New build problems - help needed (UPDATED)

[CurseTheSky]

I have the eVGA 680i SLI board, and it seems to run fine. I haven't had a chance to really get down to overclocking, because I'm running into an overheating problem somewhere... or at least, I think I am.

The computer went together without too much problem. It even worked with my 2.1v Dominator memory without a hitch. I've been running it at stock speeds, just installing programs, for about 10 hours now. While I was just beginning to install Doom 3, I got a BSOD (couldn't read it in time) and the system shut down. When I tried to start it back up, it POSTed, went to the BIOS logo screen (I turned the screen on), and immediately shut back off. Every time I tried to turn it back on after that, it would turn on for a split second, and turn right back off. I was afraid I had taken a voltage spike (even though I'm on a surge surpressor / power strip), and fried something. I let it sit for about an hour, came back, and it turned on with no problems.

This seems very typical of an overheating issue. My first guess was the processor, since it's a brand new liquid cooling system. I took the time to install it right (tightened all the connections with a wrench, put teflon tape on the threads, removed all of the air bubbles - no leaks!), but it was still possible that the water wasn't moving around. I don't think that's an issue, though. The nTune software (the only Windows program I have to measure temps right now) shows the processor at 38C idle (I think it's wrong) after about 10 minutes of being in Windows, and the BIOS showed the processor at 28C when I first started it up.

While I was monitoring the temps in the BIOS, I noticed the system and MCP temperatures were creeping up. The system temp leveled out around 40C, and the MCP climbed all the way up to 48ish. Do these seem high, or is this typical for this board? The voltages for EVERYTHING except the memory (changed to 2.1v) are still on auto, and all of my case fans are in place.

Any ideas?

Also, will the motherboard let me run a single graphics card in the PCI-E x16 2 or PCI-E x16 3 slots without any problems? The PCI-E x16 1 basically puts the card right on top of the MCP, and I'm sure that's not helping the temps.

 

[JAG87]

the temp is fine, those boards can run up to 70C on the MCP. with the fan however u shouldnt be above higher than 50C

 

[CurseTheSky]

It just did it again.

This time I let it cool for quite a while before trying to turn it on again. It STILL won't turn on. If I turn off / unplug the power supply, plug it back in, and try turning on the PC power again, everything will spin up for a split second, then quit. All of the voltages coming from the PSU lines looked fine in the BIOS.

I finally got a good look at the BSOD. It said something to the fact of turning off to prevent component damage (which SOUNDS like overheating). Keep in mind, the 49C on the MCP and 40C on system were just idle temperatures. Both times it's BSOD'd (once while installing Doom 3, the other time while playing music in Windows Media Player) the computer had been running for a few hours.

I'm wondering if it's the memory overheating. It's the ONLY thing I've raised the voltage to (1.9 stock, 2.1 raised), and before I tweaked the memory voltage, I didn't seem to have a problem. Unfortunately, I was also restarting constantly for updates, so that may factor in as well.

I just wish I could get it to start now. Clearing the CMOS via the jumper and letting it sit and cool down hasn't helped so far.

 

[CurseTheSky]

I was JUST in the middle of typing out a post about how I got it to start up again by clearing the CMOS (again), and using the switch on the board instead of the case switch when it crashed again.

This time it crashed after about 10 minutes or less in Windows. I was just in the middle of typing out a post here, and I paused to open the nVidia system monitor to check temperatures. The computer just shut off, no BSOD, no warning, nothing. I don't get it.

10 minutes shouldn't be enough time to overheat unless there's something really wrong. I suppose it's possible that my water cooling is failing or the waterblock isn't properly installed, but that wouldn't explain the 28C temperature (stable) I observed in my BIOS earlier. It also wouldn't explain how I was able to go for a few hours without a problem, and now I'm crashing in less than twenty minutes.

Now I'm leaning toward some kind of BIOS error or hardware failure. I don't know what or how, and this is REALLY starting to get on my nerves. If it doesn't start up without a hitch this time, I'll really be stumped.

