It’s been a while since the last post but I hope this might be of use to some.
After a few months
I finally gave up and opened the scope. My idea was to measure critical temperatures and characterise the fan performance based on its voltage – when the scope is fully closed. I attached 3 probes (DS18S20) to the power supply heatsink, FPGA heatsink, and one “loose” inside the case. The probes are attached using double sided heatsink thermal tape, which is quite strong. Additionally, I placed one more probe a few cm from the inlet fan.
Next, I hooked up the SLA1016 logic analyser to give the setup more power draw. I also switched all channels, analog and digital on.
I programmed the SPD3303 to 5 1h slots. Started with 8V increasing it by 1V every hour. Below you can see the graph. I also attached the scope placement. There is not much air movement in the corner.
Looking at the data I take it’s safe to under-volt the fan. I believe I’ve read the operating temperature is up to 40 deg. Here in North-East England, where I live there is rarely above 25 deg outside. Plenty of headroom in my view.
I’m settling with powering the fan with 9V. The noise is much-much lower and bearable. I can barely hear it and it’s not annoying. Besides, it's easy to stick a standard, linear regulator without risking noise from a buck regulator.
I’m planning to do similar experiment with the SDM3055 and STD1062X soon. Can’t stand them fans!
P.S.
I just spotted a mistake. The last slot should be 9V.