A large scale Variable Pitch Tilt Quadrotor implementation.

                                

 

                               A  large scale Variable Pitch Tilt Quadrotor implementation.

  

 

Here is a description of a Variable Pitch Blade Quadrotor build with some unique features. It is a large Quad, and can be quickly split into 2 halves, enabling easy packing into a carry case for simple transport. The ‘Upper’ half of the quad contains all the electronics, IMU, Autopilot, Payload control electronics, etc. The lower half carries the battery packs and the payload. The upper and lower halves slot together and are affixed by means of 4 captive spring release screws. The payload similarly mounts to the underside of the lower body half. The entire Quad can then be separated into an upper and lower body half, and payload. 

 

Each motor has a standard ESC ( not any fancy hi-speed update type…) mounted beneath the motor. This keeps the 3phase motor leads very short, ensuring little electrical noise radiation from said leads.. Beneath each ESC is a small CPU module which receives a serial command from the Autopilot. This serial command contains the desired motor speed, the blade pitch angle, and the rotor tilt angle for Yaw control. This small CPU module generate the relevant PWM signal to the blade pitch servo, the rotor tilt servo, and to the ESC, while monitoring the motor RPM, ensuring constant motor speed, as commanded.

 The upper half of the Quad has 2 Variable pitch, variable Tilt, same direction rotating rotors. The lower half has only two variable pitch same direction rotating rotors fitted. The motors are operated at commanded constant speed, and lift is varied by blade pitch, giving much faster control of individual blade lift for stabilization. This is a problem with conventional control by motor speed variation, as the rotor blades become larger in diameter. The inertial mass of the rotor increases, and it becomes increasingly more difficult to rapidly change rotor RPM, both to enable horizontal stabilization, and to control Yaw.

 The Tilt rotors are both fitted to the upper half of the Quad, with the rotors spinning in the same direction. This ensures that when the rotors are tilted ( always in the same direction) to control Yaw, Gyroscopic Precession is equal and opposite across the two rotors, thereby canceling out. Gyroscopic Precession also begins to play a bigger role in larger diameter rotors, when the tilt rotor concept is employed.

 The construction is almost complete. So far all simulations and math shows we should have around 2.5kg lift per rotor head, with an all up Quad weight of 5.5kg, inc 2 of 6cell, 6000mAH batteries, leaving around 2kg for payload. We should get around 15 to 18minutes flying time.

 If it looks like it came out of the ALIGN Factory, it did not…I used parts from the Trex600 series chopper, the tail rotor pitch assy, etc, as well as tail boom, stays, and a few more bits and pieces…The composite parts were all modeled in 3D CAD, and molds made for them. All the custom Aluminium parts were likewise modeled and CNC machined.

 Basic Specs:

Dimension from blade tip to blade tip – 2100mm

All up weight – 5.5KG, inc batteries, 3.4kg exc batteries.

Blades – Trex 450 style.

Variable pitch on all 4 blades.

Variable Rotor Tilt on 2 blades.

Max payload mass – 2Kg.

Estimated flight time with 1.5kg payload – 18minutes

Autopilot – NamPilot

 

 

I hope to fly it real soon now….!!

Just running out of space for all the aircraft here!, Guppy, Hornbill, Kiwit, all the 60 trainers, ……Need some sky-hooks!

 

Regards

Joe

‘The NamPilot’

 

Two unfinished body Halves:

 

 

 

Cutting out lightening Holes on the CNC:

 

  

 

2 halves with IMU:

 

 

 

Lightened Central Hub Assembly: (with unpainted autopilot cover)

 

 

 Lightened Central Hub Assembly:

  

 

 

 

CNC Machining of Motor Arms:

 

 

 The Pitch Tilt Arm:

 

 

 

The Variable Pitch Motor Assembly:

 

 

 

 

The Variable Pitch and Rotor Tilt Assembly:

 

 

 

Partial frames fitted Together: (floor squares are 300mm X 300mm)

 

 

Frames in Transport Mode:

 

 

 

 

 

Views: 24540

Comment by Gisela & Joe Noci on January 5, 2012 at 2:16pm

Alex,

It is very difficult to keep the weight under control with a central motive power source. Drive belts just dont cut it - teeth slip, etc, and torque rods have there own problems - the arms of the quad need to be quite stiff to ensure the torque arms do not bend much under load, as they then start to suffer severe whiplash. The required gearing to drive the rotors all add weight, etc. The only advantage for this consideration is if one can use a gas engine to overcome all the losses and get longer flying times.

