Whats the problem here?

[MSgherzi]

Last week, a customer gave me their Desktop to have it cleaned out because it ran too slow. They also complained of it intermittently freezing and just shutting off. The shutting off part is the problem. I can't do much because it shuts off when I start doing too much.

I've tested the memory and PSU, both check out fine. I disconnected both PCI cards (wireless TrendNet card and a card that has two additional IDE connectors for the 2nd HDD they have), replaced it with a working PSU, disconnected everything except the CPU, HDD, and Mobo, and even swapped out the memory for working RAM and tried a different DIMM, but still, it shuts off. I thought I had it working when I swapped the PSU because it was on for much longer before it shut itself off.

However, now I'm getting a new problem. Sometimes, instead of shutting off, it'll freeze up, the screen goes black, and then it says "Input Not Supported" until I manually shut it off and reboot.

I also noticed that this 3.4 GHz P4 CPU was getting up to about 180-190 degrees Fahrenheit when I ran SpeedFan in UBCD4Win. Too hot?

My assumption is that the mobo is gone. To my knowledge, I've tested EVERYTHING else. There's 2 additional fans that have been placed in the machine to cool it. So either the cooling is insufficient or the mobo is dead.

What are your thoughts?

 

[ATTech]

Sounds like a heat issue, first from the CPU, and now possibly from the GPU. Have you checked the BIOS logs, yet? Some systems will log overheating in the BIOS.

P.S 180 Degrees is WAAAAY to hot for a CPU. You can also try down clocking the CPU and test to see if it stops overheating, again, if the system supports it.

 

[MSgherzi]

Sounds like a heat issue, first from the CPU, and now possibly from the GPU. Have you checked the BIOS logs, yet? Some systems will log overheating in the BIOS.

P.S 180 Degrees is WAAAAY to hot for a CPU. You can also try down clocking the CPU and test to see if it stops overheating, again, if the system supports it.

I don't see any way to down-clock it. And yes, I thought that was too hot as well.

BIOS doesn't show any logs at all, actually. And the temperature in BIOS states that the CPU is a steady 150 degrees fahrenheit and the system is about 115.

There's an additional fan that was placed on the side cover of the case, and another one below the main HDD. But that doesn't seem like enough, does it? I'll try and post some pics soon so you can better see what I mean.

EDIT: I should also mention that I opened this machine up and cleaned out every nook and cranny that was dirty with dust/dirt. I also replaced the thermal compound on the CPU as well.

 

[ATTech]

150 is very hot for just simply idling in BIOS. It's no surprise that after a short time when the processor actually starts to get work, that it is overheating.

 

[shamrin]

That "video not supported" bit makes me wonder if it's the video card. Since you didn't mention changing it, I assume it's embedded, but if you've got a slot to put one in, you might try it.

And hey, c'mon guys, it's customary even here in the US to use Celsius when quoting component temps.

 

[MSgherzi]

Thanks for confirming that for me.

I actually have not really had to fix this sort of problem before. Am I just looking at getting more powerful fans here?

Here's some pictures of the system:

This is a picture of the original PSU. It has a monster output capacity of 550W.


Thanks.

 

[jscx93]

Fans aren't your problem honestly as long as that heat sink fan is on, it is cooling. That isn't like it is trying to run a full blown server or gaming machine - that is just a basic MSI board, nothing fancy. There is also no signs of 2nd hand smoke clogging the heat sink

I would just reset the BIOS to defaults and monitor the heat

If it is still cpu related - check for heat sink either broken or becoming detached from cpu

If gpu related - the board may be going

Bios update would be the only suggestion after trying all of that if you have no success and after that, I would question the temp of the sensor and be careful but you can always physically touch it to see yourself. Ive had HP machines that report ridiculous numbers but were cold to the touch - bad sensor

 

[MSgherzi]

Fans aren't your problem honestly as long as that heat sink fan is on, it is cooling. That isn't like it is trying to run a full blown server or gaming machine - that is just a basic MSI board, nothing fancy. There is also no signs of 2nd hand smoke clogging the heat sink

I would just reset the BIOS to defaults and monitor the heat

If it is still cpu related - check for heat sink either broken or becoming detached from cpu

If gpu related - the board may be going

Bios update would be the only suggestion after trying all of that if you have no success and after that, I would question the temp of the sensor and be careful but you can always physically touch it to see yourself. Ive had HP machines that report ridiculous numbers but were cold to the touch - bad sensor

I don't think the temps are wrong, because these symptoms are results of overheating. And I cleaned out the entire machine of every speck of dust just to be sure.

 

[jscx93]

I don't think the temps are wrong, because these symptoms are results of overheating. And I cleaned out the entire machine of every speck of dust just to be sure.

What I mean is, that computer doesn't have anything crazy from the pictures as far as being dirty. As long as that heat sink fan is running, it should be cooling. If not, check to see if it is loose or if one of the legs is broken. There shouldn't be any reason (unless the thermal paste is not there) why it is running that hot with that current setup (unless the heat sink is loose or not seated or something)

An older hp series is notorious for doing so, it looses its ability to monitor the fan speed and shuts off the computer as a result even though the fans are on and working just fine, it thinks that they are not and its technically a board failure, however you can turn the notification off in the BIOS

 

[MSgherzi]

What I mean is, that computer doesn't have anything crazy from the pictures as far as being dirty. As long as that heat sink fan is running, it should be cooling. If not, check to see if it is loose or if one of the legs is broken. There shouldn't be any reason (unless the thermal paste is not there) why it is running that hot with that current setup (unless the heat sink is loose or not seated or something)

An older hp series is notorious for doing so, it looses its ability to monitor the fan speed and shuts off the computer as a result even though the fans are on and working just fine, it thinks that they are not and its technically a board failure, however you can turn the notification off in the BIOS

I see what you're saying.

