GA-Z270X-Ultra Gaming showing 1,5 Vcore regardless of settings

[alyxen]

First things first - hello to everyone on the forum that was the most helpful one during my trip with PCs, laptops and such. Now I come to ask my own question, since I can't really find an answer anywhere on the Internet.

Specs:
- 7700k @ stock for now
- Gigabyte GA-Z270X-Ultra Gaming
- G.Skill Ripjaws V 2x8GB 3200MHz (currently XMP disabled, so 1,2v and 2133MHz)
- Palit Jetstream GTX1080
- Corsair RM650X
- Some M.2 SSD and 2 HDDs (not important regarding this issue)

So...
For a couple of days I've tried to lower the temps by adjusting Vcore, and so testing began as I made my trip from 1,165v to 1,19v, which seemed stable under load (AIDA stress test, didn't try Prime95). It was all nice and dandy, but today I got another bluescreen signaling instability during Dishonored 2 play session (CLOCK_WATCHDOG_TIMEOUT). So, things needed to be done - I went to bios and changed Vcore to 1,195v and went straight to Windows to test it yet again, I fired up HWMonitor, then AIDA, aaaand that's where I noticed absolutely crazy (to me) voltage of 1,512v. Absolute panic, return to bios, check if voltages are set correctly, and they were. Now it shows 1,5v everywhere, it seems I can't control it whatsoever. It even showed 1.6v when I fired PC up after a while (~1hrs)

Things I tried:
- Change Vcore to multiple different values - no effect
- Reset bios via defaults and battery reseating - no effect
- Flash bios (from F4 to F8) - no effect
- Reseating CPU - no effect
- Change Vcore with Gigabyte utility - no effect
- Change other voltages to check if they have any effect - none
- Launch stress test to see if it's a sensor issue - instant BSoD (CLOCK_WATCHDOG_TIMEOUT)

Temperatures seem to be fine while idle (~40*C on package with 1.215v is nothing new to me with this build), but I'm not able to do anything other than firing up browser and praying that it holds up somehow.
An interesting thing: when it showed 1.6v I did observe voltage gradually dropping over time back to 1.5v in bios. No idea if this helps with diagnosing.
I'm unable to get spare parts for testing, so I'm down to what I have here. All I could do is try another motherboard in a week or two, but I need this PC working as of now.

 

[jojesa]

If that is the CPU VCORE voltage listed under the motherboard on HWM, then is reading the voltage going into the VRMs...that is not the CPU actual voltage.

The CPU voltage is listed under VID, below the CPU name i7 7700k, that is the actual voltage been delivered to the CPU.
That value increases or decreases depending on CPU load.

 

[alyxen]

Oh, I didn't think it actually mattered. Here's a screenshot of all the VIDs + bios potato quality photo.

HWMonitor

Bios
https://ibb.co/kXpMMo

 

[smorizio]

also there a setting in the bios now that forces all cores to max speed all the time make sure that off so that they only turbo when needed.

 

[alyxen]

It's definitely off as it would probably fail right off the bat with such low VID @ 4,5GHz. Just fired it up a moment ago and it showed 1,592v in bios, but I still can't find anything it would be related to really.

EDIT:
It's slowly going down, I'm at 1,476-1,488 atm, but not sure if it'll go any further. I'll leave PC on for a day and see in the evening, but I'm curious if anybody would have any solution nonetheless.

 

[jojesa]

It's definitely off as it would probably fail right off the bat with such low VID @ 4,5GHz. Just fired it up a moment ago and it showed 1,592v in bios, but I still can't find anything it would be related to really.

EDIT:
It's slowly going down, I'm at 1,476-1,488 atm, but not sure if it'll go any further. I'll leave PC on for a day and see in the evening, but I'm curious if anybody would have any solution nonetheless.

What are you maximum temperatures during full load?
Are you using EasyTune or BIOS to undervolt the CPU?

