RECOIL II LAGGING FREQUENTLY

[jeakejeakejeake]

UPDATE 12/15: Solved. Piece of duct tape was stuck in fan inlet during laptop assembly. PCS and fan assembly manufacturer deserve blame for not noticing this during testing.

Since the first day, my Recoil II has a lagging issue.

When I open a (graphics-intensive) game, the laptop initially performs well. As the heatsink heats up (up to 90+ 'C), lagging events start to happen every minute or two. The frequency of the lagging seems to depend on heat - more graphics-intensive games lag more often. The lagging happens slightly less often if the fan's Turbo mode is turned on. The laptop's specs:

Processor (CPU)	Intel® Core™ i7 8750H
Memory (RAM)	16 GB Corsair 2666 MHz SODIMM DDR4
Graphics Card	NVIDIA® GeForce® GTX 1060 - 6,0 GB
1st Storage Drive	500 GB WD Blue™ 3D NAND 2,5" SSD

Here's what I've tried so far:

1. Drivers reinstalls (both for NVidia and Intel). Windows re-install. No good.
2. Thermal paste re-do (cooler master maker, which should be good). No good.
3. Replace original 150w adaptor with 180w adaptor. No good.
4. RMA. The device was returned and tested, and I was told the tests all turned out fine.
5. Extreme CPU under-volting + turning off the CPU's Turbo mode (via throttlestop). This helps by delaying and weakening the lagging. Lag still happens every few minutes. Also is a very bad solution - basically making the CPU much weaker to compensate for another issue.

Is anyone else experiencing these problems? Is this normal for Recoil II laptops?

Any idea what to do next to possibly fix this problem?

My guess right now is that there is some deep problem with the main board, possibly in how it manages power under high thermal load. I really want to avoid sending it for an RMA, since it's already been sent in twice.

Cheers.

 

[FerrariVie]

Your lagging is very likely what it's called thermal throttling. Your laptop is so hot that it's throttling the performance to prevent reaching temperature limits.

You need to try and bring the temperatures down. Have you identified if the issue is with the GPU or CPU? Because you've mentioned undervolting the CPU, but if your problem is the temperature of the GPU, undervolting the CPU might not help that much. Can you run hwinfo, play your games until it throttles for a few minutes and post back a screenshot of the temperatures and clocks of both CPU and GPU on Hwinfo? It will gives us an idea of what exactly is throttling and by how much, as Hwinfo shows min and max recorded values.

 

[jeakejeakejeake]

Your lagging is very likely what it's called thermal throttling. Your laptop is so hot that it's throttling the performance to prevent reaching temperature limits.

Yes, I am also pretty sure it's thermal throttling. But why? And what can I do to prevent it?

Have you identified if the issue is with the GPU or CPU?

I've logged limit reasons for the CPU on Throttlestop, and none of the reasons pop up when the lag is happening. However, the GPU runs at about 92 'C (whereas the CPU is at its maximum about 88 'C). At that point, there's a drop in frequency and the game lags. The laptop's PROCHOT seems to be set at 95'C, so that is not triggering (and it does not appear triggered in the limit reasons in the Throttlestop log). So it seems like the GPU automatically throttles at about 92 'C.

Can you run hwinfo, play your games until it throttles for a few minutes and post back a screenshot of the temperatures and clocks of both CPU and GPU on Hwinfo?

I'll do this later when I have time, and the info above might be enough for this purpose.

What could be a solution here? With this bug, it's basically impossible to play any graphics-intensive game.

Also: Any other Recoil IIs experiencing this? Or is this an isolated issue?

 

[jeakejeakejeake]

Update: it is indeed GPU thermal throttling. HWinfo didn't have a good display format, so I used GPUZ to log and display GPU clocks and limit reasons over time. The CPU/GPU sometimes power throttle, but that doesn't cause lag. However, when the GPU thermal throttles (at about 91'C), there is a lot of lag. Log is attached.

 

[FerrariVie]

Update: it is indeed GPU thermal throttling. HWinfo didn't have a good display format, so I used GPUZ to log and display GPU clocks and limit reasons over time. The CPU/GPU sometimes power throttle, but that doesn't cause lag. However, when the GPU thermal throttles (at about 91'C), there is a lot of lag. Log is attached.

