Lenovo Legion 5 (Ryzen 7 4800H) (Geforce 1660Ti) Thermal Issue

Hi guys, I re-pasted my Legion 5 and now I'm getting even worse scores on Cinebench R23.

Initially, I was scoring 11327 stock thermal paste, now I'm scoring 10310 after re-paste?

I used Arctic MX-4... surely this paste is better than what they apply by default?

Why am I scoring nearly 1000 points less after re-pasting?

What am I doing wrong?

I've tried re-seating and re-pasting the heat sink at least 5 or 6 different times and it's always the same result.

There's no way the generic paste is better than MX4 or am I missing something?

I am regretting re-pasting, I feel like my machine is busted now, not to mention I broke 2 plastic clips trying to get the back cover off even though I was extremely gentle.

Does anyone have any good advice on how to get my machine performing like it was out of the box?

 

Well, it would be best for you to check CPU temps whenever you are running R23, and see if there's any thermal throttling going on.

I would assume its the MX4. I used to use MX4 to re-paste laptops, and its good for the first few days and then goes to crap as it is a very light, liquidy paste. If you want some reliable paste that is very viscous and won't move around too much, I would recommend Thermalright TFX or Phobya Nanogrease.

 

It's definitely thermal throttling! I'm reaching temps up into 100C.

With stock paste, I was getting around 96C and I don't think my cpu was throttling in cinebench.

 

Yeah, the stock paste probably wasn't good, but it was firm, and it ensured contact between the CPU die and the heatsink, unlike your MX-4 right now. Again, I'd recommend buying one of the 2 pastes I told you about, and if you are really paranoid, just go with TFX, I've used it in some pretty crappy heatsink contact laptops and it helps out a lot.

 

Contrary to tenoroon's experience, my experience with MX-4 has been perfectly normal.

If you are truly getting worse temps, I would be concerned with how you are applying the paste and reapplying the heat sink

 

Had the same issue re-pasting with MX-4.

I have some Kryonaut but I thought that was even worse for laptops and longevity. Should I re-paste with that or buy something more solid?

I spread the paste as well as opposed to using a bit in the middle as it's being applied to the die and not a lid. Is that wrong?

 

The grain of rice in the middle is the best way to do it. Let the pressure of the heatsink spread the paste.

 

Even on direct die applications?

My concern is just that MX-4 is not viscous enough

 

Personally the line method works better for AMD CPUs because of the bigger dye. Im getting better temps than the first time i repasted with the rice grain method.

 

Viscosity has nothing to do with the conductivity or effectiveness of the thermal paste, only how easily it spreads on initial pressure/application. Not sure where that idea comes from.
Some people have gone to great lengths in the past to try and decrease viscosity by heating up the paste such as putting the tube in a plastic bag and setting it in warm water for a bit, to make it spread better initially.

From experience and repasting multiple systems, especially the same systems multiple times, I can assure you that this has absolutely no effect on the end result, and that if there is enough paste to cover the die, it will. Sometimes it will take a couple heat cycles to get to the very edge if you were particularly skimpy with the paste, but it has also been proven that 'too much' paste doesn't negatively affect temperatures either. If you live in a very cold climate and the paste is too thick, then just put it in a pocket for a few minutes before you need it.

I use whichever method is appropriate to the shape of the die; long skinny die, like old intel dies, long skinny line. Square die, dot in the middle. Whatever pattern you apply it will squish into a circle at the end anyway. Whether it is a bare die or not, it makes no difference. You can just get away with less paste on a bare die.


Have the people having issue ever repasted a CPU before? Because I would be more suspect of the application/reapplication method, or that you may have gotten a counterfeit product.
You may want to explain how exactly you are doing, to ensure there is nothing you are doing wrong in the method as that is the easiest to troubleshoot. Even a generic real product should not have abnormally worse performance than stock if applied correctly.

The first poster mentioned applying paste to the heatsink. I don't know if he misspoke or was actually applying the thermal paste to the heatsink, the die, or both. Don't do that.

