8500 caps

[lobust]

I just bought an 8500 sold as working, and it does in fact boot up and work perfectly. I have very little experience with macs from this generation. I never owned anything of the 604 generation at the time, and this is the first one I've bought this time around.

What surprised me is that when I pulled it apart there is ZERO evidence of any cap leakage on either the logic board or the processor card. Apart from some rust on the audio connector block shield, both boards are pretty much immaculate.

Is this normal or did I get lucky? It is polar opposite to my experience with macs up to just a year older than the 8500...

I have a G3 minitower too that is perfectly fine with its original caps, and my next oldest after that is a G4 DA also original caps.

On the other hand, every mac I've got up to and including first gen PPC has needed a recap and some of them have been a real mess...

Are these machines from an era of higher quality caps? I know people do still routinely recap macs of this vintage, but is it always necessary for function or is it purely preventative?

 

[Phipli]

I just bought an 8500 sold as working, and it does in fact boot up and work perfectly. I have very little experience with macs from this generation. I never owned anything of the 604 generation at the time, and this is the first one I've bought this time around.

What surprised me is that when I pulled it apart there is ZERO evidence of any cap leakage on either the logic board or the processor card. Apart from some rust on the audio connector block shield, both boards are pretty much immaculate.

Is this normal or did I get lucky? It is polar opposite to my experience with macs up to just a year older than the 8500...

I have a G3 minitower too that is perfectly fine with its original caps, and my next oldest after that is a G4 DA also original caps.

On the other hand, every mac I've got up to and including first gen PPC has needed a recap and some of them have been a real mess...

Are these machines from an era of higher quality caps? I know people do still routinely recap macs of this vintage, but is it always necessary for function or is it purely preventative?

Most PCI macs of that era aren't showing issues yet. Not sure if the caps are better or just their age (but that wouldn't make sense given when 68k SMD caps started going). Personally I have only recapped one machine from that era (a Beige G3 that I think had been stored badly and the caps looked a little iffy). You can if you like, but I'm not worried yet.

Others will have different opinions. I'd almost say never bring up the topic

My thoughts generally, wrt logic boards, are...
SE and older (with axial caps) - cap on condition
Early SMD electrolytics (tin cans) up to 1993 - cap NOW.
1994 onwards - cap on condition.

Never tell me "My SE/30 is fine, it doesn't need recapping" because you're just telling me that you don't know what to look for or have bad eyesight .

 

[obsolete]

My experience matches Phipli's, although I think I had a leaky 6100 years ago, and that's a 1994 machine. My 5400, 7600, G3s and G4s are all fine.

 

[Phipli]

My experience matches Phipli's, although I think I had a leaky 6100 years ago, and that's a 1994 machine. My 5400, 7600, G3s and G4s are all fine.

Yeah, I've seen Nubus macs not look great in photos. That's why I say on condition - I inspect every machine that comes in the house.

It's about the same as a 630 - some are absolutely fine, shiny like new solder on the pads. Others... Well, I've recapped most of my 630s! They're something like 1994 to 1996 or 7? (Last 68k Mac)

 

[lobust]

Appreciate the insights guys. I see no real reason to recap this one and am somewhat emboldened by your answers!

I have two 7100 boards that are real basket cases, built just a year earlier than this 8500...

 

[jmacz]

I replaced the logic board caps on my 8500. All of the caps visually looked ok. But some of them had moisture underneath and fish smell. I think the smaller 10uFs were worse than the larger 47uFs.

 

[s_pupp]

Because of a malfunctioning PCI card slot and onboard graphics accelerator, I did a complete recap of my PM8600 Kansas board. This fixed it. There was no visible electrolyte or corrosion on the logic board, but because of dozens of bad experiences with SMD aluminum electrolytics in other devices, I decided preemptively to recap the logic board of my PM9600 Tsunami.

 

[obsolete]

@jmacz and @s_pupp, that makes sense. Did you measure the capacitance of any of the parts you removed? One thing I've observed and recently begun to understand is that old capacitors that have no visible signs of failure will sometimes read significantly higher capacitance on a simple LCR meter. This is because their ESR has increased, which slows the charging, and the meter interprets the slower charging as higher capacitance. I've only observed it on power supply capacitors that have spent a long time at high temperatures, but I'm curious whether these little SMT electrolytics fail that way too.

I still don't subscribe to preventative recapping; there has to be a reason (visible or functional) for me, but I understand why people do it.

 

[Phipli]

@jmacz and @s_pupp, that makes sense. Did you measure the capacitance of any of the parts you removed? One thing I've observed and recently begun to understand is that old capacitors that have no visible signs of failure will sometimes read significantly higher capacitance on a simple LCR meter. This is because their ESR has increased, which slows the charging, and the meter interprets the slower charging as higher capacitance. I've only observed it on power supply capacitors that have spent a long time at high temperatures, but I'm curious whether these little SMT electrolytics fail that way too.

I still don't subscribe to preventative recapping; there has to be a reason (visible or functional) for me, but I understand why people do it.

Hum, that shouldn't fool a proper LCR? Are you using one of those AVR based ones (which are great, but less accurate)? If not, what frequency did you have your meter set to?

Generally I've mostly see the non-goop everywhere ones fail by just sort of drying out and the capacitance going down, sometimes to effectively zero.