ToastyMan

I think it's a flaw in the design, and has been noted in many of the reviews. The fan curve is very aggressive and reviewers have noted that the noise level is exceptional under even moderate loads. I'm looking for a new PS, not that this one doesn't get the job done, but the fan noise is probably ok if this was in a server room, but not in a quiet room of a home. It's stupid loud.

I have a GF3 1650 in a new build consisting of a 13900k and 4090 FE, water cooled. I've noticed that even when lightly loaded e.g., 350 watts power on 4090 FE, and say 120 watts on the CPU, the GF3 fan can go from nearly silent to what seems like full rpm and noise. I have smart zero fan disabled, but I tested both ways with same results. My first thought was that it's happening at about 1/3 PS load, but that seems overly aggressive to ramp the fan to 100%, and it's not always consistent based on power draw. There are times that the GPU and CPU are pulling far more power, as reported by HWMonitor, but the GF3 fan is running slower. The fan behavior is very reproducible when running the Nvidia OC tool. The CPU is almost under no load, but the GPU during testing can be pulling 450 or more watts. I've also found areas in World of warcraft where a slight movement forward/back induces the fan curve with almost negligible change in the reported CPU/GPU power draw. Besides the two required sense pins on the 12vhpwr cable for power, does the GF3 implement the optional two (card_pwr_stable or card_cbl_pres) and would the secondary function of card_cbl_pres (power budgeting sense detect) be influencing the GF3's fan curve? My PC is very quiet with the custom cooling loop, with the exception of the GF3's PSU fan curve and noise.