Now it's a bit easier for us to find a possible solution. 91° is far too high for a GPU, they will usually start throttling at around 83° and ideally should no go above 86°.

I find it weird that your GPU temperature was 67 degrees when you started logging (0% GPU load), but I don't have a way to know if you're playing something intensive before logging or if it was just at idle. If it was idle, then there is something VERY wrong with your laptop and definitely is not normal. When you got to 99% GPU load, it was drawing 50W only (I'm assuming that your card TDP is 80W) and already thermal throttling.

When you repasted the CPU, I'm assuming you did that on the GPU as well, right? And are you sure that the GPU fan is spinning ok? If your throttling was happening just by a bit or if your idle temperatures were fine, then I would suggest GPU undervolt/underclock. But in your case, it's not the best solution, as you figured out when trying to undervolt the CPU. You need to find the root cause first, or else you'll end up with a GPU that's going to perform at half of its potential, just to prevent thermal throttling.

Just for your information, my 2070 has a TDP of 115W and it reaches a maximum of 80° at 70% fans, no matter what I throw at it. I know that our chassis and GPU architectures are different, but my GPU still has a TDP that's a lot higher than yours and should run hotter (in theory). My idle GPU temperatures (with 21% fans) are around 42° 38°.

 

[jeakejeakejeake]

I find it weird that your GPU temperature was 67 degrees when you started logging (0% GPU load), but I don't have a way to know if you're playing something intensive before logging or if it was just at idle. If it was idle, then there is something VERY wrong with your laptop and definitely is not normal. When you got to 99% GPU load, it was drawing 50W only (I'm assuming that your card TDP is 80W) and already thermal throttling.

I left the laptop idle just to check, so about 5 mins no load. GPU temp stabilizes at about 65 'c, while CPU temp is about 60 'C (note that GPU is completely idle, but CPU has some low wattage background use).

When you repasted the CPU, I'm assuming you did that on the GPU as well, right? And are you sure that the GPU fan is spinning ok?

Yes to both. I've tested the fan physically though, only by checking that it spins well and it creates good airflow.

Just for your information, my 2070 has a TDP of 115W and it reaches a maximum of 80° at 70% fans, no matter what I throw at it. I know that our chassis and GPU architectures are different, but my GPU still has a TDP that's a lot higher than yours and should run hotter (in theory). My idle GPU temperatures (with 23% fans) are around 42°.

Then clearly something is wrong here... What would be a good way to diagnose or solve the issue?

 

[FerrariVie]

I left the laptop idle just to check, so about 5 mins no load. GPU temp stabilizes at about 65 'c, while CPU temp is about 60 'C (note that GPU is completely idle, but CPU has some low wattage background use).

Then clearly something is wrong here... What would be a good way to diagnose or solve the issue?

Just double-checked it now and my GPU idles at around 38°, with nothing else opened (just normal background processes, which I think doesn't consume GPU resources). It actually goes into the mid-40s when streaming video, especially at high-resolutions or refresh rate. Regarding the CPU, idling at around 60° it also a bit high for a laptop CPU. Which power profile are you using on the control/gaming center software? In theory, it should not affect the GPU, but it could be that your CPU is making the GPU hotter, since they share the heatpipes.

It looks like you've already done most of the things that I would recommend to bring temperatures down: repaste and clean fan dust. You could maybe try a different thermal paste? I never heard about the cooler master one you used, but pastes are not something that I have much knowledge about and it can actually be a good one. What I've seen most people talk about recently are the Thermal Grizzly Kryonaut and the Noctua NT-H1.

Have you also checked the thermal pads around the CPU and GPU chips? I've seen people that had issues with those, either because the pads moved to a wrong position or because they became torn with time.

Another thing to look for (if everything fails) is the surface where your laptop is on. A laptop stand or (even better) a cooling pad could help to improve your temps. I have mine on a stand that doesn't cover the back panel (it's a kind of a metalic mesh), so a lot of air can flow inside the chassis.

I actually don't know how the Recoil II looks like. Would you mind sharing some pictures? Is it thin and light or thick and heavy?

 

[jeakejeakejeake]

Which power profile are you using on the control/gaming center software? In theory, it should not affect the GPU, but it could be that your CPU is making the GPU hotter, since they share the heatpipes.

