Parametric Multi-Blade Propellor Generator

by BouncyMonkey Mar 21, 2019
Download All Files

Thing Apps Enabled

Please Login to Comment

Fist of all this generator is pure genius! Is there a way to change the turning direction from ccw to cw? I tried entering a negativ pitch it kinda woked on the blades but totaly messed up the props hub. Am I failing to finde the option to set up the direction or does that depend on the airfoil?

What about mirror function on your printer software ?

Hi bmn666
Aw shucks thanks! You know I never even thought of multirotor pilots needing CW props. Whoops.
In the current script there is nothing to do that. However I will try to find some time to add a couple of extra lines to add that functionality. All it requires is the addition of a mirror function to the script at the very end of the shape generation process.

The alternative (and easier solution) is that you can probably create the mirror image in your slicing program. I use Cura, which has an option for mirroring objects about whichever axis you like. I would expect other slicers will have the same ability.

Give that a try in the first instance.

Tanks a lot! You clearly know what you are doing since your idea works like a charm. Since I am using Cura as well I am totaly happy with that solution. I tried it with a 2 blade and a 3 blade and as far as i can tell from seeing the model in Cura mirroring them works just fine.
Thanks again for such awsome work, the ultrafast reply and for sharing your work!

I've been trying to print this but it ends up extremely thin at the base of the hub/prop. Is there a way to make that area thicker?

I had the same problem. Then I set the "Pitch Adj Hub" to something like 0.8 instead of 1 to get it stronger.

Hi CDeviant
My original design for this had a scale factor that thickened the airfoil as it got closer to the hub. Unfortunately it created some nasty steps along the length of the blade that I was never able to resolve. Something about the way OpenSCAD handles its linear extrudes when there is a twist added. It was a bit a bit nasty. So I dropped it in favour of the hub shape currently used.

The overall blade shape is a scaling of the airfoil along the length. To get a thicker section near the hub you would need to modify the function that controls the blade shape. Look for this line.

function BldChrdLen(x) = 1.392*pow(x,4) -1.570*pow(x,3)-2.46*pow(x,2)+3.012*x+0.215;

I am in the process of putting together a webpage on the finer points of running this prop-generator which goes into this. I hope to have it up by the end of the day tomorrow. I'll put up a direct link when I have it, but in the meantime here is the general area where the page will live.

Other than that, the only other way is to get in there and do some vertex pulling with a modelling package.
Sorry there is no easy fix for that.

Oh. I suddenly remembered. The other solution is to use a thicker airfoil. I have written a webpage on that. You can find it here: http://techmonkeybusiness.com/airfoil-to-openscad-converter.html

Update: 9 April 2019: The webpage on the finer points of running this script is up now: http://techmonkeybusiness.com/parametric-propellers.html

Thank for for such a detailed response. I'm pretty lost with Openscad, but I'll attempt to wrap my head around that and give it a try sometime.

I did a test with an airboat I'm building and it was incredible! All I changed was the diameter to 6". I was using an off the shelf 3 blade 4.5" prop and not getting very good thrust with a 900KV motor. The 5 bladed prop was probably 3 times or more thrust.

So much thrust it sent the airboat hurdling forward into the ground and the prop exploded, right at the base of the hub.

I printed it with PLA at .32 layer height. I'm pretty stunned at how well it actually worked.

Heh heh! Awesome! Thanks for sharing that.
One of the things I want to test is multiple blades against 2 or 3 blade props. All the information I have seen indicates that the more blades you have the more thrust you get for the same diameter. I'd quite like to get some robust data around this.

More blades of the same shape and size is more blade area and more thrust... and more watts demanded from the motor.

However each blade adds its own tip losses and they start to have the turbulence from one blade interfere with the one behind it more and more as you add blades, so you get lower efficiency.
Less useful thrust (less air moved) per watt of power into the motor.

This occurs at all scales (prop diameters)

The most efficient is a single blade prop, but the load imbalance is very hard on the shaft and bearings even with the correct counterbalance. You end up with the prop shaft bending or breaking and short bearing life.

Experiments go back at least to the mid 1930's. In appx 1939 someone flew a Piper Cub (not a model)with a single blade prop.
Single blade props are still used in some model aircraft competitions for the added efficiency. Usually indoor rubber power,but also some outdoor electric powered and glow fuel powered model competitions.

the main reason for many blades on a full scale aircraft prop is to put more blade area in a smaller diameter, reducing the length of landing gear strut needed.

A ring shroud around the prop greatly reduces tip losses and this is why you see the shrouded fan systems on modern blimps. You get the multiblade prop with reduced inefficiency compared to the far larger 2 bladed prop that would have been needed for the same thrust.

Any experiments in use of more blades will need better data than just swapping for more blades and seeing the increase in thrust.

