Hi Kelly,

Good news, I have fixed the control configuration of vane deflection to move opposite to each other for top and bottom vanes. I also did what u told me and varied the CG of the sphere. Shifting from 5cm to just 1cm below the Meridian, I was shocked by the stability of it. In fact, bringing it closer to the Meridian, improved the stability of the sphere drastically. However, a mere 1cm above the Meridian gave the same 'slipping' motion observed in all my other videos. This is in line with sam's sphere, where his CG is jus an inch below the Meridian.

PID 1 8,0 6,1 0 CG 1 cm below center No change in flap size (0 min 20 sec)

After some tests, i felt the the bottom vanes could be too small to counter the yaw effect, therefore i tried to increase the vane size. with the same position, 1cm below meridian, i tested and found that it was unstable, so i decided to lower the CG slightly. And it got better.

Bigger flaps PID 1 8,0 6,1 0 (1 min 22 sec)

I kinda deduced that, since an increase in flap size will increase the overall Cp of the ball, and therefore with the preceeding position of the CG, there was more top 'toppling' effect causing the instability. By the lowering the CG, it will then decrease the effect.

But i feel that the bottom vanes are slightly too big now, and im intending to decrease the size by bout 1cm.

Although raising the CG close to the Meridian was kind of a breakthrough after last few days of unfruitful tests, Im still puzzled by random slippings that can still be observed. Ill probably work on the balancing the craft vertically first .

Any comments or input

Hello Tim, great work there! You from NTU doing a project?

Originally Posted by tim2chia88

But i feel that the bottom vanes are slightly too big now, and im intending to decrease the size by bout 1cm.

Tim,

I would suggest that you not trim the control surface, rather reduce the deflection of the vanes.

Keep in mind that the upper vanes are much closer to the prop, thus they receiver air of slightly higher velocity >> more energy contained in each volume of air>> .

Try reducing the travel of the upper vanes first.

More shortly......

Errata:

My mistake....when you posted your orginal CG position as 5 cm I read into my mind 5mm, an order of magnitude off. I'm sorry I should have caught that "overly low CG".

For all other readers please understand that the image/description of post #289 has an error, the orange upper vanes must be reversed as Tim confirmed in last test flights.

Kelly

Tim,

Sam used two cheap piezo gyros with the P term only on pitch and roll, and a heading hold gyro with P / switchable to heading hold>>PID in some combination.

You ask about how to make the first basic test flight and your description was slight wrong, it should be flown with the "P term only" on the pitch and roll gyros gyros for the first flight and the yaw axis gyro turned off, mainly to setup the yaw control vanes off-set required to counter the motor torque. You want to have some mechanical off set into the vanes at all times since this craft uses very high power at all times with only small changes of power during any flight maneuver. After the off-set has been mechanically set to hold the yaw axis at one throttle setting you can switch the yaw axis gyro back "on" using only the P term, check to see if it has the required vane deflection to handle a full power burst, if that works okay they you may want to add the I and D term for heading hold effects.

Is it possible to set the various PID terms independently for each channel on your board? If not, this makes things much harder than using 3 cheap mems HH gyros.

Keep in mind Sam used only the P term on pitch and roll controls, and PID on yaw control.

After this setup is done then we need to chat about spoilers.

Kelly

Tim, would you please return the beginning of this thread and read my discription twice about how the spoilers work, their function, and how to set them up. This is too much data to repeat.

In brief the spoiler will dampen oscillations of pitch and roll, slowing things down and giving the flight controllers (both you and the control board) time to react.

Since the spoilers are far above the CG they impend the tipping forward or aftward pitch and roll motions, put only if they deploy downward and at the correct velocity of activation (CVA).

The spoilers surf on the in-flowing air entering the side of the propeller, when the craft tilts left, right, forward or aft the airflow is disturbed dropping the appropriate spoiler or combination of spoilers, impeding further tipping motions, thus it dampens the motion of the craft giving a feeling of solid response.

Your spoiler are too light of weight, they surf, at all times,they do not deploy downward, thus they do no meaningful work to dampen your sideways motion(s).

