nb overheated

i have a desktop asus essentio cg5290 bp009 with intel i7 on a rampage II gene motherboard. my problem is the computer randomly shuts down with a nb overheated. i put on new thermal paste and had the same result. bought a new heatsink and still the same result. is there anything else i can try?
thank you for your help.
smile
whril

 

What's an nb overheated?

With putting a new fan and paste on the CPU, that is probably no longer overheating.
What graphics card is installed? If it has a fan, is that working properly?
How about the fan in the power supply unit?

 

thanks for answering.

when i restart the computer it says nb overheated. from what i understand that is the northbridge and repasting or changing the heatsink should have stopped it?

the graphics card is a nvidia geforce gtx 260. yes the fan is working.
the power supply fan is working as well. i just started the computer, which has been off for 3 hours, to check these. the nb overheated was still showing and the computer shut down.

 

What CPU heatsink are you using, is it a top>down type like the stock Intel, is it also free from dust and the fan spinning freely?

 

the heatsink is brand new, replaced it when i could not get the other one to eliminate the error. it is a top down intel. fan is good and i made double sure there was no dust on the cpu or the heatsink.

 

Have you gently brushed clean the NB heatsink and checked that it's still making good contact?

Does the problem disappear after you remove the side case of the PC and place a household fan nearby to blow into the case?

 

yes, i have had the old and the new heatsink out many times. i have gently brushed them off and cleaned them completely with 91% isopropyl alcohol before repasting. i have had the cpu out and checked for any lint or other suspicious signs. i have gently dusted everything i can get to in the case. the heatsink seems to be making contact as when i take it off the thermal paste will be spread.

i do have the side off the case and a small household fan turned on high and aimed at the middle of the motherboard. the only difference it seems to make is only a matter of a minute maybe two.

sometimes the computer will be off for hours and when i start it the nb overheated warning is on the black screen. other times it is not. sometimes i can be on my desktop for a minute or two other times it will shut down during the boot. i tried to look in bios to check any settings but it shut down before i even got half way through the first page.

again thank you for your help.

 

How much thermal paste are you using, are both heatsinks making good, even contact?

Are you overclocking anything? If so, please set everything to defaults to troubleshoot.

 

using about the size of a half a bb of thermal paste. truthfully i am not sure about even contact. i do not recall anything that suggested to me that it was uneven.

i have never overclocked on this machine. in the morning i will check tho and make sure that everything is set to default and let you know.

thanks

 

Half a .177 sounds about right, half a .22 might be a little too much. It's used to fill in the imperfections, not to create a pad.

Even tightening of a heatsink is a lot like torquing down a cylinder head, a little at a time, in a kind of cross-rotation, so that each fastener is at the same tension and the mating surfaces are flush.

Concentrate on the NB, not the CPU. The NB might be faulting because of a PSU, motherboard VRM, PCI/e card, or other chip/connection on the 'board faulting and trying to feed, or draw, more power than should be required or is specified for.

 

okay, that makes sense. so what is my next step? i checked. everything is set to default. i loaded defaults just to be sure.

 

With the side off and the fan blowing, can you get the full temperatures and voltages from within the BIOS? We may find some clues from them.

 

took a couple of restarts but i got them.

temperatures (listed in fahrenheit)
cpu 83
mb 84
nb 161
sb 100

voltages
cpu 1.77
cpu pll 1.819
qpi/dram core 1.819
ioh 1.164
ioh pcie 1.508
ich 1.105
ich pcie 1.50
dram bus 1.508
3.3v 3.328
5v 4.974
12v 12.025

fan speed
cpu 1067 rpm
chassis fan2 13672 rpm

 

Are you sure you've checked that the NB cooler is correctly attached - it's between the CPU socket and the top PCIe slot, the heatsink seems to share a copper heatpipe with a backward 'J' shape that connects it to the VRM, between the CPU socket and the I/O shield?

 

I am a little concerned there's some confusion between the CPU heatsink and the northbridge (chipset) heatsink.

As satrow noted, it is not on the CPU.

