I'm wondering if it would be possible to build ducted fan "rocket".

My idea is to construct the duct from 18650 batteries, and to have a series of contra rotating fans along its length.

I can 3d print new blades to have a steeper angles as the air-flow/speed increases.

This is the fan I'm looking at.

A 90mm fan will let you make a ring of 16 batteries around the curcumference and it would be 6 batteries tall as it's a 6s motor.

Is this feasable/will it work? what kind of thrust would I get?

Here is a quick render I made:

enter image description here

  • $\begingroup$ Welcome to av.se. Not sure that this question is really on-topic here; may be better suited to a different stack exchange forum. Looks like an interesting project, though. $\endgroup$
    – Ralph J
    Nov 26, 2017 at 17:34
  • 1
    $\begingroup$ By stacking the fans like that, you're going to get way less thrust than what you are expecting. $\endgroup$
    – Federico
    Nov 26, 2017 at 17:37
  • 3
    $\begingroup$ That would probably be one of those things you have to file under the category of “well you could - but why would you want to?” $\endgroup$ Nov 26, 2017 at 18:06
  • $\begingroup$ I want a platform on which to build/test a rocket flight computer & control system, launch - short ballistic flight - landing. I want to mimic what you get with real rockets - have the bottom fan gimble (5~10 degrees) and stablize it with thrusters near the top (divert air-flow from the top fan). I realize its not gonna have 4x the thrust, I'm just wondering how much I'll get with this design assuming the blades are optimised etc. $\endgroup$ Nov 26, 2017 at 19:08
  • 1
    $\begingroup$ With enough engineering you could probably get some fewer stage variation of this idea to fly to a degree, but it wouldn't really mimic the challenges of controlling the boost phase of a rocket. $\endgroup$ Nov 27, 2017 at 4:37

1 Answer 1


From the link you provided:


  • Rotor Diameter: 90mm
  • Fan blades:11 blades
  • Weight: about 350g
  • Working Voltage:6s(22.2V) lipo battery
  • Motor:Brushless Motor 3553 1450kv
  • No Load Current: 4.1 A
  • Load Current: 83A
  • No Load Speed:32190 rpm
  • Load Speed:16095 rpm
  • Thrust: 3300g
  • G/A:45.16

Assuming that 16 * 6/4 = 24 cells of type 18650 would be able to deliver full electrical power for the fan, the issue would indeed be the local angle of attack of the fan blades.

Thrust at full power is listed at 3.3 kg = 32 N. This would be at standstill/hovering conditions at sea level, since measuring at that level provides the highest thrust level for advertisements. Diameter is 0.09m. Net thrust T =

$$ T = \dot{m} \cdot (V_{out} - V_{in}) = \dot{m} \cdot \Delta V \tag{1}$$

$$ \dot{m} = \rho \cdot A \cdot V \tag{2} $$

Combining (1) and (2) for the hover, with $V_{in}$ = 0:

$$ V = \sqrt{\frac{T}{\rho \cdot A}} = \sqrt{\frac{32}{1.225 \cdot \pi/4\cdot0.09^2}} = ~\text{64 m/s}$$

Impulse thrust considerations usually draw a contracting propeller wake for inducted propellers. Ducted fans work a bit differently and we can take the average velocity behind the fan for further Order Of Magnituding. Below figure is from this research paper, and shows the considerations for ducted fan flow; it contains some methods for more detailed computations.

enter image description here

Rotational speed under load is 16095 rpm = 1,684 rad/s, tip speed = 0.045 * 1,684 = 75.8 m/s. A velocity triangle at the blade tip has as angle $ tan^{-1} (64/75.8) = 40 $ deg. The blade needs to be inclined further than that, usually about 6 deg, so tip blade angle of the standard fan would be 46 deg. Purchasing the actual fan for verifying the above would be a good thing!

For the second fan, this same method can be followed: mass stream will remain the same if the hull is closed, in order for the 2nd fan to deliver the same thrust $\Delta V$ = 64 => ${V_{}out} = 64 + 64 = 128 $ m/s. Tip angle velocity triangle = $ tan^{-1} (128/75.8) = 59.4 $ deg, fan blade angle = 59.4 + 6 = 66 deg, etc.

Note that the above is valid for the hover. As soon as the "rocket" picks up speed, the local angle of attack of the fan blades will reduce and therefore thrust will reduce. So one would have to optimise time of thrust (time of amps delivered) with weight, momentary speed, and expected end speed, then average the blade angles out for the speed function.

Note that opening the hull in between the fans allows for extra air to be drawn in, increasing the mass flow. The paper cited above has results for a setup like that as well - if the increased mass flow relates in lower entry velocity, this might be worth considering.

  • $\begingroup$ wow, thanks for the great answer! I'm pleased to hear this is feasable. I may scale back to just 2 fans & 48 cells initially, but leave space for the full design. I don't really understand all the technical stuff, I was just gonna try out a range of different blade angles but I'll be sure to make some around 60 deg. If the thrust is good enough, I could even optimise them closer to forward flight. $\endgroup$ Nov 27, 2017 at 19:06
  • $\begingroup$ Blade pitch: the first fan is whatever the manufacturer has set it to to get 32 N of thrust, the second one needs a higher pitch, the 3rd one even higher, and the 4th one still higher. Best to verify my computation of 45 deg of the first blade pitch, then the above method for each subsequent fan blade pitch can be followed. $\endgroup$
    – Koyovis
    Nov 28, 2017 at 2:43
  • $\begingroup$ And then: establish total thrust at lift-off. Determine the current draw, and how long the cells can sustain this (function of amps over time gives a function of thrust over time). Compute the resulting acceleration and velocity as a function of time, and adjust the blade pitches such that the highest end velocity is reached at the end of cell life. Best done with a numerical solution, a real-time simulation of all the physics. Mathworks might be your friend here. I'm sadly lacking the time to do it at the moment. $\endgroup$
    – Koyovis
    Nov 28, 2017 at 2:50

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .