Comp turns on; won't POST

[Al Shades]

My gaming comp has been down since last May.

Specs:

P4 3.06 ghz 533 fsb H/T (Northwood?)
Asus P4T533-C (Intel 850E Chipset) socket 478
Enermax 460w PSU
1 gig Samsung RDRAM PC1066 4 x 256
BFG Geforce 6800 GT OC Ed.

When I power it up, all of the hardware component start up normally, but no display ever comes up and the BIOS beeps are never heard. I'm trying to narrow down the list of possible causes.

1. Mobo

I think this is the most likely culprit. Interestingly enough, while checking eBay for replacement boards, I ran into an auction for one that mentioned the following: "The board suddenly did not boot any more." I have not seen any mention of similiar problems related to this board from any other source. It was supposedly the best board of its kind made for the 850E Rambus chipset.

2. PSU

Could the power supply be somehow failing to make the comp POST, despite starting up all the hardware components? I doubt it.

2. RAM

I've tried booting individually with each one of the 4 RDRAM sticks I use, and that didn't work. So unless all the sticks went dead at the same time, I don't think there's a problem with the RAM. Furthermore, someone told me that a machine can POST without working RAM, so this couldn't be the cause of my problem. Is this information correct?

3. CPU

The CPU on this comp had been replaced in Dec. 04, along with the graphics card, due to a partial burn-out of the old processor (the fan had died). When the comp died in May, I tried switching the old cpu back in to see if the new one was the source of the problem, but that wasn't the case. The damaged cpu would POST, but refused to boot into Windows. Neither chip made any difference when I tried them.

I made 2 open-case videos of start-up:

media.putfile.com/MVI_060877

media.putfile.com/MVI_060918

And here are two detailed pictures of the inside of my case:

img73.imageshack.us/my.php?image=img06110ol.jpg

img73.imageshack.us/my.php?image=img06122ig.jpg

Sorry, the forum won't allow me to make them clickable.

At this point, it seems to me that the motherboard is the only possible cause of my problems. If so, I'm ready to get a replacement off eBay. Am I correct in my thinking and is there anything that I'm missing?

[DeusExMachina]

When did the P4T533-C could take PC4200 RAMBUS memory? Is it 16bit? Or did the BIOS change to accept 32bit memory? Could've swore it could only take PC-800 16bit memory. Maybe it could've take 1066 @ 16bit, but it was clear that 32bit was for the P4T533-R.

DXM

[Al Shades]

DXM, the memory I'm using is PC1066 RDRAM. I don't know if it's 16 or 32bit, although the memory reads "1066-32P". The 850E chipset accepted 1066/800 memory.

intel.com/design/chipsets/mature/index.htm

[Intel_User]

When did the P4T533-C could take PC4200 RAMBUS memory? Is it 16bit? Or did the BIOS change to accept 32bit memory? Could've swore it could only take PC-800 16bit memory. Maybe it could've take 1066 @ 16bit, but it was clear that 32bit was for the P4T533-R.

Nope, you are confused: http://www.tomshardware.com/2002/06/...bus/index.html

Officially 850E only supports 533FSB but not the PC1066, but nearly all can run PC1066, making it a moot point to most people.

Back to Al Shades

Hmm. I had a similar looking problem with my system, but obviously it was my fault. Though I reckon it can yield the same results. The problem came out to be my mobo, I somehow fried it without showing it physically.

1. Mobo: my experience says its probably likely its your motherboard. Go check if there are any leaky or faulty capacitors
2. Power Supply: Possible, its a close second, only way would be to check out with another power supply. I doubt it since your system powers up. Individual voltage rails breaking is more rare than whole PS failing.
3. RAM: I highly doubt it. RAM is a pretty reliable component
4. CPU: The most reliable component of your system, IMO. If you ever have a problem with your CPU, you would know it the day you buy one, or you screwed it up. Though there is still a itty bitty chance its your CPU.

Weirdest thing happened with CPUs. I had a motherboard I needed to test to see that it works. My friend tested with his Pentium III 667MHz and said didn't work. However, when I tested with another friend's Pentium III 900MHz CPU, it worked, and since he needed the mobo, good for him.

[Intel_User]

DXM, the memory I'm using is PC1066 RDRAM. I don't know if it's 16 or 32bit, although the memory reads "1066-32P". The 850E chipset accepted 1066/800 memory.

intel.com/design/chipsets/mature/index.htm

LOL, I bet the "32" on 1066-32P means 32-bit.

[Al Shades]

Intel, thanks for the info. According to the page I linked, the 850E did "officially" support 1066, but who knows. As for your advice:

I remember back when my system broke, it was during a week of power outages. One of my initial guesses was that there had been a power surge which damaged the PSU. I didn't think of the mobo, however. I'm not well versed in things related to power supply. Is there a specific area of the mobo I should be checking? I've already looked it over many times in the past, and didn't see anything unusual. However, I know nothing about voltage rates and such, as I've never fully assembled a computer before.

Also, speaking about cpu's: It just so happened that the cpu which I have in this comp lost one of it's pins as I was bending back 5 of them. Do you think it will still run, once I replace the mobo?

Originally Posted by Intel_User

LOL, I bet the "32" on 1066-32P means 32-bit.

