AnemoSens - SLA printed anemometer

This is my design of a SLA printed anemometer

Similar projects worth following

When I started with my own SmartHome, one of my first projects was my own small weather station (Small insight here: to monitor weather data.

Sensing the temperature, air pressure, air humidity and co was relatively easy. For example, with the BME280 you can record all values with just one sensor.

One problem that I couldn't solve until the end is how to record data about the wind, i.e. wind speed and direction. Of course, you can buy matching sensors. However, I always found them too expensive or big. If you wanted to use a compact sensor without taking up a lot of space, you almost always had to dig deep into your pocket.

That's why the decision to build my own compact wind sensor had long been made for me, but so far I haven't gotten to it.

For the construction of WinDIY_2 ( I was faced with the problem again that I needed a cheap and compact wind sensor. That's why I started to develop my own. :)


If you want to support me, you can do so by like on this project page, on Instagram or by donating a coffee. :)


Content that is not based on software/code: Unless otherwise stated, all works presented here that are not based on software/code are subject to the CC BY-NC-SA 4.0 license (attribution – non-commercial – dissemination under the same conditions 4.0 international).

You can find a summary here:

You can find the complete legal text here:

Software/code-based works Unless otherwise stated, all software/code-based works presented here are subject to the GNU Affero General Public License v3.0

You can find a summary here:

The complete legal text can be found here:

  • First build information available

    Fab10/11/2021 at 17:18 0 comments

    In the last days I added a first build guide on my blog It includes the guide for the sensor PCB and the MCU PCB for Anemosens.

    You can find the build guides at the following links:

    In previous projects I inserted individual pictures in my build guides showing which components have to be soldered where on the PCB.

    An even better way to get an overview of which components belong where on the PCB is provided by the Open Scope Project. With this you can generate very helpful HTML files in which you can see directly which components have to be installed where on the PCB.

    You can see the overview for the Anemosens Sensor-PCB here: Anemosens Sensor-PCB

    You can see the overview for the Anemosens sensor PCB here: Anemosens_MCU PCB

    Additional infos are also available in the git repository of anemosens:

  • Anemosens MCU Board

    Fab10/07/2021 at 17:49 0 comments

    In the last few days I have assembled, tested and programmed the MCU board for Anemosens.

    Thanks to the MCU Board it will be possible to use Anemosens as a standalone Anemometer that can be used for long time observations in remote areas (therefore the microSD-card slot) or just as one part of your weatherstation (therefore the RS485 interface).

    The RS485 interface is used to make the measured data available via Modbus. This way you can connect the Anemometer easily to your RaspberryPi, PC or many other platforms that support RS485/Modbus. :)

    Additionally you can activate a serial-stream of the data (packed as handy JSON) via the USB connection.

    And last but not least the integrated ESP32 also opens a BLE Server. Thanks to this you also have the possibility to pull the data wirelessly using your smartphone or computer. 

    Of course you can also write your own firmware and program it via the integrated USB-C port. The QWIIC-port makes it easy to connect additional sensors via I2C. :)

  • First prototype of Anemosens

    Fab08/01/2021 at 13:02 0 comments

    The first prototype is finished. Please see the pictures and the video for further details. :)

  • Finished the sensor PCB

    Fab08/01/2021 at 12:47 0 comments

    In the meantime I have also created the circuit board for the Anemosens sensors and completed the first tests.

    A Hall sensor (for detecting the wind speed) and an AS5048B rotation angle sensor are located on the circuit board.
    With the latter, the direction of the wind can be determined.

    As always, you can see more details in the pictures. :)

  • Slightly adapted design

    Fab08/01/2021 at 12:43 0 comments

    The design of Anemosens is now quite finished.

    I made the tip a little "sleeker" to show the wind direction. In addition, the base has also become a bit more compact.

    The circuit board including sensors will also be housed in the base :)

View all 5 project logs

Enjoy this project?



dev wrote 11/09/2021 at 13:28 point

Hi! Thanks for sharing the project. This is probably one of the very few designs with both Wind Direction and Speed. I wonder if you can share the STL files for SLA printing? Definitely looking forward to the Wind Calibration tests!

  Are you sure? yes | no

dynafem05 wrote 10/17/2021 at 11:00 point

I am new here my name is Emmanuel I am a Broadcast Journalists, Content Writer and Creator, Voice Over Artist, Comedian and lastly an MC. I  would be glad to meet people here and to share  thoughts, knowledge, Business Ideas and more. Thanks and warm regards. 

