Windows 7 Motherboard Bench Testing

seekermeister

Honorable Member
Due to the fact that the motherboard failed on my HTPC, I decided to dig through the closet to see if any of my old motherboards would work. Out of many, I found one that did...at least partially. I understand when a motherboard either lights up the monitor or doesn't, but this one is different. When I power it on, it does light the monitor in a normal fashion, but only for a few seconds. Naturally it complains about a bad CMOS checksum, but before I can change anything in the BIOS, it turns off, and can't be powered back on until I remove power to the motherboard for a couple of seconds.

Unless someone can offer a suggestion, I will probably just order another motherboard, but it would be nice not to have to make the expenditure, particularly since this system is socket 939, which has become obsolete. Which means that I would have to buy a used motherboard, the idea of which I'm not fond of.

EDIT: In case it makes any difference, the MB that I'm testing is a MSI K8T Neo 2 (MS-6702 version 1), with a fresh CMOS battery.
 
Last edited:
Pull the cmos battery out for 2 minutes and the re-install it and boot up and see what happens.
 
Did you try another monitor? It sounds to me like your monitor and your graphics solution are failing to sync up. So I would try another monitor - and if no good, I would swap out the graphics cards, or add a spare graphics card from your closet and see what happens.
 
bassfisher6522,

That got exactly the same results as before. I get the impression that heat is somehow involved because of the time factor, but it makes no sense to me why it would require powering off the system, unless it is the power supply itself that is cutting off. I don't have another PSU to test with, except the one already in the HTPC, but I was hoping I wouldn't have to actually install the MB, unless I was sure it is okay. I also swapped out video cards, just in case the problem was there, but it wasn't.

I have not actually installed the CPU fan, I just have it sitting on top of the CPU. That wouldn't overheat in a matter of just a few seconds without it being driven by the OS...would it?
 
Digerati,

Monitor??? I never considered that. the only other monitors that I have are not easy to get to, but I guess I will have to try. I haven't seen you around on the boards lately. Have you been on vacation?
 
I have not actually installed the CPU fan, I just have it sitting on top of the CPU. That wouldn't overheat in a matter of just a few seconds without it being driven by the OS...would it?
Sure it can - in just a matter of a second or two - remember 3GHz = 3 billion cycles per second! That is a lot of transistor flip-flopping in a very short time.

You need to ensure the mating surfaces are clean, then apply a fresh, proper layer of TIM (thermal interface material) then properly mount and clamp the HSF assembly to ensure (1) uniform pressure and (2) all the heat trapping air has been squished out.

Monitor??? I never considered that.
I actually would suspect the graphics solution first, but swapping monitors is usually easier. Also, I have had several monitors that refused to sync to standard VGA resolutions (which the BIOS menu uses) when connected via digital (DVI or HDMI) connection. I still use an older Viewsonic on one of my systems that behaves that way. Works great once booted into Windows, but if I want to see the BIOS Setup Menu, I have to reconnect using analog.

At any rate, I would sure make sure it was not the monitor or graphics solution before buying a new motherboard - or before buying anything, for that matter.

I haven't seen you around on the boards lately. Have you been on vacation?
Still active as ever - but I spread my time around many sites - meaning less time at any one particular site.
 
Sure it can - in just a matter of a second or two - remember 3GHz = 3 billion cycles per second! That is a lot of transistor flip-flopping in a very short time.

You need to ensure the mating surfaces are clean, then apply a fresh, proper layer of TIM (thermal interface material) then properly mount and clamp the HSF assembly to ensure (1) uniform pressure and (2) all the heat trapping air has been squished out.
Okay, I will paste and mount it. I knew that it could overheat that quickly in general use, but not just when displaying the intial BIOS screens. However, even if it did overheat, would it trip some kind of protection in the motherboard or power supply that would require that the power supply be cut-off for it to work again?
I actually would suspect the graphics solution first, but swapping monitors is usually easier. Also, I have had several monitors that refused to sync to standard VGA resolutions (which the BIOS menu uses) when connected via digital (DVI or HDMI) connection. I still use an older Viewsonic on one of my systems that behaves that way. Works great once booted into Windows, but if I want to see the BIOS Setup Menu, I have to reconnec

At any rate, I would sure make sure it was not the monitor or graphics solution before buying a new motherboard - or before buying anything, for that matter.
As mentioned earlier, I've already swapped out the video card, so that isn't the problem. However, I tend to doubt that the monitor is the problem, because it has it own power source, which is not interrupted and wouldn't be effected by powering the motherboard off and on.
 
