CPU overheating problem

[ICIT2LOL]

I have just had my Ivy Bridge crash and am trying to reinstall Windows on a new drive but when I start the machine up - it has almost completed the install the BIOS lets go with warning beeps and a message that the CPU is over temp.

In the BIOS it is set to the auto settings for any of the CPU and fans stuff and I am not really overly confident in what I should be setting. The CPU an i5 3570K is hitting 81C whenever I start the machine up and refuses to budge past the BIOS beeps.

It is a reasonably new build about seven months so I don't think it is the thermal and the Intel LC cooler block feels almost cold to the touch as does the rad.

Any ideas appreciated as this is just another set back in a horror two days.

 

[Dude]

I would re do the cooler seating just to be sure.

 

[kbrady1979]

Do you hear the regular "Water Cooler" gurgle when you turn it on? You should be able to hear the pump running if you get close enough to it, or have a stethoscope type device. If it is warming up that much and the pump area is cold, then I'm thinking it isn't working to pull the heat out. I hope I'm wrong though.

 

[ICIT2LOL]

I would re do the cooler seating just to be sure.

Yes I think I might as it came out of the blue and I have been running temps of 45C just lately when I first fitted the LC it was down around the 28C mark.

Also I might use the Noctua NT-H1 this time - Arctic Silver first install. Odd though it only happened intermittently during the crash and reinstall. At the moment Speccy is telling me it is 27C???? and I have been doing some heavy installing etc

 

[kbrady1979]

If the pump/cooler does turn out to be bad, I believe it has a 3 year warranty.....hopefully you won't have to use it.

 

[ICIT2LOL]

Don't know KB I didn't notice any gurgling sounds though before I installed it it did seem to have a fair bit of "slop" in it like it doesn't have to full but I would have thought that it wouldn't have sounded half empty either.

Problem is mate proving it is faulty eh?

 

[ICIT2LOL]

Actually KB I have been thinking about what you said re the LC cooler and it occurred to me how much coolant do these things hold and when they made mine did they put enough in it?

One way I might try to find out is to do what my Dad and I used to do when I was in plumbing, and that is to chill the until the (ie copper) storage tank and run the hand down the tank from the top and the metal warms and when you hit the fluid the level is revealed by the amount of time it takes for your hand to warm that spot - usually not at all. But how the air that I suspect is in the rad got down into the pump is a mystery but you certainly got me thinking about it.

Plus I did notice that the rad when I bought it had a very roughly made core unlike those nice even I ones you see in cars and air cons. The other thing that has always puzzled me is why do they paint the core black?? As a dark colour you would think it would attract heat rather than reflect it - odd don't you think??

 

[ICIT2LOL]

Well I marked this as solved but it has come back to haunt me. My temps flew to 96-101C this afternoon and the machine kept shutting down.

I took out the cooler / pump plate and redone the compound with Noctua NT-H1 and held the pump head down low and ran it briefly to purge any trapped air and then replaced it. Now it was making a purring noise like a cat for a while but then settled to a rhythmical ticking sound for want of a better term. The temp came down to 26C at room air temp 12C on running some scans and mail etc
So that is ok for now but I do think the config of this device is wrong. I mean by the way that the tubes are joined to the tank. Now as you know in a car the tanks are upper and bottom and the hot coolant is "syphoned" into the top tank assisted by the impeller. Now by a natural process of cooling the fluid then sinks to the bottom tank to return to the heat source for it to happen all over again.
Now in my RTS2011LC the tubes are joined into the same tank which doesn't make sense to me - how does it or would it go through the main core?? In the pic I got off a review it shows the tubes at the top which again does not make sense in that any air is liable to be drawn down into the pump as I know that my rad is not that full.
My own set up has the tubes at the bottom - opposite to the pic so my reasoning being that they would always be immersed in coolant. and that the incoming heated fluid will tend to rise to the top and sink therefore cooling it.
Now ideally I would have thought that the output from the cooler pump would go to a top mounted connection on the rad and the return tube at the bottom of the rad much as in a cars system. Or a system whereby it has to flow the length of the core. But looking around I cannot see any systems that will come even close to what I think is ideal.
If I replace this unit because I think it is going south if the sounds are anything to go by what LC system would anyone suggest as a good reliable one please.

Oh and I have another problem in that I wrecked the back fixing plate by foolishly using a screwdriver to do up the top holding plate for the pump and the small "nuts" spin around in the flimsy plastic cross piece. Any suggestion s for getting this piece out now as I can only just remove the cooler pump assembly I cannot remove the holding plate

 

[kbrady1979]

My water cooler makes a small whirring/ticking noise at the pump......you have to get really close to hear it, but it has made that noise since day one. I assume it's the pump inside. I've always had my rad oriented with the lines up top, although I have my fan on the other side of the rad, bolted to the chassis, sucking cold air in and blowing it across the fins.

As for the backplate issue, is there anyway you can take a picture of what happened?

 

[ICIT2LOL]

Yep Kelly I will do tomorrow as away from home at the moment. I thin probably if worst came to the worst I shall have to wreck the existing kit and get a new one.

I am still concerned though re the orientation of the rad I would rather keep the tubes at the bottom cos I think I mentioned before that when I was putting it in it sounded like it was not really that full and would be exposed to any air in the top.

I know the specs tell you the weight of the fluid but not the tank / rad weight all up so there is not way of checking that by weighing the thing. I have a push / pull on mine that is the Intel on the inner and a Coolermaster pulling behind so it isn't for the lack of ventilation.

