flyin15sec

I've had great success playing around with my backup Intel build. However I'm not having any luck with my AMD build.

Parts:
AMD 5400+ Brisbane (OEM)
MSI K9A2 CF-F version1 (790X chispet)
Zalman 9500A cooler
GSkill DDR2 800mhz
Visiontek HD3870
PC P&C 750W
WD 40gig SATA

The system is not in a case and is just sitting on a table. I was pretty stoked, of the all "red" components, and I was hoping to get a red case to match the parts. Perhaps make an AMD theme system. Sadly the plans are on hold currently.


I manage to get it up to 3.2ghz (230 * 14). OS is WinXP Pro. I start Prime95 Small FTT, left for an hour. Came back and the system was shut off. When I try to power it on, it tries to POST, but cuts off power right away.

I tested the PSU in another system and that is working fine. Reset BIOS several times already. Still no luck. This leaves just a few things left unchecked.

A bad CPU or bad motherboard. I'm gonna try the CPU reseat later today.

One thing I noticed is that the temp sensors are faulty on these OEM Brisbanes. BIOS will sometimes read the temp fine at 24C - 36C, depending on ambient temps. Overdrive, Core Temp 0.99.3 will read as -7C and -9C on the cores.

Once my BIOS read my temps as 246C. Windows took 10mins at the splash screen, at that point I just power cycle and reset CMOS. Kept trying for a few times, finally it came back up normally. I figure CPU must be throttling, and running at like 10mhz or something. You guys think that may be the case again? BIOS is reading high temps, so it is shutting the motherboard down?

The system as been working fine for about 6months now, but in a case and only slightly clocked at 3.0ghz. I took it out of the case, since I'm using the case for a different system now.

Attached Thumbnails

   

thlnk3r

Flyin15sec, the odd temperature readings are mostly due to the chipset having difficulties getting the right reading from the processor (on die sensor). This is well known issue for the Brisbane core. Updating the BIOS sometimes resolves this problem.

If you have a multimeter handy you could easily test the power supply. Even though it works in another system it would still be wise to check the rails to make sure they are within spec. Here is a great guide that explains this procedure: http://www.clunk.org.uk/forums/guide...ing-guide.html.

Have you tried just installing the essentials(processor, power supply, 1 stick of memory) to see if the machine POST's correctly? Disconnect all optical drivers, card readers, hard drives ect.

Let us know

Good luck

__________________
AMD Overclocking Guides · Terms of Service · My CM690
Socket 939 Appreciation Club and Knowledgebase

flyin15sec

I don't have a multi meter on hand. However thanks for the link, I'll bookmark it for future reference.

flyin15sec

NVM. After reseating the CPU, moving the PCIe video card to the other slot, unhook the HD, used only 1 memory module the Mosfet blew up on me.

Guess I'll contact MSI to see about an RMA.

Kinda explains why 3 weeks after I bought it, Version 2 was released right away.

Attached Thumbnails

 

thlnk3r

Flyin15sec, wow that is extremely dangerous. Either that motherboard came from a bad batch or you have a bad short somewhere. I understand you have it out of the case but you can still experience shorts. Try not to run that PSU on any other piece of device until you properly test it.

Here's a link that will take you to MSI's RMA page: http://www.msicomputer.com/msi_user/rmaform_p.aspx.

Good luck

__________________
AMD Overclocking Guides · Terms of Service · My CM690
Socket 939 Appreciation Club and Knowledgebase

flyin15sec

I've already gotten a reply back from MSI with a RMA number. That was pretty quick.

I use to work for a UPS company, so seeing mosfets go up in flames is quite normal for me.

I'm fairly certain, this is a design flaw with the motherboard. I've read a few reports of users putting in Phenoms on version1 and having the mosfets go up in flames too. Just didn't think it would happen with a dual core.

well regardless, I hope MSI warranty the board and send me version2. It has mosfet coolers on the board already.

thlnk3r

Flyin15sec, I did some googling and apparently this is a widely known issue with version 1. I believe version 2(MSI K9A2 CF-F V2) came with mosfet coolers (heatsinks) which helped prevent these issues that occurred on the version 1 motherboards. Some people were actually putting mosfet coolers on their V1 boards which does help.

Hopefully you receive a V2 back from MSI

Good luck buddy

__________________
AMD Overclocking Guides · Terms of Service · My CM690
Socket 939 Appreciation Club and Knowledgebase

venom55520

makes me kind of scared that my biostar doesn't have them

__________________
some people say i'm unmotivated, i say i'm just layed back
NOTICE: i am too poor to be a fanboy, and would go for different companies if they provided better performance for the cash. have a good day

thlnk3r

Venom55520, a lot of motherboards do not have heatsinks on the mosfets so it's not completely uncommon. There is a few mosfets on my DFI NF4 that should probably have heatsinks but they don't. As long as you have adequate airflow in your case then I don't think there is nothing to worry about. Having dead spots in your case could also be a culprit. In regards to the OP's case, I believe he just received a board that obviously had insufficient cooling.

Good luck

__________________
AMD Overclocking Guides · Terms of Service · My CM690
Socket 939 Appreciation Club and Knowledgebase

venom55520

thanks, that soothes my worries a little

__________________
some people say i'm unmotivated, i say i'm just layed back
NOTICE: i am too poor to be a fanboy, and would go for different companies if they provided better performance for the cash. have a good day