HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Aug 05, 2014, 05:35 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
Build Log
Quoddball Build Thread

Quoddball build thread

It's not really a VTOL but this thing has got me inspired to build something fun.

HyTAQ Robot (1 min 30 sec)


As, its name implies, the idea is to build a quad in a ball. In my case an actual ball, not a cylinder. I am hoping it will be able lean into turns and roll around in an arc, as well as roll straight, bounce around, be relatively indestructable, and of course fly.

I just ordered 100 feet of the .063 rod here for $84.60 plus shipping. That's enough for about 3 balls by my math. They don't sell it in smaller quantities, but you can never have too much CF rod so what the heck.

We will see how far I get before I have to shut everythng down to move, but it all looks pretty simple. Should be fun...
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Last edited by Ran D. St. Clair; Aug 08, 2014 at 09:23 AM.
Reply With Quote
Sign up now
to remove ads between posts
Old Aug 05, 2014, 08:14 PM
Aircraft Designer Guy
AJWoods's Avatar
Melrose, MA
Joined Jul 2005
1,379 Posts
I found this approach particularly interesting
An insect-like, crash-happy flying robot (2 min 22 sec)


Would be cool to put a quad in a similar frame
AJWoods is online now Find More Posts by AJWoods
Reply With Quote
Old Aug 05, 2014, 10:24 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
Quote:
Originally Posted by AJWoods View Post
I found this approach particularly interesting
Yes, I saw that one in the Japanes Ball Drone thread (their spelling, not mine).

There's a lot I like about it and some others, not so much.

They clearly put some serious thought into the design. The geodesic frame is probably some theoretical optimum for weight and even sized openings, but it would be hard for me to build, and I suspect the connectors detract from the overall strength. I plan to use a simple 2 hub design with continuous CF hoops except at the hubs. The crossover connection points will be lashed with thread and glued with Foamtac for flexibility.

The full gimbal design is great for decoupling the outer ball from the inner aircraft, but it is a challenge to make sufficiently stiff and light weight. They did a great job, though I don't know how well it would stand up to a serious collision. Mine will be a single axel roller which is much easier to make and should be much more crash worthy, though it will somewhat limit the degrees of freedom for ground motion. I will have to yaw to make it roll (on the ground) in a different direction.

I note that their inner aircraft is nearly perfectly balanced in all axis. Once again, that is great for isolating the inner aircraft from impacts, but it also takes a lot of careful design to get it right. Kudos to them, but I don't plan to work quite that hard. I will get the aircraft CG close to the center of the axel, but it might turn out a little bottom heavy, which means it will flip to upright when on the ground, though not strongly.

They used counter rotating props with aerodynamic surfaces. It obviously works, but in my experience such a design is not best for control authority. A standard quad is much more rock solid in flight and less subject to issues with wind.

A quad also puts the props closer to the fattest part of the sphere for more disk area per volume, and therefore a lower disk loading which should improve overall performance. Carrying around the dead weight of the cage is a negative to begin with, so anything that helps is good. Of course a quad doesn't fully fill the disk space at the equator but it is still probably more optimal than their design in this respect.

I have been doing some basic math assuming I re-use my existing motors, batteries, props, etc. It all derives directly from the 9" prop diameter. Packing the props in with just a 1/4" tip clearance leaves a 4" square hole in the middle and a 22" span from outer tip to outer tip. Adding 2" clearance between the prop tips and the inside of the cage to allow for some serious cage deformation in a heavy impact results in a 26" diameter ball. That is a pretty big ball.

I checked and it will fit in my car, just barely. It will have to sit in the passenger seat wedged in between the center console and the door handle.

This is not my first experience with oddly shaped aircraft or ball shaped aircraft for that matter. This one dates back to the days of geared brushed motors and ni-cad batteries. It flew massively better when upgraded to brushless and li-pos. Sadly, after about 10 years, pilot error did it in. A descent flight controller (or valium) would have helped that crash. (Take a walk on the wild side...)
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Last edited by Ran D. St. Clair; Aug 05, 2014 at 10:31 PM.
Reply With Quote
Old Aug 06, 2014, 10:42 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
The structure of the quad (Quodd) is underway. It's just some 3/16" plywood from the hardware store and some 5/16" by 3/8" basswood.

The U channel that hangs down below is for the battery. The cross pieces hold the fiberglass tubes that will hold the motors. If they look a little crooked it is because they are. Each tube is skewed 2 degrees for a total pinch angle of 4 degrees. This will allow the motors to tilt 10 degrees for improved yaw control and still keep the prop tips separated by 1/4". (Or at least that is what the math says.)

Another 3" square plate will go on top to enclose the channels for the fiberglass tubes. It's all held together with thin ZAP CA which is amazingly strong. It really soaks into the wood.

The motor mounts will be made in a similar very simple way. I see no reason why it should not be flying without the ball by this weekend.

