F5F Stay Refreshed Hardware Desktop Issue with RTX 4070 black screen crash—system remains active via iGPU display, but GPU-Z indicates complete power loss.

Issue with RTX 4070 black screen crash—system remains active via iGPU display, but GPU-Z indicates complete power loss.

Issue with RTX 4070 black screen crash—system remains active via iGPU display, but GPU-Z indicates complete power loss.

P
P1NG3S
Member
Find
50
12-21-2025, 08:08 PM
#1
My system: Ryzen 7 7700x, RTX 4070, MSI B650 Gaming plus wifi, 16gb ram 5600 CL36, Cooler master MWE gold 650w
I've been having issues for the last week with my GPU completely shutting down from stable readings in GPU-Z to zero. This has so far only been happening in games and not when testing with benchmarks. About 2 weeks ago the main fuse to my house blew as I was gaming, game was Delta Force if that even matters, it's an old house so I still have screw-in fuses in the attic. I replaced the fuse and started the PC back up about 5 minutes after the outage.
Unneseccerary rant:
(Fuse probably blew because I was charging my EV and my wife got home and put hers on charging aswell, then went and startet the dishwasher, the cloth washer, took a shower and then fired up her PS5 and some othe bits and bobs...)
Everything seemed fine, system booted up as normal and I went back into the game. Maybe 4-5 days later the first crash happened after 2 hours of gaming, I didn't understand what happened, shut the PC off and waited a few minutes and fired it back up and no problem. Another 3-4 days later it happened again this time at maybe 45 minutes of gaming, I at this point understand something is up but I tried the same proccedure as last. It fired back up, but 10 minutes later the crash happened again. And again, so I stopped and checked the motherboard for any error LEDs and looked at task manager that showed me the GPU usage was at 0%, fans were still spinning and the Zotac logo was still lighting.
OK this is the point were I started to feel that dread. I had to RMA the PC once before because the motherboard died on me 1 month after purchase, yes it is a prebuilt but it's a "order it to your speccs and we build it" kinda deal not that it matters I guess. Comes with 3 year warranty, great. Anyway shipping the whole damn PC is just a risk I don't want to do unless absolutely necessary, I've worked for 15 years in shipping and I know what can happen and how it happens far to well..
I've been doing a little bit of troubleshooting but I'm a absolute novice on PC hardware so I have been chatting with chatgpt to try and list all potential causes, what happened to trigger it and help read logs as well as troubleshooting ideas. I will post a summary from my chat with chatgpt because I will never remember everything as I write this post so to save on having to do multiple posts to answer questions I know you will have. I will also have chatgpt list my full setup. I've checked EventViewer and I got 8 critical ID 41 KernelPower errors in the last 7 days, but they don't seem to correspond with 1 of the crashes that I caught in GPU-Z log
This is the log reading:
2025-11-12 20:10:17 , 2850.0 , 1312.7 , 57.2 , 64.6 , 66.0 , 41 , 1529 , 41 , 1530 , 5722 , 35 , 11 , 6 , 1 , 95.8 , 62.5 , 33.4 , 12.2 , 33.8 , 12.2 , 62.0 , 12.0 , 47.9 , 4 , 1.1000 , 85.5 , 13985 ,
2025-11-12 20:10:19 , - , - , - , - , - , - , - , - , - , 5722 , - , - , - , - , - , - , - , - , - , - , - , - , - , 0 , 0.0000 , 84.6 , 13984 ,
Second 18 is missing
My chatgpt summary as I don't know what I'm forgetting to mention:
