Hey all, I built my first PC a few months ago and I’m stuck on a weird one. It’s a 7600 + 4070 Super on a 750W Gold unit. In Cyberpunk stress loops it’s fine for an hour, but if I leave a game menu open with uncapped FPS or alt-tab back in, the whole PC clicks off like someone pulled the plug. It comes right back on. Is my PSU dying, or is this a transient spike thing I keep reading about? I have the GPU on one daisy-chained cable from the PSU, which might be dumb. Any ideas?
750W Gold PSU shuts off only on sudden GPU load spikes, not in stress tests — where do I start?
- BuildNoobTess
- New member
- Posts: 5
- Joined: Wed Sep 16, 2026 6:18 am
- LoopDreamer
- New member
- Posts: 6
- Joined: Thu Oct 01, 2026 7:18 pm
Re: 750W Gold PSU shuts off only on sudden GPU load spikes, not in stress tests — where do I start?
Transient spikes! This is exactly the kind of thing that gets missed because everyone stress-tests with a nice steady load. Modern GPUs can yank hugely for very short bursts — way above their average board power — and a PSU that’s fine in a loop can still hit OCP if the spike lands at a bad moment. Menus and loading screens are classic because the frame rate goes wild and the card oscillates. Have you tried capping frame rate globally or in the driver? Also, please tell me the PCIe cable situation — daisy-chained or separate? That matters more than people think!
- PasteBotAlan
- New member
- Posts: 7
- Joined: Mon Nov 02, 2026 12:18 am
Re: 750W Gold PSU shuts off only on sudden GPU load spikes, not in stress tests — where do I start?
I had something similar on an older 650W with a 6700 XT. Mine would only shut off when I opened the map in one game, never in a benchmark. I thought it was the game, then I thought it was the PSU. I swapped to two separate PCIe cables and it stopped, but I also capped my frame rate at the same time so I’m not sure which fix did it. Is daisy-chaining actually dangerous, or just not ideal? I’m new to this and the manual was not helpful.
- BuildNoobTess
- New member
- Posts: 5
- Joined: Wed Sep 16, 2026 6:18 am
Re: 750W Gold PSU shuts off only on sudden GPU load spikes, not in stress tests — where do I start?
Thanks LoopDreamer and PasteBotAlan. I capped the frame rate to 120 in the driver last night and played for about two hours with no shutdown, so maybe that’s it? But I’m still on the daisy-chained cable because my PSU only came with two PCIe cables and one has the extra plug. I can run two separate cables if I use the other one. Would that actually matter if capping already fixed it? Also how do I know if it’s OCP versus the PSU just being old? It’s about three years old.
- LoopDreamer
- New member
- Posts: 6
- Joined: Thu Oct 01, 2026 7:18 pm
Re: 750W Gold PSU shuts off only on sudden GPU load spikes, not in stress tests — where do I start?
Capping helping is a big clue — it smooths the spikes. But yes, run two separate cables if the PSU has the outputs! A daisy chain puts both connectors on one wire bundle and one set of terminals at the PSU, so a sudden pull has to go through one path. It’s not automatically unsafe, but on a card that spikes it’s exactly the wrong place to save a cable. Three years old isn’t ancient, but cap aging and dust matter. What exact model is it? Some units have OCP set aggressively on the 12V rails and trip before the fan even spins up.
- PasteBotAlan
- New member
- Posts: 7
- Joined: Mon Nov 02, 2026 12:18 am
Re: 750W Gold PSU shuts off only on sudden GPU load spikes, not in stress tests — where do I start?
Wait, so OCP isn’t just for dead shorts? I thought it was a safety thing for when a cable is plugged in wrong. If a transient spike can trip it, how do reviewers test for that? I just tried two separate cables on mine and the coil whine changed pitch, which is probably unrelated but now I’m wondering. Also LoopDreamer, is a single-rail PSU always better for this? I keep seeing people say multi-rail is bad.
- BuildNoobTess
- New member
- Posts: 5
- Joined: Wed Sep 16, 2026 6:18 am
Re: 750W Gold PSU shuts off only on sudden GPU load spikes, not in stress tests — where do I start?
It’s a Corsair RM750e, the newer one. I’ve got two PCIe outputs on the PSU, so I’ll try two separate cables tonight. The weird thing is it never trips in a 30-minute FurMark-style loop, only in real games when the frame rate jumps around. PasteBotAlan, mine has a tiny coil whine too but I barely hear it over the fans. I’ll report back after a proper session. If it still trips with separate cables, is that RMA time or should I try a power limit first?
- LoopDreamer
- New member
- Posts: 6
- Joined: Thu Oct 01, 2026 7:18 pm
Re: 750W Gold PSU shuts off only on sudden GPU load spikes, not in stress tests — where do I start?
Before you RMA, try a modest power limit — it’s a great diagnostic. If it’s stable at 90% but trips at 100%, you’re likely riding OCP on spikes, not a dead unit. Also, this whole proprietary-cable rabbit hole reminds me of the earlier thread Refurb ex-corporate SFF desktops as a cheap emulation/HTPC base — are proprietary PSUs a dealbreaker? — Refurb ex-corporate SFF desktops as a cheap emulation/HTPC base — are proprietary PSUs a dealbreaker? — because once you’re chasing transient behaviour, weird PSUs and adapter cables get scary fast. Have you tried logging GPU power with HWiNFO while it happens? The peak column is the interesting one.