Homebuilt anemometer


Active Member
Reminds me of the late '70's when I designed a 3-cup digital anemometer. I had 3 ice cream scoops heliarc-welded to 3 AL spokes. The shaft entered a water-tight electrical box. The shaft inside had an acetate optical disk which was read by a U-shaped TI IR rcvr/xmtr IR sensor and fed to a Schmidt trigger. Pulses vs time were counted, averaged by 10 samples and the output drove 7-segment LED displays. The gov't said that it was the most accurate wind speed indicator they ever tested. It made its way into the commerical market along with a companion optically-encoded wind direction device which used a gray code disk.

If you go back that far, the product line was called MetData.
Thanks for bringing back the mems! :)


Active Member
I do somewhat remember that project from the gov't Jay. As a matter of fact I still have one of the original MetaScoops that my son uses nightly for his evening snack. :)


I'm wondering if you can make this work via CMax code in the Ocelot?? For instance you would have to somehow "count" the number of digital input transitions (from an SECU16/I) over a period of time.

Possible using an imbedded timer(s) maybe??

Guy Lavoie

Active Member
The bus latency of an Ocelot and modules like the SECU16 don't allow it to count very fast. I would say that anything that exceeds more then one pulse every 1.5 to 2 seconds risks losing some events. This is especially true if you have several modules on your adnet bus, because each individual module is polled a bit less often per second. A device like a anemometer would likely put out many pulses per second (typically at least one per shaft revolution).

This would be a good microcontroller project however!