What Components needed to measure Velocity & Spin?

[tracer01]

Hi,

I have a project where I need to determine the initial velocity and spin of a football being kicked from rest. I want to be able to predict the flight of the football using well documented aerodynamic formula. The parameters I need to be able to plug into these formula are :

Initial velocity in the X, Y and Z axis.
The rate of spin.
The spin axis.

I would ideally like to be able to contain the measurement components within the football itself rather than using external sensors. I initially thought of accelerometers/gyros but not sure if these are the best options.

Please can anyone recommend any components which will enable me to measure these parameters or enable me to derive these parameters.

Many Thanks.

 

[strantor]

I would say accelerometers. I suspect that you could get rate of spin my measuring the centrifugal force; I imagine that it would have an upward peak every time the ball crosses the earth as gravity adds to the centrifugal force, and would have a downward valley every time the ball crosses the sky as gravity is subtracted, giving you a wave that you can measure the frequency of. The rest of it should work just like an iphone or a wii remote.

 

[tracer01]

Thanks for your reply strantor. I was looking at something like this - RD3965MMA7660FC for the accelerometer which would give me the linear acceleration. My main concern is how does an accelerometer operate when the object is spinning? If for example, the object had pure backspin, would not the accelerometer measuring the X axis, at some point be measuring the Z axis due to the rotation? Is there some sort of compensation for this built in to an accelerometer to deal with spin?

You mention in your post strantor the possibility of measuring centrifugal force to measure the spin. What sort of sensor is capable of measuring this? I have only come across gyros but is there something else?

Many Thanks.

 

[strantor]

I really can't go into specifics with no experience. I've never touched a gyro or accelerometer, only read about them. I don't think I can help much.

 

[kubeek]

Say you have the accelerometer sticked to the ball, z axis going away from the ball and x and y oriented in standard way. The problem is that the ball will not allways rotate in the same way so getting the actual spin will be tricky.
For one accelerometer on the "equator" it is easy to get the magnitude of the total acceleration. Unfortunately I don´t have time nor skills to investigate how to get any results from multiple accelerometers with freely rotating ball.

Just to add a different thought to this, getting a high-speed camera and some motion tracking software should get you going much faster than developing using gyros and accelerometers.
Do you need to predict the ball trajectory in real time?

 

[Sceadwian]

I don't believe a kicked ball rotates that quickly, even a regular 30FPS camera should be enough to measure rotation spin and tumble, this would not be an automated process though.
The only way to do it real time practical would be multiple gyros/accelerometers, I think you could get away with one on the equator of the ball, and one on a point, if they're both XYZ accelerometers you can extrapolate the true rotation using trig, a math student might be able to help you with working that out.

Keep in mind this isn't very practical, the football would have to be altered to protect the chips and that doesn't even include the circuitry to either record or transmit the data from the ball. The forces involved in kicking a football are NOT small.

 

[strantor]

Not to mention that putting circuits inside the football would change it's specific characteristics. You would wind up with a bunch of data on how a footbal laden with circuits behaves, not how a football behaves.

 

[Menticol]

Sorry if my idea is non practical, I have no experience with acceleration measurement.

What about a 3 camera rig, each one filming the ball from X, Y and Z axis respectively?

Football can be painted with a dot pattern to make tracking (and hence speed and rotation measurement) easier.

 

[Sceadwian]

Given the velocity and forces involved only one camera is needed Menticol, though the colored dots are a good idea for point measurement.
I haven't toyed around with it too much, but I've started using an opensource video analysis package called Kinovea that might fit the bill.

The only way I can see this project as being useful though is in real time, but that's outside the context of school. By the time you work out the math to determine the trajectory and destination the play will be over already and the outcome known.

 

[tracer01]

Many Thanks everyone for your responses so far. Very much appreciated and some interesting points made.

Firstly, in response to @kubeek - it is interesting about your comment about measuring the magnitude of the total acceleration - i.e. the acceleration in the direction in which the ball is travelling. If I could use a single axis accelerometer to measure this value and combine this with some other measuring device to give me the launch angles that would give me the translational accelerations in all 3 axis I require. Still need to figure out the rotational values though.

@Sceadwian - Your suggestion of using 2 xyz axis accelerometers is interesting - are you suggesting that with this configuration you could measure both the translational acceleration and the rotational speed? You are right about the sensors circuitry needing to be able to handle the forces involved. Accelerometers come with 2 ratings, firstly a 'measurement' rating and secondly a 'survivability' rating. I guess it is just a case of choosing one with the necessary ratings that would handle the forces involved in kicking a football. I have even come across an accelerometer the measures 60,000g with a shock survivability of 100,000g!!! As far as any other circuitry or components (e.g. RF Transmitter, Power supply etc..) I am not sure how this is rated or affected by acceleration/shock forces.

@strantor - Again, a valid comment regarding the specific characteristics. From my research on accelerometers, I have found that the accelerometer should weigh at least 1 magnitude
less than the object it is measuring. In this case, the complete measurement system would need to meet these requirements.

@kubeek, @Sceadwian & @Menticol - All your comments about using a camera system are very interesting. Using a single camera, is it possible to obtain the 3 translational velocities along with the spin rate and angle? I am fortunate enough to have a video camera already which as a 60fps recording mode. Would this be suitable with the correct motion tracking software?


With regard to the predicting in real-time, I envisage the the data being sent from the ball back to a PC and then processed (I already have a 4th order Runge-Kutta Spreadsheet set up to perform all the necessary calculations which is almost instantaneous) to produce a graphical representation of the ball flight.

Again many thanks for all your feedback. Keep the suggestions coming.

 

[Sceadwian]

Your suggestion of using 2 xyz axis accelerometers is interesting - are you suggesting that with this configuration you could measure both the translational acceleration and the rotational speed?

You should be able to, the math is over my head but having two separate accelerometers on a single object will provide all the information to calculate all the rotational forces it's undergoing, velocity might be a little trickier as you'd need this number to be extremely accurate and between noise/resolution of the accellerometer it might not be good enough.