 

Interleaving the blades in rotation is not a good idea in this instance Alex. Unlike counter rotating props on the same axial (co-axial props), the turbulance seen by a blade half as it enters the other blades counter rotating airscrew causes very severe blade flapping, since the tip of that blade suddenly experiences accelerated airspeed, in either a forward or reverse direction, and suffers sudden increase or loss of lift, and flaps badly, with large vibrations and premature blade fatigue. Dont do it! You only gain effeciency if the one blade is fully and constantly in the other blades counter rotating airscrew.

 

With regard to the method of Yaw, pitch and roll control, please read my reply to Stuart - no differential blade pitch control needed.

This quad is also inended for Radio tag tracking of animals in the bush. The concept is that when the ground rangers wish to get to a tagged animal, they would send up the  fixed wing UAV, flying sectors over the park, to find the general location of the animal, then proceed by 4X4 vehicle to that area. The quad will be fixed to a special roof rack on the vehicle, and when they are close to the area, the quad will lift off, climb to a few hundred meters and 'listen' for the tag again, get its current GPS position, and land back on the roof rack ( guided in by and infra red beam). The rangers can then get closer and repeat the process untill they can detect the animal by means of a handheld receiver and antenna and proceed futrher till they reach the animal.. The Quad will also carry a camera and video transmitter to give a view of the area to aid the possiblity of detection.

Watch This Space.....

Comment by Gisela & Joe Noci on January 5, 2012 at 2:25pm

Hello Elison,

This is the biggest problem with this concept!  Brushless motor designers  are not yet taking this mode of operation into account, and so most motor KV ratings are way to high. My motors are 355 KV, which is at least twice what I would like. They are Hacker A40-14 V2. We will see what happens when we start flying! I am doing all the RPM versus blade pitch versus Thrust measurements at the moment, to verify my simulations for all the control loops, and the results will show if the motors are going to work....The lower KV motors, 120 or so, are all very big and very heavy - good for a monster Quad..Adding gearing just adds weight and extra bits, the whole Catch 22 scenario.

Comment by Stuart Norton on January 5, 2012 at 2:41pm

Ok, thanks I understand. It's similar to the tilt on a tri-copter. Thanks for the detailed explanation.

Comment by Alex Roup on January 5, 2012 at 3:02pm

@Joe

Re: interleaving, the CH-47 has interleaved rotors, so it's at least feasible.  Perhaps the turbulence and vibration would make it an undesireable solution.  I'll admit I haven't done my homework on that concept.

wikipedia CH-47


Re: rotor tilt, Ok I understand the tilt can generate the yaw moments required, but thought it might be advantageous to use differential blade pitch in the mix.  I would think the actuation inertia would be lower for differential blade pitch than for rotating the rotor disk.


> In such a system, motor RPM control yaw, and roll and pitch and lift, so all
> these have to be balanced against each other , each with PID control loops
> fighting for domination.

I agree that each effector (typically RPM, in this case blade pitch) is used for more than one axis of control, but I think they can be mostly decoupled, at least for small amplitude control inputs.  For example (in Matlab syntax):

% logical control order & sign convention:
%       Roll (starboard)
%       Pitch (nose up)
%       Yaw (starboard)
%       Collective (down)
%
% effector (blade pitch, all positive for increased pitch) order & rotor
% direction:
%       Forward (CW)
%       Starboard (CCW)
%       Aft (CW)
%       Port (CCW)

T_EL = [  0   1  -1  -1
         -1   0   1  -1
          0  -1  -1  -1
          1   0   1  -1 ];

effectors = T_EL * logicals;


The rooftop takeoff and landing sounds like an interesting problem.  Would you use a vision system on the quadcopter to indentify the IR source on the truck for relative navigation?  Do you have a prototype already?

