New system not stable 3

[biosol]

Hello,
I've built my last 3 systems so I thought this latest would go smoothly like
the others, but the system first started to freeze when transfering several
GB of data to it (did that 3 times), gave me application errors in IE and F
irefox, and now is rebooting under load.

The system details:
MB = ECS PT800CE-A VIA 'prescott' ready based chipset
CPU = Intel 3.0E Ghz, 1mb cache, HT prescott I believe
RAM = 2x512 Cosari DDR400 for 1GB
Video = Radeon 9200SE 128MB based
Sound&NIC = onboard
H.D. = Seagate 120GB 7,200 RPM
CDRW = Sony

PSU = was the 300W that came with the case, just replaced with a Just PC 450
W

After watching about 45 minutes of a DivX movie the system has rebooted twic
e recently, once with the 300W PSU, once with the new 450W PSU. Both times
I have entered the BIOS to check the CPU temp. which was around 38-40c. So
I don't think it's a heat problem. The case temp was cooler, maybe because
I have 3 fans in the case.

Another odd thing is that I used this motherboard for about a month and a ha
lf in a different case with a Celeron D 2.66GHz and 512MB of Mushkin RAM. I
t worked fine.

Can someone help me out? I used the power calculator at

http://www.3dcool.com

and
it came up with 299W, and I'm planning on adding 4 firewire cards and a wire
less PCI card so I need to find out what's making it unstable. Any thoughts
?

Thanks,

 

[Frank4d]

I built a system almost the same as yours about two months ago. No problems. My system details:
MB = ECS PT800CE-A VIA 'prescott' ready based chipset
CPU = Intel 3.0E Ghz, 1mb cache, HT prescott I believe
RAM = 2x256 PQI DDR400 for 512MB
Video = ATI AIW Pro 9800 128MB based
Sound&NIC = onboard
H.D. = Western Digital 160GB 7,200 RPM (two partitions)
DVD-RW = Plextor PX-712A

A few things you might try:
Load the VIA chipset drivers that come on the CD packaged with the PT800-CE.

In Windows, turn off "Automatically Restart" in "System Recovery". This will hopefully display an error message instead if just restarting the computer.

Run a memory test using Memtest86 version 3.2 (Google search for it). It is a good test for bad RAM. Memtest runs under it's own OS and does not run Windows or any drivers. If you don't find any bad RAM and no crashes, suspect other hardware or bad Windows drivers.

Boot to MS-DOS (http:\bootdisk.com). Display something like "DIR /W" so you will know if a reboot occurred. If no crashes, suspect bad hardware or bad Windows drivers.

 

[biosol]

Ok, that's all new to me. I made a mistake, I just updated the BIOS to 1.0c, 3/15/04 (because it said this fixed some 800FSB instability), but my board had 1.1 8/14/04 on it and now it doesn't restart.

Is clearing the CMOS good or bad? Will it bring back the 1.1 BIOS?

Thanks

 

[Frank4d]

If you try to boot to DOS (

http://csislabs.palomar.edu/Student/Utilities/boot622.exe)

from a floppy after a BIOS upgrade gone-wrong and it won't even do that... you are usually screwed. Best to return the MOBO and tell them "it doesn't boot".

 

[biosol]

Ok, sorry I spoke too quickly. There is/was an CMOS checksum error - defaults loaded, but I pushed F1 to continue and it booted into Windows. I guess I'll have to contact ECS about trying to get the 1.1 version of the BIOS back as they only have 1.0c and 1.0e on their support site.

I have loaded the drivers from the CD and I guess I'll try that RAM test.

Any other ideas, that's to try to test for stability?

Thanks,

 

[Frank4d]

I found this:

http://www.ecs.com.tw/ECSWeb/Produc...enuID=16&LanID=0&DetailID=423&DetailName=BIOS

I haven't tried it so good luck.

 

[biosol]

Frank, I'm about half way through the memtest and see 2 errors. What info should I write down, memory address, good/bad????

Thanks

 

[Frank4d]

If you have ECC RAM and only see a couple of errors (2-3), and all errors were corrected... you "may" be o.k... maybe. Otherwise return the RAM where you bought it and get it replaced. If you run the test several times and get the same errors replace it anyway.

 

[biosol]

Thanks for that link I've downloaded BIOS 1.1b

memtest now at 68% and 12 errors, and I think it's frozen. no don't have ECC ram, just regular.

Guess I'll tried to update the BIOS back to 1.1 and then run the memtest again. With 12 errors already that should definitely be returned right?

