Need some help with Computer Motherboard

[Victor6799]

Would someone be able to tell me how can I go about finding the fault on the motherboard which would explain several different issues. First of all the motherboard will not POST at random and other times it will not even power on. Sometimes it will power on with devices such as CPU fan, HD, DVD and power supply fan all working but no POST. Other times will signal out long beeps continously.

Is there some way that I could test this motherboard to see where the fault is originating from.

Hope the aboe is clear.

Thanking you

Victor

 

[Nigel Goodwin]

Make sure none of the electrolytic capacitors are leaking or bulging, replace any which are - massive numbers of motherboards were built using known faulty capacitors.

 

[Reloadron]

If you have the motherboard manufacturer and model number you can look up the beep codes as you mention long beeps. Depending on the age of the motherboard somethingelse you can look at is if the PSU is getting a PWR_ON signal at the main PSU / Motherboard connector. In a case where the PSU won't power on you may want to give this a read. Additionally if it does power up normal or forced have you checked the PSU voltages? This link should be helpful in checking the PSU voltages.

Also, per Nigel, millions of boards were made with defective capacitors. Visually and in detail inspect the board for leaking or buldging caps, especially around the CPU. The symptoms you describe sound very much like bad capacitors.

Ron

 

[cr0sh]

I would check caps first, too - do some googling on "defective chinese capacitors", "faulty chinese capacitors", and the like - there's a whole slew of them out there in the wild, with a variety of name; if any of them have those names, you could be looking at a bad cap situation.

I have also seen really, really weird things occur on a mobo when the power supply was defective - so you definitely want to check that out as already noted; when I mean by "weird", I mean things like "doesn't boot if the CD-ROM is plugged in with one stick of RAM, but if both sticks and the HD is in, plus the phase of the moon is right and I hold my pinkie up in the air at this angle - it'll boot" (seriously, it can have you chasing your tail - so check that PSU).

Good luck.

 

[Victor6799]

Make sure none of the electrolytic capacitors are leaking or bulging, replace any which are - massive numbers of motherboards were built using known faulty capacitors.

Thanks for the reply. I actually did check before posting this article that the capicators are all working and not bulging or cracked. I do appreciate the reply.

Victor

 

[Victor6799]

If you have the motherboard manufacturer and model number you can look up the beep codes as you mention long beeps. Depending on the age of the motherboard somethingelse you can look at is if the PSU is getting a PWR_ON signal at the main PSU / Motherboard connector. In a case where the PSU won't power on you may want to give this a read. Additionally if it does power up normal or forced have you checked the PSU voltages? This link should be helpful in checking the PSU voltages.

Also, per Nigel, millions of boards were made with defective capacitors. Visually and in detail inspect the board for leaking or buldging caps, especially around the CPU. The symptoms you describe sound very much like bad capacitors.

Ron

Thank you for the reply. I will take another look at the capacitors again to be sure. Thanks for the links. Victor

 

[Victor6799]

I would check caps first, too - do some googling on "defective chinese capacitors", "faulty chinese capacitors", and the like - there's a whole slew of them out there in the wild, with a variety of name; if any of them have those names, you could be looking at a bad cap situation.

I have also seen really, really weird things occur on a mobo when the power supply was defective - so you definitely want to check that out as already noted; when I mean by "weird", I mean things like "doesn't boot if the CD-ROM is plugged in with one stick of RAM, but if both sticks and the HD is in, plus the phase of the moon is right and I hold my pinkie up in the air at this angle - it'll boot" (seriously, it can have you chasing your tail - so check that PSU).

Good luck.

Thanks for the reply. I do have a couple of extra power supplies. Will test out to see if the one I'm using is faulty. Appreciate the reply. Victor.

 

[Brevor]

You may find some help at this forum.

https://www.badcaps.net/forum/index.php

 

[bryan]

There is also one of these test boards as long you can get it to boot.

**broken link removed**

**broken link removed**

 

[Blueteeth]

Generally, since I've been working on laptop motherboards (as well as desktop) problems can fall into three catagories. Mechanical (connectors), heat damage (over heating CPU, GPU) and most likely: failed power components.

Sometimes a failed power component will stand out, as when they go, they truely go. If you can power it up, without a CPU, or ram, or any PCI cards, basically, nothing connected to it except a PSU, and a switch for the 'turn on' jumper - check the voltages for the ram, and CPU. Usually the board will turn on its on-board switching power supplies, even without memory/CPU, but the motherboard controller will quickly turn them off if there is a fault detected.

Unless theres a massive burn mark on the board, they're generally pretty time consuming (read wasteful) to debug.

