875P-T Norhtbridge support suggestions

[MikeyLikesItSI]

Oh i believe you, maybe the microcool wasn't up the the task. I don't think the microcool did a bad job, i just would prefer my NB less then 120F. I am getting less errors and less BSOD with the VAGP down to 1.6 at 270FSB. I think when the waterblock/heatsink isn't seated perfect, the extra VAGP builds extra heat and makes it error, given the same improper seating, at least VAGP, there won't be so much heat build up. Just a thought

[Angry_Games]

120F = 49C or so...not anything wrong with that at all...the NB chips will work easily up to 70C

 

i am thinking (and no insulting your installation as im sure you are as good as i am) that maybe you just didnt have it set correctly on the chip?

 

i HATE HATE HATE the stupid paper-clip pin retention thing on the Intel P4 boards...i much more prefer the AMD style with the mount holes...the NB-1 gave me some hassle at first but it went on perfect and i would say my NB temps with fan at full speed on the 875P-T, 260x14, 1.7v vAGP were about 40C...(didnt measure but it was nice and cold when fan was full speed heh). With the fan only running minimal (about 2000rpm or so...very silent) the temp was only about 45C (it was really as cool as it was on full speed to my fingers)

 

 

anyway, not trying to push or say you stink haha...just letting you know my experience with it. DangerDen offered me a NB waterblock to go along with my cooling kit for the 875P-T but i declined as i didn't trust the stress on that weak pin-mount that the NB has

[fordf250]

Mike I have the swiftech MCX159-R chipset cooler and with no fan just the xp120 air at 1.6Vagp the base is 32C and at 1.8Vagp the base is at 36C.Putting the fan on doesn,t change it. Bumping it once will destry the thermal paste and will cause errors even once it has reseated and will then need new thermal compound applied.

[vanovich]

#1 why dont you take the white neopron the original nb køler has underneath , and stabilize it with that ? the best waterblock ive used was zchip dangerden , there was 4 smart loops that was pressured with springs . that was on ic7 boards . dont know with 2 loops though ?

[fordf250]

What about the neoprene and epoxy two more mounts to the motherboard? They would not have to much tension on them just enought to stablize the NB.

[MikeyLikesItSI]

The epoxy idea sounds pretty good. Only problem is, even if its not supposed to provide any of the load, just stabilize it, what happens if it gets bumped? you know, needs to take the load for a brief second. I'm still going to try the foam support. I got the waterblock in what seems to be the best position, i was able to clear up all errors up to 280 FSB. I could boot to the desktop at 280FSB, but would instantly go blue screen. even at 1.650 vcore didn't help. I'm not very stable at 275 either, with 1.550 vcore. I think this CPU is about at its end.

[red930]

With the foam support, make sure you pay attention to the type of foam you are using. Many different types can carry a static charge or don't have a low enough compression ratio or trap heat. Here is a link to a good resource for checking out the different types of foam:

 

http://www.bobsfoam.com/detail.html

[MikeyLikesItSI]

Sweet jesus! Where were you when i needed you 2 days ago Ace? LOL I did some quick research and read all about the ratings of foam/neoprene, electrical conductivity etc etc. enough to feel comfortable with ordering stuff. I got some in today but can't play with it due to GF birthday. I only got two different densities, but both are 1/8" thick, they are a little thicker then i wanted to use, i was going to try the 1/16" first. I'll look things over later.

[MikeyLikesItSI]

Well, the 1/8th inch stuff was too thick to place around the core itself. Fortunately the water block hangs over the edges of the north bridge chip itself, and then the 1/8th inch stuff was perfect height. Unfortunatly it doesn't do anything for the stability of my CPU, but, it did substantially change my findings on memtest. After an hour of running memtest, it only generated 2 errors on test 7, that was a far cry from the hundreds it was doing before. Either way, i'm on 1.425 vcore, 2.6 vmem, and an astounding 1.5 vagp. Perfectly stable so far. I don't know if its the CPU holding me back, or if this board is just still not stable. I'm pleased with the results i found with memtest, it did everything i expected it to do.

[red930]

MikeyLikesIT: Look at the density properties of the foam you used. Stating that you are using 1/8th inch foam don't mean (with all due respect) a damn thing to me. It is the retention properties that I am more concerned about.

 

If you take a sensor measurement from a well placed sensor (right at the edge of the the core) you will find that the type of foam you are using is not all that great given the fact that it has reduced your Memtest errors.

 

If you want, I can send you a patch of foam that WILL reduce your memtest errors as well as extrude heat as long as you have a well ventilated case.

[fordf250]

Mike

 

Did you check to see if it helped with the 5/4 divider also?

[The Stilt]

Take the original neoprene "shim" from the NB cooler off with a knife, and put it on NB. I have even used it with Dothans, and it has always worked.