Why my system died: Opinions Sought.

[aws6000]

This is a bit long winded...

Five days ago I put together my first WC system. I installed a Swiftech 8500 kit. Install went smoothly. I did have some trouble getting to a point where I felt that all the air bubbles were gone. Fired the thing up and had it running for 4 straight days. I wasn't exactly happy with my temperatures, which were identical to my temps with my PAL8942. My motherboard (Abit IT7) tends to report high temps, but I expected to see some amount of decrease. On the 5th day, I noticed my pump was making noise. Didn't know if it was the pump itself or if trapped air finally accumulated inside the pump. Powered it down (for the first time) and went to bed. The next morning the thing refused to boot. It's dead, dead, dead. I started looking into things, and while I had the pump runnnig I flipped the case upside down and some air in the system broke loose and started to move around. So now I know that the noise of the pump is caused by air bubbles and not a faulty pump.

Now the big question is... Why is my system dead? I've purchased a new power supply, but it didn't help things. The motherboard has a digital readout along with a couple of small lights. The lights are lit when the power supply is plugged in, but nothing else happens. The PS is an easy return which is why it was my first purchase. Now I need to narrow it down to the CPU or the MB. The CPU will be just about an impossible return (this is only a guess), and I don't want to buy items (expensive items) that I can't use and can't return. What are the odds that the MB is good and the CPU is dead? Is it possible that I burned out the MB due to a leak? Wouldn't the leak most likely happen when the system was running and not when it was powered off? Do small air bubbles tend to converge to form large(r) air bubbles over time? I'm fairly positive that the system is air and watertight, but Swiftech uses a few valves and connections for fill and bleed operations. I can only assume that the design of the kit is sound and that it's tight.

My PC dying doesn't make any sense unless it was just a coincidence. Any ideas, comments, or suggestions to offer a newbie?

[JimmyG]

One possibility is that when the pump was making noise, it wasn't pumping any water because it was airbound. Just as you powered it down, the cpu was overheating and burned itself out. I would check the cpu in another computer or go out and buy a cheap cpu (maybe a duron) and test with that. Or you could try to boot with no cpu. The mobo should beep or maybe even show a code that says cpu dead.

[aws6000]

I don't think the pump was air bound. The amount of air is rather small, enough to make some noise but not enough to cause failure. I will try pulling the cpu and see what the motherboard does. My ancient Asus won't boot without a cpu so it didn't even cross my mind to try it with the Abit. I guess I could try and find a dirt cheap Celeron and see what happens.

I rarely see tales of doom on this board and I was very careful with my install. This is frustrating.

[Shroomer]

Another "coincidence" that could have happened... Some boards are having trouble with capacitors blowing on them. Look around and use your nose to see if any are dead. I'm GUESSING you may have had bad contact with the core and when the air got trapped in the pump it was inefficient enough to cause overheating.

Something to try...

A total tear down. Take apart EVERYTHING. I've had a problem I couldn't figure out until I did this. Actually I never figured it out, it just worked when I put it back together.

[Freeloader]

My first guess would be that while the computer was off water leaked out . Problably the same hole that was letting air in when the computer was running. You will have to look very carefully as the wet spot may be quite small. Other than that, it could be anything. Good Luck!

[Ct. Strangelove]

what type proccessor?

AMD + no water flow + 2 minutes = CRISPY BURNT CHIP
INTEL + no water Flow + 15 minutes = possible death

Is the PC new to you? it might be faulty bios. remove all PCI cards and such. reseat yout ram, VID , CPU. we need more info to really help

*edit

Its a swifttech? don;t they have smaller pumps. I agree it could have been the air causeing no flow.

[aws6000]

It's an P4 2.53. Put together the system about 4 months ago. I don't think the chip is fried from the heat because it was working fine when I powered it down. Unless there was a tremendous amount of heat built up after shutting down, I don't see why it would be the heat. The waterblock is kind of beefy, so I would hope that it would dissipate any heat on shutdown. I think I'm going to order a new board and if that doesn't work, then I'll order a new processor.

BTW... I did completely tear down the system and it's still dead.

What bugs me the most is that I don't know why this has happened and therefore have lost my faith in watercooling.

[TyRex]

Did you try clearing the CMOS???

[aws6000]

Originally posted by TyRex
Did you try clearing the CMOS???

