Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Well, open it up and take a look. Sometimes the faults are visible like bulging or blown caps, burned transistors or other burn marks. Don't touch anything big though, caps might still have charge.
This is great advice. - Since you've had the box plugged in, give it some time to discharge. You can (very carefully) test the caps in question with a multi meter to see if they still carry a charge. I gave myself a nice little zing years back fiddling with a PowerMac g5 psu LOL. It didnt kill me but certainly not something I'd want to repeat.

When I was playing alot of rock n roll in my 20s, I also zapped myself with an old Ampeg V4 amp that had a malfunctioning ground switch. I was playing without a shirt and was sweaty and the thing zapped me so good, I went down to my knees. I liken it to getting hit by a TNG phaser set to stun LOL.

Again, didnt kill me (obviously) but absolutely something I dont care to repeat. It was not a nice experience Haha :D
 
Well, open it up and take a look. Sometimes the faults are visible like bulging or blown caps, burned transistors or other burn marks. Don't touch anything big though, caps might still have charge. ;)
Oh, I fully intend to. There's apparently a fuse, too, so it might be that simple.
 
Yes, but there is usually a reason for the fuse to blow. I mean the fault probably lies somewhere else than the fuse itself.
Indeed, there is, and one of the things I learnt in decades of electronic maintenance is that sometimes, they blow due to fatigue. Particularly if they're rated a bit too close to the normal operating current. The inrush current of a 980W power supply is quite large, so it really might be that simple. If not, the replacement supply will not last long! :D
Then the other incoming 3,1 2.8GHz 8-core will be slowly stripped and transplanted until I have a working machine.
 
A question about the 3.2GHz 8-core 3,1:
Why did Apple see the need for the fancy coating on the cpu/heatsink interface? Still can't find any information about it.
 
A question about the 3.2GHz 8-core 3,1:
Why did Apple see the need for the fancy coating on the cpu/heatsink interface? Still can't find any information about it.
Having managed to find out more about this rather silly idea of Apple's, I wonder if it's what has caused the PSU failure - it and the gasket grease used to contain it are both electrically conductive. Who thought THAT would be a good idea? If so, then I may be looking at logic board and cpu damage as well. Before I install a known good PSU, I will carefully remove this 'special coating' and revert to more normal thermal paste. If it all still appears dead, then I'll try the first set of CPUs in the incoming spare (which appears to have a GTX680!), then continue transplanting until I have a working machine. It will probably mean scrapping the 'special' heatsinks if it proves difficult to completely remove the compound and gaskets. If that is the case, I'll just go for a complete refurb of the rather tatty incoming machine and make do with 2.8GHz!
 
One of the other things I wanted for this Mac Pro was another Magic Mouse. Idly scrolling through eBay, I found one for under a tenner, beciuse "it won't right-click due to a dent". OK, it arrived earlier and the fault was fixed in 5 seconds with the aid of a pair of flat-nosed pliers. Cheapest decent mouse I've ever bought.
 
  • Like
Reactions: TheShortTimer
Having managed to find out more about this rather silly idea of Apple's, I wonder if it's what has caused the PSU failure - it and the gasket grease used to contain it are both electrically conductive. Who thought THAT would be a good idea? If so, then I may be looking at logic board and cpu damage as well. Before I install a known good PSU, I will carefully remove this 'special coating' and revert to more normal thermal paste. If it all still appears dead, then I'll try the first set of CPUs in the incoming spare (which appears to have a GTX680!), then continue transplanting until I have a working machine. It will probably mean scrapping the 'special' heatsinks if it proves difficult to completely remove the compound and gaskets. If that is the case, I'll just go for a complete refurb of the rather tatty incoming machine and make do with 2.8GHz!
Are the CPU's lidded or delidded?

If the cpu and heatsinks make contact, like they usually do in computers, I see no reason why you could not replace it with modern good quality thermal paste.

Pics?
 
  • Like
Reactions: DCBassman
Are the CPU's lidded or delidded?

If the cpu and heatsinks make contact, like they usually do in computers, I see no reason why you could not replace it with modern good quality thermal paste.

Pics?
Currently reassembled, so no pics at the moment. The one CPU I clearly saw had a normal looking lid, with the writing on. I'm assuming de-lidded would not have that. What I think happens is that there is a small gap filled with liquid metal when it all gets hot., hence the need for a gasket to keep it in place. If that has escaped and shorted something, well, that would explain the dead machine, but I would have expected some sign of that. Instaed, it worked one day, and not the next.
 
Getting hyped for this years holiday season. My 2009 MBP dressed up :D 🎄

Christmas desktop 2024.jpg
 
WHat have I done? Nothing. Too busy. And now there's a red-level storm barreling into SW England, so I'm just going to go to bed and read. I now have a spare PSU for the cMP 3,1, so I'll try that on Sunday, but I may just strip out the CPUs first and scrape off the toxic stuff used on the 3.2GHz model and apply Arctic Silver MX4 instead. Then we'll see what we have, properly. I hope.This may also mean the 8800GT is all good, and not the primary problem. If it still is, there's a 5770 just waiting to go to work. And possibly a GTX680 in the spares-or-repairs 2.8 8-core incoming.
 
An afternoon trip is now cancelled, so I will continue testing with new SSD and El Cap.
Are you going to test the original GPU which was suspect? I mean it could be that the failing PSU could not provide enough juice for all to run properly and your problems might all be result of it. But, of course the GPU might be bad too.
 
Decided I'd have a quick look at the dead PSU. It rattles. Not good. Having eased it open a bit, I could see the fuse, and it tests good. Having finally shaken out what's rattling, it's this:
20241207_135909.jpg

The front casing of a high-voltage switching IC. Can't see the rest of it, but think I can safely say this unit is an ex-PSU...
 
  • Like
Reactions: ToniCH
Well, those are readily available at around USD/€ 3 each. So, if there is not much more destroyed the PSU might be fixable.

Agreed, but extremely difficult to get into, so doubt I'll bother. Given lack of noise/smells when it failed, I'd guess this might be the whole problem, but what could cause it?
 
Still won't successfully boot anything, so time to try thr new GPU, then if no joy, I'll redo the processors. After that, I'll wait for the other one to arrive, and decide from there.
 
Can't you just take the whole board out of the PSU case? They can be cramped and a bit annoying to deal with but unless they are epoxied in they are usually quite doable. Worth it? Dunno, its up to you. You could fix it, recap it and use prime quality parts to make it last forever but unless you plan to keep the machine it not might be worth it. You never know if in 40 years it achieves some magical classic supercomputer status or something? ;)

Maybe the chip just blew because it was old, maybe cheap quality, made on monday and there was a spike or something that broke the camels back?
 
  • Like
Reactions: DCBassman
Still won't successfully boot anything, so time to try thr new GPU, then if no joy, I'll redo the processors. After that, I'll wait for the other one to arrive, and decide from there.
Now that it tries to boot ie. powers up do you get diagnostic led action?
 
  • Like
Reactions: DCBassman
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.