Propel-E 450 | A very powerful windturbine

A 2m windturbine made from corn! Completely 3D-printable and open-source.

Similar projects worth following
The Propel-E 450 is a windturbine made from corn starch, a residual product convieniently known as PLA. It can be made with easy to find parts and is completely open-source.

With it's maximum power output of 450W, it proves that real power can be made. It is designed for off-grid living but can also be used to feed into the main electricity net when used with an suitable controller.

Elaborating on the work of Precious Plastics, Scoraig Wind, Otherpower, and many more from the DIY- and maker community, Propel-E doesn’t reinvent the wheel, but adds on known and tested designs.

1. Problem: The challenge the project addresses
SuperAdobe structures have been built in 54 countries all around the world. Often in distaster-struck or remote places where there is no acces to electricity.
Being off-grid means being dependent on other sources of energy, like wind and solar. When the sky is cloudy, the production of solar panels diminishes quickly. When not complemented with batteries, there is also no power at night when using solar only. This is why solar and windpower are complementary. There is also the problem of plastic waste. In many countries plastics are just dumped into wastelands.

2. Solution: How the project will alleviate or solve the problem
The Propel-E 450 produces enough electricity for the basic needs of a small household, for example powering a fridge, an infrared heater or lighting. The Propel-E 450 can quickly and easily be mass-produced through crowdsourcing. This approach has recently proved to work, where the 3D-printing community resolved the shortage of Covid-19 faceshields. The windturbine can be printed in a few days with any household 3D-printer.

It has a very sturdy design with as less components as possible, while still ensuring maximum power output.

i. Design process & design decisions

---- Material choice ----

The Propel-E 450 is made from PLA, wich is one of the stiffest plastics (tensile modulus of 3,5MPa). It also brakes down much faster than regular plastic when burried in organic matter. If a windturbine would be disposed or broken down, the copper and the magnets can be recycled, and the PLA breaks down.

With a good solvent, other materials like PET(G), or ABS (dissolves in aceton), can also be used. For example, we designed a smaller windturbine made from recycled PET-bottles:

The Propel-E 450 has been tested in different environments, including weather as hot as 38°C. 

---- Wing design -----

We opted for fixed blades, meaning the blades are attached to the axle. Disadvantages of adjustable blades, like being prone to damage and increased material and maintenance costs do not add up to the advantage of having the blade angle perfect everytime. Also, a near-perfect blade can be made by twisting the blade along its length. Because of this, fixed blades have become the mainstream in the wind DIY-community.

To be sure we have the correct twist and taper, we used Hugh Piggot's Blade Design spreadsheet. Furthermore we choose the NACA 4415 airfoil for its good suitability for small windturbines.

We modeled the entire blade in Fusion360, with some small tweaks in 3DS Max.

To safeguard birds, recent studies have showed that colouring 1 blade black decreases fatalities significantly. One of our Patreon-supporters is currently working on this:

For bats there is only an issue if there is a vacuum behind the blades, which only is the case with big industrial windturbines having blade diameters of >60m. (Bats are, unlike birds, mammals and have weaker lungs that implode after fying through the vacuum. After a few 100 meters, they drop dead out of the sky because of oxygen depletion).
Other studies have selected purple for all 3 the blades as the best colour for both bat and birds (study on big industrial windturbines of >60m).

----- Alternator design -----

We choose for an air-core 3-phase axial flux alternator because it is perfectly suited for windturbines. The air-core makes that the windturbine starts very easy, and the 3-phase star connection reduces vibrations and makes the turbine more efficient when the current is rectified for battery charging. 

Therefore we decided to base the design of the alternator on the 2F Ferrite Windturbine from Scoraigwind. It uses ferrite magnets that are not relying on rare minerals like neodymium magnets do, which makes the windturbine even more environment-friendly. Ferrite magnets are also not prone to humidity, which is an often seen issue with neodymium magnets....