To keep the same load (as close as reasonably possible) on the motor at 10,000 rpm, starting from 10 inch diameter, changing from 2 blades to 3 blades, you trade the 1 added blade for reducing diameter by 1 inch. scaling the blades in all dimensions for the different diameter, keeping the same pitch.
the tradeoff to keep the same load on the motor as the same rpm and diameter changes.

Lower rpm props are also more efficient. Endurance competition indoor free flight models have extremely low rpm propellers and the record for rubber band indoor duration was over 45 minutes last time I looked it up.
They use helium balloons, occasionally snagging a wingtip on the string to steer the models.

Thanks for the tips (pardon the pun)
I was aware that the spitfire increased the number of blades to avoid rotary-hoeing the runway.
The prop testing I am proposing to do will be carried out in a rigorous fashion with as much good measurement as I can manage. The missing bit is a wind-tunnel, but I think my partner might object to me commandeering the lounge to set one up.

Thank you so much for including a visual representation of the customizer options.

I'm going to give this a try with my FDM printer and see how it works. Goggles will be worn, lol

I have carried out some flight tests using a 6x4.5 prop made with this parametric model. The performance was equal to the commercial prop and there may have been slight efficiency benefits, but this will only become clear when I carry out some proper bench tests.
A real advantage with this prop over the commercial one was that being nylon it was a little more flexible and resistant to damage from tip strike during landing. Tip strike with this particular plane has been responsible for many broken props and motors being ripped off and was the reason for embarking on this project. I still need to try the folding prop - https://www.thingiverse.com/thing:3506762.

I had been worried that the greater flexibility may allow the prop blades to twist under load which would cause it to feather and loose power. This didn't happen.

Parametric Folding Propellor Generator

We used to use a relatively flexible white nylon propeller for model aircraft.

Until recently you could get a modern version that used a better formula for the nylon, which does not tend to get brittle from age as fast as the ones from the 1970's did.

Flying control line models, we could watch the prop alter shape based on how the maneuvers being performed added drag to the model.

the blades tended to flattened pitch and cone forward under higher drag loads. This is essentially an automatically variable pitch prop, keeping load on the engine fairly constant.

The white nylon props had a nasty habit of shattering under the loads, because nylon gets brittle with age. There was no warning and no real way to test and ensure the prop was not becoming dangerous to use.

Putting the nylon props in boiling water for 15 minutes seemed to reduce brittleness. But it only extended the time before the prop would shatter.

We would also dye the prop as we boiled them. Rit brand clothing dye worked well.

Thanks FHHuber
I am lucky to have SLS printed nylon as an option and will be interested to see how I go with PLA and ABS filaments on my Wanhao. As you say the layer height is a diminishing return game, and I'm not all that patient when it comes to printing times. I'm just curious to see how they go.
With regards to dying nylon, yes, I was surprised to learn that lesson a little while ago with a project to print 1000 red ANZAC poppies for a Reserve Bank commemorative coin issue. We used red clothing dye on the SLS printed white nylon. They came out brilliantly with a very even colour.

Do you print this with supports?

Hi Arpruss
To be honest, I printed mine on an industrial Laser Sinter printer using nylon and so it didn't need support. But when I come to try doing them on my filament printer I expect I will need support. Because of the thinness of the trailing edge I would also print it with as fine a layer height as I can manage (which is 0.1mm for my printer)
There is a lot of testing that I intend to do with these props and I suspect a thicker airfoil may be necessary to help printability on FDM type printers.
I would really appreciate comments back on how these print.

Ender 3 will do 0.04 mm layer. Less than 0.08 doesn't gain you anything other than taking longer to print. because of "magic numbers" related t the stepper motors and their movement accuracy, its best n the Ender 3 to use layer thickness in increments evenly divisible by 0.04. 0.12 prints faster and more accurately than 0.10.

You may want to research the stepping values for your Z axis

Could you also add variables for the prop mount and add the posibility to add more holes/ variable hole size so we can make 2" props ? like these https://cdn.shopify.com/s/files/1/1285/4651/products/rush-1_1800x1800.jpg?v=1552609166

Apparently I haven't been monitoring developments in prop mounting methods. I was not even aware that this sort existed. They're kinda cute. 2" props - wow!
By fiddling with the PropShftD, PropHubD, PropHubT, and PropHubDCutterMax parameters you should be able to get the hub size you are after, but that hub>blade transition may get a bit ugly for such a small size. Now that I know about the existence of motor-shaft arrangements with extra pins, I'll look at including something like that in the next edition - but don't hold out for it at the moment.

No rush I'll wait till your latest version

Hi Enriqueeeee
I have just posted up the latest version with the added Hub pin holes. In the default version they are disabled, but when you put values into the HubPinD (diameter of the pins) and HubPinPCD (the pitch circle diameter of the pins), this option will be activated. I have included an image demonstrating its use.