Building an Anemometer

I built this in 2008 to measure wind speed. I don't have it up anymore because first a nearby lightning strike killed the transistor then I chewed up the cable with the lawnmower. I had it fairly well away from the house, about 200 feet, which is important as far as not being sheltered by anything.

calibration

Here I had it mounted on my car for calibration. If you do this it's important to mount it far enough forward that air moving over the car isn't going to hit it. If I'd just mounted it to the roof rack air from the hood and windshield would have added to the real air flow at each speed and messed up the readings.

The anemometer itself puts out TTL level pulses so inside the car I had a frequency counter reading those. I used a small voice recorder and read off the digital speedometer reading and frequency counter reading every few seconds. A slightly less fancy method would be to have an observer with a clipboard taking readings. Don't try to write while driving. I drove about an hour, being careful to go over my route in both directions, and took about 170 readings.

cutting

Most anemometer cups are shaped like part of a sphere but I did some research on the web and found that cone shaped cups are just as good if not preferred. I was thinking about filter paper in a chemistry lab, how you start with a circle, fold it in half, then fold that in half, open one side and push it down into the funnel. Use this as a rough cutting guide, just make circles then cut out 1/4, pull the edges together where the cut out is, then secure them to make a cone.

cup inside

An inside view of one cup. I used JB Weld epoxy liberally to hold things together but for balance reasons it's important to use about the same amount on each cup. The machine screw here is a 2-56 put in to hold the corners together when pulling them in to form the cone. There's probably no reason to make the cups watertight as long as they're all about the same. The cup material is aluminum flashing which is stiff and awkward but cheap and light.

cup back

The back side of one cup. The wire wrapped around the support arm is for balance. The support arm is glued into 2 holes drilled through the cup.

overall

Top view of the assembled anemometer. The support arms are 1/4 inch aluminum rod. The bearings are from a VCR head that's designed to do helical scanning so it turns the whole time the VCR is running. They're very nice ball bearings on a 6 mm shaft. Parts of the head form part of the housing, as does part of an old 1/3 HP motor. The shiny disks are from old hard drives, they're polished aluminum. The circle the cups travel in is 1 meter across.

outside

Closer up view of the outside. I didn't take a lot of pains to keep water out, figuring it was more important that any water that got in could drain back out easily. There are some holes in the head that I plugged with RTV. The holes for the arms were drilled in a lathe but not terribly precisely. The disk showing here is from an old 5-1/4 inch hard drive and it mostly just keeps water out.

inside

Inside view. What sends a signal down the wires in this are the 60 holes drilled around the edge of the inside disk. They pass through an optical detector at the left which came out of an old motor. The holes were drilled with a hand electric drill and then "adjusted" with a needle file while watching the output of the optical detector on an oscilloscope to make sure that each hole was doing what it needed to. The 60 holes is so that 1 RPM becomes 1 Hz on a frequency counter. The disk here is from a 3½ inch hard drive.

I've removed the amplifier to redesign it here but it was on a small piece of Vectorboard. I had 3 wires: ground, signal, +5 volts coming in through the DB-9 connector at the right. I also had about a 30 ohm power resistor mounted in here as a heater so I could apply 12 volts for deicing. The connector got rusty on the outside but the pins inside are gold-plated and they're fine.

plot

Plotting my data in a spreadsheet and doing a linear least squares fit to it. The slope here is 4.931 but yours will probably be different. One thing that bothered me is that there's no Y intercept. I suppose this should tell me that the anemometer has some loss and doesn't turn all the way down to 0 MPH. The real wind value was always something even on a calm day, and there was some minimum value that the speedometer would read.

With these bearings you should be able to get the static balance almost perfect. Mount it sideways and keep putting weights on until it will stop turning in any position with no obvious heavy side always down. It's sensitive enough to start turning at walking speed if you just walk along holding it out, but this has survived 50 MPH or so winds over a few years. There aren't any amplifier details here because, well, mine blew up and you'll want to make something fit your optical sensor anyway. Solar powered wireless would be a nice touch.

AB1JX / toys