My two cents

BT

 

[Victor6799]

Thanks Brevor and Blueteeth.

I will definitely take a look at the links you guys have kindly posted. I appreciate the reply.

I do have another question about how a motherboard works at the lowest levels. I would like to know a couple of things that I can think of right now. First how does the motherboard detect the hard drives that are attached to the primary channel of a motherboard. Is the BIOS that tells the Southbridge to check the Primary channel or does the BIOS have a direct channel to the IDE ports and to the drives to determine if there is a IDE drive present on that channel ? I understand that the drive itself knows that its a master or a slave based on the jumper set on the drive. Does the Southbridge do anymore then what I have already described ? Second question is can a faulty voltage regulator prevent a motherboard from beginning its POST ? I have many more questions on this topic but thought these two will be enough until later on.

Thanks again to everyone who contributed.

Victor

 

[Exo]

The bios code detects the harddisks by sending a "identify drive" ATA command on each channel. If it gets no response within a certain time it assumes no drive is attached, if there is a drive attached it will respond with information about itself wich the bios can use to set everything up correctly. Since SATA/PATA controllers are inside the southbridge on most boards the data will offcourse pass trough it.

For PATA drives it is indeed the jumper that tells the drive if it's master or slave, allowing the drive to only act on commands addressed to it, ignoring the rest.

The POST is just part of the bios code, so if the cpu is not executing code it will not POST. Many things can prevent the cpu from running. Power problems, faulty memory, flash corruption and so on.

 

[Blueteeth]

Yep, a process of elimination.

At the 'higher level', as msot PC users/tinkerers work at, they just remove memory, one stick at a time, PCI cards etc. Which is a pretty good way to work out 'whats bust'. Although that method just highlights which bit of hardware is duff, and doesn't in any way point to *how* the hardware is malfunctioning.

As an electronics guy, who moved up to 'PC's' I tend to use any excuse to get my multimeter out and start and start probing - which probably isn't the most prudent thing to do since the system is so complicated, built upon levels of hardware, firmware, interfaces and hsot software that it becomes quite dificult to marrow down issues.

That said, after being upto my eyeballs in datasheets for motherboard interface controllers, many of the 'checks' a motherboard does when it boots (pre OS, and even pre-bios loading, this is all fixed in the hardware of the mobo) are pretty straight forward so its jsut a case of either working fom the top down - starting with a full system and removing bits until it boots. Or in the case of identifying true hardware failure (and actual faulty part) working from the bottom up - that is starting with almsot nothing and adding bits to see just how far the system gets before it throws its toys out the pram.

I'm no pro at repairing computers, not by a long shot, I'm more of a general tinkerer who doesn't like to be beaten, so I'm biased towards true hardware solutions rather than experience. So, my advice is vague, sorry :/

Blueteeth

 

[Victor6799]

Exo thanks very much for the reply. That was the answer that I was looking for. I have scoured the internet using google.com but could not find which chip was the one responsible for the actual detection...so it appears its the BIOS. Many thanks. My next question is as follows (please do not ask why I am asking) do the devices located on the SouthBridge hub (as I call it) could any one of them intercept data that is being sent to the BIOS like when the BIOS is trying to detect the IDE devices connected to the motherboard ? Or do they have their own path (trace) to the SouthBridge so that the data transferred to the BIOS is safe and secure without any tampering ? Also can any device on the SouthBridge hub make changes to the CMOS memory data like the date and time ?

Blueteeth thanks for the helpful reply. Its much appreciated. I do have a question for you. My question is can a faulty voltage regulator prevent a system power up (like the fans are all spinning and what not) ? Also what parts would you suggest that they be removed as per your suggestions of trial and error ? Would you be referring to parts like PCI cards and IDE devices or SATA devices keeping the bare minimum like RAM and Video and CPU of course? Would you mean parts like chips which are soldered onto the motherboard ?

Thanks again

Victor

 

[Blueteeth]

Blueteeth thanks for the helpful reply. Its much appreciated. I do have a question for you. My question is can a faulty voltage regulator prevent a system power up (like the fans are all spinning and what not) ?

Yes. Motherboards on desktops get several voltages from the PSU unit via the big power connector (plus an aux 4 pin 12V connector) but the power for the memory, CPU, and motherboard controller come from on-board siwtching power supplies which themselves are powered by the 3.3V/5V/12V lines from the PSU.