It should also be noted it is the case's responsibility to provide a sufficient supply of cool air flowing through the case to keep the chipset adequately cooled. What is your case cooling setup? You generally want a good "front-to-back" flow through the case. So typically you want front fans blowing cool air in and rear fans exhausting heated air out.

And of course, you want the case interior to be free of heat-trapping dust.

 

i am a whole lot confused!!

it must be the cpu heatsink that i replaced. not sure where the northbridge heatsink is? is it the large radiator type attachment just below the cpu fan? is that what needs to be worked on?

the case is dust free. for fans there is the fan in the power supply, the exhaust fan in the back, the cpu fan and the video card fan. that is it. it is kept on top of a desk that has nothing else around it.

 

The NB (northbridge) is part of the motherboard's chipset. It typically will have a large heatsink (but not as large as the CPU's) and most today use passive cooling (no fan - just heatsink).

While the PSU fan does help with case cooling, it really should not be counted as part of case cooling. So it appears for case cooling, you only have one fan, the rear fan. You need to inspect your case interior to see if it supports adding more fans, either in back, or in front of the drive bays. If not, look at your rear fan and see if it has 4 more screw holes out from each corner of the existing fan. If so, it will support a larger fan and that should help.

A side panel fan might help too, if your case supports one - although I have seen where they can disrupt air flow and actually make things worse. It is a try and see thing. Some cases also support "blowhole" (top) fans.

 

there is a spot under the front bays to add a fan. question is would adding a fan be the solution or do i need to do something with the nb heatsink as well? i ask because i have had this computer for 6 or 7 years and this is the only time overheating has been an issue. seems that it is logical that another fan would be needed to prevent this from happening again after or during the problem is fixed?

thanks!

 

Contrary to what some believe, thermal interface material (TIM) never needs replacing just because x number of years have passed - AS LONG AS the cured bond between the two mating surfaces remains in tact. It does not wear out or lose effectiveness over time and will easily last 10, 15 years or longer. Even if it dries out, the solid TIM left behind is occupying the space preventing insulating air from getting inside. And keeping insulating air out of the microscopic pits and valleys of the mating surfaces is the sole task of TIM. Well, unless it is "adhesive TIM" as used on some northbridges and other heatsinks that don't use separate mounting/clamping mechanisms. Then it is used to push out trapped air and glue the heatsink to the device.

So if your NB heatsink is not being held in place with some sort of clamp or screws, it is using adhesive TIM and should be left alone anyway.

I don't know if another fan will fix your problem, but it can't hurt. You might try opening the side panel and blasting a desk fan in there to see what happens.

Note in electronics, including computers, temperatures are typically shown in Celsius. 161°F = 71.7°C which is warm but I don't consider hot. This could just be a sensor problem or maybe can be cleared by simply bumping the alarm threshold in your BIOS Setup Menu up a little.

BTW, what is telling you the NB is overheating?

 

i know the nb is overheating because when i start the computer the black screen with the info on it says nb overheated! (explanation mark is on the text not mine). sometimes it will shut down there, sometimes it lets me on my desktop for 1 or 2 minutes before it shuts down. if i go into bios it might or might not shut down.

how do i tell if it using a mounting/clamping mechanism?

i do have a desk fan blasting with the side panel off. it does not help.

when first checking the bios the temp of the nb is around 161f. if it shuts down and i check then it is around 174f to 180f. how hot is too hot?

if blowing the fan in the case isn't helping is that indicative of adding another fan not working as well or is the air flow different enough to help?

how do you tell if it is a sensor problem?

if i bump the alarm up what do you mean by clearing? would that be like resetting and telling the computer to knock it off i'm not hot yet?

sorry for all the questions. trying to wrap my thoughts around all this.

 

Okay, but that is most likely being triggered by a threshold setting in your BIOS. There may be one for your CPU too, but your CPU temp is not reaching it.

I don't like suggesting it be raised because that can be like sticking a penny in fuse box because you have too many Christmas lights on the circuit. But if the temp is not really too high (because maybe the sensor has failed) or the threshold setting some how was changed, bumping it up a little may clear your problems. Note 180°F is 82°, that still does not seem too hot to me. I note GPUs often run considerable warmer than that.