BTW, thank you so much for help me tonight.

 

[Taed]

I had a very bad experience with the new 1.1B BIOS for the PT800CE-A.

My system was working fine with 1.1A, but the new BIOS released a few weeks ago contained new microcode for a few CPUs (including mine, a 3.0E P4 with CPUID 0F41H), and allegedly fixed a problem with DDR 400 memory with a Prescott, which again would affect me.

So, I use WinFlash to program the BIOS, and it says that the program and the verify were fine.

However, I boot, the BIOS reports "1.1B", and then it HANGS. No way to get it back. So, I assume that it's a bad Flashing despite what WinFlash said.

I go buy ANOTHER motherboard, again a PT800CE-A, but it's now a revision 1.0A board versus a 1.0.

I get the system running fine again, but this time use the DOS-based AWD823E.EXE (AWDFLASH.EXE) that is on the ECS CD-ROM. It reports that the Flash and the verify went fine. But, again, I boot, it reports 1.1B, and again HANGS.

So, now I have 2 dead motherboards. Clearly, there is something wrong, and it's not the Flash process.

But, all is not lost! I put in another CPU that I have (1.7 GHz P4, not a Prescott). Both motherboards then boot fine! So, the BIOS is good, but it is somehow incompatible with my CPU!

I sent a tech support email to ECS, but they just said that I must have Flashed the BIOS wrong and they can't help me. Gee, then why does a different CPU work fine?

So, I think that the 1.1B BIOS is somehow whack when combined with either the CPUID 0F41H or with my Prescott/DDR400 combination.

Since two other posters here did manage to get 1.1B working fine, and you both have the Prescott/DDR400 combination, that seems to imply that it's the specific CPUID. Could you both check and report what CPUID you guys have? I suspect that it's one of the other 3.00E Prescotts, namely 0F33H (which did not get new microcode) or 0F34H (which also had a microcode update).

Thanks!

 

[biosol]

Hi, yes my system has been running fine on 1.1b since I exchanged the RAM. How do I check the CPUID?

RH

 

[Frank4d]

In Device Manager under "Processors" my P4 shows as two CPUs (because Hyperthreading is enabled) as Intel Pentium 4 CPU 3.00GHz and (under Details) as ACPI/GENUINEINTEL_-_X86_FAMILY_15_MODEL_3\_0 and ACPI/GENUINEINTEL_-_X86_FAMILY_15_MODEL_3\_1.

You can get the same information from various shareware apps but I prefer not to pay for them.

 

[Frank4d]

Uh... sorry my earlier post did not answer what you are looking for (long day).

EverestPro or WCPUID will tell you the CPUID. Mine is 0F33H.

 

[biosol]

Downloaded WCPUID, but can't seem to find the ID? Where is it?

 

[Taed]

WCPUID doesn't put it in the "standard" format. It's a 4-nybble hexidecimal value of the Family, Model, and Stepping. WCPUID shows those, but in decimal. Thus, you probably see something like Family 15, Model 3, and Stepping 3.

The program CPU-Z is pretty cool as well, and gives more detailed information.

With the CPUID, you can then look up in the Intel specifications which bugs affect your processor, and other super-technical information such as that. Not for your typical user!

 

[biosol]

Sorry, don't know how to convert this, but here's this info-

Family = F (15 I guess)
Model = 3
Stepping = 3

Ext. Family = 0
Ext. Model = 0
Revision = C0

Hope that helps.

 

[Taed]

I did some more whining to ECS (and I assume other customers did as well) and they released a new BIOS yesterday, 1.1d, for the PT800CE-A. I'm scared to try it after the trouble I had last time, though!

It can be found at:

http://www.ecs.com.tw/ECSWeb/Produc...enuID=22&LanID=0&DetailID=423&DetailName=BIOS

 

[Taed]

Well, I tried 1.1D, and it again didn't work for my standard configuration of a 3.0E P4 with either DDR400 or DDR266, but it worked fine with my other CPU. So, same experience as with 1.1B, but I did some more DDR testing. I remain convinced that the BIOS is bad, but only for my CPUID (and maybe 1 or 2 others that were "fixed" in 1.1B). Sigh...

However, you may want to upgrade to 1.1D since it allegedly fixes a memory issue.

 

[biosol]

Thanks for the info Taed. My system's been running fine so I'm hesitant to update the bios. What do you guys think?

RH

 

[franklin97355]

biosol, if it's running well don't fix it. BIOS updates are for those people that have a problem with the computer that was addressed in the BIOS upgrade.