Mac Pro 2,1 random reboot, could it be the power supply?

[sirio76]

My old macpro 2.1(2007 model) started to randomly reboot.
My spec are:
8core 3.0ghz
14GB ram (6x1GB DIMM from Hynix and 2x4GB from OWC)
Concatenation of 4x500GB HDD
ATI radeon 1900XT
OSX 10,7
At the beginning I've though it could be a memory issue because lately 2 of the 8 DIMM(the ones from OWC) where randomly disappearing and the computer showed only 6GB instead of 14, also the led on the memory raiser card was flashing still. I removed the foulty DIMM and replaced them with other 2 Hynix 1GB modules and it seemed to cure the problem for some time. Today the computer started to reboot with no apparent reason, the memory was fine(no led flashing on the memory slot). I removed all the DIMM installed except the first two and unplugged everything from my MP, also unplugged the monitor(24"cinema display). I was able to access my desktop from another MP but it seemed to reboot also in this condition even if less frequently.
Now, I've read here on the forum about power supply-random reboot related problem and it seem to be my case, what's your opinion?
Thanks in advance

 

[Tesselator]

My old macpro 2.1(2007 model) started to randomly reboot.

Assuming this to be a PSU problem would most likely be an incorrect diagnosis in 99% of all such cases.

It could be any number of things the very last of which would be a faulty PSU. Probably the most likely candidate (after resetting PRAM and SMC) would be faulty memory somewhere in the system - either on the GPU or
one of the system memory DIMMs.

I had this same issue with a 2006 MP1,1 w/ 7300GT GPU and the problem went away by replacing the 7300 with an 8800. The problem stayed gone with the later addition of a GTX570. The trouble has never returned so in my case I suppose it's safe to assume the 7300GT was the culprit.

I've also head of people solving this issue simply by cleaning out the system, cleaning all the card-edge connections, and reseating the various attached components.

 

[davidb367]

Check the RAM

My 1,1 started behaving badly a few months ago. Random reboots and kernel panics. A bank of RAM failed about a month later. Once I received the replacement pair of RAM modules, the problems went away.

 

[bax2003]

Sounds like RAM problem.

 

[666sheep]

Tesselator is on a good trail. You have infamous X1900XT. If you have any other card, try replacing current one. Bad graphics card could give various weird symptoms, from graphical glitches to random restarts.

 

[sirio76]

Thanks you guys for your help!
Probably the memory is not the problem, I think it is unlikely that every one of the 8 DIMM that worked flawlessly for 6 years suddenly broked. Tomorrow morning I'll clean my MP with air compressed, and some electric contact cleaner to remove the dust from the RAM modules.
Unfortunately I've not a spare GPU to try and I prefer not to invest in a new one until I'm not sure of the problem, it's an old computer and I don't want to spend much money over it.
Is there any way to diagnose a bad GPU? is it possible to see from the log file if the restart was related to the GPU?

 

[Tesselator]

Thanks you guys for your help!
Probably the memory is not the problem, I think it is unlikely that every one of the 8 DIMM that worked flawlessly for 6 years suddenly broked. Tomorrow morning I'll clean my MP with air compressed, and some electric contact cleaner to remove the dust from the RAM modules.
Unfortunately I've not a spare GPU to try and I prefer not to invest in a new one until I'm not sure of the problem, it's an old computer and I don't want to spend much money over it.
Is there any way to diagnose a bad GPU? is it possible to see from the log file if the restart was related to the GPU?

If you don't have an extra GPU then you should eliminate the other possibilities as you can. If all else checks out then you're gonna need to buy a GPU replacement. You can get a test card and resell it for as little as $25 on the bay.

Also, only one system DIMM needs to be bad and it may not even check out as being bad. It needs to be tested by the process of elimination.

 

[davidb367]

The RAM sticks in the early MP's get VERY hot. After a long period of time the chips begin to fail. Not an issue on the later MP's.

 

[666sheep]

Unfortunately I've not a spare GPU to try and I prefer not to invest in a new one until I'm not sure of the problem, it's an old computer and I don't want to spend much money over it.
Is there any way to diagnose a bad GPU? is it possible to see from the log file if the restart was related to the GPU?

If you have another Mac, you could set screen sharing on MP, remove graphics card and log in remotely. Run some taxing app (such as Handbrake) and see what happens. Try it with known good RAM sticks.

Check system.log in Console, maybe you'll find something there.

The RAM sticks in the early MP's get VERY hot. After a long period of time the chips begin to fail. Not an issue on the later MP's.

Some do, other don't. It's not general rule. OWC RAM is known to fail, though. That's why they offer lifetime warranty