Hi to all,
they brought to me this gaming machine with mainboard marked as MS-16P51 VER:1.0 that firstly didn't see the GPU at all.
I opened it up and saw that it had no voltages on GPU-core chokes and VRAM chokes.
Owner told me that it opened the back cover to take off the dust inside, and when he powered it up again the GPU was gone.
I found that PR164 (+5VSUS to PVCC of U23) was blown and that there were some shorted pins on PU23
PU23 uP1666Q (IC - unsoldered IC)
1 BOOT1 381 - 654
2 UGATE1 522 - 734
3 EN 589 -> NVDD_PWRGD - 596 - 559
4 PSI 605 - 584
5 VID 415 - 487
6 REFADJ 680 - 756
7 REFIN 672 - 591
8 VREF 661 - 596
9 FS/OC 680 - 776
10 FBRTN_VGA2 000
11 FB_VGA2 1.9 - 571 ***
12 COMP 1429 - 598
13 NVDDS_PWRGD 1.9 *** 2 - 560
14 UGATE2 497 - 732
15 BOOT2 1.4 *** 2 - 652
16 PHASE2 1.3 - 1.3
17 LGATE2 473 - 435
18 PVCC 68 - 521
19 LGATE1 474 - 435
20 PHASE1 1.3 - 1.3
After unsoldering U23, BOOT2 was opened, but NVVDDS_PWRGD was still 5 Ohm to GND, so I found that it was U19 fault, pin 1 was shorted to GND.
I didn't have uP1666Q available and searching online I found that RT8816A maybe a good equivalent, but I didn't have this either, I had one new RT8815A but it has different functions pin:
RT8815A - RT8816A - uP1666Q
9 TON - TON - Switchig freq/OCP
12 IC (to leave as floating) - OCSET - Compensation output
I verified that there were no shorts or problems on PQ34/35/36/37, then I soldered the RT8815A and changed the U19 to a new one, then voltages appear on GPU cores and VRAM (measured at startup, until 1 minute after video is on the display)
CHOKE3/4/5/6 GPU-Core1 NVDD 0.79v
CHOKE7/8 GPU-Core1 NVDDS 0.85v
CHOKE2 VRAM FBVDDQ 1.52v
On U19 SN74LVC1G32DBVR
1 NVDDS_PWRGD 458 3.31
2 GC6_FB_EN -> PCH <- GPU 458 0.00
3 GND
4 FBVDDQ_ON 479 3.31
5 +3VSUS 338 3.32
U34 NC7S08P5X_SC70
1 FBVDDQ_PWRGD 471 3.31
2 NVDDS_PWRGD 458 3.31
3 GND
4 DGPU_PWRGD 461 3.31
5 +3VSUS 338 3.32
The problem is that now , after installing any Nvidia driver the GPU seems to be recognized , but after a reboot, it gives "Error 43" .
I then ordered a new uP1666Q, changed it, voltages changed a little only on its competent chokes, i.e. 7/8 and now it measures 0.79, but still "Error 43" is there.
I thought that GPU took an overvoltage on its 2nd core when the problem was caused by the owner and then I decided to buy a new GPU (opposite of cheap nowadays...), I soldered it and...still the dreaded "Error 43" on the new GPU.
Tried different drivers and WinOSs (23.21.13.9148 and 26.21.14.4223 on w10, 24.21.13.9835 and 26.21.14.4223 on w11) I even updated the Nvidia ROM (as shown in the MSI support website) but I can't take this error away.
Can you suggest something to look for?
they brought to me this gaming machine with mainboard marked as MS-16P51 VER:1.0 that firstly didn't see the GPU at all.
I opened it up and saw that it had no voltages on GPU-core chokes and VRAM chokes.
Owner told me that it opened the back cover to take off the dust inside, and when he powered it up again the GPU was gone.
I found that PR164 (+5VSUS to PVCC of U23) was blown and that there were some shorted pins on PU23
PU23 uP1666Q (IC - unsoldered IC)
1 BOOT1 381 - 654
2 UGATE1 522 - 734
3 EN 589 -> NVDD_PWRGD - 596 - 559
4 PSI 605 - 584
5 VID 415 - 487
6 REFADJ 680 - 756
7 REFIN 672 - 591
8 VREF 661 - 596
9 FS/OC 680 - 776
10 FBRTN_VGA2 000
11 FB_VGA2 1.9 - 571 ***
12 COMP 1429 - 598
13 NVDDS_PWRGD 1.9 *** 2 - 560
14 UGATE2 497 - 732
15 BOOT2 1.4 *** 2 - 652
16 PHASE2 1.3 - 1.3
17 LGATE2 473 - 435
18 PVCC 68 - 521
19 LGATE1 474 - 435
20 PHASE1 1.3 - 1.3
After unsoldering U23, BOOT2 was opened, but NVVDDS_PWRGD was still 5 Ohm to GND, so I found that it was U19 fault, pin 1 was shorted to GND.
I didn't have uP1666Q available and searching online I found that RT8816A maybe a good equivalent, but I didn't have this either, I had one new RT8815A but it has different functions pin:
RT8815A - RT8816A - uP1666Q
9 TON - TON - Switchig freq/OCP
12 IC (to leave as floating) - OCSET - Compensation output
I verified that there were no shorts or problems on PQ34/35/36/37, then I soldered the RT8815A and changed the U19 to a new one, then voltages appear on GPU cores and VRAM (measured at startup, until 1 minute after video is on the display)
CHOKE3/4/5/6 GPU-Core1 NVDD 0.79v
CHOKE7/8 GPU-Core1 NVDDS 0.85v
CHOKE2 VRAM FBVDDQ 1.52v
On U19 SN74LVC1G32DBVR
1 NVDDS_PWRGD 458 3.31
2 GC6_FB_EN -> PCH <- GPU 458 0.00
3 GND
4 FBVDDQ_ON 479 3.31
5 +3VSUS 338 3.32
U34 NC7S08P5X_SC70
1 FBVDDQ_PWRGD 471 3.31
2 NVDDS_PWRGD 458 3.31
3 GND
4 DGPU_PWRGD 461 3.31
5 +3VSUS 338 3.32
The problem is that now , after installing any Nvidia driver the GPU seems to be recognized , but after a reboot, it gives "Error 43" .
I then ordered a new uP1666Q, changed it, voltages changed a little only on its competent chokes, i.e. 7/8 and now it measures 0.79, but still "Error 43" is there.
I thought that GPU took an overvoltage on its 2nd core when the problem was caused by the owner and then I decided to buy a new GPU (opposite of cheap nowadays...), I soldered it and...still the dreaded "Error 43" on the new GPU.
Tried different drivers and WinOSs (23.21.13.9148 and 26.21.14.4223 on w10, 24.21.13.9835 and 26.21.14.4223 on w11) I even updated the Nvidia ROM (as shown in the MSI support website) but I can't take this error away.
Can you suggest something to look for?