You were very lucky that you didn't fry/burn up your CPU. I never, ever run a system w/out installing a CPU cooler of some sort.....stock or aftermarket....just to damn dangerous.

Nice pick up Dig.....
 
Yup, that was the problem. I pasted and mounted the processor, and it now keeps running properly.
Good. :)

However, even if it did overheat, would it trip some kind of protection in the motherboard or power supply that would require that the power supply be cut-off for it to work again?
The chipset could certainly force a shutdown, or the CPU itself could go into self-protect mode. But the PSU would keep on chugging until signaled to shutdown, or (for better supplies, anyway) the PSU protection circuits detected an overcurrent condition.

For sure, a CPU "should" never be damaged by heat because the motherboard or CPU protection modes "should" kick in and force a shutdown before damage occurs. But "should" (ie "theory") and real world don't always see eye-to-eye.

Also, I have seen motherboards where the sockets were damaged (distorted/warped) by heat by a CPU that ran near thresholds for extended periods of time. The CPU survived but not the socket. This is why I don't like my CPU temps to sit above 60°C - that's my threshold to take action - usually it means I need to clean dust filters or :( tear down my systems and lug them outside to blast out with the air compressor.
 
Nice pick up Dig.....
Thanks. This is something I learned (the hard, and expensive way) a long long time ago as a radio technician in the Air Force. The importance of a properly applied layer of TIM cannot be overemphasized - but sadly so often is overlooked, even by experienced computer users and assemblers who have not had formal electronics repair training.

And like I said, a "properly applied layer" of TIM is crucial. The most efficient transfer of heat occurs with direct metal-to-metal contact between the mating surfaces of the CPU die and the heatsink. The primary purpose of TIM is to push out the insulating air trapped in the microscopic pits and valleys of the mating surfaces. Too much TIM gets in the way and is actually counterproductive to the heat transfer process - even with the best quality TIM. This is why you want the thinnest layer possible, while still providing total coverage. And this is why thoroughly cleaning the mating surfaces first is absolutely crucial - a tiny spec of dust (or old TIM) may as well be a boulder sitting in the way, allowing all sorts of bad, insulating air to be trapped in between.

While on my soapbox - I will also note that TIM does NOT wear out, go old, or go bad UNLESS the cured bond between the mating surfaces has been broken. That is, if you leave it alone, stop twisting the HSF to see if still tight, don't bounce the computer off the floor, the TIM will remain viable and effective long after the computer becomes obsolete - easily for 10, 15 years or longer.

So as a formally trained technician, I get upset when folks are (almost routinely) told to pull the HSF to replace the TIM to fix heat issues. That action is almost never needed (again - unless the cured bond has been broken) and it exposes the CPU and motherboard components needlessly to potential ESD destruction, improper cleaning of old TIM, misapplication of new TIM, and improper clamping of the HSF mounting mechanism.

Finally, while you can gain a few degrees by replacing the thermal OEM pads with a top quality aftermarket TIM, if you actually need those few degrees to keep temps below 60°C, you have other thermal issues that need to be addressed.

We (as helpers and as computer users/enthusiasts) must remember it is the case's responsibility to provide an adequate "flow" of cool air through the case. And it is responsibility of the CPU's HSF assembly to simply toss the CPU's heat up into that flow - something, I might add, the OEM coolers supplied by Intel and AMD with their CPUs are perfectly capable of doing. In other words - no aftermarket coolers are needed - even with mild to moderate overclocking!!!

[soapbox mode off]
 
no aftermarket coolers are needed

I can agree with everything that you said with the exception of this statement. Yes, the original HSFs are adequate in most cases, but the only way that I could get my desktop to run at a temperature that I felt comfortable with was with a Noctua NH-D14. The temperature difference I got with it over the stock cooler was dramatic. The stock cooler kept the CPU at a temperature too near to TJMax when I encoded DVDs. In any case, I'm a bit surprised that a person who emphasizes the bane of heat in his signature would discourage anyone from wanting better processor cooling.
 