Yep you are probably right I am reading too much into the sound as you do have to get close to hear it though that purring sound you could hear from ten feet away! I am rather hoping it was trapped air and not an impeller spinner coming adrift.

I should have suspected things were not as should be a while back now cos the temps have been hovering around 45C - not hot really but were higher than when I first built and now I have refitted the system compound.

 

[kbrady1979]

Call up Intel, tell them it's making a hell of a racket and not cooling your system and see if they will replace it. If mine went bad tomorrow, I'd go with the same unit because there isn't a nickels worth of difference in similar water coolers.

 

[ICIT2LOL]

Call up Intel, tell them it's making a hell of a racket and not cooling your system and see if they will replace it. If mine went bad tomorrow, I'd go with the same unit because there isn't a nickels worth of difference in similar water coolers.

Yep I might do just that Kelly the current model being provided by my supplier is not this one but this one Intel BXTS13X Liquid Cooled Thermal Solution I suppose just a remake of this one but is more cost though.

I shall have to mail Intel as I don't know if they have call centres out here and if they do the phone costs are just excessive for anything more than local calls

John

 

[kbrady1979]

If you have the box to the cooler, there should be an email address or toll free number to call about the product.....well, you'd think there would be anyway.

 

[Layback Bear]

I don't have a Intel cooler but a Corsair H50.
My inlet and out let tubes are at the top. I have tried them at the bottom and the temps were the same. Moved it back to the top just to get the tubing out of the way.
Their is no noise coming from the pump and shouldn't be. It runs at 1440 rpm.
Their should be no air in the system. If I remember correctly the system is filled using vacuum method.
If the cooler is mounted correctly to the cpu changing the compound should never make a difference of more than 1 or 2 degrees C. If it does their is a mounting problem.
If it was my computer and I new the monitoring programs was working correctly I would replace the LC cooler.
Noise to me coming for a pump means the impeller is hitting the housing. I would not test with it any longer and risk destroying the cpu.

 

[ICIT2LOL]

Hmm well have got a "local" number in Sydney for Intel if I need it and the rad I distinctly remember shaking it and it sounded like it was only three quarters full hence my mounting the tubes at the bottom. I just wound the CPU cooler plate around to make it neat.

Still will find out tomoor errr this afternoon when I turn it on - graveyard shift and forgot the day

 

[ICIT2LOL]

Well my friends it has gone south yet again. On start up form cold it shot straight to 47C and then on reboot to 71C. I could hear the impeller moving I think and the cooler block was warm to the touch. Another Speccy run crept it up yet further and then it just went haywire giving me that overheat warning again.

I could get into the BIOS once and it confirmed the temp of 74C on reset the thing just shuts down with the one and four beep warnings.

Now the pump is really warm and the tubes stone cold so I am guessing the pump has failed

I am going to contact my supplier first to see what the go is if I have to I shall have to contact Intel or just go a new unit. Probably the latter is going to get me a result faster and the RMA if and when I gate one I can put it in the Sandy Bridge or keep in reserve as I intend a Haswell in the not too distant future.

Now getting it out as I described yesterday is going to be a work of art as those buts at the back - see pic - I am going to have to devise a way of stopping them spinning in the retainers to allow full removal of the lock down plate and fittings on the CPU side of the board.

Been a hell of a two days with this machine with the crash and now this following on it's heels.

 

[kbrady1979]

I really hate to hear that John. Here's what I would try to get it out. If I understand you correctly, when you try to loosen the pump hold down screw, the silver insert is spinning. I would grab a pair of needle-nose pliars, put the tip in the back of the insert and the other tip right on the outside of the plastic retainer. You should be able to get a grip and hold it there. Reach around to the other side and see if you can break the hold down screws loose. If you can, you should just be able to keep the metal inserts from spinning by just applying some pressure against it with your finger while you take the screws out. Let me know how that works.

 

[ICIT2LOL]

I really hate to hear that John. Here's what I would try to get it out. If I understand you correctly, when you try to loosen the pump hold down screw, the silver insert is spinning. I would grab a pair of needle-nose pliars, put the tip in the back of the insert and the other tip right on the outside of the plastic retainer. You should be able to get a grip and hold it there. Reach around to the other side and see if you can break the hold down screws loose. If you can, you should just be able to keep the metal inserts from spinning by just applying some pressure against it with your finger while you take the screws out. Let me know how that works.

Ok Kelly I was thinking along the lines of just jamming a screwdriver tip in there but your method I think might not cause so much "damage" Mind you I could kick myself for using that screwdriver.

In the meantime I might just get another unit from the supplier as the replacement will probably take longer than just buying again and like I said the replacement if they do one (they can be a bit funny sometimes with RMA's) I can use on the new build.
I have a Coolermaster TX3 in one of the XP machines if worst comes to the worst and I need to get it going.

 

[Britton30]

John you could also use some epoxy around the inserts and let it cure before taking the screws out. I'd not try jamming a screwdriver tip in there and risk wrecking the board if it slips.

 

[Layback Bear]

Those screws doing that could of been caused by the screws bottoming out in the nut before tension was put on the cooler mounting plate.

Because of this the cooler was not tight against the cpu but the screws were tight against the nut. (Thread depth of the screws)

This could also cause uneven pressure on the pump mounting causing the noise. (Impeller hitting the housing)