Question Unstable PC peripherals on heavy load

[konigTomo]

Hi guys,

TLDR: New system decides, during heavy load, to disconnect my peripherals. It starts with bluetooth headphones, continues to bluetooth mouse and wired keyboard.
Since the prices of GPUs have droped a bit I decided it is time for upgrade.I've had always in my life Intel processors but this time I decided to go with AMD Ryzen 5600x in combo with 3070 Ti.The problem is, I have a quite unstable peripherals on higher loads and during stress testing.I'll describe the problems in more details down.

PC Components:
CPU: AMD Ryzen 5 5600X 6x
CPU Cooler: Cooler Master Hyper 212 EVO V2
Motherboard: MSI B550-A Pro AMD B550 So.AM4 Dual Channel DDR4 ATX
GPU: 8GB Palit GeForce RTX 3070 Ti GamingPro Aktiv PCIe 4.0 x16
PSU: 700 Watt be quiet! Pure Power 11 CM Modular 80+ Gold
Memory: 16GB Corsair Vengeance LPX LP schwarz DDR4-3200 DIMM CL16 Dual Kit
Storage: 1000GB Crucial P1 NVMe M.2 2280 PCIe 3.0
Case: Kolink International Stronghold
Fans: 5 fans, 120mm. 3 in front, 1 in back, 1 on a top. (One in front and one in a back came with the case, other 3 are "upHere 4PIN PWM 120 mm Black Silent Fan")

Temperatures measured with HWiNFO
Idle temps at 25 C ambient:
CPU: 36 °C
System: 39 °C
CMOS: 42.5 °C
Chipset: 35.5 °C
CPU Socket: 33.5 °C
PCIE_1: 35.5 °C
T15: 42.5 °C
GPU: 45 °C - fans at 0, they start at 55 °C. (With fans at lowest RPM, idle temp is at 33 °C)GPU HotSpot: 52 °C - fans at 0, they start at 55 °C. (With fans at lowest RPM, idle temp is at 40 °C)* with glass panel removed, temperatures are nearly identical. (+/- 0.5 C)

Benchmarking / Stress Tests:

• FurMark:
  • GPU Benchmark 1440 Preset Score: 9177 Points, 153 FPS. Max GPU Temp 69 C.
  • StressTest:
    • Fullscreen, 2560x1440, 8xMSAAafter29 min: max GPU temp was at 74 °C , average 73 °C, HotSpot max at 84 °C, average at 82 °C, CPU Temp: max at 55 °C, average 52.At that point bluetooth headset was disconnected.
    • Same test as before, but with removed tempered glass panel. I run the test for 1:30 hours, no problem at all. Bluetooth headset was not disconnected.max GPU temp was at 72 °C, average 70 °C, HotSpot max at 81 °C, average at 79 °C, CPU Temp: max at 52 °C, average 50
• OCCT:
  • CPU Stress test: 1 hour, no problems at all.
  • Power stress test: after about 20 minutes, bluetooth headphones are disconnected. After about 10-20 minutes bluetooth mouse G602 starts to "stutter". On keyboard typing sometimes i press button and it does not get it or it would just write the letter couple of times.
  • Power stress without tempered glass panel: I stopped it after about an hour, no disconnect at all.

Measures (?) I took:

• Bios is on the 1.2.0.2
• Disabled PBO
• Forced PCIe gen 3
• C-state disabled
• Memory running at 2133 Mhz
• Fresh win10 installation
• All latest drivers installed
• In "Device Manager", at every USB Device in "Power managment" tab, I have unchecked the "Allow the computer to turn off this device to save power".
• Even on all "default" settings, problem appears.