The CF rod I ordered is temporarily out of stock so that part of the build will be delayed a bit.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Last edited by Ran D. St. Clair; Aug 09, 2014 at 09:20 AM.
Reply With Quote
Old Aug 07, 2014, 10:25 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
The motor mounts are complete and ready for paint. It's a very simple design. Two pieces of 3/16" ply, 1 1/4" square, and two pieces of 5/16" by 1/2" basswood. The basswood pieces are squeezed tight around the fiberglass rod and tack glued to the back plate. Then the rod is removed and thin ZAP CA is wicked into the joint. The holes are drilled to 1/16" using the steel X mounts as a guide. The #4 self tapping screws not only attach the X mount but also pull the whole sandwich together to clamp tight on the fiberglass rod. These mounts can easily be adjusted for the correct 10 degree angle to enhance yaw authority. I have used this design before so I know it will hold. If they loosen up a bit with time and multiple impacts, I can always epoxy them to the fiberglass tubes.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Last edited by Ran D. St. Clair; Aug 09, 2014 at 09:18 AM.
Reply With Quote
Old Aug 09, 2014, 09:16 AM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
The parts were sealed with ZAP thin CA. It's an expensive way to do it but it soaks into the wood and adds tremendous strength. It also cures almost immediately so there is no time lost before applying the finish coat. A light sanding to knock off the high spots, and then rattle can paint was used, either white or black. After a couple of hours the red and green were overcoated on top of the white. Everything was baked in my oven overnight so I can keep moving forward this morning.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Last edited by Ran D. St. Clair; Aug 09, 2014 at 09:22 AM.
Reply With Quote
Old Aug 09, 2014, 03:33 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
The ugly part of the build. Everything is connected, but the wires are all over the place.

Now that I can hold it in my hand, I have decided that 2" of clearance between the props and the inside of the ball is excessive. I will reduce the ball size from 26" to 24" which should ease the transportation issues a bit.

It's turning out to be a bit more bottom heavy than I hoped. We will see What effect that has on the flight and ground bounce performance. It might be a good thing as Quoddballs wobble but they don't fall down.

Apparently the Pythagorean theorem is still valid because the gap between the props turned out as planned.

I will need to make a big foam bumper for the bottom of the battery so I can fly it until the ball is made. I left the front and back (white and black) support tubes long for now to protect the props. I will cut them to length when it is time to put it in the ball.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Reply With Quote
Old Aug 09, 2014, 06:14 PM
skumgummi dave
Gresham, OR.
Joined Mar 2004
1,816 Posts
Lurking...
foamdave is offline Find More Posts by foamdave
Reply With Quote
Old Aug 09, 2014, 06:32 PM
Registered User
Joined May 2010
422 Posts
Quote:
Originally Posted by Ran D. St. Clair View Post
The U channel that hangs down below is for the battery. .
Randy,

Why did you choose to hang the battery vertical? Did you want to have a self-righting affect from the low CG, thus aligning the the props upward when in ground contact at zero velocity?

Will this surface area of the battery box hanging downward below the hinge/roll axle cause some odd flight behavior? I.e. will it act like an horizontal stab and want to streamline aft during high-speed forward flight. Will the fight controller fight the unwanted negative pitching monument.
corocopter is offline Find More Posts by corocopter
Reply With Quote
Old Aug 09, 2014, 10:38 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
Quote:
Originally Posted by corocopter View Post
Randy,

Why did you choose to hang the battery vertical?
I set the battery vertical mainly to keep it out of the high velocity airflow from the propellers. This design has the props pushed in to within about 1/4" of one another in order to keep the ball to a reasonable size. That only leaves a 4" square in the middle for the flight controller and all the other gear. If I had laid the battery flat it would have been under the props to some degree. Also, the U channel made for easy battery installation with Velcro. Also, since the battery and my hands and arms have to fit in through gaps in the ball cage I wanted it to be easy to install.

I did not put the battery down low for any stability effect in the air. There would not, should not, be any. When on the ground, being bottom heavy will rotate the quad to upright on the axle, and to the extent that the ball can roll easily it will also turn upright in roll, though that will be iffy in grass, especially since the ball won't be a perfect sphere. Regardless, when the motors come on, the AutoLevel will bring the quad to upright quickly.

There may be some negative impact to having the CG below the axle when I "bounce the ball". Ideally the quad inside would be undisturbed by outside forces, but with the CG below the axle it will be bounced around in pitch. It will be pushed around much more in roll since I don't have a full gimbal inside, only a single axle. I expect the issues in pitch will be minimal, but part of the reason for building this thing is to find out what happens.

