Clevo P870TM1-G Randomly Shutting Off

Second time posting here!

So first things first. Specs are as follows:
CPU: Intel Core i7-8700K (Coffee Lake-S, U0)
Motherboard: Eurocom X9C
BIOS: 1.07.13REU1, 09/28/2018
Chipset: Intel Z370 (Kaby Lake)
Memory: 32768 MBytes @ 1199 MHz, 18-18-18-39
- 16384 MB PC21300 DDR4 SDRAM - Corsair CMSX32GX4M2A3000C16
- 16384 MB PC21300 DDR4 SDRAM - Corsair CMSX32GX4M2A3000C16
Graphics: NVIDIA GeForce RTX 2070 (TU106M) [CLEVO/KAPOK COMPUTER]
NVIDIA GeForce RTX 2070, 8192 MB GDDR6 SDRAM
Drive: WDC WDS100T1B0B-00AS40, 976.8 GB, Serial ATA 6Gb/s @ 6Gb/s
Drive: ST2000LX001-1RG174, 1953.5 GB, Serial ATA 6Gb/s @ 6Gb/s
Drive: Samsung SSD 970 EVO 1TB, 976.8 GB, NVMe
Drive: Samsung SSD 970 EVO 1TB, 976.8 GB, NVMe
OS: Microsoft Windows 10 Professional (x64) Build 19041.928 (2004/May 2020 Update)

I've been having a very odd problem. Whenever I fire up a game, my system shuts down. In most cases, I get about as far as the menu and it powers off. I have core temp open on my secondary monitor and it doesn't show any sort of extreme temps (50-75 at max).

I've changed the thermal paste (I'm using Grizzly Kryonaut, I've bought some Gelid, it'll be here tomorrow), and I've run Intel's processor diagnostic, memtest pro to a 100%, chkdsk'd all my drives, checked my fans, and frankly I'm one hundred percent confused about what's going on here.

 

To isolate each component try:

Cinebench R20/R23 for CPU
Uningine Superposition for GPU
It passed memtest so that's good
Then try TimeSpy for overall

Make sure you have HWinfo running while running the tests to see if anything pops up in the red

My first guess would be the PSU(s), but start with each individual component and see if one shuts down

 

GPU's the culprit for sure. System shutdown barely a minute into the Uningine benchmark.

 

Make sure that the GPU VRMs are making proper contact with the heatsink.

 

The way you worded your last paragraph in your initial post seems to point towards one day the system was working fine and then the next it wasn't but to be sure:

Was the system previously working fine for an extended period of time then suddenly this started happening? Did the problems start before or after changing the thermal paste?

Either way, you're going to have to get back under the hood and check / re-check your GPU, but if it started after a repaste that points more towards a pad moving/shifting during remount.

If it started before and was running fine one day then started crashing in games without any physical changes, it could be a dying card but either way you'll want to get up in there and re-check your pad placement and might as well do another thermal re-paste.

Take some pics when you have the heatsink off of the GPU and heatsink.

 

@yrekabakery @electrosoft It did indeed just start one day, and it happened before I changed the thermal paste. I cannibalized the thermal pads for the 2nd GPU and padded it some more. Passed an Unigine 2 4k benchmark. Opened War Thunder, shut down at the menu. Probably should just replace the thermal pads? If so, any suggestions on what to get? If not, is it indeed the card? I'll get pictures up later, it's a little late for me here.

 

Actually, now that I think about it I've had shutdowns while gaming before. I've always blamed the CPU despite it not being all that hot, 85 C-90 C at maximum usually. Which isn't quite hot enough to Tj Max, right? Even if the temp reading is off give or take 5 degrees, which afaik it shouldn't be?

I think I'll replace the thermal pads and see if that does it any good. If not, I'll have a few extra unused thermal pads. If so, excellent. I shall update this with progress on that.

 

Make sure to post some pictures to see if anything stands out

 

Do you have a spare PSU to try? I wonder if the spike load for gaming causes a draw that exceeds your power delivery capability (battery and PSU combined). I wonder if you see a marked drop in the battery level (and/or battery heat), indicating high power draw.

 

My guess is a bad VRM or thermal pad.

 

I'd go with bad thermal pad placement to ... I used K5 pro to fix this

 

Any decent pad that's nice and soft/pliable will get the most consistent results.

 

Temps seem fine for not shutting down due to temps.