If anyone has the same board and can post anything they've changed voltage wise or general setting wise, I'd appreciate it. During all of these crashes, I haven't overclocked a thing or tweaked more than a few harmless settings (the worst being changing memory timings from 5-5-5-18 to 4-4-4-12 and VDIMM from 1.9v to 2.1v - what the memory is rated at).

 

[CurseTheSky]

Running memtest86 now, bad memory is the only other thing that I can think of. It STILL wouldn't explain why the board wouldn't start back up, but motherboards sometimes do strange things I guess. Maybe it has a protection feature to keep it from restarting if it runs into a major error. I just can't believe I made it that far into my Windows / program installation without any problems, and now I'm running into all kinds. If hardware is unstable, you usually know either as soon as you start up the computer for the first time, or as soon ais you try overclocking. Neither happened for me.

 

[CurseTheSky]

No errors after one full run of Memtest86 3.2, along with four additional runs of test 5. Since it's not overclocked at all, I'm going to say the memory isn't bad (or at least, isn't completely bad). I'll run more extensive testing over night.

On to Prime95 and SuperPi... At least that rules out one possibility. I just wish I could figure out what's doing it.

 

[lxskllr]

See This thread at Overclockers.com. It looks like it could be a heat issue.

 

[CurseTheSky]

Interesting. Especially considering the backplate needed for my waterblock. I hope those components are ok...

Luckily there's a thick, soft pad between the metal backplate and the motherboard itself. The fact that it works at all SHOULD mean that they're undamaged, but you never know.

 

[CurseTheSky]

Changed all of the voltages from auto to manual settings. VCore is around 1.3125, VDIMM is 2.1, and the rest are all somewhat reasonable values. Running Orthos for five minutes now, and no problems so far. Hopefully the auto voltage settings in the BIOS was the problem...

Oh God would I love to blame this on the BIOS and not damaged hardware...

 

[lxskllr]

Originally posted by: CurseTheSky
Interesting. Especially considering the backplate needed for my waterblock. I hope those components are ok...

Luckily there's a thick, soft pad between the metal backplate and the motherboard itself. The fact that it works at all SHOULD mean that they're undamaged, but you never know.

I haven't seen that board in person, but it sounds like a crappy design. Most modern (enthusiast)heatsinks use some kind of backing plate, and to put electronics there without protection is poor engineering.

You did a fresh install of Windows right? I had major problems when I built my new system and did a repair install. That was a big mistake, and it cost me a weekend chasing down non-existent hardware problems before I figured out it was Windows giving me issues.

Good luck, I hope you get it figured out.

 

[CurseTheSky]

Fresh install of Windows XP Professional 32 bit on a brand new hard drive.

10 minutes of Orthos so far, CPU usage is at 100% on both cores, and memory usage is around 1.5GB. I have high hopes for the auto-voltage theory, but I'm not convinced yet. If it can survive Orthos all night long, or let me play a few hours of games without crashing, I'll be confident.

I should add, I haven't flashed the BIOS at all. I'm not sure if a new one is available, seeing as the board just came out. If Orthos keeps going strong, I'll probably check for a new BIOS next.

 

[CurseTheSky]

Well, it WAS going good.

I just went to bed, Orthos, CPU-Z, and the nTune system monitor (for temps and voltages) were all running. Orthos was about 30-40 minutes in, I heard the fan(s) spin down (probably the MCP / SPP fan), and it BSOD'd about five seconds later.

This time I actually got a look at the BSOD though:

***STOP: 0x000000F4 (0x00000003, 0x892CD748, 0x892CD8BC, 0x805D1140)

After the crash, while it sat on the BSOD, one of the HDDs was making clicking noises... about 3-4 clicks once every 10 seconds or so. The HDDs are right next to each other, so I'm not 100% sure which was clicking, but I THINK it was the Raptor.

Since I don't have a DMM + Thermometer handy, I checked the component temperature by hand. I can't really do the CPU well, since the waterblock is encased in plastic. The metal bracket that holds the waterblock on was decently warm. Memory was warm - almost hot. The heatpipe between the MCP and SPP was very hot to the touch, and the MCP itself was very hot. The heatsink fins on the SPP were also decently hot.

