(Urgent Assistance) EVOC P750DM2-G High PCH Temperatures

Hello all,

System specs (Purchased Feb, 2017):
BIOS: EVOC 1.06.09EVOC v2, 04/12/2017
CPU: 7700K
RAM: Kingston 2400 MHz (2x8GB)
GPU: 1070 MXM
SSD: Samsung 850 EVO 1TB
WiFi/BT: Intel 8265

Approximately 2 weeks ago, I received the Windows 10 Creators Update (10.0.15063 Build 15063) and attempted to use disk cleanup to delete the old Windows installation folders that were created after the fact because I felt comfortable with the stability.

Upon deletion, I was noticing some random freezing/lagging, so I decided to load my BIOS defaults because I felt that maybe my down-clocked 7700K might have something to do with it. After resetting the BIOS, things still felt wonky, so I decided to wipe the machine and re-install Windows. After doing so, things smoothed out with all my applications re-installed, but I was still experiencing issues with games.

Upon loading a game or OCCT GPU benchmark, the system will function normally until it reaches ~70ish°C. After that, it powers off, as if it were a thermal protection scenario. To take better precautions, I went ahead and re-pasted the GPU/CPU and replaced the thermal pads. Thermals were down from previous testing, but I'm still experiencing power downs after thermals reach ~70ish°C.

CPU/RAM Benchmarks/stress tests can run without issue. I found this to be an odd aspect to the issue, and wanted to further inspect what might be causing the power-downs. I noticed in event viewer that the Intel 8265 was crashing repeatedly on boot after these power-downs, so I ordered a replacement to troubleshoot. After testing the system with all the different RAM configuration and WiFi card configurations, I found no correlation, and that the power-downs would still continue to happen regardless. On a side note, I couldn't complete the built-in Windows Memory Diagnostic, as it would power-down mid test regardless of percentage.

I also noticed that when I run the GPU load tests, that if I limit the load, the test would run fine without issue. After updating drivers and further testing with HID Evolution technical support, we noticed the issue still existed, and will be picking back up on troubleshooting today. This morning, I downloaded HWInfo64 to further inspect temperatures and logging, and found an interesting issue.

My PCH (Platform Controller Hub) temperatures at IDLE ARE +100°C..... During a stress test of the CPU, the temperatures reach somewhere near 110°C. I swear I read somewhere that the maximum thermal limit of the Z170 Clevo motherboards was 111°C. When I run the GPU load test, the PCH thermals go above 111°C and I experience a power-down. I have no idea which IC package mounted on the motherboard, but I REALLY want to mitigate this issue with a heat-sink/thermal pad.

So, does anyone know of the position of the PCH?

Thanks in advance!

 

If I recall correctly, the PCH on this system is supposed to have a rather thick thermal pad on it, and should be located on the top side of the board (the top case has to come off to expose the chip). You might want to check with HID about the location of it just to be sure, and also to check on whether breaking the system down to that extent will affect your warranty or not.

 

Are you referring to this IC (Boxed in red)?

https://1drv.ms/i/s!AuynfBhGAZc7gcVODNVp1Ezj_Y9BKg

Thanks for the response.

 

The picture is a bit grainy, but as best I can tell that is the PCH chip to the right of the RAM. It needs to be cooled. The P870 has a metal plate there. A thermal pad with nothing for the thermal pad to transfer the heat to is not going to be very useful. Maybe others can compare their P750DM2-G PCH temps. If the high temps on this machine is an anomaly, then perhaps a motherboard swap is in order.

 

Thanks for the information Mr. Fox. I'll ensure there is a thermal pad in place that mates the surface of the PCH to the metal cover plate for dissipation.

 

also ensure that you set whatever undervolt you had before resetting that BIOS to defaults

 

That's the one, although it looks like there's a piece of thermal tape(?) over it in that picture. Usually it's bare with a thick, purple thermal pad placed over it. Also, there's no additional metal plate on this model, the pad mates it to the top case assembly.

 

After further inspection of the chip, I noticed that there was an air gap between the thermal interface and the case. I went ahead and added a 2.0 mm thermal pad and it appears to be stable under load. The temperatures are down from +111°C (and crashing) to ~99°C (OCCT Power Supply Test GPU & CPU with unlimited FPS). Shame the fans simply blow air across the heat-sink's radiators, as the PCH could use some airflow.

