Vertical Jump tester

[John Abel]

I would like to build a vertical jump tester like the one in my school’s weight room. This works by having the person to be tested step on a pressure sensitive pad. When the person jumps, the pad triggers a timer that stops when the person lands on the pad. A small computer determines the max height of the jumper from this measured time using the formula: h=(1/2)gt^2 where g is the acceleration of gravity, t = time and h = max height. This should be accurate as long as the jumper lands with his legs straight. <p>I would like to be accurate to the tenth of an inch and the difference between a 30.3-inch jump and a 30.2-inch jump amounts to only about .0007 seconds so I would need the timer to be accurate to the thousandth of a second at least. This would need a clock frequency of 1khz and some sort of counter circuit that could accurately register this rate.<p>I realize some kind of basic stamp module would probably be the easiest way to accomplish this. However, not having a micro controller, I was hoping there would be a cheaper method. I figure a crystal oscillator could serve as the clock frequency, but I would appreciate ideas of what to use as a high-speed counter for the clock pulses. Once I know how many pulses occurred I could equate for time in seconds and then determine the jump height. <p>Also, any suggestions for what to use as a pressure pad sensor would be appreciated.

[jwax]

First thought is to use a light beam that gets interrupted by a foot, rather than a pressure sensor. Easier to detect, cheaper and provides a definite on or off signal.
Position the beam right at ground level, and build a start/stop timer. Most frequency counters have a timer function, with a start and stop input. The start pulse opens a gate to allow the counter to count reference pulses, and the stop pulse closes the gate. The number of clock pulses counted (could be millions depending on clock frequency) indicates "time interval". Airtime.

[Robert Reed]

Why not just a simple mechanical switch built into the jumping platform ? This could be simply a spring loaded device (not enough to give the jumper any advantage) with narrow travel limits (1/16") . Just add two leafs--open and close. Arrange your electronics to accomadate this action.<p>[ October 11, 2005: Message edited by: ROBERT REED ]</p>

[philba]

With out a micro, I think you are going wind up with a lot more electronics. You need a timer and a counter and some sort of display driver plus the display. I think a stamp or picaxe would work well. The picaxe would be reasonably cheap. A PIC or AVR would be very cheap. I think all of them can handle the timing issues.<p>The switch thing would work though I like the idea of a simple accelerometer. It avoids alignment and stress issue with the switch as there is no contact needed. the analog devices ADXL line would work and free scale has a 3 axis one now but that is probably overkill for this application. When you see a spike in the acceleration, they jumped. When you see another spike, they've landed. Most accelerometers are easy to interface to a BS or just about anything else. A little platform over a fairly stiff foam pad (dampening) should work pretty well and make it very reliable.<p>Another sensor that may work is a shock sensor. There is a little metal ball in a chamber. the ball shorts 2 contacts. The contacts go open when there is a shock. Fairly omnidirectional sensitivity, as I recall. I see them surplus sometimes for a buck or 2. They were used in older car alarm systems.<p>You can also find pressure sensitive pads that are used in burgler alarm systems - very easy to interface to. I think they are fairly spendy.<p>Breaking a light beam would also work but it should be in a platform so they don't have to land in exactly the right spot. Vertical platform movement would break the beam. A phototransistor is simple to interface to anything.<p>Phil

[John Abel]

Do you have any recommendations for a well-valued microcontroller starter kit? Also, philba, can you elaborate on your accelerometer set up? I haven’t ever worked with these devices and don’t quite follow what goes where.<p>[ October 11, 2005: Message edited by: JohnT ]</p>

[philba]

I'd suggest either a basic stamp or a picaxe. http://www.rev-ed.co.uk/picaxe/ <p>An accelerometer measures, well, acceleration. Moving the accelerometer causes a change in acceleration. A 2 axis device will have two outputs, one for each axis (say x and y). The current acceleration on each axis is output as a voltage level. You can use an ADC in a microcontroller to read the Acc level for X and Y. If the value changes significantly from a "steady state" value, then you know that significant acceleration has ocurred. Thus, a jump happened. Similarly, a landing. The time between is the "hang time". Perhaps a better way is to start timing when you stop seeing accelerations and stop timing when you see them again. Maybe add a fudge factor for launch platform ringing.<p>In your case, you only have to deal with one axis which much simplifies the math. <p>The down side of accelerometers is that they are a bit pricey. The newer ones from Analog aren't too bad though. This one can be bought from Analog for $6. I'd think 5Gs would work for your application.<p>Phil

[John Abel]

Where would you mount the accelerometer?

[philba]

I'm no expert but I'd look into creating a platform for the subject to stand on and mount the accelerometer in that. maybe 1" high. 2'x2'? <p>The pressure sensitive pad approach may be better if you want something thin. Take a look at these<p>Phil

[rshayes]

The crudest, and possibly cheapest, accelerometer that I can think of would be a miniature speaker with a weight glued to the cone. Free if you salvage the speaker from a dead transistor radio.<p>[ October 12, 2005: Message edited by: stephen ]</p>

[jwax]

An accelerometer is used to measure acceleration- unnecessary overkill. You don't need to measure that- only the event of interest- on and off. Mechanical contacts will have a delay, maybe insignificant, but a delay nonetheless.

[jwax]

Here's a circuit with the verasatility to display whatever parameter you want- time, velocity, or distance-
http://mgc314.home.comcast.net/measurevelocity.htm

[philba]

If this is a one-off thing that isn't intended to last, I agree that a mechanical switch is ok. But they will wear - the amount of stress will be pretty high. what's the instantaneous force when some one lands? non-contact sensing is much better here. I like the speaker idea - probably want a schmitt trigger on the coil or comparator. The nice thing about an accelerometer chip is that they are really small.<p>If you can find a shock sensor cheap, that would probably be the way to go. Wish I'd bought a bunch when last I saw them. I found a module at AllElectronics that had an easy to desolder adxl202 accelerometer in it for $1.50. I bought a bunch but they appear to be all gone now, though.

[bsparky]

Paralex has a Piezo vibra sensor for $ 1.49 that will enterface with a micro real essay.

[John Abel]

These types of sensors (accelerometer, speaker, shock sensor) will be able to determine when the person jumps?

[philba]

no, the sensor only tells you something about what it's measuring. You need to create the electronics (or programming) that determines what those signals mean. In fancy terms, you need to do signature analysis to detect the jump and landing. In simpler terms, the signal from the sensor will have a "jump" characteristic and a "landing" characteristic. You can use this information to start and stop a timer. It's going to take a little experimenting, I think.<p>I believe all those sensors (and the mechanical switch, as well) will work in that regard. Your selection of a sensor should be based on what you are capable of doing as well as your form factor, cost, reliability and accuracy requirements.