Now I really don't know what to think. Everything points to the MCP or SPP overheating. I DO have the fan that comes with the heatsink attached to it, which is probably just enough to keep the SPP running. The MCP has to rely on the heatpipe only, and being near my graphics card, it's probably not an ideal location. When I switch on the PSU, the lights on the top of the motherboard come on, so I know it's getting some power. As soon as I try to turn on the computer (by the case power switch, or motherboard power switch) it instantly turns back off. I assume this is because of the still-blazing-hot temperatures on the MCP.

Now I'm sort of at a loss for what to do. EVERYTHING points to MCP heat... but what can I do about it? If I remove the heatsinks and heatpipe assembly off of the MCP and SPP, remove the old thermal paste, and apply some AS5, will it void the warranty? I want to make sure I can RMA this board if my efforts are worthless.

Even with AS5, I'm not confident that I can keep it cool enough to run 100% stable. The MCP is literaly right under the edge of my 8800GTX, so adding an aftermarket cooler really isn't possible - especially if I decide to go SLI later on.

Thanks for the responses so far. I appreciate the input.

 

[Bozo Galora]

well, HDD clicking means you have to run drive fitness test floppy for your brand of HDD, downloaded from manuf site. Stick a screwdriver tip on the tops of drives and handle end hard against your ear to see which one.
I would go to Fry's and buy a cheap REGULAR HSF and install to rule out water cooling stuff
The latest production bios is 1009
I believe there was detection probs with latest nvidia vidcard versions with older bios
And your unnamed "500W" power supply implies you are using the el crappo one that came with case
Your symptoms can apply to barely adequate PSU after warm up.
I also assume you plugged in all extra pwr molex like the CPU and vidcard

 

[CurseTheSky]

Originally posted by: Bozo Galora
well, HDD clicking means you have to run drive fitness test floppy for your brand of HDD, downloaded from manuf site. Stick a screwdriver tip on the tops of drives and handle end hard against your ear to see which one.
I would go to Fry's and buy a cheap REGULAR HSF and install to rule out water cooling stuff
The latest production bios is 1009
I believe there was detection probs with latest nvidia vidcard versions with older bios
And your unnamed "500W" power supply implies you are using the el crappo one that came with case
Your symptoms can apply to barely adequate PSU after warm up.
I also assume you plugged in all extra pwr molex like the CPU and vidcard

The parts in my signature are not my current system. I'm using a 750w Thermaltake Toughpower, which did not come with the case. Though it's generally not one of the recommended PSU brands, it has been getting good reviews.

Interestingly enough, it's the Raptor that was clicking. It's also no longer being detected by the BIOS, so I'm going to assume it's dead. There's RMA #1.

Though the voltages on the PSU were all stable every other time I've looked, it's not providing 10.03v on the +12v, 3.22v on the +3.3V, 4.73v on the +5v, and 3.02v on +Vbat. I'm not sure if the motherboard / BIOS has direct control of any of these (all voltages are set to auto at the moment), but 10 volts on the +12 rail?! Sounds like RMA #2 - PSU or motherboard.

I've plugged in the main power connector to the motherboard, the 8 pin power connector to the motherboard, and the two PCI-E connectors to the graphics card, each coming off of different lines.
This is a nightmare.

 

[Bozo Galora]

Bios has nothing to do with that
maybe the raptor going bad killed one of the 12V lines, or the TT PSU went bad, and killed the raptor

you can use a dmm (digitalmultimeter) with needle point probes on your main 24 pin molex, set the range to volts DC, put the red probe down into the yellow wire sections, and the black (ground) probe into (any) black wire section, or the metal case frame, and check that way - making sure you dont touch the metal probes of course. Then check the yellows the same way on the molex that would go to raptor SATA power

http://www.softpedia.com/get/System/System-Info/EVEREST-Ultimate-Edition.shtml

 

[CurseTheSky]

The PSU doesn't seem to turn on when the motherboard connector isn't connected. It turned on for about 30 seconds, then turned back off... like it usually does if I switch the PSU on without the motherboard / rest of the computer started.

