Strange Temp readings

[Elcs]

As some of you know, I got the stock heatsink with my AMD X2 AM2 6000+ and expected it to cool reasonably and be pretty quiet. Didnt turn out that way.

I was idling at around 48-52C on both cores with the stock heatsink and stock goop. Awaiting my Ultima 90 and bits and bobs, I decided to clean my CPU and Heatsink and reseat with AS5. I also blew out any dust that had accumulated.

After reseating (following the recent Thermal Paste roundup threads instructions on application), my temps were the same.

Oddly enough, after booting up yesterday, I was getting temperatures of around 39-42C. Played games, rose up and then collapsed back down. Prime Dual Core edition was making my load temps around 55-57C agter two hours. Turning off Prime, my temps plummeted back down to ~40C.

This morning, my idle temps are back to around the 48-52C mark and I have not touched/moved anything besides the power on/off switch in between.

Any thoughts?

All temp measuring was handled by CoreTemp 0.95.4 and Everest 4.20.

 

[Conjugal Visit]

No ideas. Room temps were consistent? That's all that comes to mind re: the temp changes.

 

[JBDan]

Your "yesterday" temps sound right on to me. But the others are high. Only thing I can offer is what happened with my 3800+X2 and what many people were getting with their 3800+X2 chips a year or so ago...poor contact with the cores and underside of it's IHS. Definitely not for the faint of heart, but I removed the IHS and "re-sat" my HSF atop the 2 cores and sure enough temps dropped over 10C to a ~normal idle/load temp for that chip. So after HSF removal grease spread was even and uniform on HSF base and IHS?

 

[Elcs]

Conjugal Visit: Although this is not scientific, it feels roughly the same temperature as it did yesterday. I was expecting all day that these ~40C idle temps would increase to 50C idle temps.... which would suggest my case isnt exhausting the heat fast enough. This did not happen.

JBDan: *shudders* IHS removal is rather nasty and something I cannot afford to do and mess up. Although it is a nice idea and core to heatsink contact is always a wonderful thing, I am more looking It is odd though that one day it is one and one day it is the other. Yes, stock goop was pretty well applied and my application of AS5 was nicely spread. Although it is my first IHS application of AS5, I have read up on the advice and my old single core application experience gives me confidence that I have applied the AS5 correctly.

What really throws me is the difference in temperatures, its crazy. In both instances (~40C idle and ~50C idle), the load temps are 55-58C. After reseating with AS5, I am still seeing no difference in temps Vs. stock goop.

Can anyone explain that? I had a friend (whom I dont believe) that told me I should email AMD support, see what they say and perhaps I will get a replacement. The hole poked in his theory is that its performance is perfectly alright, temperatures are well within prescribed limits, Im running the stock fan at 25%, im using AS5 and I dont think potentially anomylous readings from a Thermal Diode would classify as a reason for RMA.

 

[Elcs]

Mini Update:

Got a basic thermometer. Time is 9:47am, 23C ambient, 50C and 48C on the cores idle (25% fan speed, ~2300rpm). 512mb 8800GT OC idling at 56C @ 40% fan speed.

I cant grasp it.

 

[JBDan]

Originally posted by: Elcs
Mini Update:

Got a basic thermometer. Time is 9:47am, 23C ambient, 50C and 48C on the cores idle (25% fan speed, ~2300rpm). 512mb 8800GT OC idling at 56C @ 40% fan speed.

I cant grasp it.

Wow @ idle I was thinking...could your cpu fan be cutting off somehow or your case fan shutting off....*still thinking* obviously you monitor all that as I would if I had crazed temp readings as you How did your thermometer readings directly compare to 0.95.4 CT?

 

[Elcs]

Everest reports my Stock fan (4-pin hooked up to board, ran by BIOS) running at 25% with a target temp of 50C *scratching head*

Obviously my BIOS doesnt think my CPU is really over 50C otherwise I'd expect it to ramp my fan up. I am hearing no such ramping.

My exhaust fan (120mm Antec Tri-Cool @ Medium as my 120x38 Panaflo L1A died) is running off a Molex (Seasonic S12 600W) and is running fine. My 120x38mm Panaflo L1A at the front of my case is running on a Rheobus, twiddling the dial ramps it up and down so that works. I also have an 80mm L1A hooked up on the side panel (Antec SLK3000B case) where the CPU duct was.. pointing as an intake.... it makes zero difference in CPU temp whether its on or off.

