the board surrounding the caps.

So I interrupted my neatening mission long enough to get the soldering pencils and pull all the caps and clean the boards. I'll be storing the boards with no capacitors installed.

I suggest that if you're storing a vintage Mac, even if you don't have time to replace the caps, it's a good idea to remove the old surface mount electrolytics and clean the board before storing it. Removing the battery should go without saying. Over time the stuff from the caps can corrode a lot of valuable connections on the board. If you don't take this precaution, there may not be any point in storing the machine at all. Of course, my perception is colored by my first IIci, which all the way back in the mid '90s had leaking capacitors and damage so severe it couldn't power on until I bypassed the corroded traces with wires. Of course, some machines will survive long storage even after the caps leak. It's just so disappointing to see a circuit board with corroded pins and solder and traces caused by leaking caps.

The second SE/30 had no battery and the caps didn't seem to be leaking, but they looked like they were from the same batch as the first board, so I pulled them and cleaned it too.

I didn't lift any pads, but it was tempting. It took a while to get the caps loose with a pair of 15 watt pencils. A little higher wattage might be better, like 25 or so.

Still, the key to getting the caps off without lifting the pads (asuming the pads aren't rotted by capacitor goo) is lots and lots of patience and plenty of solder flux.

 

[Unknown_K]

You must have found that unit as is started leaking before thr goo dried.

All my machine with aluminum capacitors have been redone (I think anyway) so hopefully none will leak anytime soon.

 

[insaneboy]

Odd, of all the older machines I've owned, still have yet to have a cap fail(plenty of wood knocking going on right now). Maybe because I've powered up and used all my machines at least 1-2 times a year?

only machines I've had bad caps on were two iMac G5s at work.

 

[Unknown_K]

Depends on the machine and when it was made, some used tantalum capacitors and those blow up instead of leaking. I use my machines at least a few times a year, but who knows how long they were in bad storage before I got them.

 

[JDW]

"insaneboy" is truly a proper name indeed!

Your caps have partly "failed" and you just don't know it. More specifically, I assure you that the capacitance level of most of your fluid filled caps is considerably diminished as compared with what they once were when new. And just because your machine powers on as normal does not mean the caps are still at full capacity. It merely means they have enough capacity left so you don't know they are about to fail completely. With it now being 2011 and considering the age of some of these old Macs, I consider it almost dangerous to continue using a machine as normal with the original caps simply because it "appears" to be working fine. Chances are the caps are leaking and you just don't know it, and you will refuse to replace the caps simply because you perceive the machine to be working fine. That's the "danger" I speak of. But you don't have to let "insanity" overtake you. You really can remove the old caps, clean the board, and replace them with tantalums (in most cases).

Unknown_K, who here on this site used Tantalums that blew up? And when did this occur? And what voltage capacity and size of the tants were used?

Tants will blow with over-voltage, so I must assume that whomever used them did not use a properly rated cap?

 

[Unknown_K]

I have cards with factory Tantalum capacitors that have exploded.

Example:

http://picasaweb.google.com/teozenios/NubusCards#5067908435668848786 (lower right side)

http://en.wikipedia.org/wiki/Tantalum_capacitor

"Self-destruction and thermal runawayTantalum capacitors are under some conditions prone to self-destruction by thermal runaway. The capacitor typically consists of a sintered tantalum sponge acting as the anode, a manganese dioxide cathode, and a dielectric layer of tantalum pentoxide created on the tantalum sponge surface by anodizing. The tantalum oxide layer may have weak spots that undergo dielectric breakdown during a voltage spike. The tantalum then comes to direct contact with manganese dioxide and the leakage current causes a local heating; a chemical reaction then produces manganese(III) oxide and regenerates (self-heals) the tantalum oxide layer.

If the energy dissipated at the failure point is high enough, a self-sustaining exothermic reaction may occur, similar to the thermite reaction, with tantalum as fuel and manganese dioxide as oxidizer, destroying the capacitor and occasionally producing smoke and possibly flame.[3]"

And you don't have to have much over voltage to have issues.

 

[Concorde1993]

Unknown_K, who here on this site used Tantalums that blew up?

JDW, I believe you are referring to LC_575, who had a IIsi with tantalum caps on a NuBus card, which suddenly exploded upon startup.

Here's the link to the thread below:

viewtopic.php?f=8&t=14436&hilit=IIsi..+a+fire+hazard

 

[JDW]

Thanks for the link, Concorde1993!