System specs
​
Case:
Fractal Design Focus 2 RGB TG (6×120mm fans)
CPU:
AMD Ryzen 7 7700X (8C/16T)
Cooler:
be quiet! Pure Rock 2 FX
Motherboard:
MSI B650 Gaming Plus WiFi
Memory:
16 GB (2×8 GB) Kingston FURY Beast DDR5 5600 MHz, CL36 (EXPO)
Storage:
Kingston FURY Renegade NVMe 1 TB (PCIe 4.0)
GPU:
ZOTAC Gaming GeForce RTX 4070 Twin Edge OC
PSU:
Cooler Master MWE Gold 650W V2 (fully modular, single PCIe cable w/ daisy-chain)
OS:
Windows 11 Home
Other GPU tested:
Asus GeForce GTX 1650 Super TUF OC — stable in same system under FurMark
The issue
​
About 4–7 days after a
power outage
, my desktop started having
GPU dropouts only during gaming
.
It happens in multiple titles (
Anno 1800
,
Delta Force 2024
), usually after 10–15 minutes of play.
When it happens:
The
monitor connected to the RTX 4070 goes completely black
, as if the GPU loses power.
Monitors on the iGPU
briefly flicker black, then
recover after ~5 seconds
.
The
system stays powered on
— no reboot, no crash dump, all programs still running.
GPU-Z logging stops right at the moment of the blackout (sensors go “—” and 0.000V).
No errors appear in Windows Event Viewer or NVIDIA logs.
Video playback
and light desktop work never trigger the issue.
Tests performed
​
OCCT 3D Heavy Load (15 min):
stable, no errors or voltage dips.
FurMark (GTX 1650 in same system):
stable, max ~75°C.
FurMark / OCCT (RTX 4070):
passes fine while stress running.
Crash happens only in real games
, not synthetic loads.
One artifact
appeared once before a crash — half of the screen went black with green/red stripes for ~0.5 s, then returned to normal.
GPU-Z Log Data (excerpt)
​
Stable period (20:10:00–20:10:14):
GPU Clock: ~2850 MHz
Voltage: ~1.10 V
Temp: ~59°C (Hotspot 68°C)
Power: ~105–120 W total
8-pin Voltage: ~11.9–12.0 V
PerfCap: None
Crash event (~20:10:55):
2025-11-12 20:10:55 , - , - , - , - , - , 0 , 0.0000 , ...
At that point, all telemetry vanished — zero voltage, zero power, no temperature readings —
while the PC stayed responsive. It’s like the GPU power section or VRM controller suddenly shut down.
Elapsed time from stable reading to failure: ~41 seconds.
Additional context
​
MSI Afterburner
shows the card runs around
0.95 V @ 2625 MHz
, but logs indicate it boosts to
1.10 V @ 2850 MHz
under load.
Using
a single daisy-chained PCIe cable
from the PSU (8-pin) → GPU.
Tried reseating the GPU and power cable; no change.
PSU is about 2 years old (Cooler Master MWE Gold 650W V2).
GPU is about 2 years old as well (ZOTAC).
Secondary test system with weaker PSU (Corsair VS550) also boots the 4070 fine.
Key points
​
✅ 12V rail stable (11.9–12.2V) — no signs of PSU undervoltage
✅ Temperatures well within limits (~60–85°C)
✅ GTX 1650 stable → PSU and motherboard likely okay
⚠️ RTX 4070 telemetry drops to zero → power loss internal to GPU
⚠️ Artifact seen once → possible VRM or memory power-stage instability
Suspected causes
​
GPU hardware fault
– VRM controller, solder joint, or memory power stage intermittently failing
Transient power dropout through the daisy-chained cable
– less likely but possible
VBIOS/driver power-state instability
– possible but no TDR or system hang
PSU degradation after power outage
– low probability, given stable readings
​
What I plan to do next
​
Test the RTX 4070 in the secondary system for an extended gaming session (1–2 hrs).
Try a separate PCIe cable or new PSU if available (to rule out transient sag).
Reinstall NVIDIA drivers using DDU (already done once).
If issue persists → file RMA with ZOTAC / retailer.
Reply
Reply
P
P1NG3S
12-21-2025, 08:08 PM #1

