Correcting CPU diode temp reporting inaccuracies

Want to talk about one of the articles in SPCR? Here's the forum for you.
MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Thu Jun 08, 2006 4:29 pm

computergeek22 wrote:I checked my temps in the morning when the temps in my case and my processor were most likely at the ambient of my room. It turns out that at idle the mobo is misreporting my case temperature by approximately 3-4C (nxp101 says 27C while my mobo reports 31-32C). This is at ambient room temp of about 24C taken at the other end of the room.

So it seems at idle my mobo is misreporting by a lower number than at load. This seems extremely unlikely and wrong.
I'm sure Russ will respond if he's not back climbing a mountain somewhere, but I don't see why this is extremely unlikely and wrong. Nonlinearities in thermal sensing devices are pretty common. Also, unless the 2 sensors you mentioned are in exactly the same position, they will likely be subject to different secondary heat influences (on the board, from peripherals, etc) at different loads.

computergeek22
Posts: 237
Joined: Thu Oct 06, 2005 3:18 pm

Post by computergeek22 » Thu Jun 08, 2006 5:20 pm

He mentioned in the article that temps are something that shouldn't always be taken for their face value as long as the system is stable. My temps hit a maximum of 62C, measured from PCProbe; if this temp is adjusted according to my findings, it is just within my Tcasemax of 49C (the maximum temp my opty165 can reach while maintain stability?). I don't know if I should worry about the temperature too much. In one sense, I am worried that 62C seems really hot but then again the system is dual prime 95 stable for 8+ hours...

With regard to the two sensors I previously mentioned, they are not in the same place so I guess it isn't safe to assume that the mobo, at least at idle, is misreporting temps. Then again my motherboard temps are sometimes hotter than my cpu temps which seems interesting. I thought, maybe, the duct was the reason for this but that's weird that my mobo would be warmer than the processor internal diode temp by a delta of sometimes 4-5C...

I'm beginning to lose trust in all temperature readings including my video card temp which according to the nvidia thermal sensor doesn't go above 50C, even when gaming, with my vf900 @ 5v...

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Post by Rusty075 » Thu Jun 08, 2006 6:29 pm

Broadcasting from deep in......the suburbs. (I'm back from the hike...broke the left ankle for the second time, and decided that that was a sign to come home)

A couple of things that might be screwing with the results:

1. Is Cn'Q active on your system? (or Optimized Power Management, as they call it for the Opto's). That would seriously skew the idle temps.
2. Is the mobo adjusting the fan voltage?
3. What's the heatsink? There was some theorizing back after this procedure was written that it might not be accurate for heatpipe equipped heatsinks. the procdure assumes that the cooling system efficiency ramps linearly with the wattage, but that's not necessarily true for heatpipes.

computergeek22
Posts: 237
Joined: Thu Oct 06, 2005 3:18 pm

Post by computergeek22 » Thu Jun 08, 2006 8:37 pm

Cnq disabled. Fan voltages adjusted thru rheobus so mobo isn't changing fan voltages. Fan voltages were kept at the same voltage throughout testing. I have a scythe ninja so your hypothesis for the heatpipes is applicable to my setup. (Un)luckily I don't have a non-heatpipe cooling solution to test your hypothesis as the opty comes standard with a decent amd heatpipe solution.

I don't know 60C load for a Tcasemax opty of 49C doesn't seem to fit. Though thankfully, with recommendations from spcr my computer is at least whisper quiet. Not silent but almost inaudible from a meter away. Great for the performance I'm getting out of this thing...


*note* pillows help muffle hd noise in the p180!

Post Reply