I've had some issues recently with windows updates causing sporadic issues as seen in event viewer as "kernel power". Remove the updates and things go back to normal. It's a toss up between windows / intel lately with recent releases causing issues.

Just a thought / experience. It's not always a HW issue.

 

I think quite often they are not sensors and all the sensitive components that's why I personally use the k5pro so more patsy pretty awesome stuff as you don't need to worry about sizing of pads it creates it's own size up to they say up to 3 mm which is more than sufficient. One huge thing I found was updating the EC on the system made a massive difference as my system was thermal throttling with a very old EC

 

I'd be wary above 2mm gap wise for paste.

 

Apologies for the delay, guys, family issues meant my computer troubles were the least of my worries.

Update:

I read about someone having faulty RAM despite it having passed Memtest. So just to see if it made a difference, I removed the two older RAM modules I had, and used the two newer ones I had. Seems to have had some effect. If I fire up War Thunder (which I'm using as a test case, I'm not sure if it makes a difference either way), I can get into the menu and stay there now. In fact, I can also get into a test flight and fly around for about a minute before the system shuts down.

I was monitoring temps as usual, not above 65C for the CPU, GPU at 45C and HWInfo64 shows the hotspot at 55C. Just another bizarre thing, I got a random system service exception BSOD the other day.

I've ordered K5 but it's been ten days since it was supposed to have left Greece and I'm not confident about it ever reaching me. Unfortunately Amazon India's run out of K5. So there's that slight issue. On the subject of thermal pad placement, is there a handy diagram I could use for the vapor chamber heatsink pad placement?

 

Do a search for it on Google, should come up. Been done a few times on these forums.

 

@Meaker@Sager Yep, I found the one you put up a while back. New pads should come in the coming Tuesday, I'll take a picture of my pad layout and post it here when it's done. GPU picture then too.

 

Sounds good

 

Spoiler: Image

No change in the situation, here's the picture of the heatsink and the card. Below is the layout I followed.

Spoiler: Image

 

Ah that guide was for the 1080 sorry, you'll need to adjust it.

 

Yes, I'd seen that when there were parts of the heatsink that looked distinctly different. Any advice on how to go about adjusting it for the 2070? As much as I love the idea of laying thermal pads on parts of the card I'm still not quite sure about all the components. VRMs and VRAM modules are not terribe afaik. But everything else I'm admittedly not entirely sure on.

 

I figured it out by laying the card flat and figuring out for each component (cap, inductor and vrms).

 

I guess I'll try that, then...not sure how successful I'm going to be. Just er...related question. My system's been randomly powering off off late (last week and a half?). Could that be a related issue, or is that the sign of a worse problem?

 

The first suspect is the VRMs always with that sort of power off. Are the pads making proper contact to those, everything else other than the core is far less important.

 

Kind of unrelated but may be worth a sanity check on your part. Mine shuts down whenever it gets unplugged because one or more cells in the battery are borked. It jumps from 100% to 30% instantaneously, and it shuts itself down.
When it boots back up, it shows the battery as empty and charging ... some of the time. Some of the time it shows 80%, 73% etc.

The power loss is accompanied by the same critical notification in event viewer - kernel power loss.

The thing about Li-Ion battery packs is that they're usually wired in weird series /parallel combinations and therefore, readings can be somewhat unpredictable. That + coupled with the fact that our p870TM platforms are essentially "laptops" in the sense that they're compact, but can still drain 900W under load, whenever a cell fails, odds are the entire battery pack fails ... and it fails weirdly... in the sense that it can show it's charged at 80%.
It's an extremely bad idea to drain Li-Ion batteries completely ... or fast because it messes with their chemistry as well, therefore further increasing the battery level reading unreliability issue.

I suggest you check the power settings and see if its critical/low battery settings. and set it to hibernate at about 99% or something ridiculous.

Furthermore, you should assume that your battery is completely unreliable as far as reporting goes and try the following:

Put it under some load and touch or gently giggle the power connection at the back of the laptop.
If it shuts down, it's most likely due to an issue with your laptop's power socket and your battery is essentially dead as well.

Also, make sure you're using enough Watts. I have a 2x1080 SLI rig, with which it's a stupid idea to use only one 480W adapter.

The test is meant to cross check whether the power loss is due to your laptop switching to battery mode with an unreliable, broken battery.