Hi all,

I am running an Intel Atom motherboard E74516-001. It is one of the early 64 bit dual-core ATOMS with low power consumption, and no fan required - just has a large heatsink.

Today, after a couple of years of running fine, it just decided it is not going to boot. I was able to get into BIOS a couple of times and the hardware monitor says it is running at about 33 degrees Celcius, which to me does not seem to be hot enough to be triggering a shutdown, but when I try to boot Windows it always shuts down within a few seconds with an error saying that the thermal protection has been tripped. Now it won't even stay on long enough to get into BIOS before shutting down.

Is there any hope for this motherboard, or is it to the recycle bin for this one?

Thanks!No. Rather that junk it - ruin it!
I mean that if you are going to toss it, take a chance and try to  fix it. Locate the thermal shutdown sensor and replace it with something.

But if the sensor is in the CPU, you can't fix it.

What did you use for thermal compound? If you used toothpaste,iIt has to be replaced every three years.Toothpaste...here we go again...

You may want to try replacing the CMOS battery for about 4 Bucks...
I've seen them give the Temp sensors fits when they are weakining.Thanks for the tip Patio (long time no see - busy elsewhere...)

Sound like worth trying. I'll give it a try.

I wondered about corrupted BIOS too and was thinking a BIOS UPDATE might help, if I can keep it running long enough.

Sometimes it will stay running in BIOS even though booting Windows kills it. I've been able to keep it running on Hardware Monitor for 24 hours now (steady temps - 35C CPU/45C Remote), hoping there was just some corrosion or bad connection that might benefit from being burned in again.

Tomorrow I'll come back to it and try a new battery if it still won't stay up.Good to see you as well...hope all is well.

Flashing the BIOS IMHO should be a last resort.I got the new CMOS battery, but I am still getting the thermal shutdown and the associated message from BIOS telling me that is what happened on reboot.

This really has me stumped. The logical thing would be that a temperature sensor has failed and is telling BIOS the wrong temperature. But in BIOS, the Hardware Monitor SHOWS healthy temperatures, which should mean that there is not a bad sensor.

The other thing would be the wrong shutdown temp in BIOS. I have other motherboards with a better BIOS that allow you to choose the shutdown temp, but this little Intel cheapie doesn't have any option like that.

Hardware monitor shows pretty steady temps of 32 to 35 Celsius for the CPU and 42-45 C for the remote sensor.

The computer sometimes will run BIOS before shutting down, but if it does, it always shuts down when it starts Windows. Today I started a completely stone cold computer and actually got to the desktop before it shut down, but once the computer has been running, it shuts down within a second or two of the Starting Windows message, and the longer it runs, the less willing it is to even keep BIOS running.

The large factory heatsink doesn't seem loose at all. I talked to a local computer shop, and they didn't know anything about Intel Atom motherboards and were insisting that I install a CPU fan. Not much help since the CPU is showing cool temps anyway.

I'm stumped!At this point your need to prove it is a real hardware issue. When windows is in idel mode, the temps are close to what you would have running just DOS or any system with light load.

It would seem there is something in your windows start up that is over loading the CPU. But that is highly unlikely.

Try this.
1. Boot up just in Safe mode and seem what happens.
2. Or, boot in command prompt and wait.
3. Boot from a CD or DVD with some kind of diagnostic program. Maybe a memory test.

The object here is to seem if the Windows GUI is a factor. Heavy video work can run up the temps. But that could only cause overheating if the video driers were really messed up.

When  was the last virus scan?
I would suspect it to be thermal compound broken down ( dry ) on either between Heatsink and CPU or Bridge controller. If you have a hair dryer that has a cool setting you can see if you can blow cool air across the heatsinks to keep it running longer to troubleshoot -or- if you have a can of canned air you can always flip the can over inverted and spray liquid canned air against a heatsink super chilling it and see if one of the heatsinks either CPU or Bridge cause it to run longer when chilled. You then could determine location of the hot spot.

Working for Allen Bradley/Rockwell Automation in the 90s early 2000's, as an ET, we use to go thru lots and lots of canned air flipping cans inverted and super chilling components that were suspected of being a trouble spot under thermal conditions. We also had one of those IR thermometers, but that wont show you if you have an issue with heat exchanging between components burried under heatsinks. Chilling was the method that proved well for us there especially if you wanted to localize the issue to a component or smaller area of board when O-scopes and trouble codes weren't good enough to pinpoint issue.

Canned air works way better for localizing the problem vs hairdryer which would just blow cool air across a wider area. This will let you know if its truely an overheat issue!

* Be careful not to spray your fingers with the canned air when inverted it can injure you with frostbite and  cold burns to skin etc, where your skin turns white and becomes chalky ( Did that once to fingertip by accident ). ITS COLD!

Excellent Tip...nice post David...patio you are too funny ......toothpaste that's a good one hadn't heard that in a while.........gjmayWell, the strangest thing just happened. Anyone who can give a scientific explanation must be a genius.

As I was contemplating whether or not I could get the heat sink off without ruining the motherboard (it's rather large on an Atom), my other computer wouldn't start up too, as in nothing happened when pushing the power button. I imagined a dead button and a big headache changing it, but before I tore it down, I thought I'd look for simpler solutions. I swapped the power cord between it and the computer discussed in this thread, and the computer in this thread started up and has been running about 6 hours now, including some document scanning and printing I needed to do. Right now the hard drive is blazing like crazy, probably installing Windows updates it's behind on.

Could it be that a flakey A/C cord supplied unreliable power that confused the temperature sensor? Or the BIOS? Or is there another unrelated reason it started up today? Who knows!Pretty bizarre i'd say...but a flakey power source can cause all sorts of strange RESULTS so i imagine it is possible...

How are the Temps looking BTW ? ?Here they are. I've never posted attachments before, so I don't know if they will show up online or have to be downloaded.

I am curious why it shows 4 cores. I thought the Atom D510 was dual core, as Wikipedia also states. Also I thought it was an upgrade to the Celeron I had in there before, but the Windows Experience Index (3.0) is lower than the Celeron.

[year+ old attachment deleted by admin] Quote from: jkolak on July 28, 2012, 02:41:17 AM

Could it be that a flakey A/C cord supplied unreliable power that confused the temperature sensor? Or the BIOS? Or is there another unrelated reason it started up today? Who knows!