DFI NF4 Ultra-D unstable with errors!!!

[SpecialistX]

yes ive used the search button. but i still am unable to narrow down the problem enough. here is what has been going on. i recently received my dfi nf4 ultra-d board and built by ati x800xl. decided to borrow my friend's fortron 500w? bluestorm psu, the a64 s939 newcastle 3400+ @ 2.2ghz stock. thoroughly tested pqi turbo pc3200 stock rated @ 2-2-2-5. im absolutly sure that there isnt a problem with the ram, psu, and cpu, because ive been running the ram on my own rig and have had no problems since i received it like a month ago.

1) laid it out mobo on my case and plugged everything in, including the 12v pin connector, the 24pin connector to the mobo, a 4pin power connector onto the mobo, set both sticks in dual channel mode in the 2 orange slots. took off cpu from frnd's asus a8v, cleaned off the as5 with rubbing alcohol, and cleaned off the zalman alcu7000 hsf as well. placed some thermal goop onto the hsf, and spread evenly and thinly. next, placed goop on cpu and spreaded it evenly with an id card. locked in cpu and mounted hsf firmly and tightly. made sure x800xl was nicely fit in place. attached the rest of the necessary wires including the hdds. now when i formatted the first time, it crashed in the middle of the format 😕. so then i tried again, this time being able to finish installing windows. installed all the latest drives from the dfi site. what scared me a few times was that the chipset fan and the cpu fan stopped for a few moments, and then started spinning again, but at a very low rate. i notice that the fan always slows down after initial post. when i tried playing hl2, it crashed after a few mins in game. then i rebooted and ran prime95, and then it gives me the following errors (this is with the blending test btw, as it passes all the other ones).

http://www.geocities.com/mkhs_wmic/img_4038.jpg

http://www.geocities.com/mkhs_wmic/img_4039.jpg

http://www.geocities.com/mkhs_wmic/img_4042.jpg

http://www.geocities.com/mkhs_wmic/img_4043.jpg

http://www.geocities.com/mkhs_wmic/img_4047.jpg

despite the fact that i am running everything on stock settings, its still not stable. also, my 2 floppy drives just died out of nowhere. 😡 will a bios flash fix help with the errors? whats funny is that everytime i tried to install on my old 10 gig maxtor, it always gives me the BSOD during format or a little during copying setup files. this is giving me such a huge headache, i just wished that DIY systems werent so troublesome as this is my 2nd build and never have been able to get straight to desktop the first time without any problems. if for some reason the contact isnt sufficient enough on the hsf and cpu, it wouldnt cause any of these kinds of errors right? also, can everyone who has an ultra board please post everything that they did to to tweak it a little for stability even at stock specs?

thanks to all in advance.

 

[NesuD]

well i can't see the images that yo have linked but my guess would be maybe the board doesn't like the ram. try priming it with the test set to be cpu intensive rather than blended and see if it still acts up. another thing you could try is running the memtest86 utility to test just the ram. I know you said that tha ram has been thoroughly tested in another system at 2-2-2-5 timings but with those timings it kind of sounds like it was running in an Intel based system (just a guess) regardless sometimes specific memory just doesn't get along well with certain boards. Have you tried relaxing the ram timings at all? I know the DFI sli board has memtest built right into the bios not sure if yours does or not. The fan thing is is a setting in the pc health section of the bios that set temp thresholds which determine the fan speeds. At start up they probably run full speed until after post then the bios checks for the current temps (the fan pause) and adjusts the fan speed accordingly. you can adjust the fan behaviour by adjusting the fan on and fan off thresholds in the bios. Anyway back to point. I would definately put it through of memtest86 just to positively eliminate the ram from the equation. Also I would try memtest with relaxed timings if any errors occur at 2-2-2-5 timings. It doesn't really hurt Athlon64 systems to run slightly relaxed timings because of the low latency on die memory controller. Looking at the DFI site I don't see an updated bios available for it. Looks like the only bios is the original release which is what your has most likely. dated 1-25-2005. Ok now i can suddenly see the images you have linked. Definately looks like something going on with the memory to me. You don't say but if you are using 2 ram modules try running with just 1 and see how it behaves. Ok just noticed this in the manual "A DIMM'S spd is originally fixed at 1T. When DIMMs are inserted in DIMM 1 and DIMM 3 the spd must be 2T for better system stability. We recommend inserting DIMMS in DIMM2 and DIMM4". Looks like Dimm 1 and Dimm3 don't decently support a 1T command rate but Dimms 2 and 4 do so make sure you are using that configuration or adjust the command rate in the bios to 2T if using Dimm1 and Dimm 3. Hope this helps you.

 

[Dug]

Go here and make bios adjustments.

These take care of 99% of problems with the DFI

http://forums.vr-zone.com.sg/showpost.php?p=635223&postcount=1

Many people including myself have had problems with the stock settings.

 

[SpecialistX]

where can i use the memtest86 feature that comes with the new bios. i have relaxed timings to cas2.5 and cas3 just to give it a try, and still no change. i have also ran tests with a single module, but that doesnt help either, as i still get errors, using 2 diff types of ram, a pqi, and a mushkin value stick. i tried putting the sticks in the yellow slots, but doesnt post. only posts when i put it in the orange slots.

btw i recently flashed my bios to the latest beta bios (2/18), and it seems to have been able to run prime95 without errors. however, the system is unable to post after a formal shutdown. i had to reset the cmos in order to get it to post everytime. i have relaxed timings to cas2.5 and cas3 just to give it a try, and still no change.