High performance - but this shouldn't matter. The CPU isn't pulling anywhere near enough watts to have any noticeable effect on heatsink temperature. After undervolting and de-thurboing, it's using 15 watts.

You could maybe try a different thermal paste? I never heard about the cooler master one you used, but pastes are not something that I have much knowledge about and it can actually be a good one.

I used the same "premium" paste that was applied when I ordered the laptop. I say premium because I paid a little extra for it. Also, thermal paste shouldn't have such a big effect here...

Another thing to look for (if everything fails) is the surface where your laptop is on.

The problem is surface-independent. Tried this too.

Have you also checked the thermal pads around the CPU and GPU chips?

I'll check this next, when I do another repaste and fan clean, before I contact PCS for an RMA (I don't see any other solution ). I highly doubt this is the problem, as this problem has existed since the laptop arrived.

I actually don't know how the Recoil II looks like.

It's the same as e.g. Schenker Neo 15. Here's a review of that with many pictures and discussion of the laptop's chassis (also note the temperatures they report):

Schenker XMG Neo 15 (i7-8750H, GTX 1060, Full HD, Tongfang GK5CN6Z) Laptop Review

We review the Schenker XMG Neo 15 in the configuration with the Core i7-8750H, GeForce GTX 1060, and 144 Hz panel. How does the slim 15-inch system compare against the gaming competition?

www.notebookcheck.net

 

[FerrariVie]

High performance - but this shouldn't matter. The CPU isn't pulling anywhere near enough watts to have any noticeable effect on heatsink temperature. After undervolting and de-thurboing, it's using 15 watts.

I agree that this shouldn't matter if you're undervolting and underclocking, but isn't it redundant and unnecessary? It's like you're asking the CPU to be delivered the highest power possible (even above normal TDP, it's almost an overclocking), but then undervolting it at the same time? I would use gaming mode (I think that's the intermediate mode name on tongfang chassis), do an undervolt and not touch on clocks or TDP. And when not gaming, just switch to office mode.

High-performance is the same as "benchmark mode" and should not be used all the time.

I used the same "premium" paste that was applied when I ordered the laptop. I say premium because I paid a little extra for it. Also, thermal paste shouldn't have such a big effect here...

I kind of agree with you there, but we're running out of options, right? I think it's a £5 or £6 well spent to try something different (like a kryonaut).

The problem is surface-independent. Tried this too.

Have you even tried a cooling pad? From the pictures that I could see, there is already a kind of mesh on the bottom of your laptop, so I think it would indeed benefit a lot from a cooler below that.

I'll check this next, when I do another repaste and fan clean, before I contact PCS for an RMA (I don't see any other solution ). I highly doubt this is the problem, as this problem has existed since the laptop arrived.

It's the same as e.g. Schenker Neo 15. Here's a review of that with many pictures and discussion of the laptop's chassis (also note the temperatures they report):

Yeah, they're also saying that it runs hot, but they also say that the CPU never went above 90°, which is not the case for you. And if you have an undervolt, it should be even below those stock temps, don't you think? I remember hearing back in the day that the 8750H was a hot CPU, but also saw a lot of people saying that undervolting solved the problem for most of them. I think that was when undervolting started to become more common, exactly because of the 8th gen intel i7.

i7-8750H undervolting questions : intel (reddit.com)

In summary, the options that I would do are, in cumulative order of priority:

1. Kryonaut/NT-H1 + cooling pad, also checking how the thermal pads look like. Neither of those will be good by themselves, but together they might show some results;
2. If you have a bit of improvement on the below but that was not enough, then you could use gaming mode in gaming center, together with undervolting the CPU and undervolting the GPU. Not something crazy heavy, but between 125-250mv on CPU (whatever value you can get stable, but not changing TDP or clocks), while putting the GPU run at clocks to either stop the power limit throttling or just top at max boost specs (1670 MHz, I believe). I can help you with the GPU part, in case you don't know how to do it;
3. Not sure if an RMA will help you there. From what I have seen on the Reddit thread above, it is quite common for that CPU, and your laptop is a bit on the thin size for that amount of heat. I found the fans quite thin from the pictures, so not sure if they can draw and blow a lot of air. But you can try again if you want.
4. If none of the above work, then the last thing to do is to get some money for your laptop and buy a new one, with either a more efficient CPU or a chassis with a better cooling solution (usually the thicker the better, as fans will be bigger and move more air around).

