[SOLVED] Applied Arctic Silver 5 to Acer nitro 5 AN515-53 on CPU and GPU, temperatures seem higher or not changed at all

[ShangWang]

Hi all,
I recently bought arctic silver 5 and applied it after cleaning off the 1 and a half year old residue thermal paste on my acer nitro 5, it was indeed dried up. The factory paste seemed to also be arctic silver 5 from the looks of it but I'm not sure.

My fans and exhaust are clean and there is plenty of ventilation for the laptop.

I applied what I thought seemed like enough, roughly pea-sized but maybe smaller and put a line of the paste on my GPU.
After testing, the temperatures seemed to go from 40 to 45 degrees Celsius idle on CPU, but it fluctuates at times.

When playing hollow knight I get about the same exact temperatures sitting at 60 degrees Celsius.

Is it normal to have temperature increase when initially applying the paste, might I have put too little on my CPU or does it simply take curing time for it to improve?

If I put too little can I simply add more thermal paste to the already existing one or do I have to start from scratch? I've only used my laptop for about 6 hours from now.

 

[tennis2]

The "pea size" is for desktop CPUs. You likely put too much on.

Generally you'd expect to see better temps as a result of using (assumedly) better TIM. To see a temp increase is unexpected. I'm sure it's a PITA, but if you remove the heatsink again and see TIM squished off the edges of the CPU/GPU, you applied too much.

 

[rubix_1011]

Changing thermal paste alone isn't likely to make much of a difference. I tested about 15 pastes and there was maybe 1-2C difference overall for most traditional pastes.

 

[ShangWang]

The "pea size" is for desktop CPUs. You likely put too much on.

Generally you'd expect to see better temps as a result of using (assumedly) better TIM. To see a temp increase is unexpected. I'm sure it's a PITA, but if you remove the heatsink again and see TIM squished off the edges of the CPU/GPU, you applied too much.

I really wouldn't like to redo the paste right after doing it. I don't think the temperatures really went any higher, my average is 40-50 degrees Celsius at least. I will wait for a week to see if temperatures are any better, if not I will leave it. Turning off turbo boost is the only thing that keeps my CPU from going over 80 when playing high demand games, so should I be fine?

 

[tennis2]

It's considered "normal" for laptops to power/temp throttle under sustained load.

You may have some luck with undervolting. That would at least preserve some amount of >base clock performance.

 

[ShangWang]

It's considered "normal" for laptops to power/temp throttle under sustained load.

You may have some luck with undervolting. That would at least preserve some amount of >base clock performance.

I'm currently using throttle stop and undervolted CPU, GPU, and CPU cache to 150mv, the differences aren't very noticeable but it has done some good.

 

[ShangWang]

It's considered "normal" for laptops to power/temp throttle under sustained load.

You may have some luck with undervolting. That would at least preserve some amount of >base clock performance.

In the case that I did apply a bit too much thermal paste, is it really worth it to re-do the job all over again? I don't see how a little too much will make a huge difference.

 

[USAFRet]

In the case that I did apply a bit too much thermal paste, is it really worth it to re-do the job all over again? I don't see how a little too much will make a huge difference.

'too much' can be an insulating layer, rather than a heat transference layer.

It is hard to use 'too little'.
Less is more.

But if temps are the same as previous, leave it.

 

[Paperdoc]

Agree with all above, but I'll add this. The problem MAY be simply not understanding, and expecting the wrong thing. Cooling of the CPU chip is really done as TEMPERATURE control, NOT a fan speed control. The CPU has a temp sensor built into it, and the mobo uses that as its guide. It has a target temp, and will adjust the cooling fan speed to whatever it takes to get that target. Even assuming that your work actually did improve thermal contact, the result would be that, to achieve the SAME temp target as before, the system will merely run your cooling fan a little slower. This does NOT change the temp actually achieved. The fan speed reduction is likely to be small so you would not notice that effect unless you had previously recorded typical fan speeds. The more subtle effect is this: at very HIGH workloads (and heat generation), your CPU temp may run a bit cooler because the fan still has not reached its max speed and has more cooling capacity available.

 

[tennis2]

^ Very true