temp too high Celeron 800 P3V4X

[MRPiGoiL]

I've got a Celeron 800 on P3V4X using Asus S370-DL adapter.

Whether I use the Intel retail HSF or my older GlobalWin FDP32, they both give me very high temps. Running 1066MHz @ 1.7V, idle ~ 37C, prime95 ~ 70C

I'm suspecting it's the thermal paste I'm using is ****. It's that generic white stuff. That definitely explains the Intel retail HSF problem.

But I'm sure the FDP32 can do a better job, so I think contact is not 100%, although visibly it seems fine. But it doesn't seem to clip on as tight as on my Celeron 366. So perhaps the clip is not the right size?

Lastly, I heard there were problems with P3V4X and these S370/Slot1 adapters don't report temps accurately. Could this be the case? I haven't tried to run it at full load for long because I don't want to melt the damn CPU!

BTW, my room temperature is pretty cool ~ 19C-21C And yes, cost is a factor, otherwise I would have just bought the damn 1GHz Celeron without overclocking.

If you've read this far, please reply with some advice. Thanks in advance.

[funnyperson1]

yes, there are problems with that, i believe that you must update the bios, and im not sure if the slocket supports temp monitoring as the temps arent coming straight from the cpu, they have to go trhough the slocket, also dont use ASUS Probe, it reports temps too high (like 10C), use MBM5...

[MRPiGoiL]

Oh yea.... I'm running BIOS 1005, and the CPU is cC0 stepping (SL55R), and I'm using the latest MBM version for monitoring the temps.

I think there's something seriously wrong. These temps are like this whether I run it at 800MHz or at 1066MHz. I mean, whenever I run it at full load (I use Prime95), they both go to 70C. That scares me! Don't want to be cleaning up sticky silicon inside my box!

[Yodums]

False temperature reading..

I guess you can update your bios see if there are any fixes... If it was really 70 C by now it would have fried long time ago.

[MRPiGoiL]

I just found a 1006 beta 2 BIOS for the board, but it didn't have any documentation. I flashed it, and there's no difference.

But yea, I'm certainly afraid of it frying on me. Do you guys have any suggestions on how to find what the problem is? I've tried to state the factors I could come up with, but it looks like I will have to spend extra $ to test.

Originally posted by Yodums
False temperature reading..

I guess you can update your bios see if there are any fixes... If it was really 70 C by now it would have fried long time ago.

[funnyperson1]

Originally posted by Yodums
False temperature reading..

I guess you can update your bios see if there are any fixes... If it was really 70 C by now it would have fried long time ago.

thats not true, i had mine running at like 65-75 and when i upgraded my heatsink i found theese to be true numbers, also that Cumine can take torture, but i suggest upgrading your heatsink, if its aFCPGA get a CBG-38 from www.svcompucycle.com you can get it for 18 bucks shipped and its all copper!!!

[Yodums]

Originally posted by funnyperson1


thats not true, i had mine running at like 65-75 and when i upgraded my heatsink i found theese to be true numbers, also that Cumine can take torture, but i suggest upgrading your heatsink, if its aFCPGA get a CBG-38 from www.svcompucycle.com you can get it for 18 bucks shipped and its all copper!!!

Yes but there's no temperature difference whether its retail or his Global win.

[funnyperson1]

hey the FDP-32 isnt contacting it properly, look at this:
http://www.overclockers.com/articles250/
specifically this line:
" I had to modify the FDP32 to accommodate my celly FCPGA. All I did was put a standard copper cold plate under it. "

also:
http://www.neoseeker.com/resourcelink.html?rlid=12475

this will probably fix your temps....

[Pinky]

If your PC isn't crashing as windows boots, then you're not hitting 70C... lol... I get nervous at 50C !

with a 37C idle temp you're probably in the 45-48C range under full load (assuming everything else is in-line).

[funnyperson1]

Originally posted by Pinky
If your PC isn't crashing as windows boots, then you're not hitting 70C... lol... I get nervous at 50C !

with a 37C idle temp you're probably in the 45-48C range under full load (assuming everything else is in-line).
mine was completly stable when at temps of 65+.....i tihkn his problem is that the FDP-032 is built for PPGA celerons and does not contact properly....look at the links i posted...