I have done that just by leaving the clock speed alone and just reduce the voltage by -25mV at the time (1.20v, 1.75v, 1.50v, 1.25v, 1.10v) and testing for stability ever time and stop when the system becomes unstable.

 

[alyxen]

I've used both of them to try and lower the voltage, though the gradual drop that I mentioned happens on it's own. The moment I fire up the PC it starts with ~1,58-1,59v, then slowly goes down until it reaches 1,488-1,5v. It won't go any lower, though - just checked on it not so long ago and it stays at around 1,5 regardless of time spent on.
I'm unable to test it under load since it BSoD's insantly as I launch it, CPU gets far too much voltage I guess. It doesn't change, even if I decrease it by 0,035v to around 1,165v, it'll still show 1,5v.

EDIT:
Okay, I fired up stress test out of curiosity and checked on Vcore and VIDs - Vcore stayed at around 1,488 without any fluctuations, VIDs show ~1,24-1,27 at most, though it still BSoD's after a short while (needs to be shut down by holding the power button). Temps didn't go anywhere beyond 80*C (which would make sense, considering quite low ambient temperature today) and it didn't throttle at any point.

 

[jojesa]

I've used both of them to try and lower the voltage, though the gradual drop that I mentioned happens on it's own. The moment I fire up the PC it starts with ~1,58-1,59v, then slowly goes down until it reaches 1,488-1,5v. It won't go any lower, though - just checked on it not so long ago and it stays at around 1,5 regardless of time spent on.
I'm unable to test it under load since it BSoD's insantly as I launch it, CPU gets far too much voltage I guess. It doesn't change, even if I decrease it by 0,035v to around 1,165v, it'll still show 1,5v.

Where do you see that the CPU is getting too much voltage?
The highest value I see on HWMonitor is 1.3v

 

[alyxen]

Okay, there's definitely something not right going on atm.
I went to EasyTune just to try and set a bit higher voltage - went for 1,21v and LLC set to High (as it served me well before), then tried another stress test and it went on for 20 minutes without a single hitch, whereas before it didn't last even 30 seconds. Is it me, or is there really something wrong with the voltage sensor?

Link to screenshot of HWMonitor + AIDA
https://ibb.co/ktMOK8

 

[jojesa]

Okay, there's definitely something not right going on atm.
I went to EasyTune just to try and set a bit higher voltage - went for 1,21v and LLC set to High (as it served me well before), then tried another stress test and it went on for 20 minutes without a single hitch, whereas before it didn't last even 30 seconds. Is it me, or is there really something wrong with the voltage sensor?

Link to screenshot of HWMonitor + AIDA
https://ibb.co/ktMOK8

I think there is something wrong with EasyTune.
I wouldn't use it.

I see your voltage between 0.650v to 1.299V and your temps are a bit high but below Tj max.

 

[alyxen]

I installed it mainly to check if it showed anything regarding voltage, but nothing really.
Anyway - uninstalled it, went to BIOS, set 1,2v and LLC High again, went on with testing, 20 minutes passed an not a single error. But I'm still concerned about that VCore value, it never showed anything above what I set manually plus it happened after a BSoD. VID is another thing that I've never really looked at, but that aside.
Temps are quite high, that's true, though it's nothing unusual in terms of my CPU - I lost at the silicon lottery, hence no OC on it.

I'll continue testing all across the board (games, synthetics, maybe Lightroom and such), we'll see what it'll show.

 

[jojesa]

... But I'm still concerned about that VCore value, it never showed anything above what I set manually plus it happened after a BSoD. VID is another thing that I've never really looked at, but that aside.
Temps are quite high, that's true, though it's nothing unusual in terms of my CPU - I lost at the silicon lottery, hence no OC on it.

I'll continue testing all across the board (games, synthetics, maybe Lightroom and such), we'll see what it'll show.

I would keep track of the VID values since you cannot control the voltage delivered to the VRMs (VCore).
If your temps are not doing down after the voltage changes, then is not point to undervolt.