In any case, I'm a bit surprised that a person who emphasizes the bane of heat in his signature would discourage anyone from wanting better processor cooling.
Well, it is because I understand heat and electronics, and I understand the consequence for using aftermarket coolers.

Here are the facts:
1. Lower CPU temps does not automatically mean better. That is, a CPU running at 40°C is NOT going to run better or have a longer life than a CPU running at 50°C. In fact, engineers design electronics to run within a "normal operating range".

2. Motherboard designers purposely surround the CPU socket with heat sensitive and heat generating devices with the expected OEM cooler in mind so they too can take advantage of the air turbulence created by the OEM cooler - a "downward firing" cooler that disburses the air flow to those devices.

3. Aftermarket coolers like that Noctua fire sideways, in one direction only. This means some of the surrounding components will never benefit from the side firing CPU fan. After aftermarket liquid alternatives provide zero air to those surrounding devices.

4. Aftermarket coolers void the CPU warranty! You may not be aware, and for many enthusiasts it is not a concern, but it is important all readers understand that both Intel and AMD warranty their “retail” boxed versions of their CPUs that come with supplied heatsink fan assemblies as “a unit". That is, the 3-year warranty does not cover repairs if the "unit" is split and the CPU is used with an aftermarket cooling solution.

Note I am just the messenger ensuring users have all the necessary information needed to make the best, informed decision for their situation. Read the printed warranties that come with your CPUs - it's in there, or read it on-line here,

AMD is quite clear:

"This Limited Warranty shall be null and void if the AMD microprocessor which is the subject of this Limited Warranty is used with any heatsink/fan other than the one provided herewith."

5. OEM coolers are the ONLY coolers with warranties that cover the CPU too. Some aftermarket coolers, like your Noctua have excellent warranties - but many have only 1 year (or less!) warranties. And again, they don't cover replacement of the CPU in the [admittedly rare] event a failed cooler takes out the CPU.

6. It is the case's responsibility to provide an adequate supply of cool air through the case. If you could not get your CPU temps to sit safely within the "normal operating range", then your case's cooling capability is inadequate, poorly designed, or you selected the wrong case for the hardware you put inside (assuming no extreme overclocking).​
 
Thanks. This is something I learned (the hard, and expensive way) a long long time ago as a radio technician in the Air Force. The importance of a properly applied layer of TIM cannot be overemphasized - but sadly so often is overlooked, even by experienced computer users and assemblers who have not had formal electronics repair training.

And like I said, a "properly applied layer" of TIM is crucial. The most efficient transfer of heat occurs with direct metal-to-metal contact between the mating surfaces of the CPU die and the heatsink. The primary purpose of TIM is to push out the insulating air trapped in the microscopic pits and valleys of the mating surfaces. Too much TIM gets in the way and is actually counterproductive to the heat transfer process - even with the best quality TIM. This is why you want the thinnest layer possible, while still providing total coverage. And this is why thoroughly cleaning the mating surfaces first is absolutely crucial - a tiny spec of dust (or old TIM) may as well be a boulder sitting in the way, allowing all sorts of bad, insulating air to be trapped in between.

While on my soapbox - I will also note that TIM does NOT wear out, go old, or go bad UNLESS the cured bond between the mating surfaces has been broken. That is, if you leave it alone, stop twisting the HSF to see if still tight, don't bounce the computer off the floor, the TIM will remain viable and effective long after the computer becomes obsolete - easily for 10, 15 years or longer.

So as a formally trained technician, I get upset when folks are (almost routinely) told to pull the HSF to replace the TIM to fix heat issues. That action is almost never needed (again - unless the cured bond has been broken) and it exposes the CPU and motherboard components needlessly to potential ESD destruction, improper cleaning of old TIM, misapplication of new TIM, and improper clamping of the HSF mounting mechanism.

Finally, while you can gain a few degrees by replacing the thermal OEM pads with a top quality aftermarket TIM, if you actually need those few degrees to keep temps below 60°C, you have other thermal issues that need to be addressed.

