Hi Mark, > For fan 1, > my version of the datasheet shows 24 MHz for bit7 = 0. > (24 MHz / 1) / 256 = 93.75 kHz > > fan 2 matches what you have below. > go figure... Correct. I had forgotten to divide by 256 anyway so my value wasn't correct. 24MHz|180kHz makes much more sense since 24Mhz/127 (max pre-scale) is 188kHz i.e. right above 180kHz. 1Mhz|125kHz doesn't look as intreresting since both clocks cover almost the same range of frequencies. I'd say that 24Mhz|180kHz is more likely to be correct, Winbond must have corrected the values in the datasheet for PWM1 and forgotten to do the same for PWM2. Just a guess though. Thanks, Jean
