Screen goes completely cold then system powers off again
Screen goes completely cold then system powers off again
Hello, I've already attempted to talk about it here; the fix did help for a while. But explaining my issue is a bit of a story. First, let's look at my setup:
- CPU: Ryzen 7 7700X
- GPU: ASUS DUAL GeForce RTX 4070 OC
- RAM: ASUS ROG STRIX B650E-E GAMING WIFI
- Cooler: ASUS ROG STRIX LC II 360 ARGB
- SSD: Samsung 980 Pro 2TB
- Power Supply: ASUS ROG STRIX 1000W Gold Aura Edition
- Network: ASUS XG-C100C V2
- Fans: Lian Li UNI FAN AL120 V2 ARGB PWM x7
- Operating System: Windows 10 Pro
- RAM: 2x 16 GB Kingston FURY RENEGADE DDR5 RGB 6000 MHz CL32 or 4x 16 GB Kingston FURY RENEGADE DDR5 RGB 6000 MHz CL32 or 2x 16 GB G.Skill Trident Z5 neo RGB DDR5 6000 MHz CL30
- I mentioned RAM last, because it seemed to be the main suspect.
In my previous post I started with 32 GB (two sticks) of Kingston FURY RENEGADE DDR5 RGB 6000 MHz CL32. It didn’t perform well in the case (Lian Li O11 Dynamic XL, ROG certified), so I bought more. But the computer began behaving oddly with four sticks at 6000 MHz. Some suggested avoiding DOCP profile for four sticks, as official speeds are only 3600 MHz. I gradually started overclocking RAM manually and discovered that 5600 MHz (30-36-36-75) was stable—Prime95 ran smoothly for 10 hours, Memtest86 for 4 hours. It worked well for at least two weeks.
Then Windows update appeared unexpectedly, even though I hadn’t applied it. It just auto-applied and started without me choosing the option. Since then, everything has gone wrong again.
I play Runescape on one screen (or sometimes both with an alt account) and watch YouTube in Firefox windowed mode or listen to it in background. Occasionally everything freezes—mouse movement is possible, but clicking doesn’t work. YouTube audio plays for a few minutes before stopping abruptly. In the Windows Event Viewer I see “The system watchdog timer was triggered.” and “unexpected reboot.”
So far, all my attempts have failed. Using the default profile with all four Kingston sticks didn’t help. I switched to G.Skill memory because Kingston kits are XMP certified but not EXPO certified. I also bought two 16 GB G.Skill Trident Z5 neo RGB DDR5 6000 MHz CL30 sticks (two in a kit, four total). Reviews said running four at 6000 MHz was fine, so I enabled DOCP. Benchmarks passed—Prime95 and Memtest86 ran cleanly—but real-world performance failed me.
Next, I tried two Kingston sticks at 6000 MHz CL32. After a long benchmark session (almost a day), everything worked without errors. I then rebooted the PC. That attempt also failed.
Finally, I switched to two G.Skill EXPO certified DDR5 sticks at 6000 MHz CL30. It took about six hours of use before the first random crash. Now I’m running with two 16 GB G.Skill Trident Z5 neo RGB DDR5 6000 MHz CL30 sticks. They are AMD EXPO certified, and most configurations (except four Kingston sticks at 6000 MHz CL32) passed benchmarks—Prime95 for over 8 hours, Memtest86 for 3–4 hours.
I’ve updated drivers via Armoury Crate and GeForce Experience, installed the latest BIOS (version 1616), and kept everything current. I considered upgrading to Windows 11, but prefer Windows 10.
If anyone has any other ideas or suggestions, feel free to share—honestly, I’m starting to feel quite stuck.
Platform is brand new and few people have what you need to offer advice. Memory problems are likely with this setup, and the new socket isn't helping either. If you're seeing unexpected restarts, I recommend checking the Event Viewer for system logs—we'll require more detailed logging and context to pinpoint the issue. Uninstalling the Armory crate would be a good move since it matches virus definitions. Random reboots usually point to RAM or CPU issues. It seems the CPU might be overclocked or running PBO. My guess for freezing is memory, but the cause could also be CPU or motherboard power limits.
These issues don’t reveal much, and typically a CPU problem would appear in event logs frequently. It seems RAM or memory controller might be the culprit. You can use tools like HWinfo64, which tracks memory controller performance by counting FClock stability changes. Prime95 is decent for Curve Optimizer checks but not ideal for edge stability tests, and Memtest isn’t suited for subtle issues. Aida64 Extreme’s memory stress test is effective for uncovering barely stable overclocks. Try running HWinfo64 on a second screen and observe the “WHEA” column while your system runs under load or gameplay. An increasing value suggests your CPU struggles with RAM speed. You may need to adjust SOC voltage, though that’s a separate concern. Install Aida64 (trial) and execute the memory test through its tools—it usually completes in about ten minutes, but I run it for an hour to confirm. If it crashes, it indicates unstable memory operation and may require further tuning.
Yesterday it worked smoothly for about ten hours. Now it freezes after just thirty minutes. Also, the HWinfo app froze too. This happened both yesterday (around three hours) and today. Edit: it occurred in a situation where I had games and YouTube open, went to the bathroom, then restarted the video, and it froze briefly.
Download and execute Aida64 following the specified steps. Run the stress memory test in tools > system stability test > stress memory only > start, then wait an hour before checking again. You're right to suspect RAM overclock instability—AIDA64 is known for detecting such issues.
I completed the task yesterday. Initially I reverted to Kingston RAM certified for XMP, not EXPO, since those G.Skills units still accept returns (and they weren’t fully functional out of the box anyway). I also ran an AIDA64 memory stress test for nearly three hours without any issues, but after another two hours I experienced a freeze. I lowered one profile speed by one level (from 6000 MHz to 5600 MHz) and adjusted voltage accordingly. It seems I’ll have to wait now.