Read more »


Sourcefiles for standard 3D-printer with 20x20x20cm build volume.

max - 9.66 MB - 10/05/2020 at 12:23



Sourcefiles for standard 3D-printer with 20x20x20cm build volume.

fbx - 4.69 MB - 10/05/2020 at 12:22



Sourcefiles for standard 3D-printer with 20x20x20cm build volume.

obj - 4.17 MB - 10/05/2020 at 12:22


PE450 - For big printers.max

Source-files for 3D-printers with big build volume

max - 2.35 MB - 10/05/2020 at 13:52


PE450 - For big printers.FBX

Source-files for 3D-printers with big build volume

fbx - 2.15 MB - 10/05/2020 at 13:52


View all 30 files

  • 7 × 1kg PLA White Plastics used for 3D-printing
  • 2 × SS-6206-2RS bearing Bearings for the alternator
  • 24 × Ferrite magnet of 50x50x20mm - C8 grade 3 Magnets for the alternator
  • 1 × Copper wire of 1,6mm⌀ (4kg) Lackered copper wire for the alternator
  • 1 × Iron wire of 2mm ⌀ (10m) Wire to shield the magnets

View all 9 components

  • Bearing margins

    Bram Peirs @ FW2W10/03/2020 at 17:19 0 comments

    For now we had a little play on the margin between the bearing and the generator. The reason for this is that diameter of the hole for the bearing can vary a few tenths of a mm, depending on how much glue is used. Therefore we designed a custom piece that can be printed in one part so that it always will fit perfectly. It is shown below in purple:

    The bearing is placed into the part like this:

    And afterwards both are welded into the generator. It's also possible and maybe even better to wait to insert the bearing until the holder has been welded into the generator:

  • More on wind turbine controllers: to China or not to China?

    Bram Peirs @ FW2W10/02/2020 at 12:33 0 comments

    A windturbinecontroller for charging batteries needs to have 3 main functions:

    - 3 phase rectifier: to convert the 3-phase AC into DC for the battery, a diode bridge is needed:

    - charge controller: the battery needs to be charged at the correct voltage (13,5V-14,5V)

    - discharge controller: deep discharge damages the battery, so it has to be switched off at the appropriate level (around 12,6V unloaded at room temperature).

    - dump load: when the battery is full, the rest of the power needs to be diverted to prevent overcharging the battery. The most common application is to dump it into resistors. Just disconnecting with a switch is not enough as it would make the windturbine runaway with a too high RPM, leading to self-destruction.

    We designed our own controller to provide these functions, on the very handy website of (We are not affiliated)

    Although this controller worked fine, it set's you back 25EUR in parts, labour not included. That is if you don't make any mistake..  Therefore the cheap (Asian) controllers we mentioned in our previous post are also worthwile looking into. There are a lot of deviations of the same controller to be found on marketplaces like Alibaba, Amazon, Ebay,.. Often relabeled with a fancy brandname and then sold for twice or even quadruple the original price.

    We ended up buying this one for only 10,89EUR on "Zerodis Wind Charge Controller Wind Regulator Charge controller with self and hand brake function Waterproof 400W (12V)". With the use of the Google Translate app and some basic reasoning we were able to understand the labeling and how the controller works:

    It works fine and even has a button to manually control the braking (=dump load) for maintenance.

    Another controller that came in was the "Zerodis 12V / 24V 300W wind turbine generator lader controller". It's named Zerodis, but actually it was just  the cheapest rebrand we could find, underneath it's all Chinese. It came in for 28EUR, and there is clearly a quality difference: the cables are thicker, and the heatsink of the dump load is also way bigger dimensioned.

    It works just fine (for now?) and runs cooler than the other controller thanks to the bigger heatsink.

    Unfortunately, none of these commercial controllers have a function that protects the battery agains overdischarge. Therefore we added an additional unit, bought seperately for around 10EUR on Amazon:

    It does the job but relates battery charge only to the battery voltage. This means that if you connect a high power load, the voltage also drops, triggering a false 'empty battery' alarm. There are more intelligent discharge controllers on the market or to be made, that also measure the current that the load is drawing, but these cost a multitude of the price.

    This is how it looks like al wired together. (The 3-phase current from the windturbine comes in through the thick black cable)

  • Happily spinning again!

    Bram Peirs @ FW2W09/29/2020 at 17:44 0 comments

    After installing the new bearings the Propel-E 450 is spinning happily again, and has survived some strong winds!

    We are also testing some charge controllers. A cheap one and a very cheap one! More on the coming log!

