High Temps with Prescott Mach II help please

Sorry I wasn't clear

Sorry I wasn't clear before, the first temps are the cpu temps detected by MBM5 (the -15 and the 20) and the others are the evap temps. (the -57 and -39). If it were a gas problem, why would the temp sensor (evap) read like this? How much would it cost to regas?
 
Use 8 washers on each side my friend - your simply have a contact problem - I had the same with my skt 478 prescott, using additional washers really helped.

Give it a try and see how you get on.

Mav
 
Oh, forgot to mention - Don't use Arctic Silver 5 on phase. Use the Ceramique; I asked the guys at Arctic directly and they said so! :D
 
Im having huge problems

I am having huge problems and I hate to say I am on the verge of selling this unit unless I can fix it. This is ridiculous!! (I bought it 3 months ago new!!). I spent my entire day today trying to fix it:

First I shut it down and took off the head, inserted my probe next to the cpu (see pics attached) and then reattached with alot more washers (8 I think). Then when I tried to turn it on the motherboard would turn on but it wouldn't do anything (no beeps even). I then thought it was my psu but that proved wrong...Then I took it all off and made a second copper shim so I had two of them on the cpu. The problem, however, is that both weren't completely flat. Anyway it finally booted but after five minutes of windows it just blanked out completely and wouldn't restart (same prob as before)

I took everything apart and started from scratch with no luck. Finally I took the cpu out and put it under a desk lamp for a little while and them put the desk lamp focused on the cpu socket. Then I replaced the head with only one shim (like I originally had) and I am now typing from this computer! But now my temps are even worse and are not consistent at all. MBM reports between 10-48ºC which is much worse than I had before. I have the motherboard covered with clear nail polish, tons of dielectric grease, and neoprene foam.

I do, however, think I know what the problem is and I hope one of you can help me. I noticed that by adding more washers it will only bring the head closer to the metal plate. There is a pretty big gap between the top of the cpu and the top of the plate. Therefore, no matter how many washers I add it will not get closer, correct? The shims seem to be helping but shouldn’t it be able to make contact without it? The upper and lower plastic mounting brackets are flush against the motherboard as well. I have attached a few pictures that may help illustrate my problem.

Please help me!

Thanks!!

1.jpg


http://default.ga065.org/school/1.jpg

2.jpg


http://default.ga065.org/school/2.jpg

3.jpg


http://default.ga065.org/school/3.jpg

4.jpg


http://default.ga065.org/school/4.jpg
 
There should be no need for the shim and there should be no gap are you sure you have got the right mounting kit for your cpu??

Is the mounting kit screwed together nice and tight also mbm will report the temps wrong

You need to remove the shim and show us the gap between the mounting head and cpu also show us the condition of the actual head of the prom

How did you mount the cactual kit the bottom plate (not the backplate) isn't stuck on any componants like resisters or capacitors etc this may be raising it off the mobo causing the gap you are supposed to cut the small feet off or even cut the plate around components
 
Also reading back over your posts if you are still getting evap temps in the -20 to -30's under load It seems like the unit and temps are fine mbm wont always read correctly
 
The problems you are having are quite common on 478 sockets.I used to use a copper shim like you(about 1.5mm thick),but later on i started to sand away the feet of the plastic front mounting plate by about 1.5mm to get contact without an extra tim joint.You are correct in that no matter how many washers you use,it still wont get good contact without some adaptation.
 
Thanks for your replies!

I have taken more shots so that may help. Just click on the attached links. Fatty, I am pretty sure I have the correct cpu kit (see last picture). In addition, all corners of the plastic plate are flush against the mb so that means there is nothing obstructing it (I assume). I actually measured the gap to the best of my ability and I got something like 1/4 inch (5mm was a better measurement). I have also attached pictures of the head as well. Please advise.

Thanks!

http://default.ga065.org/school/5.jpg

http://default.ga065.org/school/6.jpg

http://default.ga065.org/school/7.jpg

http://default.ga065.org/school/8.jpg

http://default.ga065.org/school/9.jpg

http://default.ga065.org/school/10.jpg
 
Ok,i had this once,it seems some p4 kits(and yours looks like an early one) were once shipped with the wrong metal retaining bracket.There were two types of metal bracket,one had 7mm "wings" and the other had 9mm.You need to measure from the inside face of the bracket to the outer edge of the slot.If it measures 9mm,you will not get decent contact.If you need a 7mm plate,drop me your addy in a p.m. and i'll sort you out one.
 
Well believe it or not I decided to take matters into my own hands. I took rough sand paper and shaved off a layer of the plastic motherboard plate. Then I noticed I did have some ics and other obstructions in the way so I took out my soldering iron and melted little holes so it would fit flush. Now I need no shim and get idle temps of -15º at stock speed. I cranked it up to 4 GHz and idle is -8 with load max at 12 (after two hour burnin). I am satisfied with that but I know it should be capable of more. I got scared when my evap dropped from -48 to -26 during that test. Any more ideas to improve performance at this point? Anyway thanks for all your help!
 
The EVap temp drop shows only one thing - the sensor works very well - mine does not and contact is excellent.

Those temps are very repectable now my friend - I like the idea of melting pots for ic's and the like - in fact I actually dremmelled similar troughs now you come to mention it - hopefully you can get some more out of the cpu -2 under load seems to be the optimum to aim for so keep pushing til you reach that.

Also consider that for short bursts (like benching) you can turn the fans up to turbo to help sustain good load temps at max overclock.
 
Back
Top