UAV Coaxial Copter Uses Unique Drive Mechanism

Personal UAV’s are becoming ubiquitous these days, but there is still much room for improvement. Researchers at [Modlab] understand this, and they’ve come up with a very unique method of controlling pitch, yaw, and roll for a coaxial ‘copter using only the two drive motors.

In order to control all of these variables with only two motors, you generally need a mechanism that adjusts the pitch of the propeller blades. Usually this is done by mounting a couple of tiny servos to the ‘copter. The servos are hooked up to the propellers with mechanical linkages so the pitch of the propellers can be adjusted on the fly. This works fine but it’s costly, complicated, and adds weight to the vehicle.

[Modlab’s] system does away with the linkages and extra servos. They are able to control the pitch of their propellers using just the two drive motors. The propellers are connected to the motors using a custom 3D printed rotor hub. This hub is specifically designed to couple blade lead-and-lag oscillations to a change in blade pitch. Rather than drive the motors with a constant amount of torque, [Modlab] adds a sinusoidal component in phase with the current speed of the motor. This allows the system to adjust the pitch of the blades multiple times per rotation, even at these high speeds.

Be sure to watch the demonstration video below.

47 thoughts on “UAV Coaxial Copter Uses Unique Drive Mechanism”

1. SparkyGSX says:

I’m not sure I fully understand, but it’s very neat!

If I understand correctly, it uses the mass moment of inertia of the blades with a hinge at an angle to the shaft to change the angle of the blade, and the two blades have hinges at opposite angles (approximately +45 and -45 degrees to the shaft), so that an acceleration causes one blade to increase in pitch, and at the same time, the pitch of the other blade decreases, while a deceleration does the opposite.

I don’t know exactly where this would be on a cleverness scale, but I’d rate it HOLY CRAP THAT’S CLEVER!

1. So, the sinesodal variation is kept in sync with the spinning blade to apply a thrust vector which is not completely vertical – thus allowing the airframe move forward/backwards/sideways. Smart!

Given that this is a balancing act between motor strength (spare instaineous torque), coupling flex and propeller momentum; I wonder how tuned this is for a particular airframe and/or payload

1. PointyOintment says:

They need to do way instain mother.

1. Skunch says:

how is babby formed

2. markaeric says:

That’s what I gather, and I agree with your rating. It’s brilliant, and that’s why no one ever thought of it before.

2. vonskippy says:

Very clever.

1. vonskippy says:

Wonder if it scales under load ???

exactly what I was thinking as well. can only work with electric motors too I imagine.

1. Josh says:

Might be able to use electronic clutches with other kinds of motor, instead of speeding up and down just let it slip a bit.

2. AKA the A says:

The oscillations would become harder to manage as you scale up…

3. ka9mva says:

I think this is a significant advance in UAV technology, especially for smaller-scale UAVs.

4. It takes a lot of persistent brain power or some weird orthogonal burst of inspiration to come up with this kind of solution. I love amazingly simple ideas like this.

1. MrX says:

Indeed.. I’d say these are the kind of ideas that actually come up when you are investigating something else!

5. Daniel says:

This is truly awesome.

6. me says:

Does this rely on 2 motors, or could you get away with reducing it to a single motor and a rotating body with a fixed (in relation to the body) rotor?
Very clever system, anyhow.

1. Hyratel says:

the demo vehicle is shown with only the topside rotor with the torquepitch pivots in place, so it’s plausible that this could be applied to a Top-and-Tail design as well

1. isn’t the angle of the rotor blade proportional to the rotor speed? therefore a single rotor would be unstable because different amounts of lift would be generated for pitch and roll, aswell as a stationary hover.
I think the other rotors speed is inversely proportional to the torque pitch one, to maintain uniform lift.
But then again i know fuck all about helicopters.

1. matthijs says:

The angle if the blade is proportional tot the acceleration of the blade, not the speed. So it should be possible tot make a Top and Tail version. Not sure about the rotating body part..

1. jelle says:

not quite, (I think): The angle of each blade is determined by several factors: The initial angle it is set at at rest, the centripetal forces keeping the blade in the straightest point from the centre, the leveraged drag of the blade (determined by alpha and airspeed) pulling it away from that point. (potential for nasty oscillations?)
On top of that come the dynamic inertial effects that their system is designed for.
But the two blades do cancel each other out: one gets more drag, while the other gets less, their hinges are mirrored!

2. I think thats angular acceleration, which is proportional to speed. Thats why the video showed a different rotor pitch at 40hz and 20hz which would be

2. I think at-least two motors are required because otherwise the propeller and the body would rotate in opposite directions to conserve angular momentum. This is exactly the reason why helicopters have a tail rotor. Helicopters with tandem or coaxial rotors rotate in opposite directions.

1. At the same time most helicopters use only one motor to drive both rotors, so the answer is no. It is even sometimes possible to counter the torque from a propeller with aerodynamic surfaces and no moving parts. However usually you also want yaw control, so you do need *something* — for example a servo. But then again this article proves exactly that you can sometimes squeeze control of another degree of freedom (or two) into an actuator you already have.

7. SOI Sentinel says:

That has to be murder on that gyro axis. You’d have a strong periodic oscillation in that axis that isn’t going to help with long term stability. It probably can’t dead reckon itself out of a paper bag. Hence the full room camera setup.

