Trouble with Delta fan

[achrn]

I've found another fan with suspect pullup (c.f. https://forum.duet3d.com/topic/31272/trouble-with-cui-fan) and think this might be a trend with weak pullups. I'd like to ask Duet3D to consider it with respect to future hardware.

This is a Delta BFB0512HA-CX3D which is potentially quite handy - a reasonably powerful 12V 50x10mm centrifugal blower with four wire connections. I bought mine from Digikey, so about as reputable a source as I can manage.

Anyway, it has the same behaviour as my previous report - at spec PWM frequencies it doesn't run unless duty cycle is 100% and at lower frequencies there's more control but it howls. Stick it on the oscilloscope and it's clear the PWM signal isn't rising fast enough. At 25kHz 70% it's not even getting to 0.6V (apart from a spike to 1.36V as it switches):

Same solution as before, in this case I've put a 1k pullup to 5V, and it runs nicely at 25kHz (again, this is 70% duty):

Both these are the same fan with the same Duet board (actually a toolboard 1LC) set to 70%.

This time the data sheet specifies what the internal pullup is, as 90k:

I note that @dc42 last time explained that the Duet boards rely on Intel PWM fan spec https://www.intel.com/content/dam/support/us/en/documents/intel-nuc/intel-4wire-pwm-fans-specs.pdf which requires an internal pullup. I observe that the spec doesn't specify a strength of pullup. This fan has an internal pullup (to 4.8V) so to that extent it's compliant with the Intel spec, it's just not a useful pullup without some additional assistance.

[Note in case anyone actually tries some analysis - these traces are actually one PWM signal from the board with one added 1kohm pullup to 5V driving two of the fans in parallel.]