[ol' man]

Yeah dude your HS isn't touching your core. Let me redirect you to this thread. (http://forums.overclockers.ws/vb/showthread.php?threadid=46893&pagenumber=2)

I think the fella in that thread is having the same problem as you. You have to have the HS touching your core absolutely flat or else you will have problems like you are talking about

If you had a AMD chip this is what would have happend to it.

here (http://www.sharkyforums.com/ubb/Forum15/HTML/002200.html)

You may need to bevel the side more or something where it is cut out to allow for the socket. Look at all sock 370 HS's. They are milled on one side. Sometimes maybe they don't get it milled all the way or something. I am not sure. SOmetimes though too I have noticed that even though I am careful I still don;t get the HS on without it havine a gap. I would say it takes about 3 tries every time to get it right with the ALPHA PAL on my little tualatin core. I check it by hitting the bios immediatly upon boot to see if it is all the way on. I also do it out of the case with a power up switch devoid of any case;) and a psu also. It is harder to get the HS directly on the cpu core good in a case especially if the case is a tower and standing up right. Now if you got the IHS that is a target that you can't miss;) But at a price too!

[MRPiGoiL]

Thanks all for you help! I will try all of this and see how it turns out. I'm sure it'll turn out. Because I'm running @ 800MHz running Prime95 now, and MBM is showing 67C-73C. Damned amazing, I'm afraid by the time I get the copperplate onto the FDP32, the CPU will be calculating 1 + 1 = 2.25

So I have anymore problems I will for sure post again! Thanks!

[funnyperson1]

oil' man this is probably similar to your situation because the older ppga celerons that the FDP-32 was made for had heat plates on top....

[ol' man]

SHi# yeah that makes sence. Instead of putting a copper plate underneath the CPU I would simply sand down the bevel a little more then you temps will be lower. If you go and put another thermal joint between your core and your HS it will certinly help but not completely give the best temps either. You would need a very flat peice of Copper and will most likely have to sand it too. If you are going to sand it might as well sand the bevel a little more or maybe use a grinder. I know those things are not always avaiable too. A file would be ideal as it would be fast and also not muck up the HS as much as a grinder would. I don't care what the inside of my case looks like as long as it performs. The duct tape and cardboard would tell you that. Here is a pic of a HS. You will notice the BIG A$$ arrow is pointing to the bevel I am speaking of right. Maybe that is not positioned right or not big enough. Quite possibly your HS is not making good enough contact with the core period and the bevel is big enough and in the right position. I have done it myself as I mentioned it was just that when I applied the HS it was off center and whenthe clip was attatched it stayed that way.

Here is the pic anyway. Be careful with those cores they are not real tuff!

[Pinky]

Originally posted by funnyperson1

mine was completly stable when at temps of 65+.....i tihkn his problem is that the FDP-032 is built for PPGA celerons and does not contact properly....look at the links i posted...

It's hard to believe any cpu can run that hot for any stretch of time, I've had some of the older celerons crap out at 51C just because they were too hot, and I think those were rated to run hotter...

Hell, it aint good either way... I think ol'man, as usual ;), has the right idea(s).

[CrystalMethod]

Originally posted by Yodums
False temperature reading..

I guess you can update your bios see if there are any fixes... If it was really 70 C by now it would have fried long time ago.
I had the same problem with the P3V4X, with my celeron 800. Temps reported by MBM were way off. I tried everything I could think of, different heatsinks, reseating the heatsinks, about 6 different thermal greases, case sides off, more fans for ventilation, stacking fans, you name it, I tried it. The adapter is probably throwing the readings off. Best thing to do is get a "Compunurse", or "Digital-Doc". If you can't afford that, you can make a temp monitor from a indoor/outdoor digital thermometer. Hoot posted a great article a few months ago on how to make one.

[ol' man]

Yeah crystalmethod is right possibly too. Check to see if your thermal paste is fully squished:D

As usual I got done writing a long reply and the bug that comes and tells you you have to log back in came and stole all that I wrote.

Thinking about it I don't see how you could possibly run that chip at that speed without the HS touching the die perfectly. It must be fine and the software must be off. But as I said above check the thermal paste to see if it looks allright. If it is cockeyed then the thermal paste will have some buildup on one side.