ultrasonic anemometer

Forum 2005-2010 (read only) Hardware Interfacing
1

Hi,
I am thinking of building an ultrasonic anemometer before buying all four ultrasonic sensor I though I would ask if somebody had some experience. My idea would be to have four sensors at N S W E. First I would send a ping from N start a timer and receive an interrupt from S that would stop the timer, will this work? has anybody ever tried or I must execute a ping before I get interrupted?
The idea would be to repeat S-N E-W W-E with all the timings and knowing the distance I can calculate the speed and direction of the wind.
What do you think about that before buying all those 4 sensors.

Thanks,
-Claudio

2

Many weather bugs have done just that. However, you can save a few bucks by using just three transducers.
Circuit Cellar has a good article on this, google 'circuit cellar anemometer'.

3

With 3 flight paths arranged as the vertex of a cube you get not only the wind direction but the wind vector, that's direction in three dimensions. This is how professional whether stations measure direction. Work out how big it has to be to give you a reasonable distance to measure the time delay over. The flight time should be in the region of milliseconds.

4

I was just thinking about it over lunch and I believe there might be a problem.
Speed of sound is ~330 m/s, that means that in 1ms -> 33cm, Arduino can measure milliseconds, that means that the two sensors should really be far a part to notice a difference anyway I would have a sensibility of a 0.33m/s.
Should check if that is enough.
Thanks for the cube suggestion, but this would be done for sail boats and I wouldnt be caring about the vertical that much.

-Claudio

5

Thanks for the cellar anemometer hint, I found what look like the schematics on the link you gave me it looks really simple to build, but once I've built the hardware (which would be first time) how do I go programming the chip, and how do I program a chip like that?
do you have any links or explanation? I know it is dumb but the closest I got to hw was jtag development on ARM with linux... I am missing a couple of extra step between what I know and the soldered hw.