The photo of the blown "dipped" tantalum capacitor on that riser card is shown here:

http://img201.imageshack.us/img201/917/img1005z.jpg

It was interesting to read TechKnight's opinion here:

viewtopic.php?p=138534#p138534

Interestingly, some capacitor manufacturers disagree that tantalums "age":

http://www.kemet.com/kemet/web/homepage/kfbk3.nsf/vaFeedbackFAQ/6E593B9AE78FE87985256E5B007BAE27

But if we accept that solid tants do age, aging would vary by the type of solid tant (e.g., the thru-hole "dipped" kind vs. the SMD type that trag sells here). (We should exclude wet-slug type tantalum caps here, as they are hardly better than fluid filled electrolytics in terms of longevity.)

And as you may expect, capacitor reliability over time varies by manufacturer:

http://www.electronicspecifier.com/Resistors-and-Capacitors/Low-ESR-high-rel-tantalum-capacitors-minimize-failure-rates-in-long-life-commercial-power-supplies.asp

http://www.avx.com/docs/catalogs/trj.pdf

There's an interesting article on Capacitor Reliability here:

http://my.execpc.com/~endlr/reliability.html

When you open that web page, search for "Solid tantalum capacitors" to skip down to the right part. And if you scroll down even further, you can find a tantalum capacitor life expectancy formula.

 

[trag]

And as you may expect, capacitor reliability over time varies by manufacturer:http://www.electronicspecifier.com/Resistors-and-Capacitors/Low-ESR-high-rel-tantalum-capacitors-minimize-failure-rates-in-long-life-commercial-power-supplies.asp

http://www.avx.com/docs/catalogs/trj.pdf

Interesting. If I ever use up the current stock of caps I have on hand, it sounds like I should look into picking up the AVX TRJ caps as replacements, if I decide to restock. I think I'm going on something like three years still with the original reels I purchased. The 47uF caps came on five qty. 500 reels, and I think I'm down to the second to last reel of those. So only 1000 of those left.

OUCH! I just checked the price on those AVX TRJ caps over on Digikey. There are two models of 47uF cap, both 25V (no 16V available), and the price, depending on the ESR, is $1.63 each or $3.18 each. And that's in quantities of 500 or more. Ouch.

I could actually kind of see spending that on machines I planned to keep forever. I think. But I doubt that very many folks would recap their boards at those prices.

There's an interesting article on Capacitor Reliability here:http://my.execpc.com/~endlr/reliability.html

When you open that web page, search for "Solid tantalum capacitors" to skip down to the right part. And if you scroll down even further, you can find a tantalum capacitor life expectancy formula.

That is an interesting article. The main thing I took away from it was that one should use a 3x - 4x higher voltage rating on tantalum caps, than the expected voltage in the circuit. Unless the 47uF/16V caps are on the 12V circuit, I think Apple was already doing that with its aluminum electrolytics. 3 X 5V < 16V.

Still, while the tantalums won't leak onto our boards, it sounds like they don't last forever either. Sigh.

 

[JDW]

The main thing I took away from it was that one should use a 3x - 4x higher voltage rating on tantalum caps, than the expected voltage in the circuit. Unless the 47uF/16V caps are on the 12V circuit, I think Apple was already doing that with its aluminum electrolytics. 3 X 5V < 16V.

That's my feeling too, Trag. So long as we can spec the tantalum cap to have a higher voltage rating than any surges it might get, we should not see the life of that cap diminished at all in the short term. But then again, when we are talking about surges, we need to consider the "duration" of the surges too. For truly, if a 40v surge will kill a 16v rated tantalum cap, how long in duration would that surge have to be to kill the cap. And then the question becomes, will we ever see surges like that on the main logic board (e.g., an SE/30 board), or do such surges occur on "IIsi riser cards" alone because of some problem in how the card intermittently was connected to the main logic board?

So it's not such a cut-and-dried answer here.

 

[Peter.Howard]

Someone or a small business could probably do a bit of business selling these boards refurbished with new parts.

To people who can't solder to save themselves, or their macs!!

Maybe not now but in a few years when the original compact macs 128k, 512k become more rare in working condition.

Would not seem unreasonable to pay $100-$150 for a refurbished analog board.

 

[Unknown_K]

How many boards will be so far gone in 10 years they cannot be fixed? I don't think many people would pay the price plus shipping both ways for a board to be fixed.