You'll see it. There will be a bracket holding down the heatsink (like the CPU's), or perhaps screws in the heatsink mounting it to the motherboard.

That does not sound good but could also be due to CPU or RAM failure (unless it is telling you then the NP temp is to high). I might suggest you replace the motherboard's CMOS battery. Unplug power from the wall, open the case and touch bare metal of the case interior to discharge any static in your body. Then carefully pry out the battery, noting the polarity. It will look like a coin, about he size of a nickel or quarter. Most likely it is a CR2032 that you can get at most battery/watch/camera counters at your local discount store.

Do not touch the new battery with your bare fingers as skin oils can promote corrosion and attract dust. I put a clean cotton sock over my hand. Don't forget to touch bare metal, then insert the battery, observing the polarity (it can only go in one way, unless you step on it).

Pulling the battery will reset the BIOS. Boot directly into the BIOS Setup Menu and check/set date and time. You can take this opportunity to look at your PC Health temperature setting. When done in the menu Save and Exit to boot normally. This may fix your problem.

You cannot really tell if a sensor is bad, unless it is showing something ridiculous like -30° or 350°. But you can get an idea by using a laser thermometer.

If the desk fan, swapping the battery, checking the temps in the BIOS does not help, then with this being 6 or 7 years old, it may be time to bite the bullet and get a new computer.

 

i did the battery and checked the bios. temp was at 186 and flashing red. there is no setting for me to boost up the alarm only to disable. the alarm is set at 100f.

yes, when the bios shuts down it tells me it is from the nb overheat.

any other options?

thanks

 

Other than verifying the temps with something like the laser thermometer I mentioned above, I don't know what other option you have but to disable the alarm.

In either case, I do recommend you backup your drive(s) immediately and keep a current copy. While rare, a sudden shut down can corrupt a drive and lost data.

 

looks like what you said earlier about buying a new computer might have to be the decision in my near future. if i get a new computer can i just swap it's harddrive with my existing one?

thanks

 

While technically possible it most likely would not be legal.

I am assuming you are asking if you can swap your old Windows installation into a new computer by just swapping drives. Physically yes, legally no - assuming your current version of Windows on this ASUS came preinstalled at the factory. In that case, the Windows license is an "OEM/System Builders" license and they are inextricably tied to the "O"riginal "E"quipment. Under no circumstances can an OEM license legally be transferred to a new computer.

The terms of the OEM license also say you can upgrade anything inside your computer, but the motherboard. A new motherboard constitutes a new computer and therefore requires its own new Windows license.

And note you agreed to abide by those terms when you decided to keep using the computer when you first got it - and that makes it legally binding.

So you will need to buy a new Windows license, or go with one of the many free versions of Linux.

That said, you can install your old drive into the new computer as a secondary drive (like D drive) then you will still have access to all your old data files.

The reality is, with your computer being 6 or 7 years old, you need a modern OS with your modern hardware, and for security purposes anyway.

 

I started looking for nb overheated errors and discovered it is not always due to the Northbridge overheating!

Someone had that error and discovered it was due to a corrupt pagefile.

Others said it was corrupt memory. Have you tested your RAM?

 

no i haven't and i am willing to try anything! i really like my computer and would like very much to keep it. what is the best way to test my memory? should i do that before i try that pagefile?

i think i said it before but i will say it again..... you guys are the best!

thanks

 

Corrupt page file? Wow. I sure don't see how unless it being corrupt causes the system to work harder, which I can see happening. Not sure how RAM could do it either. But certainly worth checking it out.

To check your Page File, just set it to None, reboot, then set it to Automatically manage, and reboot again. I don't see where you ever said what version of Windows - since 6 - 7years old, I am guessing Vista. See http://windows.microsoft.com/en-us/windows/change-virtual-memory-size#1TC=windows-vista

To test your RAM, if you have more than one stick, you can shutdown, unplug from the wall, touch bare metal of the case interior, then pull all but one stick and try running with that. Swap out the RAM to check them all - making sure to follow the same procedure, especially touching bare metal to discharge any static in your body.

