AXP 2500+ getting too hot

[upsciLLion]

I have an AMD XP 2500+ being cooled by an SLK-800 using an SMB 80 mm fan which is similar to a Panaflo L1A. All of this is inside of a Lian Li PC65 with two 80 mm fans in the front, an 80 mm in the rear, and an Antec TruPower 480 just above the rear fan. This stupid thing idles at 51°C!!! The case temperature is 35°C right now at idle. I've used an air compressor to blow dust out of the case and off of the heatsink, and I've reseated the heatsink and reapplied thermal gel. Do any of you have any suggestions?

Thanks for any help you can give,

ups

 

[rgreen83]

You didnt say what motherboard you were using but most report higher temps than actual, some are more notorious than others. Either way, if your computer isnt crashing or producing errors it isnt a problem. If you are certain the heatsink is mounted as flat as possible and used good thermal material, I would let it go. Barton's are designed to handle up to 85C anyways. Mine is watercooled and my motherboard (nf7-s 2.0, a particularly bad one with temps) reports back 45C idle/49C load. But the thermal diode mounted on the side of the die reports always around 35-37C.

Long story-short- dont worry about it.

 

[ToxicWaste]

35 C case temp is not great, especially if it's winter where you are. I know my case can get that warm during the middle of summer. What's the ambient temp in the room? I put a couple more fans in the side of my case, one over cpu, one over PCI slots, and it dropped case and cpu temp about at least 4 C. I'm at 26 C Case and 42 C CPU at 100% processor usage (running seti@home in background...) but it's cold as hell in my room.

What thermal paste did you use and how did you apply it?

Are you overclocking your CPU, if so, by how much (FSB, mulitiplier, vCore)?

What monitoring utility are you using to get you temperatures?

greenie makes some valid points, but I wouldn't dismiss the high temps quite as quickly as he did...

 

[Jeff7181]

If you're reading the diode temp then that's fine... if you're reading the socket temp, your core temp is probably at least 5 degrees C higher, which still isn't a problem if you're not getting errors. I wonder how so many people were misinformed that a processor above 50 degrees C was such a horrible thing?

 

[JustStarting]

Is this the first time you've been monitoring the temps? What have previous CPU's read for temps. If you are using anything but AS, I'd get some and reapply it. Todays CPU's have very small die surfaces, and you must use the best compound possible for an interface.

If you are using MBM5, do you have the correct temperature monitoring standard applied?? There are 4 to choose from. I don't know what the 8RDA calls for. Anyone???

 

[DAPUNISHER]

Originally posted by: Jeff7181
If you're reading the diode temp then that's fine... if you're reading the socket temp, your core temp is probably at least 5 degrees C higher, which still isn't a problem if you're not getting errors. I wonder how so many people were misinformed that a processor above 50 degrees C was such a horrible thing?

Back when sktA was new platform I remember quite a few posts of people having stability problems with temps in the 50's and lowering temps 5-10c with better cooling. ect would result in a stable system. Sometimes just prying the socket thermistor up a fraction on boards that had them back then would give higher temps just by getting it closer to the chip. I think that the temps were so far off actual back then that it just became a rule of thumb to keep it under 50c to ensure stability is where that came from. I recall my t-bird 1.4@1.54ghz on my K7S5A would get flakey at about 60c indicated using a cheapo OCZ monsterII HS/F, I switched to a Alpha 6035 with 60mm delta black label and original AS and no more problems except it sounding like I lived by the airport

 

[Jeff7181]

Originally posted by: DAPUNISHER

Originally posted by: Jeff7181
If you're reading the diode temp then that's fine... if you're reading the socket temp, your core temp is probably at least 5 degrees C higher, which still isn't a problem if you're not getting errors. I wonder how so many people were misinformed that a processor above 50 degrees C was such a horrible thing?

Back when sktA was new platform I remember quite a few posts of people having stability problems with temps in the 50's and lowering temps 5-10c with better cooling. ect would result in a stable system. Sometimes just prying the socket thermistor up a fraction on boards that had them back then would give higher temps just by getting it closer to the chip. I think that the temps were so far off actual back then that it just became a rule of thumb to keep it under 50c to ensure stability is where that came from. I recall my t-bird 1.4@1.54ghz on my K7S5A would get flakey at about 60c indicated using a cheapo OCZ monsterII HS/F, I switched to a Alpha 6035 with 60mm delta black label and original AS and no more problems except it sounding like I lived by the airport

My aunt had a HP with an Athlon 1.2... she loved how quiet it was... it barely made a sound. I happened to be over once and touched the computer and the case was hot. My uncle had one of those infrared thermometers... the kind you just point at something and it reads the temp... the outside of the case was 40 degrees C. So I opened it up and the fan on the CPU wasn't spinning... it wasn't even dirty, so I concluded that it never worked since there wasn't any dust in the fins of the heatsink. I checked the temp of the heatsink... did my best to aim between the fins and record the temp of the base of the heatsink, and it was reading about 80-85 C depending where I pointed it. The computer was over a year old at that point, so it had gone a whole year at 80-85 C without a problem. I blew on the heatsink for about 10 seconds and the temp dropped into the mid 60's, lol. So yeah, I replaced the fan for her just to keep it cool, but it didn't seem to have a problem running at that temp for so long.

*EDIT* DAPUNISHER... also remember that the readings you got were before the days of thermal diodes inside the core... so you were reading socket temps, which means the core temp was at least 5C higher, possibly as much as 10 or 15 if (as you were saying) the temp sensor wasn't positioned correctly on the motherboard.

 

[DAPUNISHER]

Yeah, those old AMD's sure ran hot I also think that not only was the in socket thermistor positioned poorly quite often but that some of the board makers did an awful job of implementing the feature, heck I bet my K7S5A temp reading were off by even more than 15c for the core in that t-bird I had :Q