 

[trag]

How many boards will be so far gone in 10 years they cannot be fixed? I don't think many people would pay the price plus shipping both ways for a board to be fixed.

It's been a long time since I studied chemistry, but doesn't the reaction which causes corrosion, eventually neutralize the corrosive agent? Or does it just catalyze oxidation? Or does it depend?

I ask, because if the former, then the amount of corrosion that a leaky set of capacitors can cause is limited, even if allowed to go to completion. In some cases this may damage a board beyond repair, but in other cases the board will be repairable, if someone is willing and able to trace the corroded components and bypass destroyed connections and/or replace parts which have completely lost pins. Of course, for the custom Apple chips, that latter option is not available except through cannibalism.

Someone or a small business could probably do a bit of business selling these boards refurbished with new parts.To people who can't solder to save themselves, or their macs!!

Maybe not now but in a few years when the original compact macs 128k, 512k become more rare in working condition.

We're actually discussing the repair/refurbishment of logic boards here. The specific issue is that the surface mount electrolytic capacitors leak corrosive which can damage the logic board and surrounding components.

As far as I know, the analog boards do not have any components which will cause similar damage if left unattended for years. Well, the battery, of course, but that's not soldered down.

Would not seem unreasonable to pay $100-$150 for a refurbished analog board.

Very few folks would pay that these days. There's no telling what folks in the future will be willing to pay. There's actually reason to believe that the prices on these old things will fall. We may be seeing the maximum right now. It has to do with the collectors tending to be folks who had or wanted these machines when they were younger and going for a nostalgia fix by way of collecting. As those folks satisfy their craving, or age further and loose interest, there will not necessarily be a new crop of collectors coming along to boost the value. Of course the supply is constantly decreasing too....

Doing a good job of refurbishing these boards is at least a couple of hours of custom electronics repair, plus parts, equipment, shipping, consumables, etc. It's difficult to make it an affordable service. Phreakout manages it by (in my opinion) undercharging for his time.

 

[Byrd]

I can't help worry that capacitor removal will result in greater risk of damaging the motherboard pads and traces by keeping them there, until they fail (unless the caps themselves are visibly leaking, and should be replaced there and then once identified).

All these wonderful computers in storage though - are you keeping your older Macs as backups to working ones, or with the hope that in twenty years time they will indeed be your retirement next egg?

I recently purchased some Compact Macs from a man in the bush, who stored them in shipping/packing crate for several years (in the sun, the crate had a cracked window). I think there were eight Macs all up, and every one had cap failures, corrosion and HD/floppy drive failure. Superficially, they all looked fine - it was only with opening them up did the problems become apparent. I was only able to make up two "good" ones out of the lot, and still had to put many of my spare parts in. Therefore I'd consider choosing consistently cool and dry conditions to be significantly more important, as a whole. Without these conditions other components will fail just as quickly, and are much harder to replace and/or source.

JB

 

[Unknown_K]

I keep all my stuff in a cool and dry basement to help minimize moisture/rust issues. If you fire up a machine a few times a year you can tell when it needs work done (before tracks are destroyed).

As far as prices going up, it depends on the hobby being a long term thing or just a beany baby bubble. Some people collect items before their time while others want to have what they used when they were a kid. I collect old games I used to play, but most of the hardware outside of x86 in my collection I never used new (including macs) so they are all new to me.

The only way people will professionaly rework boards (like they do for older laptops) is if there is a steady market for that work. I reworked my stuff because its not too hard, I like to tinker, and I don't like paying ebay prices for working stuff. I did rework somebodies Classic 2 here (those boards are a pain because of spacing) and only because I wanted to help out (no real profit in it). Maybe is a 840av is worth $300+ somebody might pay $60 to rework a dead one.

Time will tell.

 

[trag]

All these wonderful computers in storage though - are you keeping your older Macs as backups to working ones, or with the hope that in twenty years time they will indeed be your retirement next egg?

For me personally, I'm just keeping the machines that I have either A) Because they are dear to me for one reason or another, or B) I have a project in mind which requires this model of machine some day, or C) It's one of the machines I thought was coolest, which may just be a variation on A, or D) Who knows why I bought it, but I haven't had a chance to get rid of it yet.

Oh, and E) I don't want it any more, but it needs some repair I haven't gotten around to yet.

I only recently finished cataloging all my Mac stuff and storing most of it in the attic (wish I had a basement). Now that I have an accurate list of what I have, one of these days I'll go through it and decide what I really want to keep, and the rest of it can go.