crazy experiments: motors, frames and propellers


  I got a crazylfie micro quadcopter for my birthday and love playing around with this remarkable open development kit. Yes, it’s not just a very small but powerful quadcopter, it is also comes with all the design files and the tools necessary to modify the software and/or hardware, which in it’s current state is already quite mature. Having said this, if you look for a ready to go quad copter with all cutting edge features – look elsewhere. While the hardware with it’s 10dof sensors, fast cpu and expandability is – considering it’s size – quite complete, there are still a lot things to be developed hard- but foremost software wise. But this is not a downside, because the product crazyflie is not exactly a quad copter, it is a quad copter construction kit with every bells and whistles whith a blog, a wiki, a forum with a whole quad copter construction community, inviting everyone to participate. And for that I’m just loving it.

In this spirit I startet some endeavors to make the crazyflie more the quadcopter I want. One thing I wantet was the crazyflie to be more powerful so that it could carry more payload and fly more stable outsides. Here are some results of my experiments with different motor sizes, frames and propellers on the crazyflie.

You can download all files emergerd while experimenting from this Git-Repository:

This is the almost unmodified cf. The battery is bigger than the stock one (400mAh vs. 170mAh). This appears to be little too much, as the extra weight takes too much of the power the cf needs to stabilize it’s flight.


Propellers                                                                                                                                                                           I wanted to know if I can use 3d-printed propellers, so I built this little test rig to measure the thrust, rpm and volts of a given propeller at 800mAh. I used an infrared distance sensor and my dsoquad oscilloscope to measure the rpm.

The stock propeller/motor combination gave me about 10g of thrust at 20250 rpm and 3.7 volts.DSC08598

This is the best performing propeller I was able to make. It can lift 7g @16560 rpm and 3.3v.

DSC08618_01  DSC08628_01

I ordered four Hubsan A23 8.5*20 mm motors (the stock ones are 6*15mm) . You can get 4 of them on ebay for about 20€.

stock motorHubsan A23 motor

To attach them to the crazyflie, I had to make some motorholders with my 3d-printer. Because the A23 motors are much more powerful (30g lift@1.6 A) and heavy (1.7 vs 4.8 g) than the original ones, I was worried the little arms of the Crazyflie could brake. So I made this frame on which the Crazyflie can be attached.


The new motors turned the Crazyflie into an uncontrollable rocked. there were several things to sort out before I ended up with a pleasing control over it. The first problem was that even the 400mAh rc-lipo wasn’t able to deliver sufficient current for the motors. I ended up using a turnigy 750mah lipo (18g) which seems to be adequate.


The next problem was the pwm noise introduced  by the high motor current, which reached about 1.2V peak to peak on my oscilloscope. This noise interfered with the rc transmission so the cf lost connection as soon as throttle was pushed. To fix this, I used 3 470u(10V) capacitors, one of them connected to the battery terminals and the other two connected to vcom and ground.


Here is a short video of testing a new frame which protects the propellers too:

6 thoughts on “crazy experiments: motors, frames and propellers”

  1. I was curious if you tested what your new possible payload weight was with your setup. And was there a particular reason you decided to go with the 470u capacitor?

    1. Hi, I didn’t measure the new payload, but the numbers say:
      One motor makes about 30g, so we have a total lift of 120g.
      The rest of the crazyflie weights about 40g, leaving 80g for thrust.
      So I think that 60g of payload seems realistic…

  2. I was wondering if you experimented with distance calculated flight stabilization. I was looking to use a helium balloon navigated with a quad copter within a room. Any help would be appreciated.

    1. Not shure what exactly you mean by “distance calculated flight stabilization”. But actually I’ve experimented with helium balloon too – it works:) the biggest problem for me was to find a baloon which doesn’t loose the helium over time…

  3. Hi Frank

    First of all, GREAT MOD! Thanks! I did the mod on a CF 2.0 and it’s working fine.
    The capacitors, used in de mod, are little big on the CF. Can you please explain why you decided to go with the 470u capacitor? Is it possible to us smaller capacitors?

    Keep up the good work!

    1. Hi Sander
      thanks! I’ve chosen this capacitors due to experimenting and because they were lying around. Those caps are rated by 10 volts so you would get smaller caps with a lower voltage and the same capacity. I suggest you to try an see for yourself how big the caps have to be in order to keep the interference from the motors low enough.

Leave a Reply

Your email address will not be published. Required fields are marked *