Ga-870A-UD3 Sensors

[CountMike]

On the MB above I have an AMD Phenom II x4 965BE processor. It's currently running at 4180 MHz  and would be no big problem to go up to 4.4 GHz but for some heat related problems with MB. On the CPUZ HWmonitor there are 3 TMPIN measurements. 0, 1 and 2 which I don't know what they represent. At idle they read  38, 38 and 86 respectively. That last one is obviously wrong as could not find any spot on MB that is that hot. Now according to Speed Fan  all 3 read from the same chip, IT8720F One problem I have is with TMPIN1 with jumps up to 60c and shuts down when processor is pushed  above 4180 MHz. I'm afraid to OC any more till I find out exactly what is heating up and take measures to cool it. (processor is not going over 55c, no matter what )
So, what do those TMPIN temps refer to ?  SpeeFan calls them "temp 1, 2 and 3" and they correspond with HWmonitor and other measuring programs. 

[Vezina]

2 of the temperatures are most probably wrong.

The CPU takes 70 Celsius just fine for short amounts of time.
The CPU VRM should also take 90 Celsius just fine.
Maybe the VRM protection turns off the CPU due to overcurrent protection ,it is reaching it s current limit ,limit implemented to avoid smoke.

If it s the first revision of the 965 you are having there over 140W TDP while overclocking way over the specifications.

[CountMike]

965 BE is C3 revision and never gets too hot. MB is rated at 140W and with the new cooler is not a problem at all, right now I'm working on NB and VRM cooling but I would really like to know what the readings of each of those 3  sensors  mean.
To understand is to know, to know is to understand.

[Vezina]

2 of them are bogus info.

The VRM stands 100 Celsius just fine ,the NB chipset takes 95 Celsius operating temperature with absolute maximum 115 Celsius ,the CPU takes even 85 with no problems if it s temporary.
So ,in my opinion , there is no direct relation between the shut down and the temperatures.
The shutdown occur due to excessive amperage draw at CPU VRM level which triggers over current protection.
It might be that the watchdog to kick in faster than it should but it protects from killing the board.

If you want real temperatures use a termometer.

[CountMike]

Thanks guys, so much I suspected too. Could not find anything that hot (at least not 86c) but I felt uncomfortable not knowing what those temperatures  represented. they might not be accurate but would be nice to know at least for reference between states when on stock and when OCed. I asked same question twice the Gigabyte Service center and did not get proper answer but to take it to a service and clean the dust Lol.
So I underclocked and undervolted everything as far as it would go and still the same temps, so I guess There's some glitch in the readings somewhere, Pretty sure that TMPIN1 and temp2 in SF is for NB, because when I overvolted and stress tested it, those temps rose slightly.
I'm realy attached to this MB as it has everything I need, from PATA port to 3 PCI and parallel and RS232 ports, 3 SATA controllers and USB3. Would be hard pressed to find a newer one with all that.