P750TM1-G shuts off at high load

Good morning all,

My laptop has been doing this for a while now, but it is starting to get worse so I was looking for some opinions if possible.

I bought a P750TM1-G from HIDEvolution close to two years ago now. The rough specs are:

Intel Core i7 8700K
nVidia GTX 1080
4K AUO display
64GB RAM (I bought this separately)
2x 512GB Samsung 970 Pro (I bought this separately)

In Windows when I play a demanding game the laptop will shut off (something like The Outer Worlds, X Plane 11, those are the most demanding games I currently have; most of the reason I got this laptop at the time was for doing quick processing and AI type workloads).

When it shuts off, the screen turns black and then after a second or two the whole machine turns off. If I attempt to turn it back on without waiting around a minute the laptop refuses to turn on.

This makes me believe that the laptop is overheating, but I'm not certain. Nothing is logged in the Event Viewer that looks related.

Since I thought it was thermal issues, I opened it up and vacuumed the dust from the fins and fans being careful not to damage anything. O noticed while I was doing this that the heatsink's screws around the CPU and GPU are basically all stripped out pretty bad. I didn't attempt to unscrew them since I don't have any additional thermal compound yet, but I was wondering which compound would be best in this case and whether I should replace anything else (foam pads, etc on the GPU?)

I gathered after buying this that the 1080 is probably too much power for this computer. Maybe I could underclock the CPU or something to help things. I'd have to look into how to do that.

I appreciate any assistance anyone can provide. It's kind of a frustrating situation and I guess a lot of it is my fault.

As it stands, I can only run it for about 5 minutes before it restarts if I'm playing one of the afformentioned games or some intense data processing.

 

First I thought it might be too much undervolt in the cpu. But maybe you are right about too much powerdraw.

Clock and volt cpu and gpu somewhat lower and see if it happens again. If this helps you can try to find the sweetspot by raising clocks and voltage again. (Also try to find out and compare how much power you are drawing)

 

Hello,

I tried to recreate the issue using furmark and prime95 and it looks like it's drawing 300-330W from the wall pretty continuously prior to the system shutting down. My initial thought was that maybe the PSU was insufficient, but it is rated at sending 330W output in DC so this should be fine, I'd think.

I feel like replacing the compound is probably past-due at this point but I'd need to buy new screws (if anyone knows what size, I'd appreciate it; it's mostly the GPU side I'm worried about; there are four around the GPU chip itself and another screw holding it in as well. All 5 are stripped out to some extent, some of them I'm not sure I can remove.

Hopefully between underclocking (and/or undervolting) and replacing the compound I can get it to reasonable temperatures. It seems the CPU runs hotter than the GPU for what it's worth.

 

Does the green light on the PSU go out when this happens?

 

As far as I can notice, the light was on. Certainly in the cases where the laptop won't turn on for a minute or so (in those cases the laptop lights turn on and then the laptop just turns back off before the monitor even turns on).

Thanks

 

Those stripped screws are probably your issue. If the heat sink is not making contact your def overheating. I'd start there. Any thermal compound won't help if it's not making contact Good luck!

 

Thanks. That is a good point. Since it came this way I'm not really sure if it is tightened down correctly or what size they are. After some looking around I haven't been able to turn up a result without buying a ton of screw sizes I do not need. Do you know what size they are?

 

It could be the VRMs are making poor contact.

 

I don't the exact size. You could buy a multi pack of different standard sizes. Then when you pull out the old ones measure them against the replacements. You can find them on eBay and such. This is one listing. I found. You can also check YouTube for CPU repasting on that model. I've seen some high quality ones in the past tell you screw sizes as they show u.
Here's a pic of the listing of common screws.

Sent from my moto g(7) using Tapatalk

 

I'll check out the videos and see if I luck out on screw sizes. Thanks for the tip. I saw the kits but wasn't sure if M2 was too big since these look to be rather small and I was hoping to not buy too many screws lol.

I also contacted my reseller but with things how they are with the virus I'm not expecting to hear back too quickly.

Thanks for the suggestion

 

Sent u a pm too

Sent from my moto g(7) using Tapatalk

 

Thanks! Found what I need for now. Will order what I need and report back

 

Good morning all,

I think this issue got resolved (and I think it was thermal-related).

I was able to purchase some Conductonaut and applied it to the CPU and GPU (as well as between the CPU die and IHS... the reseller did this initially but appeared to have used way too much so I cleaned it up).

While I was in there, I cleaned out the heatsink fins, though those were pretty clean already; I tend to clean them every few months.

Turned the computer on, noticed it was around 10C cooler under no load (first time I ever heard the computer on without fan noise; it's disconcerting!). Under full load, the temps maxed out at around 90C, whereas I noticed it was knocking on 100C before, so that's an improvement. I wasn't able to reproduce the crashing behavior yet, so I'm tentatively declaring this a success.hat

At the same time the BIOS was updated to 1.07 from 1.05 as another troubleshooting step, so it's possible this improved things somewhat, as well. Unfortunately I didn't do too much testing after the BIOS update so I'm not sure if that fixed the shutdowns but the temps are definitely improved now.

 

What undervolt and frequency are you running?

 

To be honest, I'm not particularly familiar with the BIOS options. The reseller walked me through putting the settings back to how they were when they shipped it to me. It seems like it was fairly mild, though. Temps seemed similar to what the BIOS was after a full reset.

 

You'll be in adaptive voltage mode likely and will have an offset of some kind.

 

Okay so the following options have been set:

In the "advanced" overclocking menu:

Code:

Core/Ring

Max OC Ratio: 0
Voltage Mode: Adaptive
Extra Turbo Voltage: 0
Voltage Offset: 90
Voltage Prefix: -

In the "basic" overclocking menu:

Code:

OverClocking

Platform Power Limit 1 (mW): 90000
Power Limit 1 Time Window (s): 28
Platform Power Limit 2 (mW): 118750
VR Current Limit (A): 720

Core Ratio Override

1-Core Ratio Limit Override: 47
2-Core Ratio Limit Override: 46
3-Core Ratio Limit Override: 45
4-Core Ratio Limit Override: 44
5-Core Ratio Limit Override: 44
6-Core Ratio Limit Override: 43

With exception of the voltage, none of these setting are what they were when I set them over the phone a few days ago but I stupidly didn't write them down as I was changing them. I assume that this is tweaking how the Intel Turbo Boost operates; I suppose if the performance is lacking and I still have thermal headroom I could tweak these some.

I added a thumbnail of the IHS liquid metal application. I didn't intend to mess with that, but the IHS popped off as I was attempting to clean the old thermal compound off of the top of the IHS because apparently they didn't glue it back on. When I saw liquid metal, I considered leaving it for a second, but with that much all over the place, I felt like it was a short waiting to happen.

 

Did you share that photo with anyone at HIDevolution?

 

I told them about what I found, haven't sent a photo yet, wasn't sure if they were interested. It's been a few years since I bought the laptop. I'm still awaiting a response, but since it's a weekend (and a holiday one at that) I wasn't too surprised to not hear back yet.

 

While excessive it does not seem to have shorted the gold contacts at least.

 

Yeah, it seems to have done the job somewhat.

After putting a few more days in, Im going to definitely call this one solved. Haven't been able to recreate the issue yet in The Outer Worlds which was one game that consistently caused the system to shutdown.

Thanks again for all the help

 

Excellent, good to hear