Motherboard or CPU problem?

pastymuncher said:
No doubt Mattus came to the same conclusion.
Click to expand...

Indeed. :)

westom said:
So he ignores a symptom of SATA misbehavior - marginal voltages. Other symptoms also exist that are synonymous with margin DC voltages.
Click to expand...

Firstly, are you familiar with the ATX power specification? It states that the voltages on each power line have to be within 5% either side of the nominal voltage. So as long as a 12v line is between 11.4 and 12.6 volts, it is technically within specification. These things don't operate to awfully tight voltage tolerances. The hardware in a PC is designed to cope with voltages that are rather higher or lower than ideal, probably because most people use cheap PSUs that don't regulate the voltage awfully well. I'm not sure a 'marginal' voltage really exists in the context of a PSU.

PaulGreyhead's analogy about a doctor is a good one. If I present to my doctor with a fever and a runny nose, he'll probably diagnose me with the flu. Sure, I might have pneumonia or scarlet fever, but in the absence of any symptoms pointing more specifically towards this, such a diagnosis is imprudent. One comes first to the simplest conclusion; the one which needs the fewest assumptions.
 
Last edited:
I can now clear this up once and for all. The retailer recieved it this morning and miraculously has tested it this afternoon. Never heard of that before. Anyway, all of the sata ports from the Intel ICH9R controller are dead. So it had nothing at all to do with voltage. They are sending me the newer version (rev 2.0?) of the board and it should be with me in a day or two. Fantastic service from a small retailer. :D
 
That is good service, I needed to send my DFI (975X/G Infinity) to netherlands to get repaired, took 2 weeks but thay managed to fix it.
 
It's only 3-4 months old so it went back to the retailer. Fantastic service. Did'nt know there was a newer version of the board either.
 
PaulGreyhead said:
If you have one chip or controller controlling four or six sata connectors, how can a voltage drop affect one connection and not another? With a parallel supply each channel should receive the same power input.
Click to expand...
So why did various ports on one VLSI chip work differently with the same voltage? You have assume what happens inside a digital chip is digital. It’s analog. Strange things (including metastable conditions) occur when voltages are marginal.

Also some chips work perfectly fine at some voltages while others misbehave. Again, analog conditions.

Normally, a supply controller should not let this happen. But it does.

Voltage is one of many inputs to each chip. When strange and intermittent failures are the symptoms, those 30 seconds with a multimeter is a first data collected just so other tests are not so misleading.

A marginal voltage caused a VLSI chip to change some valves. Many ports from the same IC. But only some were adversely affected. What happens inside digital chips is analog and undefined when voltages are unstable. (BTW, these type problems are particularly troublesome which is why it is better to just check obvious and simple things such as voltages AND why a calibrated motherboard voltage monitor can be so useful).
 
Last edited:
That is very good service indeed :)

westom said:
So why did various ports on one VLSI chip work differently with the same voltage? You have assume what happens inside a digital chip is digital. It’s analog. Strange things (including metastable conditions) occur when voltages are marginal.
Click to expand...

But he sent it back to the retailer, who found exactly the same problem using a different PSU and a different power source?

You're being extraordinarily stubborn.
 
Mattus said:
You're being extraordinarily stubborn.
Click to expand...
So it is stubborn to encourage bad diagnostic techniques. His board was only the example. Discussion groups are for the benefit of all. PaulGreyhead is asking questions of one who wants to learn. Why are you do fixated only on that board - not on what can be learned from that board?

Meanwhile, you have not yet learned from the experience of measuring. Whereas the spec says 5%, a voltage of 11.5 measured on a multimeter means defective. Why? We take that spec number. Then add other parameters such as how a meter works. 11.45 volts means the 12 volts is completely defective.

These are electrical concept that must be taught to our technicians. Most without training see numerous spec numbers, do not appreciate how they are interrelated, and then even get angry when told additional information. I expect many to deny this '11.45 is defective' conclusion with anamosity because a majority prefer to demean rather than learn.

You put forth a number. 11.5 volts means a defective supply. Maybe a perfectly good supply, but a defective supply in this system that probably also boots that computer.
 
(I realise this is derailing the thread somewhat, but the initial problem has been solved and I think you have to defeat this kind of attitude through argument.)

