Ball - sense that is falling

unigamer

Member
I hope I've used the correct forum...

I would like to know the cheapest way to sense that a ball thrown up is descending and send that fact to a PICAXE? I was thinking along the lines of an accelerometer but not only is this expensive but slightly complicated?

Thanks!
 
Vision processing software? Thinking maybe something like RoboRealm? A falling ball may be too fast for visual processing though.
 

Brietech

Senior Member
Are you trying to detect something is falling from *within* the ball (i.e. the picaxe itself is falling), or trying to detect an object falling *next* to your picaxe? If it is the latter, you could have the falling object break a beam of light.
 

boriz

Senior Member
Hmmm. Not sure if an accelerometer would work. From the very moment the ball leaves your hand, to the moment the ball hits the ground, the accelerometer will read a constant zero G. It cannot tell how fast it is going, nor what direction it is going.

Before you can get any useful help here, you will need to fill in a few blanks.

How big is the ball? How is it propelled? How high will it get? How far will it travel? Will it be inside or outside? Does the ball need to have the detector inside, or can an external sensor (such as has been suggested) be placed on the ground/wall/ceiling? What are you trying to do exactly?
 

unigamer

Member
Sorry, come think about it, I wasn't clear at all. This time!

I shall give an overview of what I am planning and that should be clear enough. I have a plan to build smart juggling balls. They will each contain a PICAXE and leds and will be programmable . As an additional thing I thought that having the PICAXE know if it was descending. This means the sensor must be in the ball. I did think an accelerometer might not work because of the ball spinning.

The project may come to nothing but if there are any other jugglers here they might appreciate how the idea would work.
 

ljg

New Member
Just a guess-

Place an ultrasonic beacon on the ground, then sense the falling and rising doppler pitch as the ball goes up and down. I'm not sure this can be done with a picaxe alone, but it might be done with some IC counters and timers.

nice new forum, BTW
 

boriz

Senior Member
This is a difficult problem. The best solution I have so far is this:

Each ball has a set of IR receivers in it. The TX is above the balls in the ceiling, or maybe on an overhanging pole. The RX signal strength is proportional to the balls height.

Not a good solution, but I have not given up yet.
 

Dippy

Moderator
Blimey, this is a tough one.
Great suggestions but each is tricky especially if you want to juggle in the open air.

Accelermoters/mass sensing in a spinning ball, not really a goer.
Ultrasonics + >2balls + jugglers hands/arms = a rather complex signal to process. My own work with dsPICs and FFT wouldn't even be fast enough.
IR receivers in a spinning ball. Mmm. Still wouldn't sense direction unless PICAXE code could track the signal strength cycle. The sensor would have to be perfectly omni directional.
Maybe a radio version or loop induction version along similar lines? Maybe not.


I really can't think of an easy solution (or any solution at all).
You've posed quite a project here unigamer, I predict this thread could get 'Nesbitt' in size.
Sit back and enjoy...
 

demonicpicaxeguy

Senior Member
the way i would do it is with a small gyro inside the ball that stays level and then you'd be able to use an accellerometer to detect if it going up or down it

another solution is to time how long on average while you juggle the balls how long it takes for the balls to leave your hand to when they get to the point where they start to come back down
then using a storng magnet and a reed switch for triggering a timed sequence the picaxe could do you'd probably have the most feasable solution, the assumption is that the juggling speed doesn't change by too much and of course the triggering switches trigger properly
 
Last edited:

BrendanP

Senior Member
The parrallax site has acclerometers which are easy to use with picaxe. Ive done it.

The memsic units are SMD but parralax has them mounted on DIP breakout boards.

I dont quite understand why the acclermometer wouldnt be able to be used for this app. The picaxe would read the accels output multiple times a second so it would know when the ball was stopped i.e. in the hand of the juggler then it would read acceleration/positive G's as he threw it upwards, then no G's as it stopped at its 'apogee' and then accleration again as it fell /negative G's.

You could use the accel to turn the thing on and off, it would power down if not moved for a while. So it didnt power up and down all the time as you drove around you could program the picaxe so it would take a hard, continuos shake for say 3 seconds to power it up.