Currently typing on the machine, and with high performance mode engaged in the Clevo control center, I've got 90°C on the PCH according to HWInfo64 with no load consistently. Better than 105°C on power-saving mode.

(-0.070 Under-volt[Adaptive]/Down-clocked to 40x/38x/37x/37x)

I wonder if a static under-volt would help.
Battery life is also excellent as well.

Thanks for the help!

 

105C or F???? 105C means your CPU is fried buddy

 

105°C on the PCH on the opposite side of the PCB.

My GPU CPU temps are below 75°C at load. It's the PCH that was running hot.
Right now, my temps at idle with automatic fans (Windows power plan drives it up to 100% clock speed all the time) are less than 40°C.

Edit:
On a side note, that would be amazing to still be operating at load with +100C temperatures...

 

PCH is the chipset which is manufactured on an older process usually with larger transistors running at lower speeds that can take higher temperatures.

 

Still, 100C isn't normal.

 

Yes but it's not insta death either.

 

So, since the PCH handles a lot of different hardware mingling, should I check out my BIOS settings for USB/SATA/Memory adjustments? Or maybe PCIe speeds? The PCH temperature under the full load from OCCT's power supply test is hitting 103C after 10 minutes, and it seems to be stabilizing at that temperature. If I run Overwatch/Witcher 3/Doom/Gamecube games (without any other application), it takes twice as long now to crash as it did when I didn't have a thermal pad making contact to the chassis (it currently crashes exactly at 112C).

I also have a USB hub connected via my Dell 2417DG which the peripherals crash just prior to 112C on the PCH. The computer still crashes regardless of whether or not I'm connected to the monitor.

 

A Hell lots of the Aw Bgabooks with all too high PCH temp had huge problems. And never fixed by Dell Death or other problems due high PCH temp doesn't matter

 

It could be the voltage supply circuit has an issue and is over volting it. There is not much you can do but RMA it.

 

But for it to happen after restoring BIOS defaults?

 

Could be unrelated and you had not realised.

 

Yikes. Maybe an RMA will have to happen.

 

ouch! let us know how it goes. good luck bro

 

Agree... not likely to cause any physical damage. It's not insta death to CPUs either.

But, potential insta death for performance, stability and the user experience.

 

@Spootnuke - The PCH does run rather hot, but your recordings are still on the extreme side. I think your seller will need to check it out and, probably, replace the board if they can't lower the temps.

 

Thanks for the information Prostar. I've requested the RMA, and the machine is on it's way to HID Evolution. Hopefully they can identify and isolate the issue, because I was getting rather frustrated with the thing.

I'm now starting to remember random shut-downs early on when I received the machine, but no where near as frequent as what these were.

 

It can happen, hopefully they can get it sorted quickly

 

Just received my machine on June 30, 2017 after HIDEvolution had it since May 16, 2017. After testing they had sent it off to their distributor to have the whole machine looked over, as they couldn't truly pinpoint the fault. They received it back after the distributor stated it was repaired, and then they proceeded to test it. The PCH temperatures had been corrected, but they found another fault where the machine would intermittently shut-down under CPU+GPU load. They ending up sending it back to have it corrected since all issues should have been rectified at that point. After they verified the machine was back in working order, they ground-shipped it back to me.

I was excited like a kid on his/her birthday to be honest with you. I fired it up and ran many a demanding game to verify that the issue had been resolved. Sure enough, everything was fixed. As of right now, all my temps are around 77C (PCH/GPU/CPU) or less under 100% load at stock settings.

BTW, they never gave me any information as to what they've done to rectify the issue, and I don't plan to open the machine in its current state because I'm so busy at the moment.

I do have a request for some assistance though.. I'm noticing some stuttering on loading games, and also in game. It's fairly intermittent and very in-frequent, but when it occurs, my FPS goes to "1," and it continues on for around 3 seconds. All of my drivers are fresh installs, and all software installation integrity has been verified. I've just now restored all BIOS settings to default, and will continue my testing, but I can't find a smoking gun as to why the stutters occur. Flexicharger was my main concern, and I've just disabled it. Now time to test!