Comment by Ellison Chan on January 5, 2012 at 3:02pm

Hey Joe, yes.  I can imagine the difficulty with motor selection.  We've been having a pretty long discussion on motor efficiency, and power on the Aeroquad forum.  It's still ongoing.  If you check out page 5, the user "jiankkll" had some suggestion on some low kV motors.  We're trying to figure out whether the Microdones MD4-1000 claim of flying for 70 minutes is realistic.  I have my doubts.  Apparently the MD4-1000 uses 72kV motors.  Sounds like custom wound.  Maybe one way to got is to rewind the motor with thinner wire.

Here's the link to the Aeroquad discusson:

http://aeroquad.com/showthread.php?4528-a-problem-about-power-syste...

Comment by Sean Skirvin on January 5, 2012 at 5:17pm

Very nice!  Just one thought, if you fly it and realize that you have way too much yaw control with the current system, I think you should try just doing a DCP for the yaw, and using the tilt, at a much slower setting, for trimming the yaw.  I think you will be better off in the long run if you are tilting the props as little, and slowly as possible.  You are fighting a lot more gyroscopic forces when you tilt the rotor as opposed to increasing the pitch. Can't wait to see some flight videos! Great design and build.

Comment by Gisela & Joe Noci on January 5, 2012 at 11:19pm

@ Alex

Thanks for the comments Alex. I agree that in the mathematical world the effectors are seperable, but physically any action on one rotor has a strongly coupled reaction to the others. And the situation is worsend as we scale down in size...The same applies to the CH-47. The rotor blades of the CH have high inertial mass and so a very low ( millihertz) flap frequency. In addition the Reynolds number for those blades are orders of magnitude higher than our poor little blades, so the lift stability is much better. Physics does not scale well for the scales we play at..BUT, it still works!

Comment by Gisela & Joe Noci on January 5, 2012 at 11:31pm

@Ellison,

I have done some fairly extensive measurments comparing low kv outrunners with geared high kv inrunners. For the most part, for the same shaft speed and thrust figures, the inrunner requires less input watts, at midrange thrusts up to 12% less. However, the weight of the geartrain and associated elements required 7% more thrust, which at mid thrust required 19% more input watts....Not easy to win!

I am convinced the way to go is a pancake shape, large diameter motor, of around 120 to 150kv.

I am not sure how the thinner wire will help? I think in essence the number of poles must increase, so we need more stator arms and more magnets. More stator arms take up more space so then thinner wire would help, but IR losses go up and motor efficiency goes down...As I said in my comments to Alex, scaling physics down often just does not work very well...

 

And 70minutes flying time for the MD4-1000 really needs to be accurately qualified with a proper spec of the setup, weight, altitude ( air density) , etc. To much Smoke and Mirrors in to many Vendor claims!

Comment by Gisela & Joe Noci on January 5, 2012 at 11:37pm

@Sean,

Yes, the Gyroscopic forces are large. That is also why I have two opposite rotors with the tilt mechanism as opposed to only one, as in a tri-copter - the gyroscopic precession forces cancel out, resulting in a net zero effect on pitch or roll of the aircraft. However, the force to tilt the spinning disc remains, and that is quite large! In sumulations I have rotor tilts of less than 0.5 to 0.8 degrees to stabilize Yaw in all modes of (gentle!) Flight. This Quad is not intended to be highly agile, but as stable as possible. At least it is all a hell of a lot of fun!

Comment by Squalish on January 6, 2012 at 3:01am
Let me just start by saying: my compliments on the quality and elegance of your build.

Would a large fan shroud be prohibitively difficult here? Prop collective may have benefits to motor efficiency and practical size (for some tradeoff in prop efficiency), but the larger props are somewhat dangerous as well - getting past the control problem of high angular momentum by tilting the prop will let you get to angular momentums that are very destructive.

I would love to see some empirical data on result of these efficiency tradeoffs, as far as maximum endurance goes.

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2018   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service