  • An update from our Patreon supporters

    Bram Peirs @ FW2W09/23/2020 at 10:52 0 comments

    Some of our Patreon supporters come to us with really great ideas! For example to make puzzle-like structures to ensure all pieces line up properly when glued together. This is the addition that one of our supporters made in OpenSCAD, unfortunately resulting in a non-manifold model.

    We found the idea so great that we decided to implement it in our models (made with 3ds Max). The new versions will soon be online!

    And the end-result:

    Another addition was alignment pieces for the stator mould. Unfortunately this means 8 parts for the mould instead of 4, with the advantage of decreasing the room for error to almost zero! Which do you think is best? Leave your comment below!

  • Weatherproofing completed! :)

    Bram Peirs @ FW2W09/22/2020 at 15:34 0 comments

    No wind and a sunny day here in Belgium, ideally to work 6m height to install the new parts on the windturbine!

    The new bearings that are more waterproof, showcased on the axle that is now only 30mm diameter instead of 35mm which should still be plenty of strength.

    The white part is a raincap to insure that no rain will get on the axle and the bearings:

    The raincap shoves into the big white ring:

    We also made an additional cap to protect the front bearing:

  • Broken bearing! Weatherproofing

    Bram Peirs @ FW2W09/17/2020 at 15:59 0 comments

    Some water got in and has been centrifuged from the rear all the way to the front damaging the front bearing:

    Therefore we made some extra parts that will make the backside more weatherproof. The white ring is glued on the backside of the generator, and the blue cap is glued on the axle. When installed, the white ring will be over the blue cap, ensuring no rain get's in along the backside.

    We also replaced the previous bearings (that were completely open) by sealed ones of half the price!

  • Test run!

    Bram Peirs @ FW2W08/30/2020 at 10:30 0 comments

    Some footage of our testrun! The only thing we could not test here are the gyroscopic forces, that are induced when the windturbine is rotated around it's own axis (through the use of a tail).

  • Construction of the alternator

    Bram Peirs @ FW2W08/27/2020 at 15:21 0 comments

    The 9 coils (1,6mm wire) are cast into a disk that will be stationary. Around that disk, there will be 2 rotating parts with the magnets. More on that later! Firstly, a sneek peek of the coils being cast into polyester resin. A 500ml can with red-colored hardner was the cheapest at Amazon, only 12,56EUR!

    Which becomes this after 24h of hardening:

  • More on bladedesign...

    Bram Peirs @ FW2W08/25/2020 at 16:29 0 comments

    We opted for fixed blades, meaning the blades are attached to the axle. Disadvantages of adjustable blades, like being prone to damage and increased material and maintenance costs do not add up to the advantage of having the blade angle perfect everytime. Also, a near-perfect blade can be made by twisting the blade along its length. Because of this, fixed blades have become the mainstream in the wind DIY-community.

    To be sure we have the correct twist and taper, we used Hugh Piggot's Blade Design spreadsheet. Furthermore we choose the NACA 4415 airfoil for its good suitability for small windturbines.

    We modeled the entire blade in Fusion360, with some small tweaks in 3DS Max:

  • Will these blades hold?? - Worlds first 3D-printed windturbine of this size!

    Bram Peirs @ FW2W08/24/2020 at 18:05 0 comments

    We came a long way. Developing blades of 1m length that can be printed on any household printer required quite some research. Firstly, we chose PLA because it's stiffness, printability and being environmentally friendly.

    Also, after some tests and research, we found a good way to glue the parts together. We need to say there is a lot of misinformation on the internet about this. Many people recommend MEK to join the parts, but MEK only makes the PLA soft. It doesn't really dissolve in it. DCM (DiChloreMethane) does this excellent, and you can make your own gap filling glue with it. This means you can chemically weld the parts together and they will be very strong!

    The blades ended up very strong indeed! We did several tests by putting weight on the blades and they always broke next to the joint!

View all 10 project logs

  • 1
    Print the STL-files

    We used 3 perimeters of 0,7mm, but more perimeters equalling 2mm wall thickness should also work. Make sure you are accustomed printing larges pieces as these warp and loosen from the bed easily. A heated chamber made from cardboard helped us raising the temperature to 30-35°C.