Because of the very high current CPU's use, and the sensitivity to voltage levels (for both the memory and CPU) these on-board power supply converters are usually quite complex, and always have fault protection to do their best to prevent current surges and over voltage. Each on board power supply always has a connection to the motherboard controller for checking the status - this may be a simple 'power ok' line, or a data connection where bytes are exchanged. So, if one of the switching transistors that one of these power supplies blows (MOSFETs close circuit) the converter cannot hold regulation, flags up an error, and the motherboard controller powers it down to prevent further damage to the CPU/memory. The happens all before the bios is even looked at - its pretty much pure hardware. And will halt all signs of operation, including the CPU fan. Any fans connected to the motherboard *may* still spin, because they will get the 12V from the PSU, but their speed is controlled by the motherboard controller. I would say, if they don't spin at all then the mobo has detected a fault, and shutdown.

Even if these converters are working perfectly, if they are powering anything that has shorted (example: bad RAM) - then too much current will be drawn, the converter detects this, and the above process happens. In this case, no damage is done, its just the mobo refusing to do anything because a faulty component is installed.

Also what parts would you suggest that they be removed as per your suggestions of trial and error ? Would you be referring to parts like PCI cards and IDE devices or SATA devices keeping the bare minimum like RAM and Video and CPU of course? Would you mean parts like chips which are soldered onto the motherboard ?

I've been working on laptop motherboards recently, which is a slightly different ball game, because most of the components are integrated, where-as in desktop's, where space is of no concern, its modular. So bare with me, as I haven't worked on a desktop in months.

Unless you have access to another system that you can use to test individual parts, such as: Drives (SATA), PCI cards, memory, even CPU etc.. then removing everything except the absolute minimum is a good start. If your motherboard has on board graphics - which I'm pretty sure almost all do these days, then remove all PCI cards, including the graphics card. Remove all SATA devices. Leaving only the CPU and one stick of RAM. Your monitor can be plugged into the VGA (or HDMI) connector on the motherboard.

So, with no drives, there's no OS, and no drivers, this will check just the motherboard, CPU and ram. If it fails to power up at all, change the RAM stick to the other one you have. I would be prudent to test the memory in another machine to completely eliminate that as a problem. Leaving only the motherboard, and CPU. On power up, it should happily get to the BIOS load up screen, then complain that there is no operating system found. Adding parts back one by one until things start to go less smoothly, and you should have found whats giving you greif. I would start with hard drive, then graphics card, then other PCI cards, then any other RAM.

As its an intermittent problem, and one that happens before the OS loads, it is likely to be a power issue (think I've said that before..). The above operation is basic, time consuming, but is a good methodical way of finding out whats up, and something I wish I had done more often as it would have saved me much time in trying to work out whats wrong.

When I said 'parts', I meant, pretty much everything Once you've narrowed it down to the a single 'major' part (like motherboard, a PCI card etc...) then if its the motherboard, you can probe that for power problems. Apologies for giving vague advice, as I said, my head has been full of laptop schematics and datasheets, so I get lost in the 'detail' rather than the bigger picture.

Yep, I hate computers

 

[Exo]

do the devices located on the SouthBridge hub (as I call it) could any one of them intercept data that is being sent to the BIOS like when the BIOS is trying to detect the IDE devices connected to the motherboard ? Or do they have their own path (trace) to the SouthBridge so that the data transferred to the BIOS is safe and secure without any tampering ? Also can any device on the SouthBridge hub make changes to the CMOS memory data like the date and time ?

The way the bios flash chip is connected to the rest of the system (trough the southbridge usually) depends on the motherboard, diffirent systems were used troughout the years. Most recently common practice was a special simplified version of the PCI bus called LPC bus and even more recently the SPI bus. Other devices besides the bios flash chip can be sitting on this bus, corrupting data when defective.

Your picture of the bios seems a bit off tough. The bios chip is just eeprom or flash memory holding the bios program. It does not actually do anything, no commands are sent trough it. The program it holds is just mapped into the addressing space of the system's main cpu, so the cpu can see and execute the code. Because everything is done by the main cpu wich sits on top of the northbridge, all the way on the other side of the southbridge, a defect anywhere in between can cause these errors.

Have you tested with a known working power supply already ?
What type of motherboard is it by the way ? Nvidia chipset based boards have common problems like this. Nvidia used bad solder for a few years causing connections to break after a few years. This caused massive failures in 8000 series geforce cards and nvidia chipset based motherboards.

 

[Victor6799]

There is also one of these test boards as long you can get it to boot.

**broken link removed**

**broken link removed**

Thanks for the reply. I might consider one of these cards since they are fairly affordable. Victor.

