What is a click?

I assume you mean a tick. There are approximately 1100 ticks in a wheel rotation, so there are 360/1100 degrees in one tick.
[shameless plug]If you were to use CBCJVM, we have a wheel class that could handle all of these simple calculations for you (and a MotorDriveTrain class that could calculate how to move about the board for you).[/shameless plug]

Hi Dan, welcome to the Botball Community!

Long theoretical answer:

I'm going to assume that you're referring to either back-EMF ticks or encoder transitions, both of which are often referred to as clicks. Neither ticks nor transitions are fixed in terms of how many degrees a motor has turned; it depends on the electrical properties of the motor (in the case of ticks) or the mechanical properties of the encoder (in the case of clicks). For example, if the back-EMF controller defines 1 volt millisecond (hope I have the correct dimension there) as 1 tick, then the voltage which the individual motors outputs will determine the ratio of degrees to ticks.

Short useful answer:

If you need to determine the ratio of degrees to ticks for a particular motor setup, I would recommend turning the motor about 10 rotations (3600 degrees) and note the measured position of the motor in ticks before and after you turn the motor. Some simple ratio math will give you the measurement.

Additional note:

If your ultimate goal is odometry (determining how far your robot has traveled), I would not recommend using degrees of motor rotation to find this. This is because that method only is accurate when the wheel slip is 0. Instead, I would recommend measuring the ticks per meter of travel rather than ticks per degree of wheel turn. This only assumes that the wheel slip is constant and consistent, rather than relying on it being 0. You will get better results this way.

Good luck!

PiPeep,

Yes, KIPR does claim ~1100 ticks per rotation, but only for the black gear motors. Other motors will have different averages, and individual motors vary enough that you are better off measuring your own values rather than relying on KIPR's rough specifications.

Jeremy: I use 1100 ticks per rotation, and then modify the rest through a motor efficiency variable (multiple for multiple motors/wheels), as shown here. I find it works quite well.

I try to do my math in such a way that a lower efficiency will yield a higher speed to compensate (aka, multiplying by inverses wherever necessary). So then, I can just tell the robot to move 1 meter (or 100 cm, as the CBCJVM movement library is cm based, I'm once again assuming you need this type of precision for your DriveTrain), and then I just enter the actual distance as the efficiency.

You to bring up some good points, and individually testing each wheel for a number of rotations is a great idea to ensure your robot moves in a straight line.

Oh, and welcome Dan Wang! I hope to see more of you around. You should join us on the ol' IRC sometime (see the Botballer's Chat link at the top of the page).

One other thing to note is that all motors inherently have a different number of ticks based on the mechanical variations in each motor when produced.

So while 1100 ticks approximates a good average for black motors, each black motor could have minor variations. One motor could be 1090 ticks per revolution and another could be 1110.

As Jeremy pointed out, the best way is to test each motor individually by gathering lots of data.