Wind sensor formula for mph

radiohair11 · Post by **radiohair11** » Thu Jun 14, 2012 12:51 pm

I'm using an Arduino Uno board with the wind sensor hooked up to an analog pin. Does anyone know a formula for converting the analog reading into a number in kmh or mph?

I'm also curious (if the above formula does exist) if one is known that accounts for changes in temperature.

Thanks

mndless · Post by **mndless** » Sat Jun 16, 2012 4:29 pm

I am sure there is data curve out there for the specific sensor... I'm sure there is calibration required per unit... , you can derive your own pretty easy.. depending on your needed level of accuracy.

Solicit help from a friend, have them hold the working sensor out your car window. Record sensor readings in increments of 10 mph... from that data plot, you can derive a reasonably accurate wind speed to analog out curve...

Would be interesting to compare the short wind gust curve to a prolonged sustained wind curve... I'd expect a slight difference..

If you do it, post it... I'm looking at adding that sensor soon as well...

Human · Post by **Human** » Mon Aug 20, 2012 2:48 pm

I'm getting a $30 hand-held anemometer to try to calibrate this sensor. I'll try to derive a formula for a given extreme position of the trimpot (max counter clockwise or max clockwise) where temperature and Voltage are input and wind speed is output. I'll post it if I'm successful.

Human · Post by **Human** » Thu Aug 23, 2012 3:05 pm

I have gathered a lot of data, but I have found no clear pattern. In fact, some of the data at similar temperature shows less Voltage being used at higher wind speed. Will try varying my test environment next.

Human · Post by **Human** » Mon Aug 27, 2012 2:16 pm

On Arduino, I'm starting to think that this device is a better wind detector than it is a wind speed meter. I can easily tell the difference between some wind and no wind, for normal-range room temperature and for *most* trimpot positions (270 degrees seemed to always output the same value +/- 0.1%, but 0, 90, and 180 were good). When it comes time to tell one speed apart from another, the test data set doesn't support it. I got higher Out values (at the same temperature) for slower air than I do for faster air.

Here's some of my data that shows the inconsistency. I used a hand-held anemometer to get reference wind speeds at fixed points from a hair dryer's output vent, and I measured the wind sensor's Out at those pre-measured distances, reporting the average value over 100 samples, after a 10s warmup period. I'm supplying 9V external power to the Arduino and powering the wind sensor off its regulated 5V pin.

Code: Select all

avg Out reading   humidity%   temperature in C   wind speed m/s	   trimpot in deg   Arduino PSU V
021.02                34.00	     29.84			0.0                      90                    9
206.31                33.00	     29.00			4.0                      90                    9
197.62                33.00	     29.00			5.0                      90                    9
214.49                33.04	     28.96			6.0                      90                    9

I can't fit this data to any function that I can use to determine wind speed. Am I doing something fundamentally wrong, or is this wind sensor not a true anemometer?

Post by oz » Mon Aug 27, 2012 2:46 pm

Human,

It has a log function. It's not linear. The voltage is proportional the square root of wind speed.
BUT -

It varies with ambient temp - values go up with colder ambient.
It varies a little over 10% with sensor orientation to the wind.

Varying the pot is not going to be too useful. Just set it a fairly low number with no wind (we use .5V) and go from there.
So it's not a true anemometer in the sense of being really useful (in its current state) for quantifying wind speed.
I do agree with that.

Paul

medkhai · Post by **medkhai** » Mon May 20, 2013 3:30 pm

I have the same question. I am using the RAW output from the sensor to measure wind speed from a fan.

I do not need much precision and dont really need to compensate for temperature.

However it would be nice if the output was somewhat related to MPH or m/s rather than just the analog output (0 to 1023) from the arduino.

So I understand its a logarithmic output proportional to the square root of wind speed.... but not being a math guy... how do I convert it to meters/second? (again keeping it mind I dont really care for high accuracy)

Just pointing me in what math has to be applied, then I can fine tune it from there.

Thank you
Med

Bill sullivan · Post by **Bill sullivan** » Sat Feb 22, 2014 10:10 pm

Does anyone have a simple arduino sketch of the wind sensor outputting usable values, such. Wind speed, and temp?

I'm building a synth with the hope to include a wind controller for handling note on and sustain and stop. So far the vales I'm getting don't seem to be correct, looking to see how other are getting workable values.

spiessa · Post by **spiessa** » Tue Sep 01, 2015 10:47 am

Hi

last week I got two of your wind sensors (both revisions). I want to meaure true wind speed on my bike. Because I also measure altitude (with a BME280 pressure sensor) I have a very accurate temperature reading in degrees C from this source. I think, this could avoid many problems with the calibraton of the NTC resistor.
Now my question: Do you have a formula where I can use the (known) temperature in Celsius to calculate wind speed? Did anybody else try this procedure?

Andreas

RickyTerzis · Post by **RickyTerzis** » Mon Sep 25, 2017 11:27 am

Hi...i am a new user here. In my case some of the data at similar temperature shows less Voltage being used at higher wind speed. Will try varying my test environment next.

Dudemhccv · Post by **Dudemhccv** » Thu Oct 12, 2017 5:10 am

Hi list,
how accurate is the Lux reading obtained with the Lux-Sensor in Open Signals?
Best
Alex

Modern Device Forum