We (as helpers and as computer users/enthusiasts) must remember it is the case's responsibility to provide an adequate "flow" of cool air through the case. And it is responsibility of the CPU's HSF assembly to simply toss the CPU's heat up into that flow - something, I might add, the OEM coolers supplied by Intel and AMD with their CPUs are perfectly capable of doing. In other words - no aftermarket coolers are needed - even with mild to moderate overclocking!!!

[soapbox mode off]
hey I 100% agree with the thermal paste never going bad unless the bond is separated. I am a ham radio operator and have used all different kinds of paste including the white transistor paste and have never had to reapply any of them. I still have old tubes of arctic silver 2 and 3 and it is still good. I also still and always have used stock hsf coolers and have had no problems with them cooling. yes I have replaced a few of the fans due to bad bearings. good airflow is the key along with big cases. just my experience. poorguy
 
different kinds of paste including the white transistor paste
Which we buy by the tub. Works fine, lasts a long time. Of course we use more advanced TIM during final assemblies but for testing and troubleshooting, plain ol' silicon paste is great.
 
I have always had good luck with the white transistor paste but bought some arctic silver 2 and 3 and use it when I remember. I have always used stock hsf without any cooling problems and always use big roomy cases with fans in the front and back and at the top. I am a ham radio operator and understand about cooling of electronics. I have tried a zalmon cooper radial hsf on a Pentium d and it did lower my temps but not by much for the cost of it as we know Pentium d are flame throwers. it is good to know that some others understand the way we do. don't need to overclock with the processors that are made today. still using core 2 duo myself. haven't felt the need for the icore yet too much $$$ for my budget. enjoy. poorguy
 
I am a ham radio operator and understand about cooling of electronics.
I started out in the early 70s as a radio technician for the Air Force and over the years, I maintained Collins KWM-2s and KWM-2as and 30L-1 linears at several MARS stations - which I am sure makes sense to you. ;)

I have tried a zalmon cooper radial hsf on a Pentium d and it did lower my temps but not by much
That's a point I would like to emphasis. The same holds true with aftermarket TIM (thermal interface materials).

Certainly, cooler is better, but only to a point. That is, there is nothing to suggest a CPU running at 40°C will perform better, be more stable, or last longer than a CPU running at 50°C. Why? Because 50°C is still well within the CPU's "normal operating range".

In most cases, the only thing you get out of replacing the OEM TIM pad with an advanced TIM and replacing the OEM cooler with an aftermarket cooler is bragging rights. You don't get a better performing, longer lasting computer.

There are exceptions of course, but I contend if the OEM TIM and cooler are not providing adequate cooling, you should be looking at case cooling first - to see if you can add an additional fan, or larger fans to increase the "flow" of cool air through the case.

Of course whenever you remove a heatsink, you MUST thoroughly clean the mating surfaces of old TIM and apply a fresh new layer of TIM - no harm in using advanced TIM then.
 
hey all, I agree. I have had Pentium d operate well beyond the 60 degree c mark and yes that is a concern but as mentioned earlier these are made to take the heat and seem to have the head room for that. yes in some cases cooler is the preferred but not always. some components due run at certain temps just fine. I am still running stock hsf and have no cooling issues so I will brag on that. you can only cool things as to what your ambient room temp is and that is generally as good as it will be. my experience. never noticed any great temp differences between any of the TIM from one to another, maybe 1 or 2 degrees. electronics components will run at what then are designed to and at 100% load continuously as long as good air flow is provided. remember how much heat vacuum tubes produced and all of the venting that cases had they still ran fine and lasted a long time. ah the good old days. I miss them days when you could troubleshoot down at component level instead of replacing boards or disposing of. about the most I do is recap motherboards anymore. just getting old, sounds like we started around the same time. have a good day. poorguy
 
you can only cool things as to what your ambient room temp is and that is generally as good as it will be
I find it amazing how many folks don't understand that.

You can, however, with some alternative cooling techniques, cool cooler than ambient temps, but those are expensive and not needed except by enthusiasts.
 
Last edited:
Back
Top