Aluminum Electrolytic Capacitor Leakage After 10 Years Storage—Humidity or Heat?

I've been asked to look into some field failures of some older equipment.

This equipment has been sitting unused for about 10 years since the customer's factory shut down. They are now attempting to bring the factory back up and a number of our products are failing to power up. We got some of the units back and it appears that the Aluminum electrolytic capacitors in the power supply are damaged and have leaked.

The question is why? We know of other customers who have units of the same age in service with no problems.
We suspect something in the way these units were stored (some also have LCDs that appear damaged as well) but what is the most likely cause? I'm suspecting a combination of heat, humidity and/or corrosive atmosphere.

Pointers to any information are greatly appreciated.

ElizabethAlthough this does not appear to be the prime cause of your problem, electrolytic capacitors deteriorate over time when not in use. they should be "reformed" to bring them back to meet the ratings. Here is one link to the topic.There is a bit about physical leakage here

Elizabeth, I'm surprised at you, you're normally teaching the rest of us about stuff like this (Maybe that's what you are doing....or seeing if we're awake....)Anyone who got interested in electronics during the valve (tube) era knows about reforming the power supply caps when an equipment has been idle for a few years (but you're not that old, obviously .  I was going to post a few links but Aubrey has beaten me to it.  In his second link is the  paragraph:Electrolytics do not suffer idleness well. They can cause big trouble when idle for long periods, needing periodic charging to stay "formed" and maintain the oxide layer that insulates the conducting plates. Sometimes they can be "reformed" by a slowly rising return to working voltage (see below). Even with regular use, electrolytics fail with age by drying out or leaking electrolyte following internal corrosion. If the electrolytic bulges, shows obvious loss of electrolyte, or simply can't be reformed you must replace it.
I'd get your soldering iron fired up if I were you....   The electrolytics might have been idle, but you certainly won't be
In fact if you google the title of your post you'll get plenty of information, though some of it is on leakage currents rather than physical leakage.  Electrolytics that have a voltage applied to them constantly or at least regularly stay "formed" and will go a lot longer before showing problems.
If you want a laugh about some rather spectacular leakage of a capacitor I refer you to one of my old EETimes stories:http://www.eetimes.com/author.asp?section_id=36&doc_id=1284545

I agree with Aubrey Kagan's comment. The links Aubrey provided are good discussions. Aluminum electrolytic capacitors inherently have limited life. In the 1950s, some electrolytics were packaged with plug-in octal (only four pins actually present) bases for just this reason.On many aluminum electrolytic capacitors with both leads coming out one end for PCB mount, you will see dented-in lines on the other end. The dented-in lines are places where the case is intentionally thin, so the capacitor can expel its innards when it shorts.
One old trick when re-forming an electrolytic capacitor, is to use an incandescent lightbulb as the series resistance: Bulb rated voltage must be equal to or above the forming voltage, and full-voltage rated current must be low enough to limit the leaky-capacitor current to the desired value. The bulb will illuminate the re-forming process.Elizabeth, my recommendation is to replace all the electrolytic capacitors in your old equipment. Some have already failed, so none can be trusted.When restoring long-disused antique radios, I replace the capacitors before ever applying power, as a defective capacitor can cause damage to other difficult-to-replace parts.

Thanks for the links Aubrey,Is it possible that abruptly powering up an unreformed capacitor could cause it to leak? My guess is that may have been what happened...There's a lot of information about how to reform capacitors but I haven't seen anything on what happens if you don't. Other than a strong implication that you don't want to be around.

ElizabethHere is a link with a lot of information about the so called "Capacitor Plague of  1999 - 2007" where large numbers of electrolytic capacitors were failing in motherboards and video cards.https://en.wikipedia.org/wiki/Capacitor_plague

Elizabeth, I remember reading about this in the dim and distant past but could not find any specific references to it so will have to trust my memory.  In an unreformed capacitor the oxide insulating layer is thinner and the leakage currents will be a lot higher on initial power up, ie the cap behaves partly as a (parallel) resistance.  If the IR power from that is high enough the cap will heat up and that tends to exacerbate the effect if the dielectric layer does not reform quickly.  This can then result in the electrolyte boiling with consequent leakage or explosion depending on the cap constructionAs the advice from the paragraph from Aubrey's links says, if it's leaking, replace it.  I'm sure you can find a techie to do the donkey work....

