I was using a cap as a simple filter, so fan itslelf got actually DC. All smaller 3-wire fans were working fine, I could see perfect square tacho pulses, no problem there. Now I have read (microchip AN770 and documents for TC642) that fans have actually to be driven by very slow PWM, 60Hz or less. I had dig up all docs I could find from papst, and could not find any exact info on that from them. I found that the two kinds of papst fans I have, have different kind of "third wire" - one has white wire, which should be DC 0-10V control input, but the other has yellow wire, which had to be ordinary open collector tacho signal. Docs show that I have to use a pull up resistor there, which i did, but did not get any signal there, anyway, just zero. Even more, I tried to find out signal path on one of original controller boards, that came with the fans, and it looked like the yellow wire is used that way, that 4-layer board was not easily traceable without desoldering components, though, so I can be wrong. Just the resistors around that signal wire were much larger than I would expect, in hundreds of kilo-ohms range. I tried to partially duplicate that, and did not get any better result anyway. Basically I am stuck with that problem, and seems I'll have to completely redesign that controller, to work with current sensor and not using tacho line at all. There are useful descriptions for that approach in above mentioned microchip documents. I just want to use those fans, since they are excellent mechanically, and, after all, I have a lot of them, for free :) On Thu, Jul 8, 2010 at 12:01 AM, Barry Gershenfeld wrote: > One problem with PWM'ing a 3-wire fan is that the tach signal gets chopped > up, reflecting the PWM on the supplied power. > > You and I could imagine a hold-up circuit for the tach signal, but > apparently most fan makers can't. -- KPL -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist