Input Laptop Overheating Issue

[insideman]

I recently noticed that the CPU in my laptop was running far too hot. I own an Acer VN7-592g. It has a 6700HQ.

These were my initial temperatures running Prime95 for 3+ hours (edit, I had already undervolted here, but not cleaned the fans or repasted):

I then undervolted the CPU, cleaned the fans, and repasted the CPU:

There is a significant increase in performance, and overall temperatures are lower, but it’s still apparent that my CPU is throttling. I think the culprit is core#1, running at 95 degrees Celsius. I’m not sure how to interpret this. I think it’s possible that my repaste was not ideal, though it couldn’t have been that bad, since it is an improvement over the previous results. I’m wondering if I should try to repaste again, and also whether I should consider using a more expensive compound. I’m currently in Thailand, and what I was able to find locally available was not that fancy, but seemed decent enough (this is what I used on my previous repaste):

https://www.aliexpress.com/item/Thermal-Grease-S420-computer/32363505833.html

I saw that somebody recommended using liquid metal, the results look promising:

Overall I’m not really keen to keep opening up my laptop, since it is not very user friendly to do so. It doesn’t have a back panel, so I have to go through the keyboard and layer by layer till I get to the CPU that is on the underside of the motherboard. I think 95 degree temp under load on a laptop with not so great out of the box cooling sort of normal, though the throttling is obviously an issue, and I think it should be possible to run this CPU at 85 degrees under load with a normal thermal compound, which I achieved on all cores, but one. Also the iGPU (GPU) temps are lower/low.
I guess the main issues was that my laptop was shutting down due to overheating, and that it was severely throttling (up to -60% in the first image). I think I have somewhat addressed that at least.

[tchizek]

When I have had overheating issues with laptops I have done one of two things, either cleaned the fan/exhaust (at least on the older dell’s this would get gummed up every 6-8 months and the temps would spike), or replace the fans (I did this on my wife’s Toshiba back in 2007 and it added about 2 years, maybe more except that our cat knocked it off the table, shattering the screen and breaking some of the connections on the MB.)

[MazeFrame]

I recently noticed that the CPU in my laptop was running far too hot.

Not uncommon for notebooks. When did you last clean it?

I own an Acer VN7-592g.

You already mentioned it is running hot

I saw that somebody recommended using liquid metal, the results look promising:

Don´t do that. It is sketchy in desktops that do not get moved a lot, in notebooks, it is a disaster waiting to happen.

[mkk]

Yeah this thin laptop design should not be able to do much better with a 2015 quad core CPU. Fortunately in any typical usage case it’s not going to be loaded near full for hours on end.

In case you need to have it running a long video render or something similar on a hot day, adjusting the Windows Power Options for the CPU so that the max allowed clock is just a bit below 100% can take the top off the heat. If it might end up close to throttling anyway then the loss in raw performance might be negligible.

[SgtAwesomesauce]

Here’s my 2c.

If you prepare your system properly, take your time, and are careful, liquid metal is an acceptable solution. I’m thinking about hitting my XPS 13 with some LM later this year, but I’m frankly not concerned with it’s thermal performance at the moment. It would be more of an experiment than anything else.

Don´t do that. It is sketchy in desktops that do not get moved a lot, in notebooks, it is a disaster waiting to happen.

Not really. If you apply the correct amount, it won’t leak out.

That said, if you’ve got aluminium surfaces, you’re going to be SOL because LM will corrode aluminium.

[Token]

Everyone already hit it- cleaning the fan area/fins, re-apply thermal compound (if your feeling frisky, try liquid metal), make sure the fan is in working condition.

I feel you on opening a laptop, I opened my T420 for the second time and caused a little damage doing it- but it runs cooler now haha. I have another laptop that is getting hot now, but kind of lazy on taking that on- but if it ware getting as hot as yours, its time to do it.

[insideman]

In case you need to have it running a long video render or something similar on a hot day, adjusting the Windows Power Options for the CPU so that the max allowed clock is just a bit below 100% can take the top off the heat.

Yes, that sort of is my use case. Hot country and render-like work loads. It seems the laptop just automtically underclocks when it get’s too hot though, so not sure why I would need to change that in Windows settings? You can see in my screenshots, that it is always underlocking under sustained load, just less after themal paste was reapplied.

That said, if you’ve got aluminium surfaces, you’re going to be SOL because LM will corrode aluminium.

I don’t think there is any aluminium near by where I would be applying the liquid metal. I have a copper heat sink, and I think the 6700HQ surface does not contain aluminium. Not sure where to confirm that though, but I think it is assumed safe to apply liquid metal to the 6700HQ surface.

I also noticed today that Linus made a video about applying liquid metal to a laptop:

I feel you on opening a laptop, I opened my T420 for the second time and caused a little damage doing it- but it runs cooler now haha.

Thinkpad’s should generally be much better to work in though, because they are geared more towards IT professionals where there will be frequent servicing. But yea, generally working in laptops can feel a bit sketchy compared to desktops, because of how “compressed” everything is.

[insideman]

Not uncommon for notebooks. When did you last clean it?

Yea, I hadn’t cleaned the fans for over a year. They were pretty clogged up, and also the region where I am seems a bit dustier than some other places, so that might have been a contributing factor as well.