"Real world" tests: (Glass panel was not removed)

• On iRacing: after about 30 I lost my bluetooth connection to headset. A bit later, mouse and keyboard starts to stutter. When I exit the game, after a couple of minutes (depending on how long have I played after first disconnect), I can reconnect my headset and mouse and keyboard starts to "recover". Stuttering less noticeable until it completely disappears.
• Zero hour - After about 15 minutes mouse starts to stutter. (at that time CPU: 55 °C, GPU 74 °C, GPU HotSpot: 82 °C)
• No Mans Sky - After about 20 minutes mouse starts to stutter. (at that time CPU: 55 °C, GPU 72 °C, GPU HotSpot: 82 °C). 10 minutes in, mouse starts a bit stronger. After taking down the glass panel mouse started to stabilise itselft, stuttering had after about 5 minutes completely disappeared.
• During 2-3 hours of surfing/listening to music there were no disconnections

"Real world" tests: (Glass panel was removed)

• No issues, at least not while playing No Man's Sky for about 2 hours. (at that time CPU: 49 °C, GPU 67 °C, GPU HotSpot: 77 °C)
• iRacing while playing around 1.5 hours, no problems.
• War Thunder, about 2-3 hours without any problems.

Conclusions and questions:

1. I have noticed that, by removing glass panel I have solved disconnecting issues with my peripherals. Of course, that is fixing the symptoms and not the cause.
2. Since I have no problems at all, when running the OCCT CPU Stress test, I think my GPU temps are problem. By running the OCCT Power test my my GPU temp on load are max 76 °C, average 74 °C, HotSpot max at 84 °C, average at 82 °C.Is that really that high for 3070 ti? And could it be cause of shutting USB peripherals down?
3. 2 Stock fans that my case had preinstalled, run at max aprox 880 RPMs. New ones that I bought are flexible and run at aprox max 1800 RPM. Is that enough, since there are 5 fans, 3 as intake and 2 as exhaust. Or should I invest in better fans to create better airflow?
4. Airflow. The case has close front panel and air intake on just one side. Is that even enough to have a decent airflow? Or should I just go for the meshed case, like Fractal Design Meshify C?
5. Event Viewer. There is quite a bit warnings, Event ID 10016. I tried fixing this by using fix from this link. It did not help. And on my old system (Intel i5, GTX 1060, 5 years old), I had this Event ID 10016 warnings, but never have I experienced peripherals disconnecting. Is the Event ID 10016 something to worry about?

Thank you for listening/reading and if you have any possible solutions or anything, please let me know.It is annoying having to remove glass panel just to play games..

 

[Lutfij]

Welcome to the forums, newcomer!

You forgot to mention the make and model of your motherboard. Latest drivers, being manually installed after sourcing the drivers from motherboard and GPU maker's support site? As for the case, yes that's a badly designed case, where the heat will only build up and intake of cooler ambient air is near impossible. Did you create you bootable USB installer using Windows Media Creation Tools, to rule out a corrupt installer?

Out of curiosity, were any parts recycled from the older build?

 

[konigTomo]

Welcome to the forums, newcomer!

You forgot to mention the make and model of your motherboard. Latest drivers, being manually installed after sourcing the drivers from motherboard and GPU maker's support site? As for the case, yes that's a badly designed case, where the heat will only build up and intake of cooler ambient air is near impossible. Did you create you bootable USB installer using Windows Media Creation Tools, to rule out a corrupt installer?

Out of curiosity, were any parts recycled from the older build?

Hi
Sorry, I forgot to write it down. I have updated the post to include motherboard. It is MSI B550-A Pro AMD B550 So.AM4 Dual Channel DDR4 ATX.

All latest drivers are installed from amd/nvidia/msi sites.
BIOS is at v1.2.0.2. I did not yet installed 1.2.0.3b. The users feedback isn't exactly optimistic
Win10 was installed using Windows Media Creation Tools and installation was without problems or any hiccups.
Every component is brand new, there are no recycled parts.

Yeah, I think the case is problem, since I do not have any problem when glass panel is removed.
But, is the GPU HotSpot temparature at 82 °C that high to be causing the disconnections?

 

[alexbirdie]

But, is the GPU HotSpot temparature at 82 °C that high to be causing the disconnections?

No.

And simply keep front glass panel removed.