> Hello Guys
> I've been looking around and trying to learn something about sensing
> wind speed. Apart from anemometers with three cups it appears that my
> options are ultrasonic transducers or using a pair of LM335's, one of
> which is self heated.
> Does anyone have any practical experience with wind speed? Or, has
> anyone heard of another way to measure wind speed?

I've done the anemometer type thingy, and you're not limited to three cups...
Four is just as workable, and slightly easier to make
(90 degree angles being what they are, versus 120 degree ones). I used bits of
2mm galvanised fencing wire and table-tennis balls
cut in half. Mounted them onto a small 12V motor from an eftpos journal printer
which I mounted vertically. I reduced the bearing
friction in the motor by supporting the "short" end of the shaft on a hardened
steel ball and only using the inbuilt bronze bushes
for side-thrust. I reduced the brush/commutator friction by removing the
brushes and commutator! I reduced the rotating intertia
by removing the coils and armature. Oh, and I took out the magnets too :-S Not
much of the original motor left except the shaft and
bronze bushes. Oh, and the photo-interruptor disc on the back which gave me
something like 20 slots per rev. BUT, it did
accurately read windspeeds down below 1 kmh, and over 80kmh (half a knot to 40
knots), compared to a commercial unit rated well
beyond my meagre usage requirements (and even further beyond my price tag!). It
could probably have gone higher, maybe even to
150kmh, I just didn't test it that hard.

I've also used the two-temperature-sensors and (in my case independent) heater
element, that works quite well, and is valid for just
about ANY type of fluid, not just gases.

One simple way to measure windspeed that sometimes is good enough is a large
vane which is suspended vertically by a thinish (but
rigid) wire, from a hinge. The more the wind blows, the more the vane
deflects,... if it's horizontal, time to batten down the
hatches! I don't remember how the relationship between angle and windspeed
goes, but it's probably NOT linear, whereas within
reasonable limits, both the cups-anemometer and thermal gradient methods ARE
linear with speed.

With any wind sensor, there are some issues to think about (you can decide to
ignore them, but you need to know what they are to
decide if you CAN ignore them!).

1. Turbulence. This can be from nearby objects, or your sensor device itself.
Pretty obviously, you can't tolerate much of this
from either source, unless you're purely interested in a "wind / no-wind"
sensor.

2. Direction. Wind-sensors for (eg.) sailboating might require some degree of
dependence on all three directions, since the boat
can be keeled over quite sharply, affecting the apparent direction of the wind
on the sails. Therefore some kind of "3D windspeed"
(more probably wind-velocity) sensor might be required. Meteorological wind
sensors should be designed to read wind speed
independent of horizontal direction (that is, they're essentially two-D
devices). Assuming one has handled point 1 appropriately,
and is on level ground, one doesn't normally expect much wind to be travelling
in anything other than a horizontal direction. :-)
Lastly, windspeed in a pipe is obvioulsy constrained to flow directly along the
pipe, sufficiently far away from corners and
fittings, and is pretty much one-dimensional.

3. "Dynamic range". Just how much (and how little) wind speed do you need to
accurately read?

4. Linearity. Do you need it? Can you work with a simple fudge table? Can you
measure wind speed in your own not-necessarily-linear
units and be happy with that? (A related problem occurs with thermistor based
temperature measurement,... one of our projects at
work measured temperature in what became known as Leotherms... the guy writing
the micro-code was Leo!).

I think that about covers it.

Hope you can get something useful out of all that.

Cheers,
David.