You can only imagine how many times I tried clearing the CMOS. "Ok, this time it's going to work...". :-)

I went ahead and ordered a new motherboard. It's a highly overclockable Albatron that will replace my highly overclockable Abit. I have a feeling that it's the CPU that's toast though. Apart from the Radeon 9700, the CPU is the most expensive part in the system and that's how my luck generally goes. It's only money right?

I think my next WC system will be a more traditional T-pipe installation. I'll reuse the parts I have, but try to eliminate some of the connections and possible leak points.

[Shroomer]

Being a P4, I doubt that it was heat. Although there is a rumor about S.I.D.S.

Sudden Intel Death Syndrome.

EDIT: Just want to make sure you don't think this is a joke. There have been many people reporting this. That they slowly couldn't overclock as far and it eventually just stopped.

[TyRex]

SIDS, that could seriously be it......Thats happened to a couple of my friends. Couple weeks go by with a good solid o/c with temps in the mid 30s.....then all of a sudden they just died. We could never figure out why they died, tried the chip(s) in 4 different boards, still wouldnt work. Even with the fsb jumpers as low as it could go, still no go. I wouldnt be surprised if thats what it was.

[Enyo]

i don't trust intel chips so i would't be supprissed if it was sids but from my experiance (with amd chips) i have almost never ever ever had a faulty cpu but ive gone through about 9 mobo's the only thing you could try (thou it is a little late) is putting your cpu in someone elses computer and vice versa. you may have just bought another perfectly good mobo (albatrons rock! :D) i throughly doubt it was the cpu overheating. unless you got reallly unlucky and as you turned the watercooling off and the pc the heat left in the cpu wasn't disipated and melted the cpu. i doubt it was a water leak either as usualy water evaporates before in comes into contact with your computer, unless there is some huge hole.

it may be worth investigating your warranty. there is a good chance that this problem would have occured anyway.


how old are the components?

and good luck.

[Ct. Strangelove]

wow this is sucking bad. you really should try all you componets i n antother working system. a PC wont start without ram. test each part and verify which works. then it is time to some RMAing. even thought you aready got a mobo comming why not have the spare. keep us posted

[RockerII]

Maybe when you removed the hsf there was less air moving over capacitors/nb and other components and this caused them to everheat. However, once the system was powered down, it couldn't restart because of some short. If that's the case then the new motherboard is the cure.

[Enyo]

theres a good chance its the mobo. every thing else in the system rarely breaks

[bongo]

This happens quite often.

I doubt anything is broken in your system.
First off try and reseting the bios, thru the clear cmos jumper or pressing insert before bios screen appears.

Reseating the processor is the next step.
Then reseat the memory.

If this doesn't solve your problem try swapping hardware to determine the faulty component.

Another posibility is that you might have damaged the mobo around the socket area when mounting.


Did you overclock at all?

BTW:
A little air in your tubing never killed a pump or caused for such high temps to fry a processor. If you see air moving in the tubing when the pump is on, then you have more than enough water circulating to keep the cpu at safe temps.

[Freeloader]

Enyo: i doubt it was a water leak either as usualy water evaporates before in comes into contact with your computer, unless there is some huge hole.

I've had many leaks and that has never been the case. Because the motherboard is on it's side, it has usually been the video card that has taken the most hits. Because the waterspot is small and clear, it's very difficult to see.

aws6000: Wouldn't the leak most likely happen when the system was running and not when it was powered off?

It's quite common for a watercooling system to suck air while running and leak while shutoff. This occurs when the leak is along the return line which is under negative pressure due to the pump suction. Once the system is shut off, the pressure equalizes and the system leaks

[Shroomer]

Hmm.. Bravo! I never thought of that. I'll make sure to double check when I'm leak testing.

[aws6000]

I received my new motherboard today, plugged in the processor, RAM, videocard... powered up to do a quick 10 second test. It starts, the cpu fan spins and everything seems good. I'm happy! For the hell of it I decide to give the old motherboard another shot (now that I know the cpu works). Plug in the cpu and the damn thing starts up, the fan spins the LED display lights up. I'm only assuming that it will boot, but for the life of me I was not able to get cpu fan spinning and the LED display to work until now. This is the strangest thing I've ever seen.

[BaD CrC]

I read that SIDS happened only to OC CPU because of a certain 'electron migration' due to higher voltage (than the nominal one).
Congratulations to AWS6000 :D Now you have to perfectly working Mobos.:D :D Glad you didn't buy another P4!! See? Things are not so bad after all.:rolleyes:

[Shroomer]

These things sure can be unpredictible and finicky.