You might want to consider replacing your power supply unit.
You might want to consider replacing your power supply unit.
Hey everyone, just moved from a 2700x to a 5900x resolution. I’m starting to notice my computer shutting down often while playing games. It used to happen occasionally but rarely affected me much. Now, after booting up, the system works for a bit then suddenly stops with no warning before turning back on. This feels like my power supply might be struggling, especially since I have a 750W PSU. I’ve tried standard fixes like reinstalling drivers, running diagnostics, checking temperatures, and even undervolting the GPU, but everything seems fine during gameplay. It’s confusing because I thought I was hitting close to the maximum power draw. I’m worried it could be the power delivery issue rather than a hardware problem. The setup looks solid: 5900x, 6800XT motherboard, dual 32GB RAM, Ekwb DDC pump, three Noctua fans (with case modification), Dell 1907FPV panel via MOLEX, plus internal LED indicators. I’m trying to figure out if my GPU is really using a lot of power or if something else is at play.
There seems to be a mismatch between power requirements and output. The system needs more than 5900X power, but the current build only reaches 450W on the UPS with an RTX3060ti. It might be time to check the PSU or cooling setup. Updated April 23, 2022 by An0maly_76
There’s definitely an airflow or ventilation issue (my temperature stays around 38°C–42°C, climbs to 76°C–86°C, air-cooled), but it hasn’t reached the throttling range yet—it will slow down before shutting off. The AIO pump might also be losing power. Regarding shutdowns, verify you haven’t disconnected your RAM (I did this during header cable installation and it caused a crash). It could still be a failing PSU starting to deteriorate. The NZXT H510 isn’t ideal for cooling, which contributes to the higher temps.
750W is more than sufficient. I rarely reached 650W even with OCed 5900X (200W) and a 3080 (320W) plus several components (four drives, DC pump, two AIO pumps, six fans, Wi-Fi/BT card, wireless peripherals)... Games typically don’t consume full CPU power on a 12-core 5900X, often under 100W. Unless crashes cause everything to shut down (including the PSU), then the issue likely lies elsewhere. However, a 65°C idle is quite high; it should be around 45°C, and full load at 75°C is normal.
Airflow doesn't work well when CPU and GPU share the same loop, especially with your GPU running at 65% during gaming. It doesn't throttle but reduces power instead. You can boost fans and remove the side panel without affecting temperatures. The DDC pump setup seems inconsistent—your configuration goes pump → CPU → 120 rad → GPU → 240 rad → res → pump. If true, the GPU would also get hot. Using EVGA PSUs typically fails after about four years. Another thought: your custom sleeved cables from CableMod might be deteriorating.
I know what I'm talking about—I use an 850W Corsair in a Corsair 4000X case with six 120s and a Scythe Mugen 5. My system stays at the expected temperatures without any problems. I believe your shutdown issues might stem from unseated RAM (as happened with mine) or a failing PSU. Anything can fail, particularly electrical components. For higher performance builds, I’d suggest a PSU with large capacitors and strong cooling.
Also, if your GPU is the first to cool down in the loop, it makes sense that the GPU stays cooler if the path is GPU → CPU → RAD → RAD. Otherwise, heat from the GPU would go to the CPU, but the CPU wouldn’t be able to dissipate it effectively. Something seems off. Even a 5900X doesn’t need liquid cooling, as my setup shows.
Edited April 23, 2022 by An0maly_76
Revised, more info