G16 (2024) GU605MY 2 Issues After 3 Months of Use, 1st Shutdowns & 2nd WLAN/HDMI

[donovan20055]

Hello,

After about 3 months of use, I've run into 2 issues with my ASUS Zephyrus G16 (2024) GU605MY device.
CPU: Intel Ultra 9 185H, GPU: RTX 4090M 16GB
Using latest Bios, Firmware and Drivers

1. Device shuts down without warning while gaming.
I noticed this while playing Hogwarts Legacy.
These shutdowns occurred when the GPU Memory Junction Temperature sensor reached 100C.
These shutdowns occurred before the GPU or CPU started throttling at 87C for the GPU and 100C for the CPU.The screenshot shows the following:
1st line: CPU temp, CPU power, CPU usage, P-core frequency
2nd line: GPU temp, GPU memory junction temperature, GPU hotspot temp, GPU power, GPU usage, GPU frequency, VRAM usage, Total system power
3rd line: All four are Chipset temperature sensors
4th line: All three are NVMe drive temperature sensors
Turbo profile was used with default settings.
This behavior is likely a Bios/Firmware issue relating to the GPU Memory Junction Temperature not throttling the GPU before shutdown temperature is reached.
It could also be a defect with my specific device, such as a poor heatsink mount or poor thermal compound application.
I don't know how much hotter memory is supposed to be in relation to GPU temp.
This issue is repeatable, so ASUS should be able to find the root cause.

2. WLAN 5Ghz signal is low or disconnects when connected to an external display via HDMI issue details in the link below.
https://linustechtips.com/topic/1604275-zephryus-g16-2024-%E2%80%94-the-wi-fi-issue/
This issue is repeatable, so ASUS should be able to find the root cause.

[donovan20055]

Well, I am not getting much support from ASUS.

In the meantime I am going to try do some more testing on my sudden shutdown issue. Which I believe is due to VRAM temperature. A temporary fix you could try is to lower the temperature target in either ASUS Armoury Crate or G-Helper. In my case while using the default Turbo profile I have set the temperature target to 85C instead of the default 87C. That is because VRAM is at least 13 degrees hotter than the GPU on my device. I believe from earlier testing that when VRAM gets to 100C a sudden shutdown will occur. I am countering this by forcing the GPU to throttle at a lower temperature. In addition I've applied a manual fan curve that has the fans run at 100% when the GPU get to 83C and CPU 95C. The default Turbo profile does not go to 100% fan speed even when the device is thermal throttling. I am still adjusting GPU temperature target and CPU/GPU fan curves at the moment.

I would like to take my device apart to verify that the CPU, GPU and VRAM all have proper thermal interface material application but I don't want to void my warranty at the moment. May change my mind of this later IDK.

[ElectroStingz]

Hello,

Do you run this one a flat desk or with a cooling pad or slightly lifted off the desk?

Reviews on this model mention airflow is not so good on a flat desk so if you haven't tried consider lifting slightly. 

For GPU testing have you tried Furmark? This will run high power and temps on the GPU only so if that is your problem this should crash. https://geeks3d.com/furmark

Do you also have some before temps like when it was new / working?

[donovan20055]

Hello,

I first noticed the issue while using the device on a flat desk. I then tried using a cooling pad, but the shutdown still happened. It also happened with just lifting the back. When the back was lifted or using a cooling pad, it took longer for VRAM to get to 100C. However, when it did, the device would still shut off without warning.

I tried using Furmark but the GPU would throttle since it reaches 87C (87C is Turbo profile default) before the VRAM reaches 100C. If the GPU throttles, shutdowns are not an issue. But if the VRAM gets to 100C before the GPU throttles at 87C, then a shutdown will happen.

The shutdown issue showed up as soon as I started to play games. I started playing games after having had the device for about 3 months. Before that, the GPU was not being heavily used. The WLAN/HDMI issue has been happening since day 1 of purchase.

I am currently comparing the vBios of the G16(2024) with the one in a strix scar 16(2024) and a G14(2024) all with rtx 4090s. I contacted someone with a strix scar 16(2024). His device also reaches 100C on VRAM, but it does not shut down. I have no clue why his device works fine, running at 100C on VRAM for hours while my device shuts off. If ASUS just told us that the G16(2024) is designed to turn off when VRAM gets to 100C, then I would know this is normal behavior. I could then ask them "hey ASUS, could you change the Bios/Firmware/vBios so that throttling is triggered or fans speedup before 100C on VRAM is reached". I am also going to look into g-sync and external display via USB-C some more, since the WLAN/HDMI issue and the shutdown issue could have a common root cause. I'm not sure.

[ElectroStingz]

Just note there are some differences with the laptop RTX 4090's, like your one is slightly lower on the boost compared to the ROG Strix Scar. And memory chip temperatures to some extent are self regulated so if the chip reaches the critical temperature it will switch off.

From what I can see the memory chip part number is likely SK Hynix GDDR6: H56G42AS6DX014

Usually there are 4 / 8 chips to make your 16GB so if one of them is actually running hotter that what the measured sensor shows this could also explain your shutdown compared to others. I think it's best to get the vapour chamber inspected / thermal paste checked out on the chips. (use the warranty claim / repair?)

Did you consider other methods to help like a reduction in the GPU memory frequencies? (just to see if it improves the situation)

[donovan20055]

I believe the VRAM temp shown is the hottest chip. At least that is what GPU-Z says. I'll check it tomorrow. Both mine and my contact's VRAM chips are Samsung brand according to GPU-Z. Part of why I'm looking at the vBios for the different models is to look for the differences.

I've tried to contact ASUS support via phone and e-mail. Phone support only gave very generic advice which was not helpful. E-mail support has not replied to me yet. I'll try the support section of this forum when I have some time later.

I did not consider lowering VRAM frequency, because if I have to do that to get it to work, then I might as well go through the ASUS RMA repair process. I'll ask them to check the heatsink mount and thermal interface application as well. I'm mostly just data gathering at the moment, so I will have some information I can share with the support group along with other users who have a similar device and issue to mine.