Help diagnosing unexpected shutdown

[birthdaymonkey]

I recently slapped together a rig from some older components to run a TV screen in my kitchen.

Parts:
Rosewill Legacy U3-B mATX case
Q9550 on an Acer G41 motherboard (pulled from Veriton M275 system)
4GB DDR3
Crucial C300 64GB SSD
Asus GTX 630
Seasonic M12-II 620W

Everything seemed to be working fine, but then it just started shutting down without warning. No BSOD, nothing in Event Viewer except "the last shutdown was unexpected" (ID 6008).

It's happened three times now. I originally suspected the OEM PSU (also from the Acer Veriton), so I replaced it with the known working Seasonic, which I got from a friend. The system was up for four days, but then this morning when I went to use it, it had shut down at 3 AM.

The machine was in a closed cabinet, so heat could also be the culprit, but I've been monitoring temps closely and nothing seemed unreasonably high, with CPU idling at 45 degrees max, and 50 degrees max from the hottest mobo sensor. The shutdowns also happen when the system is just idling.

Anyway, I moved the computer to a new location, although it's not as convenient as the cabinet (much lower WAF). I just moved it this morning, so I guess it's wait and see what happens.

I'm wondering if anything else could be responsible for the shutdowns. The problem didn't happen until I upgraded the memory from 3GB to 4GB (different set of DIMMs, which had been working fine in another machine), but I can't ever recall hearing of a complete shutdown due to a memory incompatibility.

Any ideas?

 

[Ketchup]

I find this part odd:

The machine was in a closed cabinet, so heat could also be the culprit, but I've been monitoring temps closely and nothing seemed unreasonably high, with CPU idling at 45 degrees max, and 50 degrees max from the hottest mobo sensor. The shutdowns also happen when the system is just idling.

A motherboard sensor is not going to tell you how hot this CPU is running. You want to be looking at the 4 CPU core sensors. I am not sure what monitoring you are using, but I would put some stress on the machine and see if you can make it reboot.

 

[birthdaymonkey]

I find this part odd:


A motherboard sensor is not going to tell you how hot this CPU is running. You want to be looking at the 4 CPU core sensors. I am not sure what monitoring you are using, but I would put some stress on the machine and see if you can make it reboot.

The core temps from the CPU sensor are always lower than what the motherboard reports as its CPU temp. The hottest they get is into the high 40s/low 50s, although I never Primed the system for more than 15 minutes or so.

I'm going to memtest overnight tonight, then I'll Prime overnight tomorrow and see what happens.

 

[CoPhotoGuy]

Are the capacitors on the motherboard leaking?

 

[VirtualLarry]

I'm unfamiliar with the Veriton system. Were BOTH the CPU and mobo from that system, or just the mobo, and you upgraded the CPU? If so, then the mobo may not be designed to handle the wattage of that quad-core. I have some Intel ITX boards that don't support CPUs higher than 65W class.

 

[birthdaymonkey]

Are the capacitors on the motherboard leaking?

I haven't checked, but I will have a careful look at them.

I'm unfamiliar with the Veriton system. Were BOTH the CPU and mobo from that system, or just the mobo, and you upgraded the CPU? If so, then the mobo may not be designed to handle the wattage of that quad-core. I have some Intel ITX boards that don't support CPUs higher than 65W class.

Just the mobo is from the Acer system; the Q9550 I got from ebay. Your suggestion made me look at the service manual more closely - and what I found was conflicting info.

At the end of the manual, in the FRU list, the Q9550 is clearly listed. This is what I based my decision to buy a Q9550 on...

http://www.manualslib.com/manual/232986/Acer-Veriton-M275.html?page=78#manual

However, at the beginning of the manual, it says "support varied Intel processors that the maximum power consumption is equal and under 65W."

http://www.manualslib.com/manual/232986/Acer-Veriton-M275.html?page=15#manual

So maybe I'm boned here. I'm going to stress test the system outside the cabinet and see what happens. Maybe it's borderline stable with the Quad, and the extra heat just pushes it over the edge.

If it's not stable (and I now have my doubts that I'll ever be comfortable with the stability), the question is should I buy a new S775 mobo for CAD 75.00, or just resell the Q9550 and put the Pentium E6600 back in...

 

[VirtualLarry]

Get an E7xxx or E8xxx CPU, and re-sell the Q9550. Or just sell all of the parts off, and grab the next G3258 combo? (Are you near a Microcenter?)

Edit: Sure that the FRU list didn't mention the Q9550S CPU? That one is 65W.

 

[birthdaymonkey]

Get an E7xxx or E8xxx CPU, and re-sell the Q9550. Or just sell all of the parts off, and grab the next G3258 combo? (Are you near a Microcenter?)

Edit: Sure that the FRU list didn't mention the Q9550S CPU? That one is 65W.

Well, the E6600 I have is the 45nm Pentium version, which runs at 3.06GHz and benches the same as the E8400.

If I do end up needing to replace the motherboard, I will just sell the Q9550, maybe take a $10 loss based on what I paid for it. Cheaper to get a G2358 (for around the same price that I can sell the C2Q for) and an H81 motherboard, not to mention all the other benefits of moving to a modern platform.

Ran Prime95 all night (15 hours) without issue:

 

[Ketchup]

That is a good sign. And this is is the new location, right? If nothing has happened since the move, I think you have solved the problem.

 

[birthdaymonkey]

That is a good sign. And this is is the new location, right? If nothing has happened since the move, I think you have solved the problem.

Yep, that's with the machine on a shelf instead of in the closed cabinet. Fingers crossed it'll continue to stay on over the next few days...

 

[wjgollatz]

Makes sure you have it backed up, fun a checkdisk, virus and malware on it.

One day my drivers were good. The next day, BSOD because a Windows Service got corrupted. I had that issue and drivers going bad on a volume that made inself corrupt. I ended up having to disable a windows service, and remove a driver to get it back. See if the problems happen when the comp is in safe mode.