Asus Probe temperature off by 10c

thanakornl

Weaksauce
Joined
Dec 10, 2008
Messages
105
I just got my p5q pro from asus and installed the PC probe. Trying to measure the temperature using pc probe and compare to core temp, hw monitor the temperature seem to be lower than core temp and hw monitor by 10 c.... is there anything wrong with my motherboard???
 
No one seemed to have any solutions here. Mine only went screwy when I upgraded from an E6400 to Q6600, and it's affecting my Q-Fan operation.
 
PC Probe measures the temperature from a sensor built into the motherboard, which is not actually capable of giving an accurate reading of your CPU temperature. To get proper temperature values, you should use either Real Temp or Core Temp.
 
And ignore idle temps as there is no calibration for anything less than TjMax temps.
The closer to TjMax you are, the more accurate the temp.
Or the further away (under) you are from TjMax, the more unreliable the temp reading.
 
PC Probe measures the temperature from a sensor built into the motherboard, which is not actually capable of giving an accurate reading of your CPU temperature. To get proper temperature values, you should use either Real Temp or Core Temp.

I am under the impression that PCProbe and the BIOS (hence Q-Fan) use Tcase, not a motherboard sensor. Obviously Tcase isn't as accurate as CoreTemp, but -17, or -10C is a little extreme.
 
I am under the impression that PCProbe and the BIOS (hence Q-Fan) use Tcase, not a motherboard sensor. Obviously Tcase isn't as accurate as CoreTemp, but -17, or -10C is a little extreme.
Tcase is not measurable short of drilling a channel into the heatspreader and inserting a thermal probe (which is how Intel measures it). Intel CPUs do not have sensors to measure Tcase.
 
Tcase is not measurable short of drilling a channel into the heatspreader and inserting a thermal probe (which is how Intel measures it). Intel CPUs do not have sensors to measure Tcase.

From http://www.tomshardware.com/forum/221745-29-core-quad-temperature-guide

Core 2 Quad and Duo processors have 2 different types of temperature sensors; a CPU case (not computer case) Thermal Diode centered under the Cores, and Digital Thermal Sensors located on each Core. The case Thermal Diode measures Tcase (Temperature case), which is CPU temperature, and the Digital Thermal Sensors measure Tjunction (Temperature junction), which is Core temperature.

From http://www.ocforums.com/showthread.php?t=525629

Tcase (On-die): This temperature is measured between the cores. It is a sensor included in every Core 2 Duo and Quads. This reading shows up as "CPU Temp" on most programs. It is a little hotter than the Tc from the Intel labs "Thermal Specification"


Can you quote a source to the contrary?
 
Can you quote a source to the contrary?
I really didn't want to go through Intel's datasheets, but if you insist...

Section 5.1.2 of this document explains where Tcase is measured from: http://download.intel.com/design/processor/datashts/318732.pdf

Sections 2.2.1 and 3.4 of this document go into more detail: http://download.intel.com/design/processor/designex/318734.pdf

Neither specifically states that there are no sensors built in to measure Tcase; however, that information would be meaningless since there is also no reason to assume that there would be in the first place. The technical documents specifically explain where Tcase should be measured from, and that a thermocouple should be used, which essentially implies that there is no in-built method to measure it.

Also, the entire point of Tcase is that it's the temperature at the surface of the CPU case. That would mean that to have a Tcase sensor built in, it would have to be integrated into the surface of the heatspreader (and if there was a thermocouple there, you'd be able to spot it pretty easily).
 
I really didn't want to go through Intel's datasheets, but if you insist...

They state there is no sensor to measure Tc. Tc is not the same as Tcase. Tc is a measurement taken by Intel while they are determining thermal maximums for the part.

From http://www.ocforums.com/showthread.php?t=525629 again:

Tc (Tcase - Intel labs): Temperature measured at the geometric center of the surface of the CPU. The only way to measure it is to crave a hairline groove in the CPU and apply/solder a hair-sized thermostat probe into the CPU's IHS. This is the way Intel measured the temperature to give the Thermal Specification of each CPU.

Tcase (On-die): This temperature is measured between the cores. It is a sensor included in every Core 2 Duo and Quads. This reading shows up as "CPU Temp" on most programs. It is a little hotter than the Tc from the Intel labs "Thermal Specification"
 
They state there is no sensor to measure Tc. Tc is not the same as Tcase. Tc is a measurement taken by Intel while they are determining thermal maximums for the part.

From http://www.ocforums.com/showthread.php?t=525629 again:
They are incorrect. Tc is just an abbreviation for Tcase. There are no "different" kinds of Tcase; if that was the case, the thermal spec would be pointless since it would be based on a kind of Tcase that cannot normally be measured (so having a Tcase sensor that gives different values would be useless).

Everything I said comes straight from Intel's technical documentation. Their documents have absolutely no mention of any built-in Tcase sensors.
 
OK, I think I know what may be going on here. You're referencing E7000 and E8000 documents only. I have had an E6400 since fall 2006 until just recently (when my temp went screwy). I'll note that when I had the E6400 PECI-disable temps were OK, PECI-enabled temps were low, and most of my C2D experience is with E6400 ownership.

The E6400 and various other processors had both digital temperature sensors AND a case thermal diode. PECI-mode used the DTS and non-PECI use the thermal diode. The thermal diode functioned as a case-wide temperature sensor and was accurate on the P5B series in non-PECI. DTS was used by the PECI mode, and is PECI is non-optional for newer CPUs, and is inaccurate on P5B-x motherboards.


From http://download.intel.com/design/processor/specupdt/313279.pdf

- page 5 states Tdiode is measured by a case thermal diode
- 4.2.10 states that DTS will eventually replace the diode, but "current" processors have both
- Figure 7-4 on page 73 states that a support system can use either DTS temps or Tdiode.

