Hey all, I've got a problem that just cropped up early this morning...
The system was rock solid since I built it almost a year ago.
C2D Q6600 (not oc'd)
2x 1gb Patriot DDR ram w/ heatspreaders
Corsair Hx620 Powersupply
Thermalright Ultra 120 w/ 120mm S-Flex fan
Coolermaster CM-Stacker 800 case (Included: 3x 120mm Coolermaster case fans. 1x coolermaster 80 case fan)
XFX Gforce 7900 GS PCI-e
1x mastor 320gb SATA2 hd, 1x WD ata100 240gb hd
Dell 17" crt moniter
The problem is:
Early this morning while working in Carrara, the pc suddenly blackscreened and the XFX 7900's fan spun up (as it does when its engaged into 3d mode) as well as the Corsair's power supply fan spun up like upon a cold boot... Monitor's green power led was blinking like if video & hd was shut off in windows power settings (have them to turn off IF after 30mins of no activity. clearly not the case here, I was using the keyboard and mouse in the application)
It took several seconds holding down the power button to shut the power off...
Rebooted, checked the bios, then went back into windows. Seemed fine.
same as first time, maybe 3 hrs later...
Rebooted immediately, got back into windows, ran HD tune 2.54 on the maxtor drive, 1/4 way through quick error scan, the same crash happened again.
Reebooted again, thinking the Maxtor drive is suspect, ran hd tune on the WS 240gb... 1/2 way through quick error scan, same crash happened...
Rebooted immediately, went into bios, and system crashed exactly 2mins :50 seconds after entering bios.
Now: Same as fifth instance, but time to crash shortened to 1min 50 seconds...
I'll be troubleshooting every component when I get back, but I've got some suspects: ram, power supply, motherboard, cpu...
Although I'm wondering if anyone else had anything like this... Or have an idea...
I cant say what you problem is for certain but i can tell you that a crash that doesn't occur at the same time is normally the sign of something degrading/getting worse. If it was a heat issue, for example, the problem would occur at around the same time each time. Therefore, I think youre biggest suspect would be your RAM or powersupply. Also, check your motherboard for darkiing/charred marks or swollen/blown capacitors... not sure how much i trust AsRock boards (actually, i dont trust them at all). Good Luck!
Thanks, its always nice to get a few guesses and opinions - often helps narrowing down the faulty component fgaster - and in this case, that 0:1:50 time window will likely be shrinking with further power-ons. I'll get to the troubleshooting after supper, starting with a voltage meter.
When I last looked at it this morning, with a flashlight, I didn't see any leaking capacitors, or any bulging but i'll double check for the bulding, as well as for any markings on the board...
From what I can tell you - before this happened...
CPU temps on idle were around the low 40's , ie: 43c (thanks to the Ultra 120 and AS5+ )
CPUZ was reporting ~1.125 vcore (can't recal exact, can't check anymore now since 1.50 mins barely gets to the desktop before it cakes out... ), while the bios reports ~2.12
And actually due to the Asrock board - they specifically state the cpu and bus may run 5% slower when running a quad core on there - so yes, my CPU's been running a bit underclocked at 2200mhz and the bus at 1000mhz (instead of the 2.4ghz cpu, 1066 bus)
Luckilly the board supports both AGP + PCI-e, as well as DDR and DDR2 ram - they're famouse for that, but it offers a nice upgrade path which is why I chose that board last fall... So I do have a couple spare spower supplies, an couple years old lag775 Celeron 2.5ghz, and a couple agp cards and ddr ram that i can swap in peice by piece to further test... i've also got 2 other systems i can test out the Corsair powersupply on...
Interestingly, I came to this very site last fall to look up recomended power supplies, and thus chose the Corsair one, as NCIX also had a nice rebate on them.
I'll be back later tonight with some testing results.
Next Step: Boot to windows. sofar, 30 minutes stable (didn't install the catalyst drivers, but will in aq few minutes, soon as I un-unstall the nforce ware), then will try a couple games and get back into Carrara for a test...
Btw, also took the multi-meter on the PCI-E connector: 12.09v stable across the 3 of 8 pins (6+2 pin connector)
Hopefully, this was the luck of the draw (or a lucky break rather), finding the most likely culprit at first test... It was either the video card or the ram I was going to pull and swap first, being the easiest to start with (before gutting the case and trying the power supplies and the cpu's swap... I know, the backplate heatsinks are a chore to merely try swapping cpu's around for testing)
Sofar 8+hrs and running, in windows, trying a few multi-threaded apps and putting the 4cores to 100% usage. Nice and stable sofar. Havn't tried Prime95 yet for a stress test.