You can also run MemTest86 on all your RAM. Let it run for several passes. You should have zero errors. But note that no software based tester is conclusive. If it says the RAM is bad, it is bad. But it can report no errors and the RAM still fails when put in circuit, or when paired with other RAM.

 

since my computer will not stay on for the hours it takes to run the memtest86 will the one ram one slot at a time work?

the os is windows 7 home premium.

thanks

 

one RAM, in one slot. Start the test and if you get any errors, you can stop the test because it indicates a RAM problem. I hope it runs for more than a few minutes before it complains about overheating.
If it complains, see what the stick does in the other slot.

Digerati, computers never cease to amaze me. Every day I read something that does not seem in the least logical. A bad mouse caused someone's computer to not boot. That was one I never would have figured out!
So if something seems unusual and doesn't seem hard to test, I post so the person with the problem computer might want to try something easy.

Years ago I had a computer running Windows 2K. After Windows loaded if I moved the mouse or hit a key on the keyboard, the computer froze and that was the end of things.
Eventually I discovered that in the BIOS either caps lock was on or caps lock was off. I don't remember which. All I did was change a setting and the computer ran fine. I don't even remember how I discovered that the wrong caps lock setting was causing a freeze! And I certainly have no idea how the setting was changed. Years ago I was so afraid to even look in the BIOS yet alone make a change.

 

I agree. They have for me since I first stood (true!) inside one in the early 70s!

@whril - did you try a desk fan?

 

yes, i have had the side cover off and a desk fan blowing on the motherboard ever since the problem first occurred a couple of weeks ago.

so far i haven't been able to run the memtest past the #7. working my way down the slots. when i first put in a ram it says .... overclocking failed hit f2 to use default settings. i am taking it for granted that is because there is only 1 memory stick in there? i also took for granted that yes i want to hit the f2?

 

okay i was able to test the ram. it all came out with no errors. while i was testing it i got tired of the large graphics card blocking the tab releases so hunted up an old graphics card and replaced it. since i replaced it i have not had any nb overheating errors and it has been running for over 2 hours.

i did check the temp in the bios and it really hasn't changed. 86c-188.5f showing in red but it is not shutting down even tho the treshold for shutting down is 100. don't know if that 100 is c of f as it does not say.

this is not making any sense to me!

by the way i really appreciate your responses, advice and ideas!

 

Hmm, the graphics card would pull up to ~75W directly from the motherboard, perhaps it has a fault which causes an overheat situation (NB/VRM); which make/model # are the 2 cards?

 

the card i took out was a nvidia geforce gtx 260. the old card i put in is a nvidia geforce series 6, not sure which model number.

 

A GTX 260 would usually require slot power + 2x 6 pin connections from the PSU to feed ~180W.

Slot + 2x 6 pin = 225W max, slot + 2 working 6 pin = 150W max.

If your current card requires less, it might only use 1x 6 pin, if it needs any at all.

So it could be that the PSU/cable had ceased being able to supply enough juice, forcing the MB to supply more via a stressed VRM - but I'm no expert in electronics.

 

You could certainly replace the graphics card again and see if the problem returns.

Has to be C because 100°F is barely warm and 100°C is the boiling point of water, or 212°F, which you never came near so I am inclined to believe that satrow is on to something and perhaps the old card was drawing too much current through the motherboard, or perhaps that PSU rail.

According to these specs, the GTX 260 is a hungry card pulling 182W max and thus requires 2 x 6pin power connections directly from a recommended 500W minimum supply. Did you have both power connections and is your PSU large enough?

 

yes, both pins were connected and the psu is a 500w. neither has ever been upgraded, came with the computer.

right now the nb temp is 25c. to see if the bios is reporting a correct temp i downloaded hwmonitor and it says the same.

i don't think i want to put back in the other graphics card again. it is huge and covers the lower half of the motherboard. i have never liked the way it sat, often pulling down on the slot.

i will not be able to run the computer for any length of time as i am leaving on a road trip to my daughters wedding tonight. will run it when i get back and see if it still overheats.

thank you all so much for your help!