    • Speed: 60mm/s
    • Rafts: Doesn't Matter
    • Supports: Yes
    • Resolution: 0.15mm
    • Infill: 8%
    • Filament material: PLA
    • Filament_brand: Anything cheap
  • 2
    Make some glue

    Dissolve ~150g scrap PLA in 300ml DCM to create a gap filling glue that will chemically weld the parts together. Let it sit for 24h to dissolve properly.

  • 3
    Weld all blade parts together

    Stack the 5 pieces that make up the blade on top of each other and add DCM-glue in between. Firmly pressing both parts together after you applied the glue. Be careful that they line up properly!

View all 21 instructions

Enjoy this project?



Serdar Ozdemirkan wrote 04/19/2021 at 08:02 point

And also wondered why it doesn't have furling mechanism? What happens in high winds?

  Are you sure? yes | no

Serdar Ozdemirkan wrote 04/17/2021 at 09:55 point

Thanks very much for sharing this awesome project and production tips!

Is that steel wires behind magnets for to improve magnetic fields?

  Are you sure? yes | no

L29Ah wrote 11/14/2020 at 02:03 point

omg why do ppl in 2020 3d print visibly-rasterized curves?

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 11/14/2020 at 17:01 point

Because non-planar printing is not that easy with a household 3D-printer. But we listen to your alternatives as you seem to be an expert.

  Are you sure? yes | no

L29Ah wrote 11/15/2020 at 01:16 point

It's not the inevitable (w/o nonplanar slicing, that is) "rasterization" that is a result of FDM-oriented slicing; it's your CAD producing chunky curved surfaces of the model, visibly made of big polygons.

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 11/15/2020 at 08:48 point

@L29Ah The model was kept low-poly because it's easier (for me) to split it into printable chunks of 20x20x20. For the Propel-E 50 we chose for a higher subdivision count but it was a lot more work splitting the parts cleanly:

Do you know of any software that can slice automatically in the required parts? I assume it exists.

  Are you sure? yes | no

L29Ah wrote 11/15/2020 at 14:29 point


The link doesn't mention splitting at all.

I don't really see a problem with simple splitting of high poly parts. slic3r can do it easily in a WYSIWYG style, openscad does that as well and can be automated, and even do something more complex like dovetail joints, as it's a programming language of its own (though it can't provide metadata like bounding box from STL by itself).

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 11/15/2020 at 15:41 point

@L29Ah I didn't say it was mentioned, just that parts of it have been splitted. (Namely, the blades). Do you have good experience with splitting in slicing software?  Our experience is that it sometimes results in artifacts in the 3D-model as the software is not smart enough. Probably OpenSCAD will do the trick.

But again, learning a new software is a lot of work, for the minimal revenue of high-poly models. Yet we are working towards this as we're raising the bar for every prototype :) Other arguments are that high-poly models put a higher strain on the software/computer and make it harder to fix non-manifold errors or other artifacts.

  Are you sure? yes | no

willyyonkers wrote 11/11/2020 at 21:22 point

What's the estimated cost for all parts & materials? I'm in the North East US & it does not look like an affordable alternative to a production unit.

  Are you sure? yes | no

DC3 wrote 09/22/2020 at 17:24 point

Why does this have a smaller carbon footprint? Smaller than what? Almost any other manufacturing method will have a smaller carbon footprint than 3d printing with a heated element. Last time I checked they used 100 times for energy than other manufacturing methods. 

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 09/23/2020 at 10:40 point

We are planning to make a complete LCA in the future, calculating the net carbon emissions from source to EOL. Older windturbines use more steel, which melts at 1500°C, compared to plastics that can be melted at 200°C with a heater cartrdidge of 40W. Also, transport can be minimised because corn starch can be produced on much more places than steel, that's dependent on mines.

  Are you sure? yes | no

mikoja8069 wrote 08/31/2020 at 01:37 point

Epic stuff

  Are you sure? yes | no

EK wrote 08/28/2020 at 04:46 point

Do the blades need infill, or could they just be printed as walls on vase mode without infill? Any thoughts on that strength-wise? (Awesome project btw!)

