Question PC crashing and fans reaching maximum speed (RTX 5070 12GB + be quiet! Pure power 12 M 850w)
Question PC crashing and fans reaching maximum speed (RTX 5070 12GB + be quiet! Pure power 12 M 850w)
Hello,
I recently assembled a PC using brand new components. Here are my specifications:
GPU: RTX 5070 OC 12GB
CPU: I5-14600K
RAM: VIPER DDR5 2x16GB 6000MT/s
MB: Z790 AORUS ELITE AX
Cooling: Thermalight Pearless Assassin 120 SE
Case: Endorfy Ventum 200ARGB
PSU: be quiet! Pure power 12 M 850w
During the installation of the 12Vhpwr cable, I noticed an issue: once the PSU was connected, the GPU wouldn’t power on and a red light appeared. After moving the cable, it worked again and the red light vanished. Over five days, temperatures stayed stable. However, two screens froze during a game with moderate requirements—one turned pinkish and the other displayed large artifacts. I thought it might be a random glitch, but both screens crashed again the next day, about 20 minutes apart.
Now the screens go dark and the fans spin at full speed. The only solution was a hard reset each time. I’ve checked GPU drivers and updated the cable twice, but the red light reappeared. A video showing the red light is available if you’d like.
Could there be something wrong with the cable? What should I do next?
The link appears to be unreliable and may lead to failure. Inspect the connections on both ends of the cable for any signs of harm, discoloration, or melting. Swap out the cable if anything seems incorrect.
In disaster? What type of disaster could that be?
There is no sign of any damage, blackening melting. The cables are pristine and are seated in the ports (GPU and PSU) as much as they can.
I completed the entire process of re-plugging the cable, but it seems less responsive now. However, even while still connected, the red light comes on when I push it to the side. By the way, I don’t have another cable to test. After moving the cable to a more stable place, I started the PC and used it briefly. It looks like the GPU fans turned off once I began playing a game. Then I immediately shut down the PC through the shutdown option, and just before it fully powered off, the fans briefly activated again.
The 5070 doesn't pull huge amounts of power but people with 4090/5090 cards have experienced melting connectors and fires due to poor high resistance connections.
Even with a 5070, if the red light blinks on and off, it implies you have a bad connection. Modern 12VHPWR and 12V-2x6 cables have six 12V wires working in parallel. If one or more of these 12V connections is bad (high resistance) extra current flows in the remaining connections.
If several 12V pins go high resistance, the current in the remaining pins can exceed the current rating of the pins, which get hot and in extreme cases, melt the plastic shroud.
Thermal cameras show plug/socket connections on GPU cards running warm/hot during normal operation.
You system is unstable and you need to sort out what appears to be a bad connection between PSU cable and GPU.
There's a vague chance you have a dry solder joint on the GPU card where the power connector is soldered on to the GPU motherboard, but a loose power lead plug/socket is more likely.
It is completely acceptable for enthusiasts to halt GPU operation and this does not signal a malfunction. When powering up a computer, the GPU's cooling fans briefly activate for a short duration as the BIOS verifies proper rotation. If the GPU isn't consuming much energy, the fans will cease to run, helping preserve the longevity of the bearings. Only during demanding tasks like running high-performance games or rendering software do the fans engage to maintain optimal temperatures. This behavior is also observed in many contemporary ATX power supplies—a characteristic I personally find undesirable. The PSU fan activates briefly upon startup and then stops. It only becomes active when the system draws several hundred watts from the supply. I prefer purchasing power supplies equipped with an ECO switch that enables a steady, slow rotation of the fans. As an engineer, maintaining cool PSU components is important; I’m content with bearing wear if it occurs sooner. Consistent airflow helps keep the electrolytic parts cooler, extending their lifespan. For every 10 degrees Celsius reduction in temperature, the expected service life doubles. This issue can be fixed by investing in a quality cable. It’s essential to ensure you buy a cable that is fully compatible with your PSU—this prevents potential short circuits. The safest approach is to purchase directly from the manufacturer, then test the connections using a multimeter to confirm they match the original pin configuration exactly. If the red LED flashes when you gently move the cable into the GPU port, it suggests the 12V-2x6 connector isn’t seated properly. Loose contacts in the connector pair can lead to overheating and failure. A replacement cable may be necessary. You might have inadvertently shifted pins by a few millimeters or more if you bent the wires too tightly to fit under a side panel. Generally, avoid bending the 12V-2x6 cable beyond 35mm from the connector’s rear.
The link between the cable head and the socket on the card appears to be shaky. Is the plastic component firmly seated in the card socket? If yes, one of the wires might be loose; pushing it sideways could cause the connection to fail.
As mentioned before, an unstable connection can cause problems because the current flowing through those wires and pins in power sockets is very high during heavy usage, such as gaming.
Could you try temporarily securing the cable with a strap on the left side of the case? Observe whether keeping them fixed at the expected connection point makes the red light stop working.
If you need to replace the cable, obtain one that is specifically approved by the manufacturer from a trustworthy supplier. These wires and socket pins handle significant power under load. An incorrect connection can lead to serious consequences: increased current would have to pass through to power the load (like a graphics card), resulting in higher temperatures. This could damage the card, harm the PSU, or even trigger an electrical fire if the PSU fails to cut off power in time during an overcurrent situation.
Based on what you shared this year, these kinds of problems (black screen and fans spinning loudly) usually point to a GPU power cable issue. Most people on forums say fixing it by swapping the cable works. From your description, it seems likely something was damaged during the first connection. It doesn't have to be obvious damage—just that the cable wasn’t handled properly. In my experience, the cable itself still looks fine but caused crashes every 2-3 days.