For Example:
1. remove heat sink from die
2. clean off heat sink and dies from all old paste (may use alcohol and tissue, or cloth, or paper towel)
3. apply appropriate amount of thermal paste to cpu/gpu die
4. line-up heatsink and place directly down (no sliding) and torque screws in order, half turn at a time to ensure even application of pressure.
5. retest temperatures (before you put the back plate on the laptop if you are not sure if you did it right)



Some issues (don't laugh, I have seen all this, and some people just don't know)
1. not cleaning off paste from heatsink also, only die,
2. sliding heatsink around over die so all thermal paste is scraped off.
3. torqueing down one screw completely at a time (unless you have constant pressure on the die) resulting in the heatsink getting torqued down one side at a time pushing out the thermal paste.
4. taking off and putting back on the heatsink without changing the paste after checking it spread right, (doing so causes bubbles to form and air is not a good conductor, it is insulator).
5. misaligning heatsink so it does not make contact with the die
6. misaligning thermal pads so heatsink does not make contact with die
7. using not enough paste
8. using way too much paste
9. applying paste to both heatsink and die (only put it on the die)
10. not taking any paste off anything and just putting new paste onto old paste and putting heatsink back on
That is all I could think of off the top of my head

 

Not sure about OP but I've re-pasted previous laptops before (mostly Dells) and I've seen fairly large improvements.

The reason I mentioned viscosity is because I thought the thinner and more easily spread paste is more likely to succumb to the pump out effect.

It should be genuine as I bought it direct from Amazon and the packaging looks legit.

I ordered some SYY 157 so I'll try that, and if I don't see any improvements I'll try some Kryonaut I have because why not.

I don't care about big improvements, I'm just annoyed that I seem to have f'd it up and wish I'd left it as it originally was

How hot does the CPU in your Legion 5 run if you have one?

 

Repasting is useful if you're experiencing high temps and thermal throttling. If you've good temps even with the stock paste then leave it alone and just clean the fans. Repasting is a trial and error process you're not going to get it right every time. SYY-157 is a underrated thermal paste which work great for my legion 5 as i got it right this time. Make sure you use enough to cover the dye though.

 

Just repasted with SYY-157

Couldn't get it to spread at all, it just transferred back to the spreader.

Ended up putting (rather generous) lines on, and then hoping the heatsink would spread it.

After booting back up, Cinebench score of 4600+ and max temps of 85C (without the fans spinning up fully)

No thermal throttling like before either, so I'll have to see how it is longevity wise, but already a big improvement over before.

 

I have never experienced or seen this 'pump out' which I assume to mean that the paste eventually completely moves out from between the die and heatsink.
I used to experiment with ultra thin applications of thermal paste using bare dies and heatsinks back when the theory was that too much paste would harm temps. This turned out to be false. I think the pump theory is the same.

The purpose of thermal pastes is only to make up the difference or unevenness caused in manufacturing the heatsink, torquing down, and die. Ideally there would be no interface material needed with perfectly machined surfaces. This is not practical. So a thin layer of thermal paste makes up the difference. Therefore we can assume that any imperfections will 'trap' the paste where it is needed, and any location that has perfect continuity is best with direct contact and no paste needed.

But if you have a good application, leave it be. Any excess from the application will ooze out over time with use and what is needed will stay.

My temps under standard benchmark load were about 90/79 CPU/GPU for the Legion 7 with Arctic Silver 5 (not cured)
Same test on the 5 Pro was 90/76 - Arctic MX-4
This is after running for 1/2 hour so all temps should be stabilized.

5900/3080 got close to 15% better performance.


EDIT
So apparently my suspicions were not wrong on fake MX-4 paste being out there now

 

That good to hear. It should have good longevity like normal paste so dont worry about it.

 

I just ordered some Thermalright TFX. I hear it's really good for laptops, I will apply it and post my results here.

Hopefully the TFX will solve my problem.

 

Just wanted to update... I applied the TFX and I am now scoring 11,000+ on cinebench R23 and temp is maxing out at 96-97C.

Problem has been solved, wasn't a heatsink issue.

Do not use Grizzly Kryonaught, AS5 or Arctic MX-4 on a legion laptop.

Use SYY-157 or Thermalright TFX.

Thermalright TFX looks to perform well with low mounting pressure. Perfect for laptops!

 

One thing I should mention is my idle temps are unchanged between the 4 pastes but the Thermalright TFX seems to draw heat into the heatsink faster.

 

Update on the Thermalright TFX.

The paste seems to have pumped out.

Right after repasting I was scoring 11k+ on Cinebench R23 and a few days later I'm only scoring 8,900.

Never should have taken my laptop apart the stock paste was fine... Jeeze.