Duron question............STRANGE things going on

Asha'man

Senior member
Oct 15, 1999
773
0
0
Hi all!

I have a little problem. Me and my friend just bought a pair of duron 650s, abit kt7s and super orbs. We did the L1 thing on both of them and I actually managed to convince my friend that I am the one to get the fastest setup. My friend was the one who also bought a new case so we started out by putting everything in his case. So we got the first one stable @ 1030 @ 1.85. Temperature readings were max load 54 C and it idles at 36 C or so. He has a middle tower, 250W PSU.

Then I put the other setup in my old case, about the same size but a 235W PSU. I hardly get it stable @ 1000 @ 1.85. I did however not reinstall windoze as my friend did. Dunno if that matters. Anyway, it idles at 52 C and disappears beyond 60 C at full load. So I think that there is probably too little contact between the cooler and the chip so I press the cooler down. Tada the temperature sinks a little, full load stays at 56 C. I take the heatsink off and put some more grease on the thing, after cleaning the chip and cooler poperly. I get the same resutlt, high temp and won't go beyond 1G. So we swith mobo, chip and cooler. Not taking the cooler off. Then the strangest thing happens. My friend easily get what just was my rig to 1050 stable and almost 1060. And not very hot either, goes down to 36 C idle and 52 C at full load.

And the one he got stable and quite cool @ 1030 is HOT in my case.

so we switch back. Then I put in bent buttons made of metal (I don't know what the metal is called in english) between the clip thing that holds the cooler on the socket and the coolplate that presses against the chip. So I get it a little cooler, no problems w/ 60+ C. But still, it idles high as hell. I install rain. All of a sudden I get idle temps at 36 C!!!

But still it is not stable at 1050 in my case, which it was in my friends non-rain system.
I can take it to 1030 but it gets really hot w/ the case closed so I run it at 1G. There it is very stable. I took it through a LANparty, some 60 hours and no problems @ 1G and 1.775 Volts.

But why WHY WHY can I not get it to 1050 and WHY must I use rain to get it cool while idling? And why does my friend not have to use rain. We both use WinME btw.

HEEEELLLPPP!!!!

Thanks

/Asha'man
 

Asha'man

Senior member
Oct 15, 1999
773
0
0
NO no no.

Don't get me wrong. I am of course not disappointed about my Duron hitting 1030!
I am concerned though that it is so damn hot! I don't want to find my little baby fried one morning....

So I am trying to find a reason to why it gets so hot in my case and not in my friend's case.

I must help my poor chip...!!!

thanks

/Asha'man
 

Fish54

Senior member
Nov 19, 1999
253
0
0
Well, here are a couple thought and anyone is more than welcome to criticize these ideas. My first thought was that perhaps your friends case has better cooling all around...at the very least that 250W P/s may have a better fan in it than your 235. Also, if his case allows more space around the cooler or just sets up better airflow. My second thought, and this is where my theory gets a little sketchy ;)... was that perhaps the p/s is causing the heat problems....There is no doubt that the p/s could cause stability problems since it cannot deliver as clean of a signal or as much juice as the 250, but what about this idea... IF your p/s is delivering less voltage...ie. cannot deliver the same change in potential as your friends, it must work harder, which will cause heat, but it will also need to push more current to get the same total amount of juice into the chip... I know its far fetched...but maybe ;)
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
I'm hip to the power supply theory. I really can't see a case making such a huge difference in temps.

Yeah...less voltage means more current. Current = heat. Think about what you have isolated it to: Power supply or case. Try your experiment again with case open and a room fan blowing on the motherboard. This should take the case out of the picture and leave you with only the PS to blame.

spidey

edit - My .02...overclocking to the limits is fine art. Don't skimp on the power supply, it IS important.
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
Not to mention that orb style heatsinks cause higher "under-reads" of cpu core temp by the socket-thermistors.

I would be very, very wary of a c-orb/s-orb kt7 combo reading above 50C. Not htat the heat is a problem, but you could be running 60-65C core temp, and I would personally prefer to keep it lower than that.


