System unstable after water cooling setup

[toddfx]

I just installed my first water cooling loop tonight. I have the XSPC Rasa 750 RS360 liquid kit on my Q9550 @ stock 2.83 ghz, and also added in a Swiftech MCW80 VGA block for my dinky Asus nvidia GT-240. I have an EVGA 780i board.

I bled it with the motherboard unplugged for about 45 minutes, then booted into BIOS. Temps looked good there, with CPU and board temp at ~30C while the MCP was at ~40C. I monitored the temps for 10 minutes and all remained steady so I walked away for a few minutes (still idling in BIOS). When I came back, I was welcomed by a black monitor and the horrid beeping alert coming from my motherboard speaker. I quickly powered it down.

After repeated tries (about 20, through troubleshooting), I continue to get lockups and BSoDs within minutes. The lockups happen everywhere: in BIOS, during Windows boot, and while idling in Windows. Here is what I got the first time I made it into Windows:

(I also find it strange it was not reading my fan speeds during this screenshot, but on later attempts it did end up displaying the speeds)

I have tried re-seating CPU and RAM, and tested with an alternate graphics card with it's stock air cooler, just to rule out any GPU issue. Same result.

My water tubes have bled for about 2 hours and i’ve done a fair share of tapping and tilting, but some fairly sizable air bubbles remain. Could this be the problem? Lingering pockets of air causing temp spikes?

Longest run i’ve had so far was about 8 minutes. Shortest about 20 seconds. It seems like a temperature problem, as the longer I let it sit, the longer it lasts when I turn it back on. But why are all my temperatures reading so low? I don't get the "overheating" beep alert when Windows locks up, but on a couple of occasions I got it (or at least 3 long beeps, anyway) on reboot AFTER a lock-up. Just to reiterate though, the crashes happen while I am staring at low temp readings of 30 - 40 C across the board.

I should also mention that I removed my CMOS battery before installing water cooling, as I wanted ALL electicity out of the board. I get no errors on post, but could some incorrect CMOS settings be a culprit? I’m unsure what to look for there.

Thanks a lot guys, it is unfortunate I am having these issues on my first water cooling build! All went so well during setup...

 

[Tsumi]

Sounds more like a hardware problem than a problem with your watercooling setup. I suspect that there is something up with your motherboard, since you have ruled out the GPU. It can also be your CPU or RAM causing the trouble. Do you have any spare CPU's or RAM to swap with the ones you have? Or another computer you can test those components in? You just need to start ruling things out one by one at this point. It is likely you may have damaged your motherboard in some way you did not realize.

Removing the CMOS battery is going a bit far, might I say extreme paranoia? However, lockups in bios indicate some sort of hardware problem, especially since everything is at stock. A bios update might be worth trying, but if it's freezing in the bios... not recommended.

Bubbles in the stream will not cause any sort of temperature spike. It takes time to heat up the waterblock itself, and the waterblock is always in contact with at least some water. Any air bubble in the block is quickly displaced by water.

 

[DeathPrincess]

Check the connection to the CPU fan header on the motherboard isn't loose and is inserted properly. Some boards don't work if they sense the CPU fan has stopped working to stop overheating (from not having a cooling meathod). Or that something is connected to the CPU fan header.

Check all your connections, including ram/other components, that you haven't knocked something out of place, or that something is loose.

I would also spend time bleeding the loop. All loops take time to settle, and for tiny bubbles to work their way out the coolant, so it will need refilling during the initial stage. Make sure you remove as much as possible each time, and look for cloady sections where tiny bubbles lurk. Flicking these will make bigger bubbles come out. What kind of coolant are you using? With distilled water there is low risk of damage to components unless you leave it on the surfaces for a while (to pick up contaminents).

 

[thax]

you didn't get any water on your mb did you? i would rule out ram go down to 1 stick and test one at a time, also how tight did you put your water blocks on? On my system if i go to far it randoms bsod and fails to post.

 

[Cruiza]

Check the connection to the CPU fan header on the motherboard isn't loose and is inserted properly. Some boards don't work if they sense the CPU fan has stopped working to stop overheating (from not having a cooling meathod). Or that something is connected to the CPU fan header.

I definitely would check this as well, as I had a mobo that did the same thing if the fan speed wasn't high enough.