You need to glue small steel washer at the midpoint to the span of the spoilers on the outer lip edge closest to the prop, or even better glue two at a time, one on each extreme end,both of the span and distance away from the hinge line. You need just the right amount of weight to have the spoiler break downward with the slightest tipping or translational flight.

Kelly

Tim,

Would you please perform a simple smoke test for me and post a video of the results? This will be off great value for you and many others.

Part 1

Bundle 3 or 4 sticks of incense with a rubber band, light them all at once. Place your sphere on a table in a very calm room, set the power to hover speed, hold the smoke sticks up to the sphere in line with the prop, about 30 cm out and slowly move towards the prop, next lower and then raise the smoke to show the flow into the side of the propeller disk.

Part 2

Have an assistant hold the sphere while you hold the incense, place the incense about 10 cm or less directly in line with the prop on the right side of the sphere. The smoke will flow into the side of the prop, now both of you must move together to the left and then to the right, repeat this several times; going to the right the smoke will enter the disk, going to the left the smoke will not not enter into the propeller disk. You may have to adjust the distance that you place between the sphere and the smoke stick, but you will be able to see exactly how the spoilers work, this will help you and many others to see the holding force that makes the spoilers surf, and the resultant force that allow the spoilers to deploy downward.

Thank you,

Kelly

I try to build a model like this, but how should I programme my transmitter. Which servo's should be given to which channel. I try to do this with a hobbyking multirotor control board V2 or 3.
I could take my OpenPilot CC board out of my drone, but I would not like to do this.
Can you please advise me?

Hi from Oklahoma, USA! ( I'm from Malaysia)
I have been reading this topic for a while and finally decide to join this forum. I have started to work on my flying sphere in Oklahoma State University for the past three months and called it Unmanned Flying and Rolling Orb (UFRO). I really enjoy reading all the comments here and especially made by Kelly and Sam!

I have built 16 versions for the past three months to test out different construction materials like foam board, composite using fiber glass and carbon fiber, and etc to build a durable sphere. I have been trying to use the Ardupilot (APM2.0) to provide stability ( and autopilot in future) and finally decided to quit yesterday after many failed attempts to achieve hovering stability. Instead, I then started to use the rate and HH gyros like what Sam did on his 1st Taiwan Ball.

This is what i've got so far using the same servo configuration as shown in post #176. I will add "skirt" and spoiler later.
Currently, i am still having issue on maintaining hovering stability. It tends to drift and spiral downward after few seconds of hovering.

With Gyros - UFRO version13 A
No manual control yet. Just testing the hovering stability using the gyros

UFRO v13A (1 min 50 sec)

Diameter : 20 inches
Propeller : 11 x 4.7 R (i am using one of the contra rotating prop and attached to motor, this prop fits into the 5mm shaft and that is the reason i am using this prop! )
Weight : 824 grams.
CG : currently 1.5 cm ABOVE the meridian. Ideally i want it at the center of the sphere.

With Ardupilot (APM) - UFRO version 13

UFRO v13 FT5 (2 min 7 sec)

CG: 1.5 cm BELOW meridian.
Weight : ~ 960 grams.

With Ardupilot (APM) - UFRO version 8.4 - just flying around and testing the durability of the sphere~ , not good stability from using the APM. Just turn on the throttle and let it fly away to see what it does.

UFRO -version 8.4 (2 min 45 sec)

UFRO - Version 8.4 (0 min 45 sec)

Hope to hear some comments regarding UFRO version 13A. Again, I will add the "skirt" and spoiler later.
Sorry if i have confuse you guys with different versions here!

I will post some pictures later~

Very impressive
guys tried to scratchbuild their flying ball drones,but mostly failed dued to the weight awareness,people did not realized the weight is the key to the success of the flying sphere,suggest to keep RTF weight as low as possible,ideally below 400 grams.
hope COROCOPTER,my dearest friend (Kelly) is still around because HE IS THE ONLY GUY KNOWS THE ROOTCAUSES
regards
btw,loved KL
Sam