You could use a 433 Tx in the ball and recieve it into STAMP DAQ so you could graph/study the accels output as you juggle the balls, you would see a pattern in its output and could then program the picaxe accordingly. Or drop the ball off a ladder and see what the output looks like when its falling and then throw it upwards and see what that output looks like in STAMP DAQ and go from there.

This isn't complex to do.
 

Dippy

Moderator
"This isn't complex to do" - what would the acclerometer output be when balls are going up and down AND spinning?
 

Jeremy Leach

Senior Member
Something simple that might be worth considering : Forget trying to work out the decent - instead time the duration between catches ... the decent will simply be the second half of this cycle. To detect catches might be a lot simpler because of the sudden shock of hitting your hand?
 

BrendanP

Senior Member
I dont think it matters, all he needs to know is when the ball starts moving each time, when its thrown up and when it stops at its peak. The hard stop when its caught in his hand on will show high G forces because of rapid deaccleration so he can use that point to have the picaxe to then look for accleration as its thrown up into the air again.

The G forces as the ball spins will look different than the rapid forces as its thrown into the air and the hard stop as its caught.

In the abscence of any other workable solutions I think its worth a try.
 

Jeremy Leach

Senior Member
I was thinking that because hitting hand is so sudden then there might be a simple mechanical switch solution to detect this rather than any complex accelerometer.
 

BrendanP

Senior Member
The acclerometer is very simple to use, you simply connect 4.5V and ground to it and connect one or two or three of its outputs to the picaxe's inputs and use the pulsin command.
 

Wrenow

Senior Member
Brendan,

I think you forgot one point of the physics. From the time the ball leaves the hand until it is caught, the ball is effectively under 0 G. So an accelerometer should be reading -0- throughout the throw. Unless it is sensitive enough to register the deceleration due to atmospheric friction. Also, most accelerometers seem to have an axis. I am not sure how easy it would be to read a spinning ball, even if it is sensitive enough to read the friction from the air.

Unless I am missing something. ???? Or my understanding of ballistics is flawed. ????

Wreno
 

Michael 2727

Senior Member
If you had a heavy object inside the ball (at the core) like a mouse ball, they are clay or steel coated with rubber, and micro switches which were triggered by the weight of the object. When the ball and weight were falling the weight would not appear to be as heavy as when resting. The same would apply once the upward inertia was lost on the way up.
The tricky part would be the switch alignment and tension to activate them.
 

unigamer

Member
Thanks everyone for their input! I suspected accelerometers would have problems because of the whole spinning thing but wasn't sure having never used or researched them. I'm going to keep thinking about it, I have a feeling there is a very simple and cheap solution...

I had thought of timing descents instead of detecting but I never thought of coupling that with a shock sensor of some sort. That actually could work pretty well.
 

flyingnunrt

Senior Member
Are you wanting to use it as an aid to teach others to juggle. ie. to flash at the apogee so they can get rid of the next ball???
If so then how about a ball that has contact segments similar to say a soccer ball (football to some of you)... capacitively coupled.?? when released it would start a count which would be stopped at the next catch.
The picaxe could then divide this count by 2 whist being held, and at the start of the next throw/count it could turn the leds on at the half count point.
It wouldn't be totally accurate as you would in fact be 1 throw behind all the time but this woudn't matter much if the throws are reasonably consistent, you could trim up the math a bit if the throws are not to high.

PS I didn't even own a computer in 1970
 
Last edited:

unigamer

Member
I wasn't thinking of it as a teaching aid but it would be an interesting experiment.

I like the idea of just being one throw behind, you are right, throws tend to be fairly consistent. I will research myself but does anyone have any suggestions for touch sensitive pads or whatever just now?
 

leftyretro

New Member
I got it. Hire an assistant to watch the balls and hit a switch when each reaches it's top of travel. See that wasn't so hard... ;)

Seriously this is a very challenging application. I would think nothing short of some external scanning laser that could track and calculate top of flight. Certainly not something a Picaxe could do in time. As far as self contained sensor that could deal with spinning rotation as well as vertical travel I can't think of a thing..

