OK you guys are in trouble now
Here is a few of the answers to the questions you are asking about.
These are direct quotes from Perry ( Dam it, did it again!! )
Brians256 :
For instance, does your design turn the fan full on and full off when it reaches the highpoint temperature? How much hysteresis does it have (we don't want the fan turning on and off constantly as it hovers around the setpoint)?
Answer :
Hysteresis doesn't come into play because it's a variable speed controller. If it was simply on/off, then hysteresis would be a factor.
Brians256 :
Does it use PWM to allow very low fan speeds and high efficiency or does it use linear regulation that is smoother but puts out more waste heat?
Answer :
It uses PWM to allow variable speed with good efficiency and to avoid having to use a large
heatsink for the switching transistor.
Airspirit :
If I may make a suggestion, (if it would be possible), it would be nice to be able to set a
minimum system voltage for the loop, allowing your fans to run constantly at 5V-7V, but ramping up
if the temps get above a certain mark. That would be VERY useful, indeed, and might draw some
serious attention in the fan-cooler crowd, since this would be a virtual godsend.
Answer :
This would be easy to do with the addition of one more potentiometer.
Airspirit :
worried about spiking temperatures due to slow or no fan cooling.
Answer :
Since the water is continually circulated, the thermistor would sense the increase in temperature and would start the fan as soon as it was needed (assuming that it was set up properly). The change in fan speed for a change in temperature is adjustable (as is sensitivity and the set point).
MadDogMe :
maybe just a single LED to show it's 'kicked in'?...
Answer :
Easily doable. It would involve another inexpensive IC set up as a comparator. You would set it to come on when the fan was turning a certain speed (you'd have to calibrate it for your particular fan).