Wind Fisher

Some links and informations:

From : (image below)

In french language:

Scientific publications:

Control of a Magnus Effect-Based Airborne Wind Energy System

Ahmad Hably (1) , Jonathan Dumon (2) , Garrett Smith (3) , Pascal Bellemain (2)

This publication is available on

For what I understood the subject of the above publication includes the vertical trajectory (Fig. 1.24).

But it seems that Wind Fisher develops sideways crosswind trajectories (eight-figure, Fig. 6) as explained on

Hello @WindFisher , please could you confirm or rather put your insight on these elements? Thank you.


1 Like

Here’s our AWEC2021 poster which shows horizontal trajectory.


Hi Garrett,
I took a closer look at this horizontal crosswind trajectory, and ended up seeing an interesting detail:
Magnus effect kites avoid turning by only reversing the direction of rotation after each crosswind trajectory.

This feature could facilitate scaling by stacking superimposed units or by scaling the units themselves. What do you think?

Does anyone else have the vague impression that enthusiasm over the magnus effect is mostly about the unexpected nature of a different way to achieve aerodynamic lift - the novelty? Almost as though the novelty itself is interpreted as a compelling advantage? That just because it is unusual or unexpected, that makes it more effective? Well, I guess you never know until someone tries it. But could it be that the 100-year-old magnus effect is unusual precisely because nobody uses it for anything, because it is more of a pain-in-the-ass than it’s worth?

Magnus effect is worth it if you need to save space (for any reason) for the same force (it is why it has some success onboard cargo ships with limited deck space see for example already discussed here, but will probably be outperformed by suction wings). For a helium-filled balloon, I think you might expect to save weight as well.


I think the novelty attracts a few minds. The notion of airborne magnus effect wings seem initially a bit farfetched to me, due to the mechanical and power consumption requirements. This of course has mass consequences. Still not too farfetched that its not worth considering at all though.

I agree with @batlabat that suction wings seems a better option, and maybe even better just a plain wing.

For the ship case magnus effect sails have the other advantage that they dont need to be oriented to the airflow. I think this aspect should not be underestimated, because you entirely remove advanced control from the equation. Thats maybe why we are seeing magnus sails on ships and still not any «airplane» style wings as sails

This is an essential point. The fact that the Magnus kite doesn’t make any U-turn after each horizontal trajectory helps to maximize the use of space. You can see it in the video: the direction of rotation is reversed, it is all, and the Magnus kite starts again in the opposite direction. If you end up with very long or stacked Magnus kites for very large AWES, this can be a decisive advantage.

I agree. I tend to want to give anything a fair chance. Kite-reeling in general looked so promising according to the published numbers, but seemed to stop all progress after that. Why? Just as I would never throw away kite-reeling as “proven” ineffective, i would say it is difficult to “prove a negative” and magnus-reeling could turn out to be an answer, if someone can make it work. Hard to say, really.

The magnus/wind boat/ship sailing thing - yes, good point about no need to trim the sails. Still, it’s been tried in the past and never caught on. Kind of reminds me of vertical-axis turbines. Long proven as less effective, more expensive, and less reliable than standard turbines, they nonetheless enjoy a steady parade of new advocates who then fail to get the hoped-for results out of them.

Well for Magnus Sails I think Norsepower is still alive and kicking and hopefully has progress enough to go more mainstream. Time will tell

Foe AWE, worldwide interest is still increasing according to google trends

I tried to understand how it was possible to ensure the rotation of the cylinder, the variation in the length of the cables according to the power and retrieval phases, as well as the generator function. To do this, I used the photo, the poster, the video and the patent FR3123317A1.

From the poster:

The system includes the following elements :
• The Magnus effect kite.
• The two winches of the ground station.
• The two motors of the ground station.
• The two gear reducers placed between the winches and the motors.
• The cable, wrapped around the kite and the two winches.

The solution found amazed me for its economy of means. For each of the two cables, the two winches of the ground station make it possible to turn the cylinder in one direction (winding on one winch and unwinding on the other) and then the opposite direction (the reverse), and in the same time to vary the length of the cables (use of wind power and lift during the generation phase (motors as generators) to unwind the winches, a side faster than the other to keep the rotation; then use of the motors during the retrieval phase to wind the winches ?).The two motors of the ground station drive their respective winches via gears. As well as ensuring rotation by alternating the direction of rotation, and the power and retrieval phases, they also work as generators. Please @WindFisher, correct me if I am wrong or not precise enough.

Garrett’s explains are clearer on the following video, at 17: 56:

1 Like

I like this method, it is a Kiwee winder type system combined with a magnus system. The method is novel but perhaps difficult on a small scale… If Im not wrong it needs to rotate to fly, this means the transmission/steering lines need to rotate. If this is the case, how does the reeling method work when you want to pull the kite in or let it out?

Kiwee uses a continuous rope drive transmission, as for Magnus balloon with motorized belts (but where the motors are aloft, allowing the single tether to make the reeling-out/in alternate power and recovery phases).

Wind Fisher rather uses a transmission with two winches settled in the ground station. As a result the amount of rotation in the same direction is limited by the length of the cable, that is not a problem since in reeling operation, the winches are alternately winding and unwinding to supply reel-out and reel-in phases.

I asked myself the same question, and couldn’t find an answer.
Let us take an example during reel-out power phase (the kite pulls on the winches as it moves away): wind speed 10 m/s, reel-out speed 5 m/s, spin ratio (tangential speed / wind speed) only 1; that means that the two winches are unwound at 5 m/s while an additional unwinding of 10 m/s of one of the two winches occurs, leading to an unwinding of 15 m/s; but in the same time the other winch should be winding the same cable at 10 m/s while it is unwound at 5 m/s. It doesn’t seem possible, or have I forgotten something?

1 Like

Hi Pierre, We may look at stacking at a future date. Right now, two cylinders is what we need to be successful with.

There are several reasons we have gone the Magnus effect route and it’s not about the novelty of this 19th century technology :wink:
The key reasons are: 1) the wing at scale has enough volume to be lighter-than-air which enhances safety during control failures as the wing won’t fall, 2) Magnus effect wings do not need pitch control thereby simplifying the airborne architecture, 3) gust response is more benign than profiled airfoils with much lower force extremes resulting from gust induced increases in angle of attack or apparent wind speed, 4) cylindrical wings are easier to design and build than streamlined and twisted airfoil wings.

1 Like

You may enjoy this video as well: Wind Fisher mobile wind tunnel - two camera angles on Vimeo

1 Like

Hi Garrett: I have a question which is not about the video where perhaps a rope drive transmission is used instead of the transmission with the two winches which act the respective motors-generators via gears. But I am not sure.

By what I understand (rightly or wrongly) on your animation, poster and video, the rotation of the cylinder, alternately in the two directions, is ensured by a cable wrapping the cylinder, then ending to the two winches.

When there is no variation of length of the cable (so no pumping mode use) the principle is understandable: one winch unwinds while the other winds up until the rope runs out, then vice versa to rotate in the other direction. But for pumping mode I don’t see how it is possible to ensure both rotation and variation of cable length in the same time. If one side winds up (for rotation) faster, the other side has to unwind as fast for balancing. As a result that prevents any rotation.

Let’s take an example during the power phase of reel-out (the wing pulls on the winches as it moves away): wind speed 10 m/s, reel-out speed 5 m/s, spin ratio (tangential speed/wind speed) only 1 for ease of calculation, leading to 10 m/s; this means that both winches are unwound at 5 m/s by the pull of the kite which drives the winches and the generators, while an additional unwinding of 10 m/s of one of the two winches occurs to ensure rotation to maintain the lift force, leading to an unwinding of 15 m/s; but at the same time, the other winch should be winding the same cable at 10 m/s while it is unwound at 5 m/s as specified above. This doesn’t seem possible, or have I missed something? Thanks.

1 Like

If I understand the concept correctly, I don’t even know if they can be called winches? Seems that the line is wrapped around a pulley wheel, and the energy generation works by means of regen braking… In this case you are right that the lines must be static length (obviously because the transmission is a loop.)

My observation is, how can one reel in two lines wrapped around a flywheel style pulley? How do you “reel” a loop, especially if that loop needs to continue to rotate in order for the kite to stay afloat. Is there a clutch on rotor aloft? I don’t know if it can be referred to as a winch if it can’t pull in the line.

1 Like

On the last video Wind Fisher mobile wind tunnel - two camera angles on Vimeo, it looks like an endless rope drive (Kiwee style) is working. And the Magnus kite flies by figure-eight, making half turn at each side of the trajectory.

In contrast, on Wind Fisher Energy Balloon Animation on Vimeo it looks like two winches (per cable) are implemented, allowing the cable wrapped around the cylinder to rotate it in one direction then the opposite direction. And the Magnus kite flies also by figure-eight, but without making half turn: the direction of the rotation is reversed. These animation matches other material such as the information on the poster and the patent.

In both methods (rope drive or winches), I don’t know how rotation (to produce lift) and reel-out power phase (by extending the length of the cable) can be solved in the same time. That said by using two winches, and during reel-in retrieval phase, it would be possible to shorten the cable while stopping rotation.

So I can’t wait to hear Garrett’s explanation.

They are definitely winches. We are not using a rope drive with pulley but a double winch drive on both ends of a single cable. A rope belt does not give enough reel-in / reel-out distance to be power productive. The prototype is currently using a pulley and belt since we are only using it for aerodynamic and control studies at a fixed distance and not for power production.

Here’s the architecture from the patent to help you understand the system design for power production.


There are two simple equations which govern the reel-out speed (r_dot) and the cylinder rotation (omega) based upon the reeling speeds of winch line 1 and 2 (V1 & V2) and the cylinder radius (r_cylinder).

  1. (V1+V2)/2 = r_dot
  2. (V1-V2)/r_cylinder = omega

So yes, reeling and rotation are possible at the same time depending on speed and direction of the pair of winches. And I can confirm that the energy required to rotate the cylinder is significantly less than the energy produced during the generation phase (as hinted at in the poster).