XMG NEO 17 (M22) / Tongfang GMxRGxx - Backlight Circuit Fault causing EC Power-State Lockdown (dGPU & WLAN issues)

[TheElectronicus]

Hello everyone,

I am troubleshooting a Schenker XMG NEO 17 (M22), which is based on a Tongfang chassis (likely GMxRGxx series, 2022 model). Specs: AMD Ryzen 9 6900HX / NVIDIA RTX 3080 Ti (16GB).

The Incident: The original LCD panel developed an internal short circuit (measured shorts across almost all rails on the panel side). This caused a catastrophic failure in the backlight circuit on the motherboard:

• The backlight fuse was blown.
• The P-Channel MOSFET which switches the 19.5V backlight rail was destroyed. Measured nearly 0 Ohm between Source and Gate and 240 Ohm in both directions between Gate and Drain and Source and Drain.

Current State of Repair:

1. Replaced the faulty LCD with a brand new panel.
2. Replaced the blown fuse.
3. Temporary Fix: Unsoldered the damaged MOSFET from the board and bridged the Pads (Drain to Source) with a wire to test the panel.

• Result: The display and backlight work perfectly fine now.

The Problem ("Limp Mode"): Despite the working display, the system seems to be in a severe "EC Lockdown" or safety state:

• dGPU (RTX 3080 Ti): GPU-Z recognizes the card, but reports 0 MB VRAM, 0 MHz Clocks, and the Bus Interface is stuck at PCIe x8 1.1 (instead of 4.0). But i can switch on dGPU only, and also have Display output, so the dGPU is working in principle.
• WLAN/BT: The module is completely missing from Device Manager. The M.2 slot has power, but there is no communication.
• Right-side USB Ports: All USB ports on the right side of the laptop are non-functional.
• Right-side Card Reader: Is on the same Sub-Board as the USB Ports on the right side, but works
• S-States (Sleep/Hibernate): The laptop fails to enter S3 or S4. The screen goes black for a second and then immediately returns to the lock screen.
• OEM Control Center: Cannot communicate with the EC (Embedded Controller), setting things there have no effect.

Technical Observations:

• Gate Signal: I checked the Gate of the (now bridged) backlight MOSFET with an oscilloscope. During the boot sequence, the EC/Driver correctly pulses the gate between 19.5V and ~9.6V.
• My Theory: Since I have bypassed the MOSFET (Drain - Source) with a wire bridge, the 19.5V rail is "always on." When the EC tries to toggle the backlight rail during its power-on self-test (POST) or power sequencing, it sees that the rail doesn't respond (it stays at 19.5V even when the Gate is pulled high). I suspect the EC (ITE IT8227E or similar) detects this as a hardware failure ("stuck rail") and consequently disables the dGPU, WLAN, and the USB controller for the right side as a safety measure.

Questions to the experts:

1. Does anyone have a Schematic or Boardview for this specific Tongfang board (GMxRGxx)?
2. Is it common for Tongfang/ITE ECs to partially disable high-power peripherals (PCIe, USB hubs) if a power-sequencing feedback (like the backlight rail) fails?
3. Are there any known "persistent error flags" in the EC that need to be cleared via a specific tool or flash procedure once the hardware (MOSFET) is properly replaced?

Any insights or documents would be greatly appreciated!

[TheElectronicus]

I'm interrested, if this behaviour is normal or there is a more severe fault/damage.
As i wrote, everything else works on this laptop. And it is working stable, no bluescreens or sudden shutdowns.
I also assume, that the shorted LCD Panel has not damaged any other components, like the chipset and GPU itself, because the new display panel works just fine without any signs of problems. Also the Webcam, which is on the same EDP cable as the display itself is working. So my thought is, that if there was high voltage feeding back from the shorted panel on the data lines of the display or the 3,3V rail, the damage would be significant and permanent. But as the display and everything else, except the things mentioned above, are working i assume that it's some safety state of the EC-Chip.
As i need to order an appropriate P-Channel MOSFET, as i don't have one with the necessary specs or in the necessary housing, i wanted to ask the experts here, how my chances are, that this fixes my problems.

[john49884]

dGPU (RTX 3080 Ti): GPU-Z recognizes the card, but reports 0 MB VRAM, 0 MHz Clocks, and the Bus Interface is stuck at PCIe x8 1.1 (instead of 4.0). But i can switch on dGPU only, and also have Display output, so the dGPU is working in principle.

You have a graphics card reporting 0 MB and no errors in device manager?

