Motherboard repair

[l4322]

Hi, I just got my electronics degree and would like to start building projects and repairing stuff for fun. I thought I would start with a couple of broken motherboards I got for a pound off e-bay, however they didn't teach us repairing at university and I don't really know much. I was wondering if there is a good book or (preferably) site that goes through the basics of repairing procedures, whether for motherboards or other stuff.

Thanks

 

[mattg2k4]

One of the biggest problem with recent motherboards is bad capacitors. Apparently some major company got the formula wrong, creating a large number of motherboards doomed to cap failure. You can read up on it here.

I don't know of anything else on a motherboard that would be easily fixable though, most of the components tend to be surface mount. I have heard of voltage regulators going bad every now and then, so check the boards for that.

 

[Dean Huster]

Times have changed.

In 1963, you could repair anything. These days, because of SMT and the low cost of manufacturing, things are mostly replaced vs. repaired. Not that they can't be repaired ... it's just that it's not a cost-effective operation. In the case of the motherboard, repairing it might cost more than a new one and the new one will be better and faster than the original.

About the only time something like a motherboard is "repairable" is if you consider your time to be free. To make such a thing a commercial operation would be a financial disaster. You may notice that there aren't nearly the number of radio/TV repair shops around that there were 30 and 40 years ago. Unless a TV is an expensive one, such as a big-screen or digital model, it's usually cheaper to buy a new, improved model than it is to fix the thing.

Troubleshooting skills are learned mostly by doing and experiencing. It's something that's difficult to teach from a book, but at least that would be a start. If nothing else, I'd suggest learning to draw and read block diagrams to help you learn signal flow. That can help you learn to isolate problems to specific areas.

Dean

 

[l4322]

Well.... I do have an electronics degree.... I can read block diagrams... I'm not sure how that helps though. I would need to get the block diagram
from the datasheet or something.
I believe I could fix motherboards if I had their datasheet or repair manual, but I can't seem to find anything on the net. One person I contacted at a company named compal said those things are confidential... that's why I posted on this forum.

Does anybody know where I could get repair manuals?

 

[Madmartin]

First check your processors power. 3.3v 2.1v, 1.6v whatever is specified. Try getting your CPU's voltage and pinout from its datasheet (lucky one ! ) or some web references. use an oszilloscope to check for messy, unclean power. a little noise is ok, but usually no more than 10 percent. I agree to the caps problem. I fixed my Pentium 3/500 three weeks ago with this problem. Be sure to get 105 °C caps - they last longer. Most regulators are either waste-as-heat types for AGP (2.5v) and RAM (1.8v) and a buck converter (you should know about) for CPU voltage that is regulated by the cpu power id pins. Datasheets sometimes available.
Every thing else is IMHO out ouf your iron's scope.
Try weather your mainboard will power up without cpu in place - this makes it safer during repair. Also try removing memory and all other cards, HDDs, FDDs etc. before plugging on. Make your measurements, repair / work around if possible, then fit in CPU, memory and *PC-speaker* and plug on. your bios will tell you if it works - beeeeeeeeeeeeeeeeeeeeeeeeeeeep - no graphics card found. plug everything else in and there you go.[/u]

 

[lavenatti]

https://www.repairfaq.org/

This site has unbelievable amounts of info.

 

[riteet]

a dying breed

Yup, Dean Huster brings up a great point, of how this new technology becomes irrepairable. I myself am a student, but have been working for a telecom firm in thier Warranty Division. I've seen technology go through that dept from 1980-current day board designs. The newer stuff can usually be repaired simply replacing/reprogramming an FPGA, while older stuff is usually a game of signal checking and discrete device testing.

As I see it, sooner or later, as we see more digital solutions, and single chip designs, I'm pretty sure my job will be replaced with some untrained person (non-degree) who will just run a test fixture, and replace whatever the test says to. It's just a dying breed of technicians.

For now my job is safe... And I'll hope to get out of it soon after I aquire my BS/MS

Dave