One thing still irking me is - is it really the video card or the motherboard? Still can't find anything visually abnormal. Even on the video card - its matt black pcb looks fine, and the capacitors are solid state caps.
I don't have any other systems on hand with PCI-E slots, nor any other PCI-E video cards to verify for sure its either the XFX7900GS card or the PCI-E slot on the motherboard. Just may have to pickup an el-cheapo card being the least hassle of the 2 (else 2nd motherboard would go into this celeron case i've been using to access the forums here with. Although, that'd be a cheap upgrade, grab a newer asrock hybrid board for this case, then a bit later swap in a newer cpu onto it.)
Whichever I do, i'll post my findings back here, and let this experiance add into the searchable treasury of knowledge ;)
One thing I do have to say however, is the quietness from the Radeon's fan is nice, vs the 7900's fan revving up to a higher rpm as the 3d engine kicks in. I gotta wonder - this 7900 has been my first and only nvidia card (aside from this incident, its been a pleasant run much like any other video card i've owned in the past). Its kinda making me wonder which way I'll upgrade next for a newer video card whenever that may be later down the line. I'll have to pay closer attention to the video card reviews for noise levels
Last edited by Jadeonar; August 11, 2008 at 02:46 AM.
AFAIK if the x1600 works fine then there's probably a problem with the PCIe slot or perhaps even the northbridge with handles it. I'd still have to say the problem is alllllll Asrock. Might want to invest in a newer board to put that 7900 to good use (its much more powerful than an x1600).
Could be worthwhile just running a 3dMark a few times over the x1600 to get it doing some work this may help identify if its the PCIe slot, then even repeat with the 7900. No local hardware shops willing to chuck your 7900 in a mobo for a couple mins to run a 3dmark?
Actually - I did take the 7900 to a local hardware shop that only builds custom pc's and sells hardware, where they put it into a new rig they were building for a customer. Already had Vista installed on it. Well, after booting up and around 2 minutes, the screen went black for them and the system became unresponsive. So all fingers are pointing at the card.
Note: I was lucky and the hardware store didn't charge me for the test, otherwise it'd have been cash out of my pocket plus/or possibly buying a cheap ati x2400 pci-e to test in the slot.
I went to XFX's site - help desk. Thankfully I registered the card soon as it arrived. I described the problem via their ticket system, and it took several days for a reply in which required another reply and again several days wait... But the XFX Support / Technician who did respond did suspect the card was at fault by me having tested the card in another machine getting exact same results. So they gave an RMA and the card's on its way. XFX's Doublelifetime warrenty is great, no fuss, as long as you don't mind shy of a week's wait in ticket messaging.
I've been working stable without lockups or crashes in a 3d modeling app using the x1600. The ATI driver has a wierd glitch in displaying the 3d scene in fully textured view, but its fine using the garoud shaded view (that or its typical of having uninstalled the nvidia drivers and installed the latest catalyst drivers and windows isn't happy about it). If you take a flat surface plane, and pan the camera around, you get some funky triangular slopes in that flat surface - which is supposed to be textured as pavement :P Then again, I usually switch to Garoud shading view mode anyways even with the 7900, when the scene starts getting fairly complex and resource heavy.
As long as I can continue working on 3d modeling with my quadcore, I'll live with it til I get a replacement from XFX. (the other alternative being this PentiumD "Prescott" Celeron 2500mhz (w/ 256kb L2) that is easily 4x slower for 3d rendering in a multithreaded app. a preview / test render usually takes me 3-10 minutes, and a full quality render 15-25 mins on the quad core. Multiply that by 4.5 - 5x+ the time to render on this Celeron and you get the picture.)
Last edited by Jadeonar; August 21, 2008 at 02:14 AM.
The XFX rma replacement arrived today. Thanks to Fed-Ex, who dropped it off on the doorstep, didn't knock on the door or bothered to ring the doorbell. Just about stepped on the package on my way out.
The replacement was a 8600GTS 512mb.
Later this evening, I got back, swapped out my ol' x1600pro agp, with the 8600GTS. Powered it up... No video... I'm not surprised. I picked up a 8800GT in last week's NCIX sale for 99bux which was PCI-E 2.0 and didnt like the Asrock board either. Guess it'll be motherboard upgrade end of this month...
And tonight just got even better. Plugged in the x1600pro , fired up the machine... monitor light keeps flickering trying to get a video signal to no avail... Maybe I can get the darned thing going again... I do have an older 17" onhand as a spare... Oh what fun.......