Problem with asrock z77 Pro3

[Tom_Brady]

Hello.
I’ve got a problem with my new mobo, Asrock z77 Pro3 (and i5 2550k, 550W corsair).
When I set multiplier @ 3,4 everything runs fine, offset is set to -0,140. The problem is, that when I set multiplier higher than that (for ex. 40), when I do a stability test on LinX, after 5 minutes, the speed is throttled down to 3.4Ghz for a few seconds, and then jumps back to 40. As the test goes on, the throttling rises more and more, to the point where it is throttled to 3.4 for 4-5 seconds before jumping back to 4.0. What can be wrong? Oc @ 4.0 should not be a problem… right?
My temperature on CPU@4.0 is 60 degrees max, voltage is 1,24V (cooler Noctua NH U12P with backplate for s1155). Power limits are maxed. Eist, c-states disabled.
What could makes throttling?
Bios is actual, 1.90. Turning off thermal setting in bios did not help.
Program ThrottlingStop did not help.
I put a 80mm fan directly on power phase radiator on motherboard – it still throttling.
When I choose everything on default (also with cmos) and pick optimized@ 4.0 it also throttling…
Any suggestion?

 

[abbadon_34]

The physical thermal throttling does not kick in for an i5 until 95-100c , so obviously this is not the case if you run 60c max. I have not used LinX so I would defer to others on that issue. Is it truely throttling, or is turbo mode switching on and off ?

 

[Tom_Brady]

But why is turbo mode switching on and off under load (same thing happens, when I use prime95 or other heavy load program)? Afaik, it should stay at 40 (all cores are set to 40) and drops only to 16 when processor@ idle (when eist and c-states are enabled).
Is there a way to check why multiplier is getting lowered? Some program or logs in windows?
Maybe some sensor is broken, and it throttle even when temps are ok? Or bugged bios?
I have sent support ticket to asrock nearly 2 weeks ago and still no response...
Should I downgrade bios?

 

[abbadon_34]

Is this happening one all cores or just some of them? Use a program that records the clock speeds and temps for each core simultaneously. It may be clocking some cores lower to keep a single core higher, which would be normal.

 

[Tom_Brady]

It's happening to all cores. And I don't think it's normal. All cores are set to 40, all are under load so why it should be reduce?
Also I made some test yesterday.
It happens ealier when multiplier is set higher (for example it happens after 3 minutes of testing with x42 and 40 seconds when ratio is 46; @3,8 it took almost 8 minutes) Thats why I think its because of throttling. I put a sensor into radiator on power section, and it shows 55C so temperature on mosfets belows must be higher. But where I can find temp limits for VMR? Or how can I chceck temperature on it?

 

[sp00n]

In case you haven't found the solution by now, this sounds very much to be a problem with VRM overheating. The VRM elements on the Pro3 seem to throttle at around 65C (NOT CPU temp!).
Try tucking a fan on top of the heatsink next to the CPU socket, it should help to a point.

 

[ihog]

In case you haven't found the solution by now, this sounds very much to be a problem with VRM overheating. The VRM elements on the Pro3 seem to throttle at around 65C (NOT CPU temp!).
Try tucking a fan on top of the heatsink next to the CPU socket, it should help to a point.

He already did, but he should record and post his VRM temps.