Lefty
 

flyingnunrt

Senior Member
You could have a piezo bepper in each ball with a slightly different frequency in each of the balls, in addition to the leds..

Can you bounce juggle?
 

steirny

Member
I think flyingnunrt is close.

I have used piezo crystals to detect ticking in a pendulum clock movement for the purposed of balancing the whole device. Think 1 ton tower clock here. These devices will create voltage spikes proportional to mecahnical stress. This effect is also used in piezo cigarette lighters.

It could work well in the catching shock mode with suitable pulse limiting and stretching for the picaxe to detect. Diode and cap for stretching. Resistance and protection diodes to both rails for limiting.

Much cheaper than accelerometers. Worth a try
 

boriz

Senior Member
Demonicpicaxeguy.
“time how long on average while you juggle the balls how long it takes for the balls to leave your hand to when they get to the point where they start to come back down”

GENIUS!.

I had not thought of this ‘timing’ method. It’s definitely the solution.

-

BrendanP.
Gravity acts equally on all matter. Every atom in the ball is experiencing the same force. When the ball is in your hand, it is experiencing two forces. The force of gravity and the push of your hand. (your hand only works on the bottom surface of the ball). These two forces exactly cancel, so the ball does not move. If the ball was hollow and contained a ball bearing , the bearing would be at the bottom. It is ‘trying to seek the ground’ but is prevented by the inside surface of the ball, which itself is ‘trying to seek the ground’, but is prevented by the force of your hand.

If you drop the ball from a height, while it is in freefall, there is only the force of gravity acting on it (and everything inside it). So it accelerates. And any instrument inside the ball (every atom) experiences the same accelerative force. The imaginary ball bearing just floats inside the ball. It, and the inside walls of the ball are experiencing exactly the same force. Check out the Vomit comet

-

Jeremy Leach.
YES!

-

I think DPG and Jeremy have solved it. A simple shock sensor inside the ball to register when it is caught. After the first two ‘shocks’, the pickaxe should be able to estimate the interval and base any subsequent light effects on that. Keep a reasonably constant rhythm, and your in business.
 

BrendanP

Senior Member
My work with the memsic 2125 accel involved it use in monitoring and evaulating animal movement. I didn't get into the physics involved.

Im reading the data sheet that came with the parralax DIP mounted memsic 2125 two axis unit.

It says "the memsic 2125GL is low cost dual axis thermal accel capable of measuing dynamic acceleration (vibration) and static accleration (gravity) with a range of 2 g's"

"measure 0 to +-2 g's on either axis; less than 1 mg resoloution."

It seems to me the only time the ball is at 'neutral g's' is. is floating is the split second in time when it is neither rising or falling which is when it is at the apogee of its trajectory.

At all other times in between launched for the jugglers hand and then being caught again it is either at postive g's (rising) or negative g's (falling). There fore it seems to me (at least in theory) that the accelerometer can do the job.

I conceed the problems of the ball spinning are problematic.

However as a practical solution you simple need to know when the ball is at rest in the hand as a start point. The accel would then sense it being launched upwards.

As other posters have suggested the LED color change could be done with time from that point.
 

mikek

Member
If the mesmic accelerometer works similarly (from a data output point of view) to Parallax's 3-axis accelerometer, then the static acceleration part causes trouble if you're trying to measure dynamic acceleration. Unless you can keep the device in a stable position, the "static" part measures tilt. For example, on the 3-axis device, if you lay it flat on a table, the X and Y values are 0, and the Z value (down through the chip) is 1. If you rotate the device 90 degrees around the X or Y axis (basically standing the chip on its edge), the Z value goes to 0, and the X or Y value goes to 1 or -1 (depending on the direction of rotation). I believe you'll see similar results with the mesmic, but I don't have one to try it with.
 

boriz

Senior Member
It seems to me the only time the ball is at 'neutral g's' is. is floating is the split second in time when it is neither rising or falling which is when it is at the apogee of its trajectory.
It's floating the whole time! From the moment it leaves the hand, 'till the moment it is caught, the force acting on the ball does not change. It’s a constant 1G. The earth does not know or care what direction the ball is going in. The force of gravity does not change. It’s constant.

