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: 16733

Comment by Tomas Soedergren on January 10, 2012 at 8:58pm

Inventive spirit! Looking forward to the demo videos :)

Comment by Helitrasher on January 11, 2012 at 6:27am

I think noise and mechanical complexity are the only issues with the Heli´s, plus a bit of fear factor !.

And I would love to see a auto autorotation, I know from experience that it takes years to master.

Comment by Helitrasher on January 11, 2012 at 6:39am

Wieght is also a problem, but the day when a multi rotor can do a Tick Tock or a inverted backflip pyro roll are only in the hands of the variable pitch machines.

Comment by Patric Millar on February 5, 2012 at 12:16am

Hi Joe.  Looks like a very interesting project.  As someone with a foot in both the UAV/Quadcopter camp and the tracking wildlife camp I'm waiting for more updates. 

I'm also interested in the mechanics of your tracking rig?  What type of antenna do you plan to use on it?  Are you planning to feed the VHF signal back to the ground for tracking purposes? What is your anticipated VHF range?  I've tracked collars from a helicopter before in Australian forests with small mammals in Australian forests and done some trials in remotely downloading GPS data from the collars.

P

Comment by Gisela & Joe Noci on February 5, 2012 at 3:35am

Hello Patric,

See you amd Michael are in cahoots...

 Are you the fellow that Michael mentioned might fly the quad for him with a tag tracker on board?

 

 Patric, one of the planes we used was the Hornbill one, which you have seen on this forum. The tags are UHF around 430MHz, which has both positive and negative aspects. The positive is that the antenna are smaller, nearly a third less than the VHF tags, which is great for embedding in Rhino Horn, etc, and also great for fitting inside the fuselage. The downside is that there is slightly more signal attenuation in dense green bush, so through the bush signals are difficult - don't really want to give ranges here; maybe we can tie up off-forum, if you want more info.

 

Tag TX power is as you know very low in the interest of battery life. Time between transmissions is also a compromise. Then to get the best detection range one really needs to use a higher gain antenna on the detecting receiver. However, higher gain means narrower beamwidth, so you need to scan the antenna in flight. On UHF it is easier, due to the smaller antenna; VHF is a bugger...Although the traditional VHF tag and receiver is a narrow band RF design, so the signal to noise ratio is much better than the wide band hi-speed digital transmission of the UHF didital tag. This give already a 4 to 6db signal to noise gain on the narrow band system which directly equates to either less gain required on the antenna for the same range, or greater range for the same gain antenna..

The UHF antenna we use in the A/C is a derivative off the MOXON antenna, a 2 element type of Yagi, but the MOXON principle reduces the size even further - gain is around 5 to 6dB. The antenna is mechanically scanned using a model aircraft servo, driven by our autopilot.

So all these conspire to make tag detection very difficult from a foward flying platform, with low TX power, TX maybe every 30seconds, and the antenna beamwidth requiring beamscanning, etc. Image flying at 20m/s with a TAG TX time of 30sec - the A/C will be 600meters away from the position of last TX. If the antenna is pointing in the wrong direction, and the tag is quite far, you do not detect it. 30 seconds later you are 1.2km away, and so it goes..

Our trial showed that the best way is to have a long endurance A/C - at least 2hours, and fly ever-widening concentric circles, till a limit, and then ever decreasing circles and repeat; ie, saturate the area of interest. This is for large area coverage. If in a ravine or such like , it is a little easier, since the animal can only be in the ravine so you fly the ravine outline repeatadly. This is interesting, since you need to be in 'view' so must follow ox-bow curves, etc, and the flight planning for that needs to be good - a high rocky outcrop, or a tall tree overhanging the ravine edge into the flight paths are pretty good pro-gravity aids to anti-flight.

We are still doing work on all this, but our solution for the mainstream tag detectors today are stationary (but mobile) poles  fitted with a small solar panel, battery, GPS, and a tag detector(s) with either a single omni collinear hi-gain antenna ( 11dB), or multi tag detectors with multi antenna, eg, in a ravine with one ant looking down and another looking up the ravine. Each pole has either a GSM cellnetwork (GPRS) link, or a sattelite link, back to the database. The poles are placed in plains, along known animal paths, near water sources, in ravines, etc.

What a long story...

