tl;dr - Is driving brushless DC fans using PWM unhealthy for the fan compared to variable but steady DC voltage? If yes, why and how?
The super simple input [PWM] -> MOSFET driver [PWM] -> DC Fan
to adjust the speed of DC fans is well known. The DC fan receives a PWM of the same frequency as the input, with sufficient juice from the MOSFET at higher voltage. Ignoring everything else, assume the fan gets a high-current PWM signal swinging between 0-12V at some duty cycle to vary the fan speed, and that the 0-12V levels are clean (no spikes etc).
Say I have a bunch of 0.5A, 12V brushless DC fans. These are not PWM fans (i.e. they only have 2 wires). I now drive them using the above PWM circuit to vary their speed. Assume the PWM frequency is around 25 kHz, and that the MOSFET can easily switch at that frequency.
I've read that adjusting the speed of DC fans using a variable DC level voltage is "more healthy" for the fan than using PWM like above, but didn't give details.
For the experts out there:
Are there dangers to driving DC fans like the above using PWM (pulsed) instead of steady voltage levels? If yes, what are they, exactly, how do they manifest? How important is the PWM frequency for the fan (assuming the MOSFET is fast)?
p.s. I can build a PWM-to-DC circuit (e.g. source follower, etc) but here I'm interested to understand the dangers of driving DC fans with PWM.
p.p.s. My personal experience (edit: excellent explanation for this by Tony EE below)
I'm pretty sure PWM may affect them (though here I'm asking a general question). I've heard clicks coming from all my DC fans when driving them with PWM, particularly at low-duty cycles. More importantly, after a little less than 1 year under continuous operation at various duty cycles (50% being the most prevalent), some of the fans no longer respond to low duty-cycle PWM. Specifically, they still spin at 100% speed when the duty-cycle is 100% (practically 12V steady) but any lower duty cycle results in the fan spinning at a very slow speed, regardless of the actual duty cycle value -- as if the fan had 2 speeds only: full and very-low. All these fans responded well to PWM before. They still spin freely when pushed by hand, I feel no extra resistance compared to the other fans (so it's not the rotor).