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 08/29/2020 at 16:56 point

Thanks for the kind words! The forces will mostly be on the outside of the blades indeed, but to prevent warping/twisting along the length of the blade, we choose for 8% infill. Also we'd recommend a minimum of 2 perimeters because if there is one layer that lets go a tiny bit and water comes in, you're in trouble.

The biggest improvement would probably come from changing the print direction (print lines parallel with the length of the blade), but this has the disadvantage of having to use support and room for error cause of the need for post-processing. Another improvement we are thinking of is 'true' 3D-printing, or non-planar slicing:

  Are you sure? yes | no

Jules Anderson wrote 07/14/2022 at 19:42 point

I've ( successfully, i think ) printed the blades on a belt printer. 2 pieces, sliced lengthways and then bonded with the solvent glue.  Time will tell :)

  Are you sure? yes | no

Jacob David C Cunningham wrote 08/25/2020 at 17:44 point

There's no internal spar connecting the blade sections?

Also thanks for linking that spreadsheet/the NACA airfoil, interesting part. Did you do any simulation/modeling like fluid/expected rpms/power output, etc?

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 08/26/2020 at 11:26 point

Hi Jacob, the infill generated by the slicer serves as a spar. This picture gives a clearer view: We used 8% cubic infill in Cura, which is a pattern that is strong in all directions. It can best be described as triangles getting smaller and bigger, making octahedrons along the way.

  Are you sure? yes | no

Jacob David C Cunningham wrote 08/26/2020 at 16:10 point

Sorry I guess to be clear, it seemed concerning to me that you just glue the edges together... wasn't sure if they interlocked or something so they don't separate at high rpms... but I don't know, you guys probably tested/considered that.

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 08/26/2020 at 16:55 point

The parts are welded together (chemically), not glued. With welding, the weld is as strong or stronger as the part, and more or less the same material. Glueing is applying a different material that adheres the parts. We have done tests in storms of 15m/s, and the blades held out perfectly. So strictly seen, interlocking is not necessary for strength, and it makes that the blade has to be split up into 6 parts instead of 5 to print with a regular printer. It would be handy to make the welding more dummyproof, so that the parts are welded into the correct positions. We will implement this feature in future builds.

  Are you sure? yes | no

Jacob David C Cunningham wrote 08/28/2020 at 05:02 point

Well that's great, good luck with it, it looks nice the end build

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 08/26/2020 at 17:00 point

We didn't do simulations, but the Scoraigwind spreadsheet calculates the blade angle and chord width according to a given tip speed ratio (TSR) and angle of attack, which is a well known and proven method. The power output is verified with real world tests. We will update the details with the power curve in a minute.

  Are you sure? yes | no

Dan Maloney wrote 08/24/2020 at 20:46 point

My wife spotted a small wind turbine on our neighbor's property the other day. "How much is something like that?" she asked. I answered "A lot," but now maybe the answer is the price of a decent 3D-printer.

Nice design, looking forward to more information. Good job!

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 08/25/2020 at 09:30 point

Thanks Dan! If you can print at a Fablab in your neighbourhood, you might even save up more. And PLA is only 6$/kg overseas:

  Are you sure? yes | no

Dan Maloney wrote 08/25/2020 at 16:39 point

Well, sure, but then I wouldn't have an excuse for a new printer. You can see the bind that would put me in ;-)

  Are you sure? yes | no

rlarge wrote 08/24/2020 at 18:49 point

You should print one in black, there are some interesting study's that indicate that it significantly reduces bird impacts.

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 08/25/2020 at 09:33 point

Black, I always heard about purple being the best. Can you share the studies? We are planning to make special versions of each windturbine (for example recycled carbon reinforced), and black or purple can certainly be one of them!

  Are you sure? yes | no

Tom Nardi wrote 08/24/2020 at 02:03 point

Very interested to see more information on this project. What was this printed on,and how many parts total?

  Are you sure? yes | no

Bram Peirs @ FW2W wrote 08/24/2020 at 16:40 point

Hi Tom, all our windturbines are designed to be printable with any household 3D-printer. We used an Anycubic i3 Mega S as it is cheap, but still a good printer. There are 27 parts in total (17 uniques + 2 * 5 for the wings). We'll post more very soon!

  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