My system: Ryzen 7 7700x, RTX 4070, MSI B650 Gaming plus wifi, 16gb ram 5600 CL36, Cooler master MWE gold 650w
I've been having issues for the last week with my GPU completely shutting down from stable readings in GPU-Z to zero. This has so far only been happening in games and not when testing with benchmarks. About 2 weeks ago the main fuse to my house blew as I was gaming, game was Delta Force if that even matters, it's an old house so I still have screw-in fuses in the attic. I replaced the fuse and started the PC back up about 5 minutes after the outage.
Unneseccerary rant:
(Fuse probably blew because I was charging my EV and my wife got home and put hers on charging aswell, then went and startet the dishwasher, the cloth washer, took a shower and then fired up her PS5 and some othe bits and bobs...)
Everything seemed fine, system booted up as normal and I went back into the game. Maybe 4-5 days later the first crash happened after 2 hours of gaming, I didn't understand what happened, shut the PC off and waited a few minutes and fired it back up and no problem. Another 3-4 days later it happened again this time at maybe 45 minutes of gaming, I at this point understand something is up but I tried the same proccedure as last. It fired back up, but 10 minutes later the crash happened again. And again, so I stopped and checked the motherboard for any error LEDs and looked at task manager that showed me the GPU usage was at 0%, fans were still spinning and the Zotac logo was still lighting.
OK this is the point were I started to feel that dread. I had to RMA the PC once before because the motherboard died on me 1 month after purchase, yes it is a prebuilt but it's a "order it to your speccs and we build it" kinda deal not that it matters I guess. Comes with 3 year warranty, great. Anyway shipping the whole damn PC is just a risk I don't want to do unless absolutely necessary, I've worked for 15 years in shipping and I know what can happen and how it happens far to well..
I've been doing a little bit of troubleshooting but I'm a absolute novice on PC hardware so I have been chatting with chatgpt to try and list all potential causes, what happened to trigger it and help read logs as well as troubleshooting ideas. I will post a summary from my chat with chatgpt because I will never remember everything as I write this post so to save on having to do multiple posts to answer questions I know you will have. I will also have chatgpt list my full setup. I've checked EventViewer and I got 8 critical ID 41 KernelPower errors in the last 7 days, but they don't seem to correspond with 1 of the crashes that I caught in GPU-Z log
This is the log reading:
2025-11-12 20:10:17 , 2850.0 , 1312.7 , 57.2 , 64.6 , 66.0 , 41 , 1529 , 41 , 1530 , 5722 , 35 , 11 , 6 , 1 , 95.8 , 62.5 , 33.4 , 12.2 , 33.8 , 12.2 , 62.0 , 12.0 , 47.9 , 4 , 1.1000 , 85.5 , 13985 ,
2025-11-12 20:10:19 , - , - , - , - , - , - , - , - , - , 5722 , - , - , - , - , - , - , - , - , - , - , - , - , - , 0 , 0.0000 , 84.6 , 13984 ,
Second 18 is missing
My chatgpt summary as I don't know what I'm forgetting to mention:
System specs
​
Case:
Fractal Design Focus 2 RGB TG (6×120mm fans)
CPU:
AMD Ryzen 7 7700X (8C/16T)
Cooler:
be quiet! Pure Rock 2 FX
Motherboard:
MSI B650 Gaming Plus WiFi
Memory:
16 GB (2×8 GB) Kingston FURY Beast DDR5 5600 MHz, CL36 (EXPO)
Storage:
Kingston FURY Renegade NVMe 1 TB (PCIe 4.0)
GPU:
ZOTAC Gaming GeForce RTX 4070 Twin Edge OC
PSU:
Cooler Master MWE Gold 650W V2 (fully modular, single PCIe cable w/ daisy-chain)
OS:
Windows 11 Home
Other GPU tested:
Asus GeForce GTX 1650 Super TUF OC — stable in same system under FurMark
The issue
​
About 4–7 days after a
power outage
, my desktop started having
GPU dropouts only during gaming
.
It happens in multiple titles (
Anno 1800
,
Delta Force 2024
), usually after 10–15 minutes of play.
When it happens:
The
monitor connected to the RTX 4070 goes completely black
, as if the GPU loses power.
Monitors on the iGPU
briefly flicker black, then
recover after ~5 seconds
.
The
system stays powered on
— no reboot, no crash dump, all programs still running.
GPU-Z logging stops right at the moment of the blackout (sensors go “—” and 0.000V).
No errors appear in Windows Event Viewer or NVIDIA logs.
Video playback
and light desktop work never trigger the issue.
Tests performed
​
OCCT 3D Heavy Load (15 min):
stable, no errors or voltage dips.
FurMark (GTX 1650 in same system):
stable, max ~75°C.
FurMark / OCCT (RTX 4070):
passes fine while stress running.
Crash happens only in real games
, not synthetic loads.
One artifact
appeared once before a crash — half of the screen went black with green/red stripes for ~0.5 s, then returned to normal.
GPU-Z Log Data (excerpt)
​
Stable period (20:10:00–20:10:14):
GPU Clock: ~2850 MHz
Voltage: ~1.10 V
Temp: ~59°C (Hotspot 68°C)
Power: ~105–120 W total
8-pin Voltage: ~11.9–12.0 V
PerfCap: None
Crash event (~20:10:55):
2025-11-12 20:10:55 , - , - , - , - , - , 0 , 0.0000 , ...
At that point, all telemetry vanished — zero voltage, zero power, no temperature readings —
while the PC stayed responsive. It’s like the GPU power section or VRM controller suddenly shut down.
Elapsed time from stable reading to failure: ~41 seconds.
Additional context
​
MSI Afterburner
shows the card runs around
0.95 V @ 2625 MHz
, but logs indicate it boosts to
1.10 V @ 2850 MHz
under load.
Using
a single daisy-chained PCIe cable
from the PSU (8-pin) → GPU.
Tried reseating the GPU and power cable; no change.
PSU is about 2 years old (Cooler Master MWE Gold 650W V2).
GPU is about 2 years old as well (ZOTAC).
Secondary test system with weaker PSU (Corsair VS550) also boots the 4070 fine.
Key points
​
✅ 12V rail stable (11.9–12.2V) — no signs of PSU undervoltage
✅ Temperatures well within limits (~60–85°C)
✅ GTX 1650 stable → PSU and motherboard likely okay
⚠️ RTX 4070 telemetry drops to zero → power loss internal to GPU
⚠️ Artifact seen once → possible VRM or memory power-stage instability
Suspected causes
​
GPU hardware fault
– VRM controller, solder joint, or memory power stage intermittently failing
Transient power dropout through the daisy-chained cable
– less likely but possible
VBIOS/driver power-state instability
– possible but no TDR or system hang
PSU degradation after power outage
– low probability, given stable readings
​
What I plan to do next
​
Test the RTX 4070 in the secondary system for an extended gaming session (1–2 hrs).
Try a separate PCIe cable or new PSU if available (to rule out transient sag).
Reinstall NVIDIA drivers using DDU (already done once).
If issue persists → file RMA with ZOTAC / retailer.