 

[Victor6799]

Yes. Motherboards on desktops get several voltages from the PSU unit via the big power connector (plus an aux 4 pin 12V connector) but the power for the memory, CPU, and motherboard controller come from on-board siwtching power supplies which themselves are powered by the 3.3V/5V/12V lines from the PSU.

Because of the very high current CPU's use, and the sensitivity to voltage levels (for both the memory and CPU) these on-board power supply converters are usually quite complex, and always have fault protection to do their best to prevent current surges and over voltage. Each on board power supply always has a connection to the motherboard controller for checking the status - this may be a simple 'power ok' line, or a data connection where bytes are exchanged. So, if one of the switching transistors that one of these power supplies blows (MOSFETs close circuit) the converter cannot hold regulation, flags up an error, and the motherboard controller powers it down to prevent further damage to the CPU/memory. The happens all before the bios is even looked at - its pretty much pure hardware. And will halt all signs of operation, including the CPU fan. Any fans connected to the motherboard *may* still spin, because they will get the 12V from the PSU, but their speed is controlled by the motherboard controller. I would say, if they don't spin at all then the mobo has detected a fault, and shutdown.

Even if these converters are working perfectly, if they are powering anything that has shorted (example: bad RAM) - then too much current will be drawn, the converter detects this, and the above process happens. In this case, no damage is done, its just the mobo refusing to do anything because a faulty component is installed.




I've been working on laptop motherboards recently, which is a slightly different ball game, because most of the components are integrated, where-as in desktop's, where space is of no concern, its modular. So bare with me, as I haven't worked on a desktop in months.

Unless you have access to another system that you can use to test individual parts, such as: Drives (SATA), PCI cards, memory, even CPU etc.. then removing everything except the absolute minimum is a good start. If your motherboard has on board graphics - which I'm pretty sure almost all do these days, then remove all PCI cards, including the graphics card. Remove all SATA devices. Leaving only the CPU and one stick of RAM. Your monitor can be plugged into the VGA (or HDMI) connector on the motherboard.

So, with no drives, there's no OS, and no drivers, this will check just the motherboard, CPU and ram. If it fails to power up at all, change the RAM stick to the other one you have. I would be prudent to test the memory in another machine to completely eliminate that as a problem. Leaving only the motherboard, and CPU. On power up, it should happily get to the BIOS load up screen, then complain that there is no operating system found. Adding parts back one by one until things start to go less smoothly, and you should have found whats giving you greif. I would start with hard drive, then graphics card, then other PCI cards, then any other RAM.

As its an intermittent problem, and one that happens before the OS loads, it is likely to be a power issue (think I've said that before..). The above operation is basic, time consuming, but is a good methodical way of finding out whats up, and something I wish I had done more often as it would have saved me much time in trying to work out whats wrong.

When I said 'parts', I meant, pretty much everything Once you've narrowed it down to the a single 'major' part (like motherboard, a PCI card etc...) then if its the motherboard, you can probe that for power problems. Apologies for giving vague advice, as I said, my head has been full of laptop schematics and datasheets, so I get lost in the 'detail' rather than the bigger picture.

Yep, I hate computers

Thanks Blueteeth.

A very informative and helpful reply. Alot to digest for now. Will study it some more and hopefully I will be better off even if the issue that I am having continues. Victor.

 

[Victor6799]

The way the bios flash chip is connected to the rest of the system (trough the southbridge usually) depends on the motherboard, diffirent systems were used troughout the years. Most recently common practice was a special simplified version of the PCI bus called LPC bus and even more recently the SPI bus. Other devices besides the bios flash chip can be sitting on this bus, corrupting data when defective.

Your picture of the bios seems a bit off tough. The bios chip is just eeprom or flash memory holding the bios program. It does not actually do anything, no commands are sent trough it. The program it holds is just mapped into the addressing space of the system's main cpu, so the cpu can see and execute the code. Because everything is done by the main cpu wich sits on top of the northbridge, all the way on the other side of the southbridge, a defect anywhere in between can cause these errors.

Have you tested with a known working power supply already ?
What type of motherboard is it by the way ? Nvidia chipset based boards have common problems like this. Nvidia used bad solder for a few years causing connections to break after a few years. This caused massive failures in 8000 series geforce cards and nvidia chipset based motherboards.

Thanks again for the helpful reply. The motherboards that I am having issues with are P4 Socket 478 motherboards. One is a P4B533 and the other is a P4P800-X. I've tried a different power supply but that made no differnce. The CPU which are on each motherboard are different. One is a P4 2.53GHz and the other is a 2.8GHz FSB 800MHz. I don't remember the FSB on the 2.53GHz CPU. I've tried different video's card but still no difference. Thanks again. Victor.