 

Regards

Joe

Comment by Patric Millar on February 5, 2012 at 4:32am

Hi Joe

Thanks for the response.  The answer is yes.  I've been discussing with Michael the supply of the collars, VHF and UHF download system around the 50g mark.  We already have an automatic UHF download system which can run as a self contained unit, but this doesn't suit Michael's application.  It works great where your target species has a known roost, watering point ot foraging area, as you've already described.  However, Michael's target species is very mobile and has a large home range. We do have a satellite upload model which is around the 800g mark, but that's way too heavy for the target species at hand.

I fully understand the trade off between batery life and performance.  In the GPS collars I'm involved with everyone wants a 10g collar that'll take 24 readings a day for 90 days and cost $200. :-)  We are always looking at new batteries and are tweaking the software choices to allow for increased battery life and increased data collection regime. I'm digressing here.

The current challenges we need to overcome are :

  1. Size, shape and bulk of our UHF system.  At the moment we use a small UHF transceiver that needs to be connected to a netbook and a 3 element yagi.  This is not really suitable for a UAV system at this stage.  We are looking at shrinking this down, but not for a couple of months at the earliest.  We are also looking at antenna design so your comments are much appreciated.
  2. Flight time is the other issue.  We need to get around a 1 - 2 minute window for the systems to handshake and exchange data.  A multi rotor will probably give us 10 minutes max, especially with the weight we are looking at.  So the issue is what happens if the systems are mid handshake and the multi roto needs to return to base and the target species moves away.  Your system sounds much better in terms of flight time.

Again congratulations on the great work.  I'll keep in touch with you on this one.

Happy flying

Patric

Comment by Mick on February 10, 2012 at 6:33am

Nice design.  Can you tell me what you're using to control the ESC's and pitch/tilt servos?  Are they off-the-shelf components or something you have custom-designed?

Mick

Comment by Ron Poynter on February 9, 2013 at 10:17pm
Joe

I see this post is a year old. Have you flown this aircraft and has its lift capacity and flight time worked out as planned?
Ron
Comment by Elias-Jose Vasquez on April 27, 2013 at 2:01pm

are there any videos of this quad?

Comment by Gisela & Joe Noci on May 6, 2013 at 11:13pm

Report back on my Large variable pitch and tilt Quadrotor.

A VERY Belated report back on this quad concept and its failure...

It did not fly,; It did not even get past the test phase! Physics are against one in this I fear. Each motor and blade assy was capable of almost 6.7kg lift at 2/3 throttle, but that at a motor rpm of around 2200 RPM. At 1000 RPM the lift reduced to 4.2kg, with the blades at 15deg AOA. Motor amps on a 6 cel Lipo pack was 16A. The max RPM for the Trex blades should not exceed 1500RPM, and lift was 5.9kg then , amps =20A.

However, the mechanical forces on the tail rotor assy used for the tilt mechanism , fitted with 580mm main rotor carbon blades are huge, and the changes in pitch do cause a slight buckling of the 'disc' formed by the rotating blades - this at the end of the long arms induces slight vibrations, and the tail rotor assy simply fails - the assy is held by grub screws on the motor shaft, etc, and these are simply not good enough. they fret loose ( locktight does not help - just means you cannot get the screw out again..) and then the assy moves on the motor shaft, the vibration increases, and the process leads to rapid self destruction. You can see from the manufacture process and pics, I did not skimp on the design or materials strength, so that cannot be improved. The bottom line is that the concept does not scale well. The mechanical strength need to swing large diameter blades, and carry the weight, and deal with the buckled disc effect ( as you have on a normal heli, with the swashplate in action) , etc, simply means the head needs to be mush stronger. Thats why the TREX main rotor head is stronger, bigger, etc...But to place 4 such heads on the quad, the extra weight, etc, just means a case of severely diminishing returns, hence the poor scaling related to size increase - OK for smaller quads!

So, it was an interesting exercise. I should just have done one arm and tested it first, before going the whole hog, but sometimes , when the CNC machine is set up and jigged to do the job, it is easier to do the whole lot, rather than set up again to do more!  

Anyway, there are fa too many quad related posts now, so I am having fun with our SurVoyeur plane, new wings, new motor/prop combo, IR videocamera, etc!  I will post a blog soon on the results ..

Joe...

The Nampilot.

Comment

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

Join DIY Drones

© 2014   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service