[insideman]

Just completed a successful application of liquid metal. At least for now my laptop started normally and is running stably. I’m running without undervolting for now, just to have a comparison to the factory settings, though I might undervolt again in the near future if I think I can significantly improve performance. I stress tested using OCCT and it ran for about 1 minute before stopping due to reaching a threshold of 85 degrees celcius. I know the T-junction for my CPU is 95 degrees, but I just wanted to run that test quickly to see what would happen. Originally the test wouldn’t even started, because the CPU would immediately run too hot.

I’m going to run a workload now that stresses all CPU cores and see what max temps are like, and I’ll also probably fire up something on Steam, because I applied the liquid metal to the GPU as well. I’m not much of a gamer, but would be interesting to see how gaming performance is now (not that I have a very good comparison for aforementioned reasons).

Obviously only time can really tell how good my application was, but I feel fairly confident about it. I’m insulating using nail polish, which was recommended by der8auer, so I should be alright. Still feels a bit scary though for some reason. I guess liquid metal in a laptop is always a bit scarier than in a desktop, because of moving it around.

[SgtAwesomesauce]

Liquid metal application should be thin, and the surface tension is high, so this means that the chances of it leaving the application area, while non zero, are extremely slim. I’m a tiny bit wary of it myself, but I wouldn’t lose any sleep over it, especially since it seems that the laptop isn’t working well without the application.

[insideman]

I definitely think that some people over-emphasize the risks. I watched so many tutorials, including the one from Linustechtips where they accidentally shoot liquid metal all over the motherboard and manage to clean it up and get the laptop to work, that I wasn’t really too concerned.

I think it’s just the idea of a cooling agent bricking your entire machine that seems really risky. That can’t really happen with thermal paste. Though even if it doesn’t start after the application, that doesn’t even necessarily mean that your laptop is no longer usable.

There are a lot of ways you can mitigate risk by practicing applying liquid metal away from your CPU first. I actually managed to release far too much liquid metal like Linus, but I was practicing on toilet paper first so I didn’t have to clean up any of the electronics. If using Thermal Grizzly’s liquid metal, I would definitely recommend doing that, because the syringe seems quite stiff at first.

[insideman]

I’ll run the same benchmarks I ran here tomorrow for comparison.

Here is the only picture I have of the application. Flash was very reflective on the GPU.

[Calculatron]

Coincidently, I was thinking of cracking open my laptop and re-applying the thermal compound. The stock TIM they use is usually something selected on the basis of its robustness of lifespan and cost, rather something a bit more suitable for the form factor.

I did this on my previous laptop, some variant of an HP Envy x360 with some higher-tiered variant of AMD’s Bristol Ridge, and it was the best thing I ever did. It dropped the temperatures by 15C, and the CPU was both a lot more consistent and frequent with its boost clocks.

I still have some Prolimatech PK-3 which I plan on using. Previously, I had some Grizzly Cryonaut, which I really liked. The liquid metals aside, honestly, any decent name-brand TIM used today is going to perform fairly close to one another, as we all know.

[insideman]

If any thermal issues whatsoever are experienced, it’s probably a good idea to repaste with a high quality compound. The paste job they did from the factory was really messy and had a lot of excess thermal paste. It actually took me quite a while to clean that from the transistors and capacitors.

[insideman]

I’ve completed some benchmarks now for my performance with liquid metal:

Liquid metal with no undervolt:

Liquid metal with undervolt:

Liquid metal with undervolt and Acer Cool Boost:

All test were done using a Prime95 blend test on 8 threads. The screen shots are just a snap shot of the clock at the time it was taken. Most clock speeds would vary by +/- 5%. The tests were conducted in a room with ambient temperature of approximately 30 degrees celcius to reflect the working conditions that are often present when the laptop is working.

FIrst I ran with no undervolt to see how the CPU would perform with the stock settings. This resulted in a clock speed of approximately -15% from the base frequency. This already outperforms my previous best of -30% using my Shenzhen Alseye S420 repaste.

Core #1 was the highest again on all tests. I made sure that my spread was even on both the CPU and the area where the heat sink makes contact with the CPU, so I don’t think I should have hot spots. Not sure what’s causing this, it seems to just be unlucky.

Next I reapplied my undervolt of -0.14v to the CPU Core and CPU Cache using ThrottleStop. This seems to have quite a large impact, as my performance improved to approximately 0%, which is still throttling, but at least not underclocking.

Finally I found out that my laptop, Acer Aspire VN7-592G Nitro Black Edition, can take advantage of Acer’s Cool Boost utility. It does seem to noticeably ramp up the fans and initially seemed to be making a difference, but after 3 hours, it doesn’t seem to outperform not having the utility.
I think there might be some benefit however if the fans are that much more active. The margin for error is also quite large using my methodology (varying test length and ambient temperature), so I can only really measure large differences. I feel like the cool boost might be adding 1-2 extra percentage points in performance.

All in all, I would say the paste was a success. It’s a shame I couldn’t prevent the laptop from throttling, as it seems a lot of people were able to do with other brands. I think I also had much farther to go than others, seeing as I started at -60% overclock (perhaps even more as my CPU was already undervolted at that point) under load.

I’m going to run some longer work loads now to test for stability, as I had issues with a previous undervolt under full load, but I think overall my performance is much better.

[insideman]

I just started my overnight work load, which pegs the CPU at 100%. At the moment I am also using my AC, so ambient is at about 22 degrees celcius.
My laptop doesn’t appear to be throttling under these conditions, and is running at under 90 degrees celcius on core #1 with an overclock of approximately +20%.

I thought I would share this, as this represents a more realistic situation for most people. It also demonstrates presumably the difference ambient can make.