Reply
Reply
H
HiImAnnabel
Member
Find
238
12-21-2025, 08:08 PM
#2
Welcome to the forums, newcomer! It’s important to note that the TDP of both cards differs, meaning the RTX 4070 will consume more power than the GTX 1650. Your PSU is approximately two years old (Cooler Master MWE Gold 650W V2). Consider getting a higher quality, higher wattage PSU and check if the problem continues. Your motherboard is an MSI B650 Gaming Plus WiFi. What BIOS version are you running? When using DDU, perform it in Safe Mode to eliminate all GPU drivers (Intel, AMD, Nvidia), then restart and manually install the newest driver from Nvidia's support site via an elevated command—right-click the installer and select Run as Administrator.
Reply
Reply
H
HiImAnnabel
12-21-2025, 08:08 PM #2

Welcome to the forums, newcomer! It’s important to note that the TDP of both cards differs, meaning the RTX 4070 will consume more power than the GTX 1650. Your PSU is approximately two years old (Cooler Master MWE Gold 650W V2). Consider getting a higher quality, higher wattage PSU and check if the problem continues. Your motherboard is an MSI B650 Gaming Plus WiFi. What BIOS version are you running? When using DDU, perform it in Safe Mode to eliminate all GPU drivers (Intel, AMD, Nvidia), then restart and manually install the newest driver from Nvidia's support site via an elevated command—right-click the installer and select Run as Administrator.

Reply
Reply
L
leleinator3000
Member
Find
66
12-21-2025, 08:08 PM
#3
Yea, I get it. It's all I've got at the moment, unfortunately. I actually own a 20-year-old entry-level gaming PC from when I was 16, but honestly I'm not sure about it. I'm thinking of getting a used one to start, though—I don't know anyone nearby who could lend me something useful. I don't have friends or family close by (about a 12-hour drive away) with a gaming desktop. Would K2 be what you were looking for?
Reply
Reply
L
leleinator3000
12-21-2025, 08:08 PM #3

Yea, I get it. It's all I've got at the moment, unfortunately. I actually own a 20-year-old entry-level gaming PC from when I was 16, but honestly I'm not sure about it. I'm thinking of getting a used one to start, though—I don't know anyone nearby who could lend me something useful. I don't have friends or family close by (about a 12-hour drive away) with a gaming desktop. Would K2 be what you were looking for?

Reply
Reply
K
kolja1234
Junior Member
Find
17
12-21-2025, 08:08 PM
#4
Considering a used PSU might not be wise since it's hard to know what tax status or lifespan it had.
1.K2 would that be the right question?
You should check the BIOS updates and ensure you're using the newest chipset driver for your system before updating the BIOS with the latest version.
Reply
Reply
K
kolja1234
12-21-2025, 08:08 PM #4

Considering a used PSU might not be wise since it's hard to know what tax status or lifespan it had.
1.K2 would that be the right question?
You should check the BIOS updates and ensure you're using the newest chipset driver for your system before updating the BIOS with the latest version.

Reply
Reply