OT. Quadrocopters.

Quite impressive!

I agree. Seems to me though a quad copter is a lot easier to control than your basic chopper with a tail rotor. At least I can visualize the control system for a quad copter.

Actually, it's the other way around. A conventional helicopter can perfectly well be flown without any electronic assistance. A quad is absolutely impossible to fly without gyro stabilization.

Not sure why this subject would be considered off topic (OT). After being away from electronics for a long time, these quadcopters have captured my interest, again. Got an arduino, dug out my old Fluke and soldering station and am having a ball reliving the excitment of my earlier days when it was all still new. Might buy an ARDrone quadcopter before I get around to building my own. Check these lil' mini q-cops flying formations:

nb

Without thinking about the problem too much, why would there even need to be a cyclic for control? Wouldn't independent collectives be enough for control? Cyclics may be needed for forward motion (equalize lift on opposing blades) but are they needed for control?

Torque matching isn't really needed, unless you're trying to match RPM (which would be good for vibration ;). Lift matching can be done with collectives.

be a

trol?

s)

no you don't need cyclic, my point was that effect of cyclic is different lift at each sector of the circle described by the rotor. independent collective, or in small electrics trottle, on each of the rotors in a quad should have the same effect, theres just only four sectors instead of "infinite"

hich

ives.

I think you still need torque matching or it would rotate around its own axis

I believe the principle of quads is that each pair of rotor rotates in opposite direction. Each pair must then match the torque of the other pair to prevent it from rotating. If you have collective on each rotor everything just need to rotate at the same speed

If you control speed an increase in speed on one rotor must be matched with a decrease one the other rotor in the pair, so the resulting torque matches the other pair (I think)

-Lasse

My point was I could visualize the control system easier. Fly forward would be turn the rear rotors faster. Stuff like that. For sure you need a gyro.

Hello,

When I was a post grad in physics 1981/1986 we toyed round with the idea of a system to balance an umbrella on it's tip. To this day I don't know if this can be done.

Regards,

Peter Nolan. Dublin.

I think these microchoppers have fixed pitch blades and do all manouvering by varying the rates of the 4 fans. I'm guessing two fans clockwise (opposite corners) and the other two anticlockwise so that torque cancels, to yaw you'd speed one pair of opposite corners and slow the other. to pitch or roll (and thuse move laterally) speed one side and slow the other, to climb or descend speed or slow all 4.

Yes, the quad is mechanically incredibly simple. That is probably one of the main reasons it has gained so much popularity. In fact, the only moving parts are a pair of bearings in each motor.

On a conventional helicopter, the key to the magic is a device called a swash plate. The mechanical stuff is quite complex and fragile, but the control inputs are actually as simple as could be: Tilt the swash plate in one direction, and the helicopter will tilt in the same direction. Move the swash plate up or down increase or decrease lift.

This does not seem correct, that would make it tumble. You want a temporary increase back and decrease front, so the whole thing tilts forward, then it will want to fly in that direction, but you then likely will need more speed on all rotors to keep it at altitude.

quadcopters such so do helies, as they have only very limited flying time on a battery charge I would go for a huge glider with electro motor. Fill it with nukes and drop where you like. Oops,they already have drones...

I don't see that it much matters. The thrust differential has to be pretty much continuous. The "cyclic" and "collective" paradigm seems to work for both.

Yes, I wasn't considering yaw around the vertical axis.

That works for hovering but I can see where forward motion will get complicated.

Tandem rotor helicopters (like the CH-47) have similar problems. Obviously they still have cyclic.

Not sure I follow.

Right, the "cyclic" is controlled by varying the thrust of the different rotors.

"Inverted pendulum." Your quad-copter guys' website has a video of their kit doing it.

-- Cheers, James Arthur

eed

Hello James,

I was just looking at it on YouTube:

This however is just in two dimensions. Their machine throws the pendulum into the inverted state which to me is a great feat in itself where it is obviously very stable. I see, having googled, that they also use the quadrotor to balance an inverted pendulum in three dimensions in another video I was just looking at. Revelatory. I should have known this was called the inverted pendulum. As the Americans say every day is day spent in school.

Regards,

Peter.

You can make an inverted pendulum stable open-loop, by vibrating the pivot sideways. That creates a potential well at the vertical position, where the AC kinetic energy is zero to leading order.

Cheers

Phil Hobbs

te:

can

ad is

ward

ou need

Hello Phil,

Without seeing exactly what you are saying I would have a feel for it nevertheless. When I was a boy we amused ourselves by balancing what I think you call a broom handle in the States what we call over here a sweeping brush on the top of a finger with the palm facing upwards and I just about imagine what it feels like to actively vibrate one's finger to achieve greater stability of the broom handle. I must investigate.

Regards,

Peter.

is

need

My freshman physics prof had a demo that used a motor and crank to do that. Jiggling the pivot vertically makes the straight-down position unstable too--there are two stable positions, one on either side of straight down. How far away they are depends on the amplitude and frequency of the vibration.

Cheers

Phil Hobbs

This may be of interest:

Start at page 15 (English text):

Experimental implementation of linear and nonlinear control laws