Problem with DP67GB and/or 2600k (Overtemperature/sensor error)

[idata]

Hello.

I've just build my new computer some days ago and am having some problems since then and don't know what to do.

First, the processor (it's a 2600k) seems to overheat a lot. It stays around 45~47°c in idle and goes up to 85+°C under full load (video encoding, intel extreme tuning stress test, prime95, etc.) Ambient temperature at ~25°c. This is at stock clocks with a stock heatsink and fan, which stays around 2000~3000 rpm. I've tried everything: Reseated the heatsink in all four possible positions making sure that I heared the two "cliks" when pressing the push pins, all types of thermal compound (applying varying amounts on processor and heatsink) and so on. The heatsink is always firmly in place, as I can't even tilt it, and, when removing the heatsink, the termal compound is squashed, meaning that there was pressure applied. What's more strange: even with the heatsink firmly in place, untiltable, I can keep my hand on it even when the processor is at 85°c! There's no way it would remain cool with the pressure that is applied on it, if the temperature readings are correct.

What's strange is that not only the processor readings seem wrong, the motherboad ones seem too. Mainly because Tcase always seem to be 10°C or more higher than Tjunction and, by everything I know about thermodynamics, it's impossible, since the heat is generated at the cores, how could the package be hotter? There are also some strange alarms in Intel Desktop Utilities that happened two times in about a week saying that the memory voltages had gone over 1.9v! This seems only a spike, because upon checking on any monitoring utility, the voltages are back at 1.5 and nothing is registered. But I'm afraid it could damage the processor or the memories.

It doesn't seem like a PSU (Sapphire 625W) problem because the very same lines must be used for generating other voltages and only the memory voltages showed an alarm (and only two times.) Also, some motherboard sensors sometimes stop working. Some temperatures and fan rpms will start showing up as 0 until I reboot. Maybe my motherboard AND processor are defective?

I'm using the latest BIOS and latest drivers/programs. Even if these are only mistakes on the sensors, wrong readings could cause wrong throttles and lower my performance, not acceptable. :/

Thank you for reading through such a long post, I hope this is solvable without an RMA.

[idata]

Try not using IDU, uninstall it for testing, you can always download it again. Try another monitoring program, but don't run them at the same time. IDU in particular does not like that, and that causes false alarms, I've seen that myself.

Also, the i7-2600k (I own one) reports the CPU temperature (Tcase, CPU Package) and the individual Core temperatures differently than previous Intel CPUs. Now the single CPU temperature is always greater or equal to the core temperatures, instead of less, and it will never be less than the lowest core temperature. IMO, this is not a bad thing, as the CPU's thermal safety features, or any CPU function that depends on temperature, has always used the single CPU temperature as the trigger for whatever needs to be done. Core temperatures have always been ignored.

Frankly, I have never liked the sensor/monitoring chips used on Intel mother boards, the one I have never worked correctly, apparently.

The CPU does not send it's temperature reading to the sensor chip in order for the sensor chip to tell it what to do. The CPU decides by itself whether or not to throttle due to high temperature. The CPU can adjust it's voltage regulators by itself, which it actually does all the time. You won't lose CPU performance due to a bad sensor monitoring chip, but only due to a truly over-heating CPU.

[idata]

Yes! The false alarms on IDU stopped when I closed HWmonitor, no more false alarms for me since the original post! :-)

(Running 24/7)

But, with or without IDU installed, my CPU temperature is still very high! VR, PCH and MEM are all around 40° or less, but CPU still likes to go from ~44° on idle to 85°+ on full load, don't know what else to do, guess it's RMA for me. Even if's just a faulty sensor/monitoring chip, it still worries me and Intel Support told me to ask for a replacement from the retail store from where I bought it.

What really bugs me to no end is that the heatsink is always only mildly warm, even on 85°C+ on full load and I've checked its seating numerous times. This makes me wonder: Is the Tcase sensor on the motherboard and are the Tjunction sensors on the core themselves? If so, how could two independent sensors err at the same time showing me high temperatures? Maybe the chip is really overheating or there is something wrong with my socket/heatsink/processor surface with impedes full contact?

Didn't want to RMA it, probably a week or more without a computer. :/