Asus GL702ZC throttling

[ProtosAngelus]

Hello Guys, I have a question, don't know if anyone experienced anything similar.
Specs: Asus Rog Strix GL702ZC

• RX 580
• Ryzen 5 1600
• 8 GB RAM
• Win 10x64
• SSD primary, HDD secondary
• Updated BIOS
• Memory diagnostic no errors
• dxdiag no faults

So about 2 weeks ago I had my GL702ZC cleaned out and repasted because the idle temps were 65+ deg C (CPU and GPU). After the local service center returned the laptop the temps were cooler but anytime any GPU intensive app was started, the computer would shut down in an instant (Google earth, 8 seconds, video games, 2 seconds, GPU Novabench, 1 second...) if it were on any profile but battery saver since battery saver didn't boost the memory clock to 2000MHz (games still shutdown laptop on bat sav). So I gave it back and they repasted it again. Now every game/3d program can be launched perfectly without shutting down the laptop.
However, new issue arose. The moment the GPU reaches 75-78 deg C, the GPU core clock starts throttling. https://i.ibb.co/NZkpJF1/HWlog.png You can see where the game is launched (temp spike)
If I start a game from cold boot, when the temps are low. Eg, Dota 2 right as the Windows boots (SSD). It will run smoothly and perfectly on highest settings 60 FPS for a short time. As soon (about 10 seconds into the actual gameplay on the map) as the GPU reaches ~75 deg C, the FPS will be all over the place... The "minus" is half a second. (maybe something else is throttling?)
60-20-20-60-60-20-20-60-20-20 FPS and so on and so forth. Jumping wildly ( https://i.ibb.co/hWMMPtg/Log.png this is another log of another test which I started logging as the GPU was throttling).
Next week I'll be getting the laptop to a top class service center (that often fixed soldered GPU/CPU/mobos) for them to do a proper re-paste.
Things that have caught my attention.
- temp jumps by 5-15 degrees in an instant of plugging in the power cord if the laptop is under load (no jump if no load, I'm guessing the boost warms up the chips which should be normal)
-eg. Running Furmark for 10 seconds on battery, lower FPS, ~ 60 C temperature. Plug in the cord. Temperature jumps to 75 in an instant, higher fps, fluctuations and stutters and lags.
(I never ran Furmark for longer than 20 seconds cause it's more stressful than watching the World Cup)
- VRM GPU temp1 is constantly 26 deg C, under any sort of load
- VRM GPU temp2 is 56 C idle and 65 under load, always caps at 65, not a .1 degree more
- GL702ZC had thermal paste on all the VRMs, VRAMs, CPU and GPU out of the factory, but the service crew put thermal pads (too thick/thin?) on VRMs and VRAMs and paste on CPU and GPU.
Any idea why the GPU started throttling so soon and why it's spitting the core clock all over the place? Could it be that the GPU VRMs are badly conducted by pads or an inefficient thermal paste on GPU?

Laptop bought last year.
Before this sh**storm, the laptop could play Witcher 3 on high/ultra at 60 FPS steadily. GPU would reach 80-85 degrees and CPU 80-86 degrees while idle temps were 60+
Idle temps now are 48-53 GPU and 50-60 CPU (it will boost itself when opening programs and stuff so warms up from 55 to 60 in a second and then drops).
Dota 2 (highest settings). GPU 75-78 degrees. Sometimes it won't go over 75.000000 deg C, sometimes not over 78.000000 deg C.
CPU 70+ under load.

Any thoughts before I turn in the laptop to the service center? My warranty is expired so no-no on the RMA's or anything. It all started after the opening of the heatsink to re-paste. As I'm writing this I'm noticing that the keyboard above the GPU is quite hot (only thing I've been doing past 2 hours is studying physics so nothing was strained except me) and on the left of it, above the CPU, it's quite cool.
Right now GPU is 60 degrees and CPU is 58-66 (fluctuating depending on clock, 1500MHz->3400MHz) degrees Celsius
GPU VRM temperature1 is 26
GPU VRM temperature2 is 57

 

[Kyron71]

Find a board specific for that laptop.

Off hand I believe the temp sensors are telling the graphics card to start throttling to early. I had a Alienware with this problem. Was able to adjust temp to run a little hotter before throttling.

 

[ProtosAngelus]

Board specifics? What do you have in mind exactly? I did do a BIOS update and the UEFI is really weak on modifications. Only securty, USB, SMART, Virtualization and such. Nothing about clocks or voltages or temperatures to adjust in the BIOS. I could do a GPU Bios update if necessary. Is it possible that the contact is bad or the thermal paste is inadequate or something of sorts because I ran Dota 2 after 20 minutes of browsing the web. The GPU was at 50 degrees C. I launch Dota, GPU temp starts climbing from 50 to 55 in 20 seconds of Dota 2 opening (installed on HDD). I immediately open a Bot game so it starts to render heavier stuff for the load to kick in. In about 5 seconds the temperature jumped from 55 to 75 degrees C on the GPU and I'm pretty sure it would keep jumping if not for the throttle. Before this problem I had a problem where the laptop would shut itself down in a second of heavy GPU load (prob temp went from 55 to 105 in a second). Something is overheating just don't know what.