Agree! thanks to Kelly expertise and advise, plus your Taiwanese Ball, it sure give me the encouragement to build this sphere. This is actually my first RC experience even though I am an aerospace students design and built UAV. I have experience in composite material, in-flight deplorable inflatable wing technology, and Space Habitat. I have built UAV for the Annual Design, Build, and fly competition, just never flown any RC plane before
UNfortunely, I am not from KL lol
Btw, weight is not an issue in my case. Stability is
Since this is my very FIRST RC plane, I was very scare everytime the sphere took off away from my hand and have tendency to turn off the throttle immediately.
Thanks to my ZERO Rc experience, I actually severely injured my right hand last November while testing a himax contra rotating prop when it suddenly spins in full throttle because I accidentally turned off the TX before unplug the battery from the ESC. This incident sure gave me "propeller phobia" for a while now, I am fully recover and getting better on dealing with spinning propeller. However, i am still treating the sphere like a baby, afraid to let my hand go and let it fly away. Hehe.
Well, what didn't kill me sure make me stronger! Lol
So , building the sphere is a strong statement to make the point that an exoskeleton can sure protect object and person from propeller blade!

anyway, The heaviest version I had was about 1200 grams (overbuilt) , still have sufficient thrust and most importantly the sphere can survive an 5' - 7' height drop I only used 60% thurst on version 13A

UFRO v13A - Flight Test #4 (3 min 7 sec)

Fixed the spiral issue still have to adjust cg to see how it affects the stability.
No spoiler and skirt here
I am just adjusting the throttle here and let the sphere fly.
I am waiting for another v-mixer to have manual control

Hello Sam, and Ben,

Thank you for the kind words, I watch this thread with keen interest.

Ben, Welcome and thank you for posting your vids and details. I'm impressed with your workmanship and perseverance.

I have several suggestions or thoughts that maybe of help, but first can you tell me if your are use a single propeller or a counter rotating pair on version 13A? I was not sure from your description.

Two thoughts:

A single large diameter slow turning prop will work to increase thrust and stability.

Idea #1.

First thrust: ..... you want to slowly accelerate a very large mass of air (i.e. a large diameter slow turning single prop) as apposed to rapidly accelerate a small mass of air (smaller diameter twin fast turning props)

Idea #2

Second stability: The large diameter prop will create a great deal of gyroscopic stability that combined with the mechanical off-set of the control vanes will create a sort of localizing stability, much like can be seen when you view a full size or large glow powered helicopter viewing it from the tail end in hover. What you will notice is the fact that the helicopter will be tilted downward on the right hand side, this is due to the resultant vector forces of the tail rotor forces interacting with the main rotor torque. In the spheres you don't see this tilting since the control vanes are on opposite sides of the craft yet there is still a resultant force vector that created a stabilizing force.

I'm expecting some very astute poster like Brandano to take me to task on idea #2.......but the idea is much like standing on polished ice with leather soles, if you have counter rotating props; the lightest motion and you skid all over the place.

On the other hand a single prop with off-set vanes is like the above person but imagine the person is holding onto the handle of an old fashion kids wagon, the person will have a very slight increase in stability, not much, but just the amount that slows things down so the human sense of balance can react against the wagon handle and allow the person to move across the ice in small steps.

Please post photos of all of your versions and some details to allow us to learn from your hard work.

Many thanks,

Kelly

It all depends on what you expect from the plane. If you want it to be self leveling it needs to know where "up" and "down" are at any given time, and you can't do this with a pendulum because the plane will experience all sorts of accelerations in addition to the gravitational pull whenever it is disturbed from level flight. What you can obtain with rate gyros is generally just dampening on one or more axes, which means you will still need a pilot to bring the craft back to level when it has crept to an angle over time. With an inertial platform each acceleration is integrated over time to calculate the current position of the craft (if I got bumped this way and then got bumped this other way I must be here) but again errors do creep in over time, depending on how good the instrumentation is. the best bet is to implement some form of feature tracking on a camera that will recalibrate the inertial platform while it runs, like cruise missiles do when picking up terrain features to adjust their intended trajectory. This is a bit complex to add to a lightweight model. Helicopters, expecially the coaxial ones, do seem to manage to correct their position (within limits) relying only on a pendulum, but even then they will drift from level flight over time. Their configuration though limits the translational speed in a way that makes them self righting through aerodynamic forces.