So if 11.5 is defective, is 11.6 defective? 11.9? 11.99? Where do you draw the line? And why should that take precedence over the specification?

Why are you do fixated only on that board - not on what can be learned from that board?
Click to expand...
What I've learned from that board is that its power supply was perfectly adequate and that my intuition, like that of many people in this thread, was correct.

Meanwhile, you have not yet learned from the experience of measuring. Whereas the spec says 5%, a voltage of 11.5 measured on a multimeter means defective. Why? We take that spec number. Then add other parameters such as how a meter works. 11.45 volts means the 12 volts is completely defective.
Click to expand...
Why? 11.45 is within the spec just like 11.5 is. 'Completely defective' means nothing; something is either defective or it's not.

You're using lots of elaborate and condescending words but not really saying anything of substance.
 
Last edited:
Mattus said:
So if 11.5 is defective, is 11.6 defective? 11.9? 11.99? Where do you draw the line? And why should that take precedence over the specification?
Click to expand...
As I stated early. Those numbers contain significant information - too much to detail here. But if those numbers are not provided, then the many components of a power 'system' remain undefined. That spec of 11.4 (and other numbers from that spec) means a multimeter must measure above 11.7VDC. Obvious for reasons beyond the scope and technical grasp of this discussion.

Had the OP taken 30 seconds to measure and post those numbers, then something moves from a list of unknowns to a list of either 'definitively good' or 'definitively bad'. That is accomplishment with near zero labor. IOW what the Japanese mean by, "Work smarter; not harder."

That you did not know why 11.5 is defective only confirms to me insufficient knowledge of some basic electrical principles. So why so much animosity? I am not the one doing personal attacks and condescending. I have only provided technical knowledge and diagnostic recommendations. I have challenged popular urban legends. Those who are too attached to myths will routinely take offense rather than learn. Few who actually know this stuff get attacked for simply exposing popular myths; for challenging bad habits such as shotgunning?

Normal is for a defective supply to still boot a computer. Then weeks or months later, the system does strange things such as booting slowly, strange shutdowns, and intermittent crashes that get often blamed on heat. First hardware to verify is a supply 'system'. Once those components are confirmed good in as little as 30 seconds - without doubt - then a hardware diagnostic procedure moves on to other suspects.

Unfortunately, a popular myth says if the computer boots, then its supply is good. The concept of 'follow the evidence' means moving things from a list of 'unknown' to some 'definitively ... ' category. If one suspect remains, the problem is solved the first time. Then numbers further confirm the problem really was solved. Solutions that are faster, done right, and cost less money. "Work smarter; not harder".

A troubleshooter need not know why 11.4 is defective. We make it simple for him. A response will only be as useful as the numbers provided. 30 seconds to measure voltages. With numbers, a next reply has definitive answers.
 
pastymuncher said:
In this case a big stick may be better. :D
Click to expand...

Or this
desert-eagle-01.jpg
 
The new board has just arrived and about to be installed. That is amazing service by any standards. The board looks the same to me but the part number is different. No obvious rev 2 markings on the board. Will let you know if the problems sorted.
 
All rebuilt and working perfectly. Tried a few restarts/shut down cycles to make sure and all is well. Let's hope this board last's longer than the other one.

The best bit? Absolutely not a psu problem which most of us knew it was'nt anyway!! :D
 
Glad it's all sorted! :)

westom said:
Do you know that? You just changed components of the supply 'system'.
Click to expand...

I'm starting to lose my patience.

The motherboard was faulty. We know this because he's changed the motherboard and all is good again.

You may continue to use woolly phrases and quotation marks like supply 'system' which don't really mean anything in the context of fixing a PC, but the rest of us want to know which component in our PC is faulty, so we can replace it. This is what threads like this are about. Solving people's PC problems. Not becoming mired in this kind of breathtaking irrelevance.

If we'd all followed your advice, the OP would know more about the power system in his PC than he'd ever need to, but it still wouldn't work.
 
Perhaps if we ignore him he will go away. ;)

Still have'nt found any difference between the boards. Running perfectly so i am happy again. :)
 
You must log in or register to reply here.
Back
Top Bottom