Yes, the speed of the ball changes, and so does its direction, but the force you are measuring (gravity) does not change. So at no point during the balls flight will an accelerometer feel any change, regardless of it’s speed or direction.

See freefall and Microgravity
 

BrendanP

Senior Member
The device measures G forces as the ball accelerates and deaccelerates.
At terminal velocity or rest it cannot meausure.

If what your putting forward is correct we would not feel G forces when a car accelerates or brakes.
 

boriz

Senior Member
The device measures G forces as the ball accelerates and deaccelerates.
Yes. When the ball is going up, it is getting slower and slower. It’s decelerating because of the only force acting on it, gravity, 1G. The reason it decelerates is that the force is acting in the opposite direction to it’s movement.

A moment later, when the ball has slowed to say half it’s start speed, has gravity changed? It the force now 0.5G or 2G? Of course not.

When the ball stops moving at the top of it’s trajectory, it is neither decelerating nor accelerating, it is stationary. But the force acting on it has not changed, it’s still 1G.

During it’s descent it is accelerating, getting faster and faster towards the ground, but the force has still not changed, 1G. The reason it accelerates is that the force is acting in the same direction to it’s movement.

The ball has experienced changes in speed and changes in direction, but the force has not changed. It has been a constant 1G the whole time.

But when the ball comes to rest on the ground, or in your hand, THEN the forces acting on it HAVE changed. It now has two equal and opposite forces acting together that exactly cancel each other out so that the ball does not move. The force of gravity and the force applied by your hand. If one of these forces was slightly stronger than the other, the ball would begin to accelerate. For example during the throw. You apply a force (upwards) that is greater than gravity and so the ball accelerates upwards. Until the ball leaves the hand, there are two forces acting on it, but once it has left the hand, it’s only gravity again.


If what your putting forward is correct we would not feel G forces when a car accelerates or brakes.
No. The forward and backward push you experience in a car are nothing to do with gravity. Any speed change of the car must be the result of an applied force. In this case the engine or breaks. But your body is not connected to either. You are connected only by your back on the seat and your hands on the wheel. When the car accelerates or decelerates, the forces pass through these points of contact to accelerate or decelerate you.

Simple Newtonian laws of motion. A body at rest tends to remain at rest unless an external force is applied. A body in motion tends to remain in motion unless an external force is applied.

In the case of the car, the force is applied by the engine to the wheels, from the wheels to the car chassis, from the chassis to the seat, then from the seat to your back. Each one of these steps takes time and causes physical stress at the points of contact (bending/heating/crushing etc.). In the case of the thrown ball, the only applied force is gravity. It acts equally and simultaneously on every part of the ball inside and out. Until it is caught.
 

pbunyan

Member
what has been said about the accellerometer is true, but the idea for a ball inside with a microswitch may still work. as both the ball and the ball inside are falling at the same rate, the inner ball is effectivly weightless relative to the outer ball. just like in the international space station, your not really weightless, your just falling at the same rate as the space station (except with the ISS, its going so fast that when its fallen, it fallen beside the earth in an arc, at which point the pull of gravity is in a different direction, so it effectivly falls around the earth constantly in a circle)
 

boriz

Senior Member
You and I and everything on the planet are experiencing an accelerating force of 1G. But we are not moving. Why are we not zooming away at rapidly increasing speeds?

If we were in a space ship that was accelerating at 1G, we would be pressed against the floor and the ship would be getting faster and faster. From the occupants point of view, they would feel normal. Like they were at home on earth. With no window, there is nothing that would indicate they were moving at all. They would feel exactly as you feel, at home in your living room. After a few weeks, they would be travelling at half the speed of light, yet inside the ship, there would be no evidence of movement at all and the accelerometer will still read a constant 1G.

A sensor on the surface of earth, stationary on the ground would read exactly the same as a sensor on the space ship. The two sensors read the same 1G, yet one is stationary and the other is zooming away at ever increasing speeds. How can this be so?