 

Pulling on the slot does not sound good. I would not re-install it either.

Congrats to the daughter and watch out for idiots on the road.

 

thank you... the wedding was great. the 2,748 miles of driving wasn't too bad either.

came home and turned on the computer. has been on for 9 hours, the last 2 which has been hard use. the nb temp is holding steady at about 30c. no shutdowns with overheating so i would say that it is safe to assume that the graphics card was the culprit?

thank you all so much for your help, advice, and ideas. major geeks is a great bunch of people!

thank you!
whril

 

The change of graphics card might have dropped the load on the PSU, at this point, it could be either the PSU or the old gfx card. I'm still suspicious of the PSU.

 

good suspicion. if i encounter any more trouble i will switch out the psu.

thanks again!

 

Glad all went well, including the wedding.

 

okay things have been working fine til early this evening when the nb overheat started again. after the computer shut down i reached inside the case to see if anything felt really hot. when doing so i noticed that the silver finned box is separated from the mb. i noticed that there is 2 places that look like screws attaching it to the mb. 1 of which looks like a screw of some sort the other is just a hole. the hole and where it is loose is the bottom left hand corner. am i right in assuming this should be screwed down tight? is this the nb heatsink? is this something i can replace either the screw, if it is supposed to have one, or the heatsink, if that is what it is? i uploaded a pix of it.

thanks
smile
whril

 

Yes, that's the northbridge heatsink and it should be making good contact with the chip underneath it. Can you access the rear of the 'board to check what fixing the remaining screw uses?

 

the only way to access the rear of the mb is to take it out. only thing shown from the back is the screw attaching the mb to the case.
i can see the front of the screw well if that helps.looks more like a push pin than a screw.

thanks
whril

 

Yeah, plastic push pins are often used to retain those heatsinks.

Look in the bottom of your case. It is not likely that push pin was never there. Hopefully you can find it and re-insert it and the retaining end has not broken.

But since this heatsink has come loose, the TIM's bond has broken so you will need to remove the heatsink completely, clean the mating surfaces and reapply a fresh new layer of TIM.

I use and recommend cleaning with 91% isopropyl alcohol, available at your local drug store. I use a cotton swap. Just be sure to unplug from the wall and touch bare metal of the case interior before reaching in to discharge any static in your body.

Then apply as thin as possible layer of new TIM across the chip. I snip off the end of a plastic shaft cotton swab and bend the end to form a hockey stick and then spread the TIM like icing a cake.

If you cannot find or obtain a new push pin, you can use thermal adhesive.

 

there is no sign of the push pin anywhere. i also doubt that it is unlikely that it lacked a push pin to begin with. until i think. how would a push pin fall out? where would it go. it is not as if the computer was moved. anyway there is no push pin.

am i understanding correctly that the only way to get this push pin out is to remove the motherboard?

when i take the heatsink off... clean....repaste.... i can then use the thermal adhesive to replace the heatsink instead of the push pins? since i have 1 push pin should i use it and then adhesive for the missing one?

thanks!
whril

 

If not secured properly to begin with, it could have popped out long ago. Of if it had a defect, it could have just broke recently and fell out - perhaps when you cleaned your system or upgraded it and you just did not notice. Maybe the broken piece is just stuck somewhere and you cannot see it.

It was a factory assembled motherboard and factory assembled computer transported to your house with several stops along the way.

Most likely you have to pinch the end so you can pull it back through the hole, so yeah, you have to remove the board. Or you might be able just break the other pin and use adhesive TIM. Being plastic, the part left under the board won't hurt anything.

Yes and no. Yes, you can use adhesive TIM instead of regular non-adhesive TIM. But no, you cannot use just one push pin because that will apply uneven pressure. You either have to use two working push-pins or no push-pins.

 

thanks for the info. i will let you know how it goes!

smile
whril