1. Dead reckoning navigation is quite hopeless anyway for anything that flies, especially non-fixed wing. I’ve never seen any working dead reckoning setup based on inertial sensors.

But I don’t think this setup makes it much worse — consumer gyros and accelerometers have bandwidth on the order of 1kHz, the rotor RPS is 1 or 2 orders of magnitude lower. Also consider that other aircraft that was demoed some time ago that was basically a single big propeller blade (electronics inside), spun by a tiny propeller mounted on its leading edge. The whole aircraft was spinning at about 20 RPS and it was able to navigate & record video.

1. Tim says:

Do you have a link for that video? Sounds incredible

2. AKA the A says:

Dead reckoning is out of the question for anything that costs less then a couple of grand, the cheap gyros simply don’t have what it takes to do dead reckoning on any useful scale…

btw if it was bigger (and a whole lot more expensive), it could carry laser gyros which actually need to be shaken about :D

8. andrewjhull says:

“Does this rely on 2 motors, or could you get away with reducing it to a single motor and a rotating body with a fixed (in relation to the body) rotor?” – as shown it looks to rely on two motors, *however* it *may* be possible to simplify the design to use only one motor…. but that will require another bit of lateral thinking….. I’m sure we are all open to suggestions on how to achive this… I suspect some sort of (magnetic?) clutch or cone drive gearbox might provide a soultion… but thay is just a first pass of the old mental compiler… any other ideas are welcome.

1. andrewjhull says:

You might also find this interesting.

2. acidrain says:

I think you’re referring to this: https://www.youtube.com/watch?v=ixLWIFdUj1c I have one of these on my desk at work. It’s interesting, but the lack of directional control is disappointing.

The problem is that it doesn’t really know (or care) what direction it’s facing in, making it difficult to determine when to pulse the motors to cause the blades to tilt.

9. Leithoa says:

What sort of impacts does this have on the airframe? You can see in the spars some pretty significant vibrations that could become problematic as you scale the vehicle.

1. andrewjhull says:

Certainly you might need to dampen those vibrations in some way, however that is a sepate mechanical problem which could be overcome with mechanical isolation and tuning. It would be interesting to see how welll this particular design does scale up (or indeed down).

10. fm` says:

I wonder if the induced axial vibrations even couple to the frame at all. Most vibes are from unbalanced props.

11. zaphod says:

does someone have a simpler explanation for how this works?

1. Josh says:

The top motor speeds up and slows down slightly during each revolution it makes (This is where they talk about a sinusoid waveform) and as the blades are connected to the motor by small hinged joints, the blades either begin to lead the motor hub when it slows or lag behind the motor hub when it speeds up.
Simply the blade keeps spinning while the actual motor hub is slowing down and because the hinge is not parallel to the motor hub but instead at 45 degrees, when the blade isn’t pointing straight out from the motor (as it does when the motor is at a constant speed) it starts to twist.
This twist ends up the same as blade pitch in a normal helicopter only you don’t need an extra thing to move the blades and some linkages to connect the blades to the moving mechanism, just the same motor that is spinning the blades in the first place.
Hopefully that helps. Are you Zaphod from PA?

1. zaphod says:

thanks for the explanation. I kind of thought that varying the motor speed was how this worked, but what I don’t get is why the craft doesn’t revolve around the z-axis, shouldn’t the varying speeds of the rotors cause the vehicle to yaw like the how cheap rc helicopters turn? I’m not sure what ‘PA’ is so probably not.

2. AltMarcxs says:

Imagine you are standing in a train, when the train accelerate you can feel it and even fall back, but once the train is at constant speed (doesn’t matter if slow or fast) you didn’t feel anything and it’s easy to stand straight.

What they do with the additional sin wave is to accelerate the motor speed a little bit within a quarter turn (revolution), this acceleration make the blade to “fall back” aka change pitch, with a good timing the quarter turn is always on the same spot, kind of standing wave so it can be used to pitch or roll.

12. CRJEEA says:

I was thinking what if there where two shafts with interlocking wave shaped ends. When the motors are synchronised the shafts stay still. The propellers are driven. With the shafts attached with a groove allowing liniar motion to alter the pitch of the blades.
Just a thought to get around the inertia aspect without adding too many more linkages.

13. markaeric says:

Am I correct in assuming that this couldn’t make quick changes in altitude without causing the aircraft to pitch/roll? Or would the relatively long duration of the accel/decel forces just cancel out, perhaps only inducing a slight wobble?

14. I’m not 100% sure, but I think I’ve seen a similar mechanism some decade ago in an RC helicopter named “the Revolutor” from Keyence. It was a sort of technology demonstrator with multiple miniature gyroscopes(actually spinning ones, there were no MEMS gyros back then), and a main rotor loosely connected to rigid driving axis with a rotary encoder.

The helicopter changed blade pitch by accelerate or brake it at the right moment, I think, which is very similar to how this one in article works.

1. markaeric says:

Oh, wow. That does seem to use the same principle for changing blade pitch. Thanks for the heads up!

15. Smarter one, would be with only one motor and the lower propellers linked only by air flux.kinda rudimentary copter

16. If this sort of thing appeals you to, you might also like “1Star”, the 1-motor running robot out of the Biomimetic Millisystems lab at Berkeley.

17. John says:

What is the word at 53 seconds. I have listened to it loads of times and the best I can come up with is “passive poo strider”

1. a passive hinged rotor responds to a pulse driving torque with a controllable cyclical response