When the ships engine is switched off, the acceleration stops, the speed becomes constant and the accelerometer reads zero. Everyone onboard will start to float around like astronauts. Imagine this ship is heading towards the earth. At what point does the accelerometer begin to read anything other than zero? Gravity stretches out into space a great distance, the outer planets orbit the sun, dragged around by the suns gravity. The outer stars of the milky way galaxy are around 50,000 light years away from the galactic centre, yet they too are dragged around in orbit. So when does the accelerometer begin to read anything?
 

BCJKiwi

Senior Member
I have some specific information for you. This is not based on any theory but recordings from accelerometers.

I have a gizmo which normally lives in my car mounted on the windscreen. It incorporates accelerometers and can be used in two modes - either for Drag racing (which I don't do) or road racing (which I don't really do either but do do some laps on the local track). When driving, it shows acceleration/deceleration and Lateral G's from cornering. It can also be set to record these g forces as well as RPM and time of day. It can therefore be used to analyse and compare different laps, cornering forces etc to improve ones driving and lap times.

If you have it in your hand and rotate it gently on any of its accelerometer axes, the indicators shift about as the g forces change with respect to the calibrated reference orientation - i.e. if it is tilted in relation to its axis then it will register a change.

If it is suspended on a cord so it forms a pendulum and is swung back and forward in the direction of one of its axes, it records no changes even though it is changing direction and accelerating and decelerating through each cycle of the pendulum.
However if you reproduce the equivalent action while holding it in your hand you need to change direction quite rapidly so you feel the mass against your hand resisting the change. Only then is the change of direction registered.

As for its application in a juggling ball, I don't see how it could be utilised effectively.
1. You would have to use a three axis accelerometer and integrate all the changes adding to the complexity.
2. All you would record is the changes of orientation of the accelerometers, not their movement in relation to the earth. If it bangs about against the inside of the ball you may register these 'hits'.
3. The movement of the accelerometer(s) inside the ball is unknown and variable and therefore there is no way to predict which of the accelerometers are giving you the information required - up/down or rotational. This would be the same if the accelerometers were free inside the ball or attached to the ball.
4. If a single accelerometer is used in a gimbal arrangement inside the ball then the only reading you will get is when it is stopped with sufficient force to register. The change of direction at the top of the flight would not register.
My casual observation of jugglers is that the 'catching' force is minimised with fluid motion - i.e. its not as if the ball hits the ground. The catch is soft and as close to the movement at the top of the flight as the juggler can manage so I'm not sure that the change of direction at the 'catch' would register either.
 

springer

New Member
A very tricky problem….

One solution could be taken from GPS technology
Place a triangular (three is the minimum, 5 is optimal) array of detectors on the floor (or wall or ceiling). These detectors are a known distance apart and referenced to a single clock signal.

Within each ball (at the center) is a transmitter, of whatever types you wanted, but I would go for radio. Each transmitter is set to a different frequency, so that each ball can be individually identified by the detectors, and no ball interferes with any other.

Each ball then has a fast free-running clock signal, which is constantly transmitted to all the base detectors.

The software then compares the ball id, transmitted clock id and base clock received time.

The time difference between base stations of a unique signal will tell you where, relative to each station, that signal originated from. The shift (Doppler) in time signal will tell you how far you are away.

With a bit more maths, it’s then a ‘simple’ timed / trigonometry problem to fix any ball in a three dimensional space.

This all works fine for cars n boats n plans, where accuracy is measured in meters, but for close up detection, its all down to the speed and accuracy of the clocks.
Hope this helps….
 

boriz

Senior Member
Unigamer.
Anything that can detect the impact of the catch should suffice. An accelerometer may be able to do this, but as has been pointed out BCJKiwi, the orientation will be random so this makes things a little more complicated. Personally I would keep things a simple as possible. That means making my own impact sensor with a bit of springy wire. Guitar string would do. Just solder two pieces of springy wire into the circuit board with one passing through a loop made by the other. Any impact should cause them to touch and complete a circuit. Time the interval between triggers, divide it by two and that’s the apex of the balls trajectory.

BrendanP.
Sorry, I do tend to waffle on a bit sometimes. My intention was only to inform. The physics are somewhat counter-intuitive and having now discussed this with some other friends, you are not alone.
 
Top