[TheElectronicus]

Thanks for the replies so far. Here is an update on the current state of the machine:

1. Backlight & Power Rail Fix I have replaced the MOSFET for the backlight circuit with an IRLML9301TRPBF. Its specs meet or even exceed the original component. I also replaced the fuse with the correct 1.5A type (I had previously used a 2A fuse from my parts bin as a temporary measure just to get it running).

2. The GPU / HWiNFO "Ghost" Issue The issue where the GPU reported empty values wasn't a hardware failure after all, but a software limitation. I was using an outdated version of HWiNFO64 that didn't fully support this chip. After installing the latest version, the GPU is detected correctly and reports all values as expected.

3. GPU Bus Width Question (x8 vs x16) I noticed something regarding the PCIe bus: HWiNFO reports the GPU as PCIe v4.0 x16 (16.0 GT/s) @ x8 (2.5 GT/s). While the speed correctly jumps to 16.0 GT/s under load (FurMark), the bus width stays at x8 instead of the maximum capability of x16.

• Question: If anyone has the same laptop (XMG Neo 17 M22 / Tongfang GM7AG8P), could you check if this is normal? I'm trying to determine if the XMG connects the 3080 Ti via x8 by design, or if this is a sign of a hardware error/throttling due to the previous incident.

4. The "Miracle" USB Recovery Yesterday, something strange happened. While I was working normally, the screen suddenly went black. The laptop stayed powered on (keyboard backlight remained lit), but it was unresponsive. After a while, the XMG logo and the Windows loading circle appeared, but it stayed stuck there for over 5 minutes. I eventually forced a shutdown by holding the power button. Upon restarting, the right-side USB ports—which had been dead since the short circuit—suddenly started working again. They now detect devices and transmit data perfectly. I'm happy, but also a bit wary about why they suddenly "resurrected" after a system hang.

5. Remaining Issue: WLAN The WLAN module is still not detected, even though all necessary voltages are present at the slot. I suspect the module itself might have been fried. I will try to test it with a replacement module as soon as I can get my hands on one.

The laptop is basically coming back from the dead, one piece at a time. If anyone has insights on the PCIe x8 lane width or the mysterious USB recovery, I’d love to hear them!

Best regards, Gregor


PS.: I know it is the wrong forum, but was not able to move the thread after posting. So i will leave it here for consistency. If a mod wants to move it to the laptop forum, it's fine with me.

[TheElectronicus]

I measured all the EC-Chip lines, and everything looks good i think. I Used the GM7MxxP schematic for that because for GM7Rxx that i have i found none. But i assumed that the EC-Chip pinout would be mostly the same regardless if an Intel or AMD system

Laptop was laying on its top to access the EC chip for measuring, so the Display was closed, so also all the backlight enable signals make sense. Battery was disconnected, BIOS battery was disconnected, laptop was only supplied by AC adaptor

[TheElectronicus]

Booted a live Fedora image. The WLAN and BT are also not found under linux.
Therfore it leaves only the hardware as the source of the problem.
I'll investigate further this evening, maybe i find something. If so, I'll let you know

[TheElectronicus]

The WLAN module is working again.
A Load Switch named KAN18 directly in the vicinity of the WLAN M.2 slot has gone high resistance.
I bridged the switch, suppying 3,3V to the M.2 Slot and the WLAN module. The WLAN and Bluetooth work now flawlessly.
I don't know how the fault on the display propagated to and killing Load switches all over the board.
But with this quick fix, the laptop is pretty much in complete working order.

The last question which remains, is the bus width of the Nvidia graphics card. It is still reporting to be x16 capable but only runs on x8. I've confirmed that with multiple programs, not only HWInfo. So i assume it is true.

Although when gaming, i do not have any problems.
So if somebody can confirm, if the x8 bus width is normal or not, it would be highly appreciated

Thanks,
Gregor

[john49884]

The last question which remains, is the bus width of the Nvidia graphics card. It is still reporting to be x16 capable but only runs on x8

Normal, that GPU doesn't need more. Weird fault and you can also check how many pcie lanes you have counting how many coupling capacitors are between the CPU and the GPU(two per lane).

[TheElectronicus]

Thanks for your answer.

Yeah, i wanted to avoid that. Because those capacitors will be underneath the huge heatpipe and cooler assembly. I tried removing all the screews of this assembly an trying to lift the cooler from the motherboard. No chance. Its somehow stuck. Maybe because of the huge area of thermal pads all over. But I also do not want to force it, maybe it is somewhere still mechnically fastened down, and destroy the whole thing for good.

So the easiest would be, if somebody has the same laptop and could have a look what HWInfo says on his machine.

Or i contact XMG, if they can say what should be the correct bus width. If they say x8, then all is good, if x16, then i may investigate further or leave it as is.