 

[fire400]

sounds normal. you probably got a crap board after they repaired it.
additionally, to save the manufacturer money, they probably modded your board to lock/shutdown after certain temps, they can implement any firmware they want.
and if the board is pre-defective or refurb, not all manufacture testing software can catch bad boards.
Does Asus have modding software that you can adjust?
ideally, with gaming laptops, the cooling sucks in the first place. most gaming hardware should be cooled adequately like most desktops. laptop lifespans are hit and miss.
sorry you're having the issue.

if you've tried a new SSD and reseated the RAM and loaded a fresh copy of win10 x64, upgrading or downgrading BIOS, and fresh drivers on the hardware, then those are pretty much the basics, with or without the Asus modding software (from Windows environment).

you're not alone. I'm testing Starcraft Remastered on an MSI i5-H-9th gen CPU w/Geforce 1650, the game stutters after playing for a few hours on intense maps.
starcraft.exe (set to realtime), performance set to MAX, computer is about 40c, variable cooling, "HSF blast mode does nothing and is annoying"
I had a desktop with i5-4th gen and GTX1050 playing SC:R, no issues with stutters at all like the MSI laptop.

it's hit and miss with laptops. sometimes you get lucky and land a nice one with the game you want.

 

[ProtosAngelus]

sounds normal. you probably got a crap board after they repaired it.
additionally, to save the manufacturer money, they probably modded your board to lock/shutdown after certain temps, they can implement any firmware they want.
and if the board is pre-defective or refurb, not all manufacture testing software can catch bad boards.
Does Asus have modding software that you can adjust?
ideally, with gaming laptops, the cooling sucks in the first place. most gaming hardware should be cooled adequately like most desktops. laptop lifespans are hit and miss.
sorry you're having the issue.

if you've tried a new SSD and reseated the RAM and loaded a fresh copy of win10 x64, upgrading or downgrading BIOS, and fresh drivers on the hardware, then those are pretty much the basics, with or without the Asus modding software (from Windows environment).

you're not alone. I'm testing Starcraft Remastered on an MSI i5-H-9th gen CPU w/Geforce 1650, the game stutters after playing for a few hours on intense maps.
starcraft.exe (set to realtime), performance set to MAX, computer is about 40c, variable cooling, "HSF blast mode does nothing and is annoying"
I had a desktop with i5-4th gen and GTX1050 playing SC:R, no issues with stutters at all like the MSI laptop.

it's hit and miss with laptops. sometimes you get lucky and land a nice one with the game you want.

There was no repairing done or motherboard changing.
I bought a new laptop 1y4m ago. The temps back then were normal. 2 weeks ago the idle temps rose to 65+ on both GPU and CPU. I gave the laptop (warranty expired, 1y) to a service crew for re-paste and clean. They changed out the gooey paste on the VRMs and VRAMs with thermal pads, and changed the stock paste with some other thermal paste.
This was the first paste job: https://i.ibb.co/hZ60SYK/IMG-20190727-172032-1.jpg Look at the dried out paste on the right-most upper chips, left of the exhaust.

Because that paste job was done poorly, the laptop would shut down any time GPU went into full load (bench, video game). I gave the laptop back last week to have it re-pasted again. Now, they taped down the thermal pads on VRMs and VRAMs (I think they still missed the dried out chips upper right corner) and maybe?? re-pasted the GPU and CPU.
After the 2nd repaste (I haven't opened the heatsink cause I have no paste or alcohol to clean and repaste myself) every game and benchmark can be run without the laptop shutting down instantly. However, now every game throttles after about 10 seconds. Eg. Dota 2. In main menu the GPU temp would not increase fast. Inside a BOT game (rendering world and heroes, etc.) the temp would increase much faster. I'd have about 10 seconds of fluid 60 FPS before the heavy throttling/stutter/fps lag starts (75-78 deg C).

The heavy throttling starts in any benchmark after ~1 second.

Novabench GPU only benchmark with HWmonitor in background. Can you guess where the bench starts?

https://i.ibb.co/ZJfK2Ck/Novabench-GPU-only.png

As you can see, the temp increases by 20 degrees in 1 second of running the benchmark. The increase is similar for Dota 2, albeit it's not 1 second but 10 (inside those 10 seconds the FPS is fluid and the GPU clock is constantly at the boost clock, not OVERCLOCK, there was no overclock of any kind, ever).

 

[fire400]

another easy way to clean off old, crusty paste, is to use a post it note to scrape off old compound and use canned air to spray off the debris or shake it off of the motherboard by tipping it side ways and blow it off.
dump the thermal pads and use copper heat sinks instead, ones that properly accommodate the elevation (slightly more than) the gap, apply 'arctic silver 5' on both sides of the copper sink.
thermal pads were never truly amazing, they were just a way to save money on cloning and process-manufacturing to name a few.

example;
eBay: "Copper Shim Thermal Pads Heat Sink Sheet for Laptop GPU CPU"

AS5 is one of the most balanced thermal compounds for laptops, especially works well for copper sink modding.

DIY is better personally than non-high-maintenance modding service centers - many of them do not specialize in attention-to-detail, as they have SLA (service level agreement) to finish jobs in a very short time period before moving to the next project/job/ticket/task.

also, can you can-air-blow or replace the cooling fan(s)?
would it bother you to use an external cooling pad as well?