Is that usual? If so, how can I measure the lines?

 

[Bozo Galora]

No, it wont "turn on" disconnected unless you bridge 2 certain wires with a paper clip, which is too dangerous for someone not familiar.
I meant WITH the connector plugged in.
You have to insert the needle probes from the top of molex and work them down until they touch connectors inside.
And the everest HW monitoring software above should give you accurate voltages

FWIW - 12V lines should be plud minus 5% idle, 10% under peak load, or 10.8V min., 11.4V idle min

tester
http://reviews.pimprig.com/miscellaneous/frozencpu_atx_2.0_psu_tester.php
http://www.frozencpu.com/psu-165.html?cid=prf

Edit: dunno if this is relevant, but 3 of 24 newegg buyers found their 750TT dead
http://www.newegg.com/Product/CustRatin...eld=0&SummaryType=ALL&Pagesize=&Page=2

 

[CurseTheSky]

Very interesting. I'll try measuring voltages again in a minute.

I figured I should mention this:

On the motherboard, the SATA ports are labeled 1-6, from the top of the board to the bottom. I plugged my HDD's into SATA ports 1 and 2, with the Raptor in #1. When I started up the computer, the BIOS had them listed as Raptor in #6, and Seagate in #3. I had to switch them around to install Windows without copying files onto the other HDD. Later, to give myself more room, I moved them down to lower SATA slots. They were then listed as slots 2 and 5, or something like that - completely out of order. Each time the drives were right next to eachother in SATA ports.

Does this sound strange, or does the BIOS reorder the ports for some reason? Since it was one of the HDDs that blew, I figure it was worth mentioning.

Edit: from a Newegg review on page 3:

Pros: none anymore, burned out a month after buying it. im not even using half of if its so called "toughpower". i wasnt even playing nor doing anything that'll draw power, i was just chatting and watching tv on it, all of a sudden the whole system turned off. tried to boot up again.... nuttin. all i know is that the mobo is getting power (green light thanks to asus) but when u hit the switch nothing happens. luckily i had my enermax psu which i almost sold, and put it in and the whole system is up and running again. now i got to contact newegg and see if they can switch it for another one that hopefully wont die on me.

Sounds a bit similar. I always thought that it wouldn't turn back on because of overheating. My motherboard was also still getting power the entire time. Very interesting.

Too bad its been over 30 days. I can only return it for a replacement and hope for the best.

 

[Bozo Galora]

SATA ordering means nothing
you are gonna have to look in your manual to see how SATA numbering goes
I dont know how loud raptor clicking is, might just be smart monitoring, wouldnt jump to conclusions there

EdiT:
EVGA seems to be pretty decent board
http://xtremesystems.org/forums/showthread.php?t=122354

 

[alphatester]

Hy Curse The Sky,

yesterday i baught the same mobo evga 680sli, together with the new gtx 8800. I have the same problems. First start, and the lcd indicator on the board shows Code 25. I thaught the VGA Card maybe is corrupt, and i changed to Nvidia gx2 card. no change. i changed it back. next time i removed the dvb-card. no change. usb cable-no gigabit ethernet from port 1 to 2 and whow the board starts to post screen.

I have a similar configuration with thermaltake 700w psu C2D 6600. The both PCI Express Power Lines are on different lanes...

Well I made my changes in the bios and restart...

I want to update my existing win xp 32 bit... but after a a few minutes or seconds there is a bsod, or the system falls down...

The system worked before with an asus P532N SLI deluxe. and two Asus 7950 GX2 in Quad SLI mode. without any problems...

Some any ideas?!?

 

[alphatester]

Problem solved... Got a new EVGA 680i sli...
Great performance... My e6600 works with 3.8 Ghz. With the GTX 8800 12900 pt. in 3D mark 2006!!