Elizabeth... you also said... "(some also have LCDs that appear damaged as well)"LCDs going so dark as to be unreadable is pretty common, I have had DMMs that have done that.Is that the problem?  There's really not much you can do expect replace the LCD.There's a good write-up on the problem herehttps://telement.wordpress.com/2012/05/16/calculator-lcd-display-goes-dark/

@ElizabethAs probably one of the oldest of the old timers, the failure of electrolytic capacitors has been a constant battle when re-enrgising a power supply. In fact the Apple lle and BBC computers I repair, 30 plus years old and stored in lofts or attics, now have their electrolytic capacitors changed before powering up. If there are X2 suppression capacitors across the mains then it may well be advisable to replace these, as they often lose their magic grey smoke when re-powered for the first time, Listen for the sharp crack, they split, smell awful but usually do not catch fire. Are there any soldered in NiCads in the circuits, these have a terrible result in long time de-powered equipment, if these leak you can almost certainly say good bye to PCB tracks.Surprisingly I still have some ex war department 8uF oil/paper capacitors that have never failed in 50 years of very sporadic use. but they are huge.Not much help on the why but certainly had a lot of experience replacing.

Thanks for all the responses...It's really to bad that someone just powered up all this equipment without bothering to check it first...
At this point, replacing the units (or at least the power supplies) may be easier and faster than trying to repair.
There are a couple other related mysteries in this whole thing. Like why on some units the cap on the bootstrap circuit failed and the main bulk cap didn't. Not to mention how a power supply from 1995 got into a unit with a 2007 serial number.The joys of trying to troubleshoot problems remotely.

i would suggest you to switch off the  power to the pcb and disconnect the connections to the pcband short all the cap in the circuit using   a 60 watts bulb to discharge the charge of all the caps.after this wash both sides of the pcb using a tooth brush and isopropile alcohol after this wip it with a cloth and dry it in the sun light for 5 minutes. now by  using a analog multimeter  check  all the caps one by one replace the leacky caps  .now connect the  the pcb  wires as it was connected earlier  and switch on the power to the pcb and check the voltage across the caps for example if the voltage across a cap was 25 then the it is better to keep  the voltage rating of the cap double  to the  supply voltage in this case 50v if you go by this theory cap will not fail. the pcb was washedwith iso-propile alcohol to clear the electrolite solution  leacked on the pcb pl try this it may help you.  Sambath

Sambath mentioned the electrolyte solution. Capacitors used to use a weak boric acid solution as an electrolyte.  It was not strong enough to cause burns, and a little baking soda in water would neutralize it. So now I'm wondering- do aluminum electrolytics still use boric acid as an electrolyte?Also, a comment on Sambath's post- I agree that it is wise to check for leakage, but in many cases the surrounding circuitry will make it appear that the capacitor is leaking when in actuality it is not.  If a capacitor appears to be leaky it may be necessary to lift one of the connections so you can get a good "out-of-circuit" reading.
-Rick 

Hi  Mr Rick,It is not necessary for the solution      leaking from the cap to make burns on the track  if the solder pads are very close this solution can conduct through the solder pads and can create a short circuit i have come across few cases like this in my 40 years of experience and out of theseSome of the cap externally looked intact cleaning the Pcb with isopropile solution  cleans the Pcb very clearly and drys fast than baking soda and water .

The electrolytes used in aluminum electrolytic capacitors have changed a bit over the years and become more proprietary. Cleaning the board is a good idea if we intend to repair (rather than replace) it. I'd be tempted to replace ALL the electrolytics on the grounds that they have been potentially stressed and are more likely to fail in the future but that is not my call. My job was to analyze the circuit and give some reason for the failure. In particular, why did the bias cap leak electrolyte even when the main bulk cap didn't.  I wrote something up and sent it out late last week. I haven't heard anything back so my answer may have satisfied everyone.

Elizabeth
In a reply to Max's blog  that was repeated on Planet Analog there was a comment by industry expert John Dunn. It was:Restoration of Electrolytic Capacitors
Sometimes, one can resurrect an age-deteriorated electrolytic capacitor, but the key is to do that safely. Please see:

http://licn.typepad.com/my_weblog/2010/12/alu...tion-john-dunn-consultant-ambertec-pe-pc.html