Your newer documents state:

- "older" CPUs had diodes
- DTS is used for everything


So, depending on your CPU, you may have a case thermal sensor (in the form of a diode) as well as digital sensors, or only digital sensors.
 
Perhaps. I'd like to see some documentation of that from Intel's actual datasheets though, rather than a specification update document. The PDF you linked to does not actually provide any explanation of what this Tdiode is.
 
I just got my p5q pro from asus and installed the PC probe. Trying to measure the temperature using pc probe and compare to core temp, hw monitor the temperature seem to be lower than core temp and hw monitor by 10 c.... is there anything wrong with my motherboard???

No. Long story short temperature monitoring by software isn't the exact science you'd think it should be. Basically the method they use the calculate temps varies from one board maker and software vendor to the next. ASUS is consistent in that their software and BIOS ROMs always show temperatures 10C cooler than you see in readings taken by Core Temp and other software. ABIT always read about 10C warmer than ASUS's software does. Intel varies but is somewhere in the middle. Other brands do their own thing as well.
 
ASUS is consistent in that their software and BIOS ROMs always show temperatures 10C cooler than you see in readings taken by Core Temp and other software.
That is not correct. Based on my current temperature readings according to Real Temp, PC Probe II is between 5C and 15C off. There is no relationship between its readings and the actual DTS temperatures as far as I can tell.
 
That is not correct. Based on my current temperature readings according to Real Temp, PC Probe II is between 5C and 15C off. There is no relationship between its readings and the actual DTS temperatures as far as I can tell.

Wait, isn't somewhere between 5C and 15C, 10C? That's the way its always been in the past. Always about 10C on the cool side according to PC Probe II. (Give or take.) As for accuracy, all the software and temperature methods can vary to a large degree from one another.
 
Wait, isn't somewhere between 5C and 15C, 10C?
Since you specified 10C instead of a temperature range, I thought you meant that the software had some sort of fixed offset compared to the core temperatures.
As for accuracy, all the software and temperature methods can vary to a large degree from one another.
Right now, the DTSes are the best we've got. They can actually be very accurate at higher temperatures. Unfortunately, there really isn't any way to test exactly how accurate they are, since it isn't possible to measure the temperatures within the cores any other way.
 
Since you specified 10C instead of a temperature range, I thought you meant that the software had some sort of fixed offset compared to the core temperatures.

It does vary some, but it always seemed to read about 10C off from other boards. ABIT's used to read about 10C in the other direction.
 
It does vary some, but it always seemed to read about 10C off from other boards. ABIT's used to read about 10C in the other direction.
The problem is that their software reads from different sensors, so the accuracy is generally much less reliable than the DTSes. Programs like Core Temp and Real Temp are useful for gauging how far you are from Tjmax, which is really the important thing (aside from the thermal spec, but you can't measure Tcase without drilling a channel into the heatspreader, so it's useless for nearly everyone). The sensors that motherboard software read from on the other hand are useless for telling you how close or far you are from the critical temperature.
 
The problem is that their software reads from different sensors, so the accuracy is generally much less reliable than the DTSes. Programs like Core Temp and Real Temp are useful for gauging how far you are from Tjmax, which is really the important thing (aside from the thermal spec, but you can't measure Tcase without drilling a channel into the heatspreader, so it's useless for nearly everyone). The sensors that motherboard software read from on the other hand are useless for telling you how close or far you are from the critical temperature.

I realize this. This is also why Intel even tests by cutting a channel into the heat spreader and placing a probe inside the groove.
 
Perhaps. I'd like to see some documentation of that from Intel's actual datasheets though, rather than a specification update document. The PDF you linked to does not actually provide any explanation of what this Tdiode is.

I think I pasted the wrong link, the document I intended to past had several references to Tdiode. If that's the case I'll try to fix it soon.
 
I realize this. This is also why Intel even tests by cutting a channel into the heat spreader and placing a probe inside the groove.

Dan, Since you already have an intel CPU with the diode machined in, have you done testing with to see if ASUS/ABIT, CoreTemp, RealTemp reflect the closest readings to what your Probe shows?? Then at least we should know whats right/wrong and possibly by what margin.

My M3A78-T with PC Probe II reads low (usually 7-10c difference), but CoreTemp and AMD Overdrive also read about the same. Its probably a fault in PC Probe II by design.
 
My M3A78-T with PC Probe II reads low (usually 7-10c difference), but CoreTemp and AMD Overdrive also read about the same. Its probably a fault in PC Probe II by design.
It's not so much a design flaw as it is the fact that PC Probe II doesn't read the proper sensors.
 
I think the actual problem is that PECI on these Asus boards isn't done right.

On my E6400, equipped with both a diode and the digital thermal sensors, you'd get good readings from the diode and inaccurate readings from the sensors when PECI was turned on. With my Q6600 (and most current C2D/C2Q) there's no choice but to use PECI because there is no diode sensor and, again, the results are wrong.

The PCProbeII software is just going to pass along the same (incorrect) values which the BIOS calculates. What I don't understand is why it's do hard for these guys to get it right.
 
Dan, Since you already have an intel CPU with the diode machined in, have you done testing with to see if ASUS/ABIT, CoreTemp, RealTemp reflect the closest readings to what your Probe shows?? Then at least we should know whats right/wrong and possibly by what margin.

My M3A78-T with PC Probe II reads low (usually 7-10c difference), but CoreTemp and AMD Overdrive also read about the same. Its probably a fault in PC Probe II by design.

I don't have the CPU configured like that. I believe Marc Adams the heat sink and cooling editor has that. I don't think we have a Core i7 configured this way yet either.

It's not so much a design flaw as it is the fact that PC Probe II doesn't read the proper sensors.

Right.
 
Back
Top