This will not be a particularly fast aircraft so any aerodynamics of hanging the battery down below will be negligible. I also expect it will have little or no effect on the flight dynamics. Quads often carry cameras on gimbals hanging down below and it doesn't seem to bother them. It probably increases the rotational inertia of the aircraft slightly, and that is a negative, but the impact is minimal. The rotational inertia of the ball cage in roll, not pitch, will probably be more significant. It remains to be seen what, if any, impact this asymmetry between pitch and roll will have. I expect that I might need slightly different PID values for pitch and roll to get the best stability.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Reply With Quote
Old Aug 09, 2014, 10:51 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
The Quad part of the Quoddball is done. I found some tall landing gear left over from another project so I temporarily attached it with some cable ties. I have flown it in the driveway and it seems to fly fine. Of course it is just a basic quad so there is nothing special about that. I will need to test fly it in a larger place to tune it further.

I will admit to waffling about reducing the ball size. If I later decide to incorporate a full gimbal then the larger ball will provide the needed room. It looks like I will have a week or two to think about it until the CF arrives.

Since this is just a quad and not a VTOL I am using the KK2V1_17S1Pro firmware by Steveis, not the OpenAero2-VTOL firmware. This multi-copter dedicated code seems to work very well, and can be found here:

http://www.rcgroups.com/forums/showthread.php?t=2144803
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Last edited by Ran D. St. Clair; Aug 09, 2014 at 10:58 PM.
Reply With Quote
Old Aug 10, 2014, 04:06 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
The Quoddball inner quad flew at my regular flying field today without incident. Jason flew it and gave it the CrazyJ seal of approval so it must be good. It actually flies better than my regular quad.

Being a plus quad instead of an X quad it has a different visual presentation in the air. I have decided o add some features to help make it easier to see orientation, including a white flat plate disk in the front, and a black vertical tail in the back. Neither one of these is intended to have any aerodynamic function. I also decided to leave the front and back motor mount tubes equal in length to the tips of the props plus 1/4". This will give a place to mount my visual aides and also help protect the props even when it is inside the ball.

The "wheel hubs" on the ends of the axel shafts will be 2" in diameter and painted red and green as well. Even with all these additional visual aides I expect it will be advisable to fly it in AutoLevel mode except when it is up close. I have tuned the AutoLevel so it still has pretty good maneuverability. The stock AutoLevel settings in this firmware were so conservative that full stick input was only about a 20 degree tilt angle which felt really awkward compared to what I am used to.

I decided to go ahead and cut down the axel shafts by 1" on both ends, so it will now be a 24" ball instead of 26". That is still quite large.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Last edited by Ran D. St. Clair; Aug 10, 2014 at 04:31 PM.
Reply With Quote
Old Aug 10, 2014, 09:47 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
Ball cage hubs made from 1/14 aircraft ply, 2" diameter, drilled 12 places, 5/64" dia. to a depth of 1/2". The center hole is 1/8". Plywood sealed with thin ZAP CA including inside all holes and then holes re-drilled.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Reply With Quote
Old Aug 11, 2014, 08:09 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
The front and rear visual orientation aides have been installed, the white 3" disk in front and the black fin in the back. Both are 1/16" plywood. The fin in the back is glued on with Goop. The disk in the front is mounted with a screw to the 1/4" O.D. aluminum standoff with a 4-40 internal thread glued into the end of the fiberglass tube. The side tubes have the same standoff for mounting the cage wheel hubs.

The cage wheel hubs have also been painted and installed, red on the left and green on the right. I used two 4-40 internal hex head bolts with a 1/4" unthreaded shoulder as a sloppy bearing surface inside the 1/8" hole of the cage wheel hubs. A washer under the hex head was used to provide additional support. These bolts will be installed with removable Loctite when the time comes.

The big unknown now is how stiff the cage will be. It needs to be stiff enough to protect the quad, but compliant enough to bounce without breaking. An individual hoop of .063 CF would be way too flimsy, but with all of the hoops tied together they should reinforce one another. I don't begin to have the skill to model something like that mathematically, so I will just have to hope I guessed about right. I can always order more CF with a larger diameter if necessary.

Meanwhile it will still be a few more days before Aircraft Spruce gets stock of the CF I ordered, and then it will be a week or so for delivery. There's nothing to do but wait, and perhaps have some fun flying it without the ball.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Reply With Quote
Old Aug 11, 2014, 08:33 PM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
805 Posts
A picture of the prop tip spacing.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Gaui X3 Build Thread TucsonFlyer Mini Helis 228 Aug 15, 2014 11:52 AM
Data WWI Forum Index With Build Threads & Reference Information Thread Links Murocflyer WWI Aircraft 21 Jul 08, 2014 11:11 PM
Build Log X-UAV Talon build thread Thermicdude FPV Aircraft 56 Apr 13, 2014 09:56 AM
Discussion My first Quadrotor build. (Changing this from a Q&A thread to a build thread) Deserteagle83 Multirotor Talk 10 Feb 26, 2014 12:21 PM
Build Log Spirit of St. Louis plan build Edit: Moved Thread piperjoe Giant Scale Airplanes 0 Sep 18, 2013 08:03 AM