Well, this machine has taken a turn for the worse. I went to go turn it back on, and nothing. Fan spins for about half a second and shuts off. Another PSU does the same thing. My gut feeling still tells me that this whole problem is with the motherboard, IMO.

 

[jscx93]

If it is truly bare bones and does that with multiple power supplies, then most likely it is dead.

 

[Gmckinnell]

heatsinks

Check the heatsinks? Make sure they are connected and securely fitting.
You may need to remove them, re-apply thermal paste and re-attach.

 

[MSgherzi]

What probably happened, IMO, is that the sensors were gone, as you mentioned, but that was probably one symptom of an overall failing motherboard to begin with. I just feel bad for taking an entire week to figure that out and now I have to order a new motherboard and make the customer wait even longer. I'll probly upgrade the shipping at my expense.

Anyone else is welcome to chime in. But thanks to those who did!

 

[MSgherzi]

Okay, I've finally figured something out. And I need some more input.

It's the CPU. Everything works fine and turns on as long as the CPU is not plugged in. I actually plugged it in while it was on, and the machine suddenly shut off.

I pulled the heatsink off and re-applied thermal compound yet again and no luck. The heatsink locks are kinda loose, but I tightened them.

Does this sound like it's the CPU or motherboard? I can't seem to narrow it down to one or the other.

This machine is also like a ticking time-bomb with one thing after another breaking.

I need some advice from someone more knowledgeable on how I can narrow it down to one or the other without having to put the CPU in another machine, because I don't even know which machine has the proper socket for this type of processor.

Thanks in advance.

 

[PcTek9]

I think it's the cpu. the overheating... but... pluging a cpu into a running motherboard is just not what you should be doing, especially a damaged cpu. some cpu's are 90-110 watts or more... you could quite easily damage a motherboard by plugging in a cpu while the motherboard is 'on' LOL wow. Maybe now it is BOTH ???

 

[MSgherzi]

I think it's the cpu. the overheating... but... pluging a cpu into a running motherboard is just not what you should be doing, especially a damaged cpu. some cpu's are 90-110 watts or more... you could quite easily damage a motherboard by plugging in a cpu while the motherboard is 'on' LOL wow. Maybe now it is BOTH ???

How can you damage a mobo by plugging a CPU in after it has been turned on? Especially while it's bare bones.

Besides, don't you realize that if you were to find another mobo laying around that had the same socket that you would disconnect practically all other peripherals first? Especially if you had a 550w PSU like I already mentioned.

 

[jscx93]

wait .. are you suggesting hot swapping cpus?

 

[ATTech]

Okay, I've finally figured something out. And I need some more input.

It's the CPU. Everything works fine and turns on as long as the CPU is not plugged in. I actually plugged it in while it was on, and the machine suddenly shut off.

I pulled the heatsink off and re-applied thermal compound yet again and no luck. The heatsink locks are kinda loose, but I tightened them.

Does this sound like it's the CPU or motherboard? I can't seem to narrow it down to one or the other.

This machine is also like a ticking time-bomb with one thing after another breaking.

I need some advice from someone more knowledgeable on how I can narrow it down to one or the other without having to put the CPU in another machine, because I don't even know which machine has the proper socket for this type of processor.

Thanks in advance.

Doesn't necessarily mean it's the CPU. Yes, the CPU is overheating and that is causing the shutdowns, but what is causing the CPU to overheat is the question. With the CPU not in the system, there isn't really any way for the CPU to overheat, is there? If the system has proper ventilation, air flow, heat dissipation, and all settings are correct, then then next likely culprit is a short that is causing the CPU to run hotter. This could be internal to the CPU, or on the motherboard. Unless you're a damn good hardware engineer with the proper equipment, then the only way to tell would be a known good processor and/or motherboard.

On the subject of hot swapping a CPU. Look at it like this: Processors now-a-days have a few hundred to over one thousand little pins at the bottom of them, all condensed within a few square inches of real estate. Imagine these as little I/O paths. When plugging in the CPU, being just a few millimeters off can send an input signal through an output pin frying an internal component. That's why you don't hot swap a CPU.

 

[dronzer]

i have not read your full post but last week i had a pc which shutting off any time,
i figured out that its a power button which was fulty.
so, check the power button ,may be that will work ?

 

[ATTech]

i have not read your full post but last week i had a pc which shutting off any time,
i figured out that its a power button which was fulty.
so, check the power button ,may be that will work ?

A shorted power switch could cause it to shut off a few seconds after it's turned on, but it wouldn't cause the initial issue of running OK for a while and then shutting off (Well, theoretically it could, but the system would first try to shut itself off manually before it cut off, and that would be seen).

Improper contact between the motherboard and CPU is probably why it doesn't even boot up anymore. Bent pins, foreign object in to socket, etc.