Also, what kind of beeping pattern do you get? They actually mean something, haha. Usually you can google the patten (for example, "long short short beep" or something) for some help.

 

[toddfx]

Thanks for all the quick responses.

Initially I had one of the radiator fans plugged into where the CPU fan had previously been. It is only taking up 2 of the 3 prongs, so just out of curiosity I plugged my original CPU fan back in just to see if that would make it happy. No go. Same thing happens no matter what (or nothing) is plugged into that CPU fan port.

The mounting bracket that came with my XSPC CPU liquid block was kind of strange on installation. It has a sticky adhesive that presses directly on the rear of the motherboard (as per instructions). I just now sandwiched a layer of paper in there to act as a buffer in addition to the adhesive pad, thinking it was somehow causing a short. No go. Seems worse now, honestly.

I re-sat the CPU block twice more, tried 4 different sticks of RAM, and tried my alternate video card again. I also cleared the CMOS. No luck. I was considering a BIOS upgrade, but don't want to risk that since it is so unstable even during post.

This morning I have yet to make it into Windows. It posts about 10% of the time, does nothing about 50% of the time, and gives me a long beep the other 40% of the time. The long beep/buzzing is what I have always known as a temperature warning. None of the Phoenix AwardBIOS beep codes listed discuss long beeps.

I haven't even made it into BIOS the last 10 tries or so. When it does manage to post, the posting text does appear on screen, but appears to lock up even before I can make it into BIOS.

I don't have an alternate CPU or alternate motherboard to troubleshoot them.

My faith is quickly diminishing. Could it already be Sandy Bridge time for me?

 

[toddfx]

One more note.

During my water cooling installation, I was very careful. I can't imagine what I could of done to inadvertently kill the motherboard. I can pretty much 100% guarantee I did not get any water on any component. I was very careful in making sure my hands were dry when touching things, and had hawk eyes on all liquid ports during testing. Not a drop of tigerblood was leaked anywhere.

The only thing I cannot rule out is greasy fingers. It was a warm day in Chicago yesterday and I was excited yet nervous about my first water cooling setup. I do recall being sweaty at times, but cleaned my hands often and am always careful not to blatantly touch any circuitry.

There was one point (with the PSU totally unplugged) that a metal washer fell onto the motherboard. The CMOS battery was still plugged in though. Would that be enough to cause a short? In the past I have had things like that happen with no illl effects. So I dunno, just spitballing!

EDIT: One VERY strange thing that has come up twice now, is that during post it claims "Hyper-theading technology detected! Hyper threading ENABLED" which is complete BS, as my Q9550 does not even have hyper-threading support. Obviously either the motherboard or the CPU is very confused.

 

[Cruiza]

Did you try just a single stick of RAM at a time? Different slots?

 

[toddfx]

Did you try just a single stick of RAM at a time? Different slots?

Lol. Wow, so my newb colors show. I have been operating with just a single stick of RAM, but what I failed to test was the different slots. Apparently the first 2 slots are dead or something.

I moved the stick over slot by slot, and when it got to #3, the thing posted and booted flawlessly. I have sticks in slots 3 and 4 now, and it has been running without issue for 20 minutes now (knock on wood). Unfortunately I only have 4 gigs now instead of 8... BUT at least the damn thing runs.

Did I seriously manage to break two RAM slots? Or could it be something BIOS (or power?) related? I'll check my motherboards specs as far as RAM sequencing goes.

Thanks, guys.

 

[Cruiza]

I don't know what the deal is with mobos and RAM slots, but there seem to be so many seemingly random issues between the two. Based on your chipset, I assume your board isn't under warranty any more?

 

[Tsumi]

I've had boards where a RAM slot failed out of nowhere. Honestly, anything could have caused that, the most likely one being from old age and usage. I believe you when you say you didn't do anything wrong, you're way more careful than what I did with my watercooling setup. Well, not saying that I'm not careful, I'm just barely careful enough to make sure nothing gets damaged

 

[R-Type]

I've had boards where a RAM slot failed out of nowhere. Honestly, anything could have caused that, the most likely one being from old age and usage. I believe you when you say you didn't do anything wrong, you're way more careful than what I did with my watercooling setup. Well, not saying that I'm not careful, I'm just barely careful enough to make sure nothing gets damaged

With random failures like that, the first thing I do is check for bulging capacitors.