I thought it was a reference to the parity. I don't know all that much about RAM. Just the basics/

[DeusExMachina]

Originally Posted by Intel_User

Officially 850E only supports 533FSB but not the PC1066, but nearly all can run PC1066, making it a moot point to most people.

Yep, as Intel never designed the 850E for 1066mhz (they didn't make such a board -- Asus and 2 other m/b manufacturers did).

But the P4T533-C wasn't designed as a 32bit board.

Samsung made 16 and 32bit 1066mhz memory. But the P4T533-C was originally designed for PC-800 16bit memory (2 modules required to work) -- it's why it had 4 slots, compared to only 2 for the P4T533-R.

This I know because I picked the 32bit P4T533-R over the P4T533-C over that issue (only one module was needed to operate the board).

So, 32bit wasn't designed for that board -- it clearly states 16bit.

ADDED: Reread the review again...

Asus P4T533: The Only Board With 32 Bit RDRAM Support

http://www.tomshardware.com/2002/06/...us/page10.html

DXM

[Intel_User]

Also, speaking about cpu's: It just so happened that the cpu which I have in this comp lost one of it's pins as I was bending back 5 of them. Do you think it will still run, once I replace the mobo?

You mention that you tested two CPUs. The one you are saying is the broken one or the not broken one?? Or is it both broken. So now you have a thermally damaged and physically damaged CPU??

Whether it was a power outtage or not, I would still point to your mobo as the culprit. It seems another person on the Intel threads have similar problem?? Sometimes the damage could be internal so you can never check from outside which is burnt out.

Is there a specific area of the mobo I should be checking?

If there is any area, look for the components near the CPU socket. Maybe you can see a brownish component.

[Rickwell]

Your mother board is using PC 1066 RDRam32 nano seconds that's 16bit .You couldn't put 32 bit PC 4200 rdram in because it has more pins and won't fit.You said you tested the ram.You have to run RDRam in pairs only.You can't just put one stick in at a time like ddr ram.I kind of dought there's to much wrong if it's at least starting.But sitting around can make the contact pionts on your hardware get crusty.Reseat the ram,video card,hard drive pugs and cable,any power cables to the mother board.Do all that with the Power Off.You can try to reset the BIOS.I have that board and if I let it set for a long time it wouldn't start either.I had to be calm and try to make sure to reseat all the hardware two or three times to get it to work.That got all the crusty crap off the contacts.but eventually it would go again.Try those things first before you decide that it's a dead part or replacing stuff.Good Luck

[Intel_User]

Samsung made 16 and 32bit 1066mhz memory. But the P4T533-C was originally designed for PC-800 16bit memory (2 modules required to work) -- it's why it had 4 slots, compared to only 2 for the P4T533-R.

Hmm, you could be right. I only read reviews of the board. I thought both 16 and 32-bit were compatible.

Ah, I see what 32 means now. Old datasheet for RDRAM in Samsung's site says "32P" is latency in ns. There is 40 and 45. So it just means you got fast RDRAM .

[Intel_User]

Go try cleaning your components for dust. I fixed a problem by cleaning once. I never knew it would be such a simple problem.

[DeusExMachina]

Originally Posted by Rickwell

You have to run RDRam in pairs only.You can't just put one stick in at a time like ddr ram.I kind of dought there's to much wrong if it's at least starting.But sitting around can make the contact pionts on your hardware get crusty.

What I like to know is how he got the 32bit memory module to fit a 16bit slot?? The notch on the bottom is placed differently in a 32bit.

DXM

[Rickwell]

Originally Posted by DeusExMachina

What I like to know is how he got the 32bit memory module to fit a 16bit slot?? The notch on the bottom is placed differently in a 32bit.

DXM

He didn't and can't put 32 bit ram on that board.It only can use PC800 or PC 1066 16 bit RDram.Click on the link to see what I mean.I know it can be confusing lol.
http://www.memorysuppliers.com/sam5180184pi.html
http://www.memorysuppliers.com/sam51pc53ram.html
http://store.yahoo.com/pctekonline/25pcramrdram2.html

[Al Shades]

Okay guys, lots of new info here. I'll answer the questions and ask a couple of my own.

Originally Posted by Intel_User

You mention that you tested two CPUs. The one you are saying is the broken one or the not broken one?? Or is it both broken. So now you have a thermally damaged and physically damaged CPU??

Correct, I have a thermally damaged and a physically damaged CPU. However, the latter only became damaged yesterday. Previously, I had tested the comp with the undamaged chip as well as the thermally damaged one. The thermally damaged had, in the past, never failed to POST, as I mentioned earlier. This, coupled with the fact that the undamaged chip was relatively new, led me to believe that I had enough evidence to eliminate the CPU as a possible source of the problem.

On cleaning:
Am I correct in thinking that the following Isopropyl alcohol rubbing solution would NOT be appropriate for cleaning the parts, due to the reason indicated?

putfile.com/pic.php?pic=3/7301561946.jpg&s=x11

How many posts do I need on this board before it will allow me to post full links?

[DeusExMachina]

Yeppers. The notch would prevent seating 16bit in 32bit slots and vice versa.

And you nailed it on the one module testing -- won't work, 16bit must be used in pairs.

DXM