I'll take any advice!

 

A good way to check this is use HWiNFO64 and RivaTuner for on-screen display so you can watch what is occurring in real-time. You can use the configuration option in HWiNFO64 sensors to change the layout and control what values show in your on-screen display. It does not work with some of the crappy Windows 10 XBOX/Store UWP game filth, but it works with just about anything else on Steam/Origin/Uplay.

You can find some help with that in this imgur album: http://imgur.com/a/UYC3s

 

What's you storage setup?

 

I'm strictly running a Samsung 850 EVO SSD (1TB).

After verifying some things in HWinfo, I've noticed I'm hitting the following:

Yes - GPU Performance Limit - Power
Yes - GPU Performance Limit - Utilization
Yes - GPU Performance Limit - Reliability Voltage

I also monitored wattage entering the CPU and GPU. The limiter comes into effect whenever I hit 120ish watts on the GPU. Also, I've seen some evidence that leads me to believe that 160ish watts (GPU/CPU combined) also causes the GPU performance limits to come into effect.

I don't recall having these limiters come into play when I had the machine previously. Is there any way I can limit power consumption? I purchased (2) 230 watt power supplies at order and I'd hate to be forced to upgrade.

Thanks for any suggestions in advance!

PS - GPU Performance scaling is enabled in the BIOS if that matters.

 

Soooo... I think it's definitely related to the limited power output from the 230 watt AC adapter. After reviewing power consumption per component, I'd have to say that 230 isn't enough for a 7700K/1070/cooling pad combo. Just ordered (2) Delta ADP-330AB D AC adapters and will sell the 230 watt Chicony AC adapters on Ebay I suppose.

Maybe that had something to do with the performance of the PCH and random shutdowns. Maybe the Summer heat and stronger cooling requirements from the system's fans is too much for the 230 watt AC adapters. Either way, I wish to do some mild overclocks, and 230 will not be enough to meet the requirements of this little beast.

 

You say you are running a pair of 230W or just a single one?

A 330W can give more ocing headroom.

 

Strictly a single 230W AC adapter during operation.

 

After receiving my 330W AC adapters, I noticed that the brick was cooler under load, but I still had the periodic stuttering. When playing CSGO, anytime I would press escape to access the menu, there would be a lengthy pause and stutter even though I was running the game at a ridiculous FPS on 1440p without much hardware stress. After some thought and frustration, I decided to disable G-Sync to my external monitor since I only ever experienced the stuttering during my sessions on the Dell monitor. After doing so, all of my issues were gone. Not sure how to rectify this situation with the monitor, but I'm sure there's some kind of conflict. I'm pretty busy at the moment, so I haven't been able to sit down and play with the configurations, so if anyone has any ideas, let me know!

 

The stuttering is likely due to moments when the game's framerate drops below the monitor's refresh rate, which activates G-Sync. Assuming you have a 1440p monitor with a 144 Hz refresh rate, use RivaTuner Statistics Server to cap CS:GO's framerate to 136, re-enable G-Sync, and see if it goes away.

 

I never thought about that... Pressing tab or escape causes a stutter in CSGO no matter the hardware configuration, thus causing G-Sync to enable. I should have remembered that. I guess I was having a hard time linking that together with other games I've played where I've had stuttering occur because those other titles rarely achieved 144+ FPS. It makes perfect sense when I look back on all the scenarios where the stuttering has occurred.

I wonder if I can globally limit all titles to 140 FPS.

Thanks for this tip! I'll try to confirm it when I get back to my desk.

 

Due to all the different engines etc a global limit would be tricky to enforce.

 

Which is why you have to do it on an application-by-application basis. RTSS makes it easy. All you have to do is point the program to the game's .exe file, set your framerate limit and any other settings, and you're good to go. I don't have a ton of games installed at the moment, but all of them I've tried have worked flawlessly.

 

Sounds like a decent solution.

 

Not even with Nvidia Inspector?

 

That's another way, I suppose. I've not tried it because RTSS does what I need.


Sent from my iPhone using Tapatalk

 

Well, my PCH temperatures on an Origin Eon 15-X based on the same chassis are ranging between 85 to 95 degrees Celsius.

Glad to know the relatively high temperatures are not an anomaly.