GK7MR0R Does not start

[master3112]

Hello everyone.
I have a GK7MR0R (XMG Core 17) The device does not start. When I press the power button, the device tries to start and draws 0.8A for a second. According to the schematic, the following voltages should be present in the connected power supply unit and these are also present:
+5VALW
+3VALW
+3VALW_PCH
+1.8V_PRIM
+1.05VO_OUT_FET
The battery is being charged.

First I tried to rewrite the BIOS, same behavior.
Another user had the same behavior and his chipset was defective. I also changed the chipset, same behavior.
I disconnected the GPU and CPU inductors as a test, same behavior. I have not found a short circuit anywhere.
I have no idea what to do next.
Can anyone help me?
THX

[master3112]

I have made further measurements. I have noticed that when the power button is pressed, the GPU drivers Q166,Q170,Q171,Q172 and the GPU RAM drivers Q173 and Q177 are for a short time active (thermal camera). 3V3 is for a short time present on all drivers at PIN 20. I measure 0.8V at output PIN2+3+4. A GPU is already low impedance, but I measure at inductors L63,L64,L65,L66 =0 ohm. Only at inductor L67 and L69 (GPU-RAM) I have 32ohm. I would say that the GPU is defective, right? How low can the resistance of a GPU be? Under 1 ohm? However, when the GPU inductors are disconnected, the device does not start because the output after the inductors is monitored by the U1903 and U28.
I have also measured the CPU driver U13 PIN 20. There, 3V3 is also present at PIN20, but the driver does not become active.

[master3112]

I have taken further measurements.
I tested the functionality U104, U105, U106, PU1, PU2. These are switched via RUN_ON signal from the EC (U5 pin118). Signal is present and all work very briefly and supply the corresponding voltage.
- +5V
- +5VS
- +1.05V_CPU
- +1.05VS_CPU
- +3.3V
- +3.3VS
- 1V8_AON
- +1.8V
The strange thing is that these only remain active for a very short time, less than 1 second. Even stranger is that I have to press the power button several times quickly (10-15 times) to initiate the start. The device tries to start, the power button lights up briefly, but it goes off immediately. I have looked for short circuits in the mentioned rails, but have not found anything conspicuous. Does anyone have such behavior with a repair device, or an idea?
THX

[master3112]

I need to correct a few things after further measurements.
I don't have to press the power button several times. Pressing it once is enough, it just takes a while until the start sequence is initiated. Then the device tries to start. I would say the GPU is fine, because the power controller briefly activates the GPU drivers. I can also measure typical voltages for a GPU on the GPU driver. I think this is OK after all.

However, nothing happens with the CPU. I think that's where my problem lies. I have looked at the CPU controller U3. The U3 is activated at pin37 via R565 of EC/Pin3 (CPUCORE_ON_EC). I briefly have 2V at R565/pin1, but only 0.2V at R565/pin2. For 2V to be present at U3/Pin37, 2V must also be present at U24/Pin4 at the same time.
From EC/Pin84- (VCCIO_ON), 2V is also briefly present at U24/Pin1 via R3820. If 2V is also present at U24/PIN4, the VCCIO driver U18 (1V05 voltage for the CPU) and also the CPU controller are activated. However, 2V is only present at pin4/U24 if pins 1+2=2V, i.e. VCCIO_ON and SLP_S3#, and this is where I have a problem. The SLP_S3# signal is missing at pin2/U24.

I have read on the Intel page:
„The SLP_S3# output signal can be used to cut power to the system core supply, since it only goes active for the Suspend-to-RAM state (typically mapped to ACPI S3). Power must be maintained to the PCH primary well, and to any other circuits that need to generate Wake signals from the Suspend-to-RAM state. During S3 (Suspend-to-RAM) all signals attached to powered down planes will be tri-stated or driven low, unless they are pulled using a Pull-up resistor.“

However, this means that the SLP_S3# signal is actually used to cut the power supply. Since it is not present, it should actually work. I have looked at the functionality of IC U24 on another pin on the mainboard and there is definitely only a signal on PIN 4 when there is a signal on Pin1+Pin2. So I definitely need SLP_S3# signal to activate the CPU controller. Can anyone help me?