Mac Classic Capacitor Failure Cause

[superjer2000]

I just finished pulling the analog boards from my 5 Classics/Classic IIs so I could recap the analog boards. I had done the logic boards before and had varying degrees of SMD leakage. Same with my other Macs, different amounts of leakage from the logic board cans and zero leakage from my SE30 and prior compact mac analog boards. 

What surprised me with the Classics is that every single one of them had analog board leakage by the "tank farm" and wherever else the Nichicon low impedance (brown) caps were used. 

So my question is:  Was there an issue with the way Apple used these caps (maybe the same mistakes made as in the leaky cap LC PSUs) or was there an inherent problem with his batch of caps that seems to have a 100% failure rate?  I have been using Nichicons in a number of recap projects (shifting more to Panasonic lately) but I'd hate to think I should have been avoiding Nichicon from day 1.  Any theories as to what is causing these failures on such a consistent basis?  

Also, as others have noted the two rectifiers in the area DP5 and 7 seem to wick up this electrolyte causing the plastic rectifier shell to lose the labelling and make it appear as though the rectifiers themselves are leaking.  I found this odd But I'm sure the rectifier aren't leaking fluid. (I hope). Anyways, You can really tell Apple cheaped  out on these macs versus the Plus series and SE series. The 3+ different versions of the Classic analog boards made ordering capacitors a real treat too. 

 

[Franklinstein]

I think the big issue with these is age: the Classic/Classic II are from 1993ish, which makes them about 25 years old. If people kept their other consumer electronics from that same era, those devices would likely also require a new set of capacitors by now. Go pop the top (or bottom) from a portable CD player or even a PC laptop of the era and you'll notice the same problems in most of them.

The other problem is the miniaturization and modern manufacturing processes: old (>1980s) capacitors are usually much larger than their modern counterparts at the same values, or the old ones don't even have values that some modern capacitors provide. These newer caps are engineered to very tight initial tolerances and then usually stuffed into tiny, hot enclosures in close proximity with many other components. They're not designed with longevity in mind: once their service lifetime is up (it's usually something rated in hours that averages to around a 5-year lifespan with average use) then they're said to have fulfilled their design requirements; any additional service is a bonus. No manufacturer is going to engineer a long-life or fail-safe/leak-proof component because they will be expensive and nobody will buy them; most people throw our their consumer devices well before the average service life has been exhausted anyway.

If you want an example of bad capacitors, look at the industry-wide exploding counterfeit capacitor fiasco circa 2004ish. No major electronics manufacturer came away from that unscathed, except perhaps those who were very discerning and chose only the best capacitors, such as those from WIMA, Sanyo/Panasonic, or Nichicon. That being said, I'm not sure Apple dictates the use of, or even really cares about, discrete components used in their products; they likely leave those supply decisions to their OEMs like Foxconn.

Personally I tend to use as many solid or polymer caps as possible; liquid electrolytics are generally relegated to power supply use when very high voltage and/or capacitance are required or when other types are not available. I don't see many complaints regarding Nichicon or Sanyo/Panasonic electrolytics while others, like Rubycon, are hit-or-miss.

 

[superjer2000]

I agree that 25+ years is well outside of design expectations, but it doesn't change the fact that even on the Classic analog boards there are a number of caps that don't leak. It really is just a group of caps, of a particular type, in a particular area of the board that always leak. So that tells me there was either a design issue or a cap issue.  Again, time is the key element here but if I can pick a cap with slightly different properties to reduce the chances of this happening prematurely I'd do that. 

 

[Dog Cow]

What surprised me with the Classics is that every single one of them had analog board leakage by the "tank farm" and wherever else the Nichicon low impedance (brown) caps were used. 

Me too. I just replaced capacitors on a Revision A Mac Classic Analog Board. But after replacing the capacitors, now it has a short somewhere, probably caused by electrolytic fluid.

 

[Paralel]

It's almost certainly a combination of age, a given manufacturing process, and chance. I made the mistake of discounting caps as the issue with a type of PC laptop that had a power problem. Some of the exact same models work, no problem, but some, like the one I have, are completely dead. Someone tried replacing the caps on another one of the exact same model that wouldn't power up and it came back to life, showing that caps were indeed the issue.

Once something is outside of its specced age, the reason or mode of failure doesn't really matter, because it could be just about anything. It is something of interest to us, because we want to try keeping these machines running, but, in general, they did their "job" according to spec, and that's all they were meant to do.

 

[superjer2000]

Thanks all for the feedback. I guess I was just surprised as most vintage systems I've seen might have some leaky capacitors but not every system, every time.  While I've heard of leaky caps in SE and SE/30 power supplies I've never seen it personally and I've never seen leaking caps on an SE or earlier analog board.   I was just surprised by the 100% failure rate of these Nichicons. 

 

[techknight]

It also depends on the amount of "hours" that is on the system. The more heavily used and abused the capacitor is, the worst it leaks. 

There are probably chemical reasons why this happens, like electrolyte turning acidic when current is applied, or the bottom seal getting hard and loosing its ability to do its job, etc... etc... 

Whatever the cause, SMD caps had it the worst. there were many many many products of that time period. And even some that were still in active USE in that time period had failing SMD caps. the Sony Handycams from the early 90s were notorious for this, Sega GameGear, Even some Panasonic DAT machines were full of bad SMD caps. 

The saga continued even into the early 2000s. Gen1 Mitsubishi DLP TVs had 3 PCBs that were sandwiched together which had over 100+ SMD caps, all fail causing power up problems, color faults, etc.