  Are you sure? yes | no

julian wrote 10/11/2021 at 21:57 point

Hi, how you get the bearing without much friction?. I made myself some 3d anemometers and the typical 608 bearing didn't work at all I would like to purchase something similar to a spinner bearing but no clue

  Are you sure? yes | no

Fab wrote 10/16/2021 at 10:10 point

Yeah, thats indeed a problem I didnt solve yet. Especially at lower wnd speeds it needs some initial torque to start the rotation. Hopefully today I will try to remove the grease from the bearing and see if this reducees the needed intial torque. :)

  Are you sure? yes | no

Tim wrote 10/29/2021 at 16:02 point

I'm working on my own weather station now. What seems to be working well for me are ceramic bearings, although they are much more expensive than your standard 608s. I liberated a bunch of fidget spinners that were about to go into the trash, but found that their bearings were too small and not enclosed. I am currently using two of these (one for the anemometer and one for the wind vane), with the hope of finding something a bit more affordable down the road as I'd like to make a bunch of stations:

They work quite well, just be sure not to lubricate them since that adds additional force required to start rotation. Hope this helps. 

  Are you sure? yes | no

john.r.sheahan wrote 10/09/2021 at 21:40 point

Hi Fab

I've got a Davis instruments anemometer in my home weather station. It seems to have pretty good low speed performance, I assume its a point and a bush internally, I didn't disassemble. Trying to cal that on the roof of a car was rather inconclusive. It didn't seem that linear.  

At least its worked for a few years. Just reworking the code in the attached arduino+radio to reduce supply current. Its solar doesn't like multiple overcast days.

  Are you sure? yes | no

Fab wrote 10/11/2021 at 17:21 point

Hi John,

sounds like a cool approach to calibrate it with a car. :D Didnt thought about that.

My idea (so far) is to build a tiny wind tunnel and use a commercial available anemometer as reference. Also thought to ask one of the universities here in my area if they are willing/interested/open for a possible cooperation about this. 

But first: next mechanical design iteration. :)

Best regards from germany

  Are you sure? yes | no

john.r.sheahan wrote 10/12/2021 at 00:22 point


Making a wind tunnel sounds like work. If you can borrow one, would be  a good solution. And you need an anemometer with a cal you trust.  I suspect you just need a reasonably laminar airflow, so a mast to get maybe a metre above a car roof is likely sufficient. You need a calm day and a very quiet road obviously. I'm in rural Australia so a quiet road was available, but it was neither still enough, nor was attaching the anemometer to the roof rack stable enough. So I used the stated Davis cal. The fastest gust I've seen in the last year was in the vicinity of 55Kph, so a car cal could give you both acceptable range and a few points within the range.    And a gps to get accurate car speed is trivial.

Not that happy about using the Davis cal, I have an older Davis tipping bucket rain gauge, I checked its cal by adding a known water volume slowly - it was off by 25%. Although I don't think the plastic bucket is long term stable, probably needs to be stainless sheet, or maybe chrome plate. 

FWIW, softwarewise I log peak gust speed by taking distance (# interrupts * scale) over 8 seconds,  average speed as a vector sum over 12 minutes update rate, and average direction weighted by vector direction.  Took a few goes to get something acceptable, and a little work to get linear enough approximations for the vector maths on a lowly 328P which is mostly sleeping.

  Are you sure? yes | no

john.r.sheahan wrote 10/08/2021 at 00:24 point

Any plans to calibrate this, and check its light breeze performance too?

  Are you sure? yes | no

Fab wrote 10/09/2021 at 10:45 point

Yes and yes, light breeze performance is not that great since the bearing needs relative high inital torque to start rotating. It's getting better when the bearing achieved its first 100 turns but yeah, still not perfect. Actually I want to finish the electronics and firmware. Then I will concentrate again on a small redesign using a 603ZZ bearing instead of the current one. Hopefully the smaller bearing diameter will decrease the needed initial torque. 
After that I plan to do a proper calibration. :)
Best regards

  Are you sure? yes | no

john.r.sheahan wrote 10/12/2021 at 00:36 point

just a thought, removing the bearing dust shield, washing out the grease and replacing with a light oil might help. Not like you are loading the bearing heavily. I guess the design then needs to keep the bearing clean.. which becomes a maintenance problem. 

  Are you sure? yes | no

Fab wrote 10/16/2021 at 10:10 point

Removing the grease is an awesome idea. I will try that as a next step. Hopfefully this already solves this problem. Thanks for the hint. :)

  Are you sure? yes | no

john.r.sheahan wrote 10/16/2021 at 20:42 point

sorry, its no awesome idea, its an ugly, ugly hack. Hopefully it provides useful information.    

If you want some possible ideas for wind speed + direction filtering, try here  Apologies for the crap code, I'm an ee hardware designer.  

  Are you sure? yes | no

Fab wrote 10/17/2021 at 10:48 point

Sorry, but its not an ugly hack. From my point of view its an adaption of the available material to a special usecase. The main reason for grease in the bearings is the reduction of friction and wear, carry away generated heat and to avoid corosion due to humidity. So if I remove the grease and replace it with some more light oil (for example that one used in sewing machine) I will probably loose the high load capability the bearing was orginally designed for (which I do not need in my usecase). Bu as a tradeoff I will probably win a much lower initial torque. If this works properly and realiable over a longer period I dont get why it shouldn't be a compromise to solve this isse.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates