OT: Computer Repair 101: Why the Mother Board?

[Paul Alciatore]

For the computer gurus here. Not looking for advice, just a better understanding.

So, I am puzzled. My laptop had a fan that went bad. I cleaned it and it seemed to run OK so we used it while the new fan came in. I thought this would be safe enough. Bad call, it wasn't. The laptop died a few hours later. Live, learn, and pay $$$s. I installed the new fan but that did not bring it back. I moved my work to an older tower and started to build a new tower which is now up and running. Due to lack of time, I decided to take the laptop to a local shop for at least a professional diagnosis. They said it was the mother board because changing the processor did not help. They wanted big bucks for the repair so I paid for the diagnosis and took it home. I ordered a used mother board for about 1/3 of their repair cost and it arrived today. Installed it and presto, it boots. So the local shop was right: it WAS the mother board.

So, my question: When a processor overheats why is it that the mother board goes bad? I have looked very carefully under magnification and there is no sign of any heat damage on the mother board, neither near the processor chip or anywhere else. The processor is OK because the original one is in there now and running. This is apparently a common type of failure because others have told me that the mother board was likely bad.

So, does anybody know what it is about the mother board that goes bad so quickly? Is there some kind of heat fuse somewhere on it? Or some component near the processor chip that is unusually sensitive to heat? Or perhaps the processor socket? Or what? What is the most common mechanism for this kind of failure?

[dp]

So, does anybody know what it is about the mother board that goes bad so quickly? Is there some kind of heat fuse somewhere on it? Or some component near the processor chip that is unusually sensitive to heat? Or perhaps the processor socket? Or what? What is the most common mechanism for this kind of failure?

Everything on a computer depends on a healthy buss. There be more than one buss on a computer, so they all have to be healthy. Multiple points of failure. Then computers need clock signals. Lots of them. Some are principle, others are derived. All have to work. Some are single points of failure, some are multiple points of failure. Interrupts control the pulse of the system and interrupts are real time events. There is a controller for managing IRQ events, and several chips that manage them. Multiple points of failure. There are co-processors on the system that use the busses, IRQs, and clocks. Video, IO, networking (part of IO), disk (part of IO). Multiple points of failure. RAM contains a bazillion PN junctions. Bazzilion points of failure. ROM contains a bazillion PN junctions - another bazillion points of failure.

It's amazing it works at all.

[Jaakko Fagerlund]

If it heated up, then it could be any number of things - bad caps, bad diode, bad transistor, anything that doesn't like heat. Heck, I've even seen a damaged CPU socket, the overheated CPU unit had cracked one of its mounting points loose and whaddayaknow, it didn't work after that.

[wagnerite]

which fan died? CPU fan or an auxiliary fan?

since you said you ran the laptop for a few hours without the fan, that leads me to believe that it was NOT the processor's fan that died. So i'm gonna venture to guess that it was a GPU fan (Graphical Processor Unit, probably was built onto the motherboard). Hence replacing the motherboard along with the built-in GPU worked.

Side note: If you had said you ran the laptop for 1 to 10 minutes before it died, than I'd guess it was the CPU fan.

[Paul Alciatore]

It was the fan attached to a heat sync that was attached to the CPU via a copper tube. Probably a heat pipe of some kind. I guess it also cooled the interior of the laptop so the heat damage could have been anywhere. I was focusing on the CPU because that is obviously the main heat source in there. It is the only fan in the lap top. It did operate for several hours, perhaps 5 to 10 after I cleaned the fan. The fan was working, but perhaps not moving enough air.

I am asking the question because I got the impression that this is a common failure in loptops and I find it strange that the CPU chip would survive while the mother board did not. Perhaps there is an element that commonly fails on them and I am curious to know what it may be. The GPU may be the case as the power did come on, but I got no display at all. Part of my reason for asking about this is so I can better understand it and perhaps avoid future trouble. I guess I am looking for some kind of experience factor here.

[mrriggs]

I had a motherboard die in a laptop. A new motherboard cost as much as the laptop so I Googled the problem and found out that it was a common failure for that model. It wasn't any components burning out but broken solder connections. The solution was to put the motherboard into the oven, 7 minutes @ 385° (if I remember correctly). Sure enough, put it all back together and it worked.

[ptjw7uk]

I guess you will never know what fried on the motherboard as I suspect no one has repaired one at component level , just not cost effective! Let alone the difficulty of changing components on a multi layer board.
The industry norm seems to be to go for the easy option, change the motherboard after all the customer is paying!
I'm surprised that a slow fan caused the problem as my laptop seems to have buckets of heat in it even with the fan going full tilt. Mine gets quite warm on the bottom that its uncomfortable to have the thing on my lap, I think the battery is getting hot as well!

Peter

[Peter.]

I've found the easiest way to unintentionally kill a laptop is keep using it on mains power when the battery has lost it's ability to hold charge. It fries the voltage regulator circuit or something. One machine had a separate power board but most I have owned were integrated on the mainboard.

[macona]

A laptop motherboard is no different than a board out of something like and XBox or PS3 and those are repaired all the time. You need special equipment to work on them.

My guess is during the changeout of the fan there was too much stress put on the board in someplace like the cpu or gpu ball grid and a few connections came loose. Blame it on RoHS solder. The stuff is brittle and does not take heat cycles well.

Also power sockets are a common failure since they see so much stress.

mrriggs's "reflow" in a oven will only last a short time. The only way to really fix these problems is to remove the component, usually a BGA, reball it and replace it. I went though that with my PS3, had someone reflow it and it fail less than a year later. Took it to someone who removed the CPU, reballed it with lead solder balls and replaced it. It is still going strong.

The problem with laptops is that they are such a commodity item that by the time something goes bad the labor to repair one exceeds their value. I have a nice Dell, Core2Duo, Nvidia graphics, 1080p screen... they sell for about $100 to $150 on ebay used working. Game systems are a bit different, the PS3 I have is still worth ~$300 since it is a first gen machine that can play old PS1/PS2 games without emulation so they are worth repairing. Same thing with a lot of industrial electronics.

[beanbag]

One thing I have learned is that if a device is of sufficient complexity beyond your comprehension, just "mucking around with it" can cause it to fail. You won't understand why, so the only lesson is to not do it again.

[martik]

I just dealt with a dying cpu fan on my 6 year old Toshiba A200. It was making a scraping noise and running very slow for a year or more lol. I did not replace it as the new one was integral to a metal housing and cost almost $100, so I just took mine apart and cleaned and oiled the shaft. It was not an easy task! The M/B should be protected as the BIOS monitors temps and will shutdown when overheating. Good idea to replace the thermal paste as it dries up after a while - on mine there was paste BOTH on the cpu and some other IC's under the metal housing that were dried up. What make/model is your laptop?

Also a good idea to blast the vent ports with compressed air every month or so.

[terry_g]

I had a Toshiba laptop that had the fan fail. It would overheat and shut down.
It shut down several times and start up and work properly after it cooled down.
I could only use it for about 20 then I had to shut it off and let it cool down.
I took it apart and fount the fan worked when connected to an external power source.
Apparently the fan control circuit on the motherboard had failed.
The laptop was old and not worth the cost of a new mother board.
I connected the fan to one of the USB ports with a resistor so it would run at about half it's
normal speed. It never shut down again after the repair but battery life was noticeably affected.

Terry

[saltmine]

I found out a long time ago that most laptops are 50 pounds of crap stuffed into a 30 pound sack. Cooling is so critical it's almost a shame. Until engineers can come up with some kind of nano-technology or figure out a way to miniaturize the components we have, I'm not going to dump any money into a laptop. My present computer is a tower, with an Intel quad core processor and 8Gb of RAM. To insure I don't have to endure a "heat event" I installed a liquid cooling system on the main processor, and a heat transfer pump on the graphics card. I've been running the liquid cooling system ever since I experienced a "heat event" three years ago.(I live in Arizona, where air conditioning is a matter of life or death). A lot of people I know continuously have problems with laptops. In fact, quite a lucrative industry has grown up around dead laptops and replacement parts & labor.

[Evan]

Most probably the voltage regulators for the CPU. They are placed immediately adjacent to the CPU and if the CPU overheats so will they. When they go out then so will the CPU.

[Black Forest]

The new Haswell processors are much more efficient than previous processors.