On Sun, 8 Jun 2008, Bruno Pr?mont wrote: > I will look if I can find a way do determine the real fan speed in > order to compare with the values both chips report. The exactly right instrument to use for this is probably an optical tachometer. It looks for reflections from a piece of reflective tape or a white stripe painted on a rotating machine, and tells you the speed in rpm. If you're in university (or know someone who is), the physics or mechanical engineering labs will often have such a device. http://www.extech.com/instrument/products/451_499/461700.html is one example but they are made by several companies. There is another kind of tachometer that has a small wheel or probe that you hold against the rotating machine; the machine spins the wheel and the tachometer displays the rpm. A computer fan motor is not very strong, though, and making the measurement may tend to slow down the fan. http://www.extech.com/instrument/products/451_499/461750.html is an example, but again, several companies make them. The previous way to do this was with a stroboscope - a flashing light with adjustable flash rate. You point the light at the rotating machine and turn a knob until the flashes of light "stop" the motion of the machine. You then read the flash rate from the knob and compute the rpm. http://www.ietlabs.com/Genrad/1531.html is a classic example. If you have a function generator, you might be able to connect a LED to its output and use it as a stroboscope. (Make sure the output of the function generator is strong enough to drive a LED; some of them are not designed for this.) Another way to do it is to use a plain old fluorescent lamp (with magnetic ballast) or neon glow lamp. Either one will flash at 100 Hz (Europe) or 120 Hz (North America). It may help to be in a dark room with the fluorescent or neon lamp as the only source of illumination. You can't adjust the flash rate, but you can use it to get an idea of the fan speed. For instance, if you mark one fan blade, and it appears to "stop" in the same spot every time, the fan is probably turning at 6000 rpm (Europe): 100 Hz * 60 sec. (It could also be turning at 12000 or 18000 or 24000 rpm, but most computer fans don't go that fast.) If the mark appears to "stop" in two places, 180 degrees apart, the fan is probably turning at 3000 rpm, and so on. Standard disclaimers apply; I don't get money or other consideration from any companies mentioned. Matt Roberds