My CPU is at stock speed, undervolted at 1.35V compared to 1.4V stock.

Time 4:10pm.... Ambient 24C... Cores 50C and 47C... GPU 57C. Everest echoes the core temps... GPU is incorrect due to me using an older version of Everest which detects 8C below what it should be.

Next idea? Run with side panel off? It doesnt seem to be related to heat recycling I dont think as even if I stress it hard (Prime Dual Core) the heat disappears and the CPU returns to similar idle temps within a few minutes.

Thank you for your assistance JBDan. I wish I had more help as I am out of ideas.

EDIT: Everest Screenshot - Please note as above, the GPU Core Temp is 8C too low on Everest.

 

[JimiP]

After you removed the waxy pre-applied thermal pad, did you use rubbing alcohol and slowly rub out as much substance as you could from the base? That's the thing with the pre-applied thermal pads... once you use them for a length of time, the thermal "wax" melts into the porous metal.

Not that this is really important, considering you are getting good temps right now as it is. But it wouldn't hurt to make sure the base of your HSF is cleaned off with rubbing alcohol and then tinted with AS5, then apply the AS5 to the CPU and re-installed it. You might even have lower temps.

 

[Elcs]

Originally posted by: JimiP
After you removed the waxy pre-applied thermal pad, did you use rubbing alcohol and slowly rub out as much substance as you could from the base? That's the thing with the pre-applied thermal pads... once you use them for a length of time, the thermal "wax" melts into the porous metal.

Not that this is really important, considering you are getting good temps right now as it is. But it wouldn't hurt to make sure the base of your HSF is cleaned off with rubbing alcohol and then tinted with AS5, then apply the AS5 to the CPU and re-installed it. You might even have lower temps.

Yes, I used high % isopropyl to clean it all off, gave it a twice over like i do with any heatsink. Also made sure i gave the CPU a good look at.

Both mating surfaces were 'fresh' so to speak before AS5 was applied (no tinting as you know I have just discovered this in another current thread).

48-52C idle and 57C'sih load temps with an ambient of 23-25C and on an under volted, stock speed chip arent exactly great

 

[Bluefront]

I would discount any TIM problem......the stuff won't change from day to day.

A fan might though.....are you monitoring RPMs? Maybe the CPU fan is slowing down for any number of reasons.

More likely......a software problem in your temperature monitoring program, or perhaps a conflict between two different programs. A flaky temp sensor on a chip or a board can return inconsistent results. The physical temperature might not be the problem at all.

 

[Elcs]

Originally posted by: Bluefront
I would discount any TIM problem......the stuff won't change from day to day.

A fan might though.....are you monitoring RPMs? Maybe the CPU fan is slowing down for any number of reasons.

More likely......a software problem in your temperature monitoring program, or perhaps a conflict between two different programs. A flaky temp sensor on a chip or a board can return inconsistent results. The physical temperature might not be the problem at all.

My GPU and CPU fans are hooked up to their respective boards so RPM's are monitored. No fluctuations as GPU is manually Rivatuna controlled and the CPU is BIOS controlled, effectively staying around the same speed consistantly. My other system fans are hooked up to my Rheobus and simple dialling gives a noticable change in sound, so those fans are working fine.

The heatsink itself gets quite toasty which leads me to believe that the contact is good and that if id boost the fan speed (waahhh noise) then Id get lower temps.

I have to say, considering the speed of heating up and cooling down, its phenomenal. I just turned off Prime95 and both cores dropped from Load temp to Idle temp (~7C delta) in under 60 seconds.

Can I conclude a dodgy temperature sensor?

 

[JBDan]

It is looking that way geesh what else could it be Seems so rare since I've been in this forum. Wondering is a butched temp sensor rma qualify?

 

[Elcs]

Originally posted by: JBDan
It is looking that way geesh what else could it be Seems so rare since I've been in this forum. Wondering is a butched temp sensor rma qualify?

It probably is. The product is defective and detracts from the safety of said productl; it could in theory burn a hole through my socket, fry my motherboard and take out anything from there.

Proving that its fudged, is another thing. Really, all we have proven so far is that we dont know what is wrong with it and our best diagnosis is.... Its the temp sensors fault.