Timer Test Car
Posted: Wed Feb 06, 2008 3:49 pm
I know folks have talked about using paper or other objects to pass by a timer's sensors, looking for lane bias, occlusion factors, etc.
I am wondering if there is value in making a car designed to test multiple lanes?
It would be a basic flat board, the width of the whole track. Wheels in the outermost lanes. Probably feasible up to 4 lanes, probably too difficult to keep perpendicular beyond that. One end could be straight across to help see how 'ties' may look to your timer.
You could run the car backwards and the other end could be cut in a stair step pattern. Equally spaced so that each lane finished say 1/2" apart from each other.
Now this doesn't do anything to test overall accuracy from the start pin. But what you would expect that whatever the difference in time between 1st and 2nd, is a constant thru 4th place. Assuming you run the test car thru the finish gate at a constant speed. Maybe some lag starting to show between 3rd and 4th as the car deaccelerates.
The stair step pattern could be a separate piece of masonite and be screwed to the main car deck. Thus allowing for a collection of the 'profiles' with different finish orders and spacing gaps.
I guess the question is, what would it prove or show? It would only be good at showing relative time differences. And what is the likelihood of those being off...
What if you could go further, and add an independent timer to the car. Something where it actually runs from the starting pin like normal. Perhaps with a push button engaged at the pin. The gate drops, the switch opens and starts a high resolution timer running from some PIC and 9V on the car. But how to detect finish line? Most folks go sensors in the track with lights above. This means trying to put sensors on the car and ignoring ambient light until passing under the IR or LEDs at the timer.
Now the two times would likely never match exactly. The difference in just the two starting switches could be a millisecond. But again over a few runs, the car's time should be the same relative difference from your timer each run.
Fun design candy, but would it really be practical or useful?
We have four 'ol cars we use to test the track with when we setup. After only a few runs, one can learn the expected finish order, as they each run pretty consistent. This got me thinking to making the car interval fixed and known, and not just helping verify that finish order was being recorded properly, but that indeed the times were 'in the right ballpark'.
If nothing else, it helps visually show your audience that you are making attempts to detect inaccuracies.
I am wondering if there is value in making a car designed to test multiple lanes?
It would be a basic flat board, the width of the whole track. Wheels in the outermost lanes. Probably feasible up to 4 lanes, probably too difficult to keep perpendicular beyond that. One end could be straight across to help see how 'ties' may look to your timer.
You could run the car backwards and the other end could be cut in a stair step pattern. Equally spaced so that each lane finished say 1/2" apart from each other.
Now this doesn't do anything to test overall accuracy from the start pin. But what you would expect that whatever the difference in time between 1st and 2nd, is a constant thru 4th place. Assuming you run the test car thru the finish gate at a constant speed. Maybe some lag starting to show between 3rd and 4th as the car deaccelerates.
The stair step pattern could be a separate piece of masonite and be screwed to the main car deck. Thus allowing for a collection of the 'profiles' with different finish orders and spacing gaps.
I guess the question is, what would it prove or show? It would only be good at showing relative time differences. And what is the likelihood of those being off...
What if you could go further, and add an independent timer to the car. Something where it actually runs from the starting pin like normal. Perhaps with a push button engaged at the pin. The gate drops, the switch opens and starts a high resolution timer running from some PIC and 9V on the car. But how to detect finish line? Most folks go sensors in the track with lights above. This means trying to put sensors on the car and ignoring ambient light until passing under the IR or LEDs at the timer.
Now the two times would likely never match exactly. The difference in just the two starting switches could be a millisecond. But again over a few runs, the car's time should be the same relative difference from your timer each run.
Fun design candy, but would it really be practical or useful?
We have four 'ol cars we use to test the track with when we setup. After only a few runs, one can learn the expected finish order, as they each run pretty consistent. This got me thinking to making the car interval fixed and known, and not just helping verify that finish order was being recorded properly, but that indeed the times were 'in the right ballpark'.
If nothing else, it helps visually show your audience that you are making attempts to detect inaccuracies.