Don't skip step 1, as step 2 will only be good enough if you have some results on the previous. You could indeed skip 3, in case you find good money for it or don't want to wait some weeks without the laptop for the RMA process, especially knowing that PCS might just reply saying that it is expected behaviour.

 

[jeakejeakejeake]

Thanks Fevieira, I ordered a kryonaut and will try that to see if it helps. According to PCS support, Recoil chasses are very good at dissipating heat, so this behavior is strange. The CPU isn't the problem - undervolting it doesn't prevent the GPU from heating up (and GPU-only benchmarks create the same thermal throttling).

If the kryonaut paste doesn't fix this, I'll ask for an RMA.

 

[jeakejeakejeake]

The problem is solved. Here is what I did:

1. Removed Cooler Master Maker thermal paste, replaced with Arctic Grizzly Kryonaut.
2. Clean fan intakes and outputs and fans.
3. Remove a stuck piece of duct tape from the fan inlet on one side. This is the duct tape that is normally surrounding the connection between inlets and fan assembly, ensuring that their connection is isolated. Instead of surrounding the inlet, it was stuck in the gap between inlet and fan assembly, causing almost complete blockage for the fan inlet air flow.

After all these steps, the laptop is performing excellently. I load-tested it multiple times, pulling about 100 watts shared between CPU and GPU assemblies on full fans, and its final temperature after 30 minutes was 75 'C. This is far far below the 96 'C that were seen before. Moreover although power limits were triggered here and there, there were no thermal limits triggered for either GPU or CPU. So I'm fairly certain the issue is solved.

I am also almost certain that the last point, the stuck piece of duct tape, was the cause of the problems. The thermal issue was a problem that has existed ever since the laptop was opened, and the fan setup is clearly something that is fully set up in the Chinese factory. There was no glue or residue anywhere around the duct tape, so I'm fairly certain the fan, since assembly, had looked like this. I am surprised that a faulty component and laptop could go through testing and come out the other side. The Chinese factory's mistakes aside, it's appalling and disappointing that PCSpecialist has not only verified the laptop as "100% working" before shipping it out, but also ran an entire RMA on its thermals and said everything was fine. I later realized this is why I didn't even consider in these two years of use that the laptop's internal assembly might be faulty and I should personally inspect it.

This was the first time I bought something from PCS and am now very hesitant about buying from them again, in a few years when a new laptop is needed. I certainly won't be recommending PCS to anybody.

 

[FerrariVie]

My thoughts are that it is hard for PCS to identify that extra duct tape on the fan when doing the assembly. Even on the initial testing, not sure how much time they spend doing it and how easy would be to identify that issue. You have repasted the laptop yourself and didn't find it the first time, right?

However, I do agree with you that they should have spent more time investigating it on the RMA, as that was the right moment for them to find the issue and when the real a lack of quality assessment happened (in my opinion).

 

[AgentCooper]

I would suggest that you inform PC Specialist of this and include the link to this thread, it’d be useful for them to know.

 

[jeakejeakejeake]

My thoughts are that it is hard for PCS to identify that extra duct tape on the fan when doing the assembly. Even on the initial testing, not sure how much time they spend doing it and how easy would be to identify that issue.

I've asked details about their testing procedures and stress-testing the laptop under high load is one such test. There is no way (other than potential negligence) that 90+ ('C) temperatures would not be noticed on a 30-minute stress test.

You have repasted the laptop yourself and didn't find it the first time, right?

Indeed, I had not. I didn't expect there to be a problem with the fan assembly, as the multiple prior tests (before shipment and during RMA) left me ensured that at least specific components are working correctly. My guess at that point was software error, or potentially poor pasting on my part (neither of which was the case).

I would suggest that you inform PC Specialist of this and include the link to this thread, it’d be useful for them to know.

I might at some point. Right now though I'm exhausted by all the hours I've spent on this and will try to put the frustration behind me before I contact PCS. Also, you might want to watch this.

 

[AgentCooper]

Also, you might want to watch this.

Yowzer! When I’ve got four and a half hours to spare, I am all over that! 😀👍