Mike


 

monotone

Member
Dec 16, 2000
103
0
0
You should be careful with that 235w psu. The fuse in my enlight psu died from the stress. Now i got a aopen 300w and things are running smooth.
 

Asha'man

Senior member
Oct 15, 1999
773
0
0
Thanks for your thoughts guys, the thing is that I tried using his PSU in my case since I thought that that might help but well, it didn't.

But maybe a 300W should do the work. Going to look for one of those.

I've heard other people talking about the Orbs underrating the temperature of the cpu but how can they do that? Since the thermistor is underneath the chip itself??? ???

Hmm what else. Yeah..there is actually one thing that differs his case some from mine, he has a hole in the back of the plate holding the MOBO, a hole right under the socket. So I'm thinking, could that be affecting the temp. reader or could it actually make the CPU cooler?

Thanks for your answers!

/Asha'man
 

Fardringle

Diamond Member
Oct 23, 2000
9,200
765
126
If he has a hole in his case right next to the socket, then yes that could most definitely be affecting the temperature, as his case is allowing more air to circulate directly on(under) his CPU...
 

natedog

Member
Dec 19, 1999
175
0
0
HAHAHHAHAHA. It's Mike AGAIN! Dood, get a life. You choose to point out how this guy is having problems on HIS cpu, but what about his friends? Working pretty damn good there? What do you think of that? And it is under reading?? Then it is really running 70+ degrees? You're on crack my friend. It's not because of his orb, and it's not because of his power supply (whoever said that). Sometimes man, you just get a cpu that can only handle so much, and your friends can handle more.

Telling it like it is.

Natedog
 

Asha'man

Senior member
Oct 15, 1999
773
0
0
Hey

Nja it's not really like my friend's chip is better, actually, when I put my motherboard, my CPU and my Orb in his case, it runs @ 1050 without a hitch. Temperatures (idle) at 38C or so.

When being in my case, it idles above 50C.
Also I do not take off the CPU when changing case, of course.

That is the strange thing about the whole thing.

I am having a hard time imagining that the PSU (235W) maxes out when my CPU is @ 1GHz.

Thanks for your thoughts.

/Asha'man
 

McBoo

Junior Member
Aug 23, 2000
6
0
0
You might also consider if the Motherboards are of different types -the position of the CPU in relation to the rest of the case could affect CPU temp's. An inch or two different in MB placement will make a bit of difference to how the air is circulating in the case. Also, Temp readings from board to board vary quite a bit. I wouldn't be disappointed with >1 gig performance from a $45 USD processor though... I can only get a stable 978MHz out of my 600 --but I'm pretty happy with that.
 

lepper boy

Golden Member
Nov 2, 1999
1,877
0
76
he said they were both using KT7s... my .02cents is... that both the temperature gauge things are screwy on both motherboards. if both sensors aren't in exactly the same place under that little cpu you got, then the difference in temp could vary.......
 

GaryTcs

Senior member
Oct 15, 2000
298
0
0
Mike might be playing his same old tune, but it is the right one. Orbs suck, because they don't have enough clamping pressure. Speed is the sum of all parts. one bad one stops you. Try the H/S first, so you don't make a keychain. (your friend should too) Try a taisol, or g/w fop 32. Use some arctic silver under them [mike says]. Then increase airflow in your case, get a couple of fans.Then if you are still stuck, try the p/s. My last question is; is 50mhz really worth the hassle? I mean we're talking about a 0.5% increase in performance, for a .75v increase, and a heat problem to boot. your already at 1ghz, at a relatively low 1.75v, leave it there.

Natedog, it's Dude, not dood. And most certainly the hole under the socket in his friends case affects the sensed temp by the mobo. It doesn't however affect the core temp (ok maybe very marginally), just the readings are screwed up. Amd says that the Socket A processors are good well into the 90c range, so 70c is not out of the question.