Rover 5 Encoder Output Measurement Notes

Wheel Startup

This shows me matching the wheels' encoder values using my Java program. This view is from the back of the robot.
I start each wheel off fast (~220 PWM/sec.), then slow it down to around 25 PWM/sec.
I found that a PWM/sec. change of 2 gives an EncoderCount/sec. change of around 8-12:

Encoder Settling Time shown in Java Program

This shows how I adjust the PWM/sec. in my Java program, and observe how long it takes the EncoderCount/sec. for each wheel to settle down to a stable value (the EncoderCount/sec. values are updated once every 5 sec.):

LF, LR, and RF wheels' Encoder values set with +/-2 of each other

This shows three wheels' EncoderCounts to be as close to each other as possible, i.e., over a spread of 5 in this case, and how they vary (+/-1) over time, even with no PWM/sec. change:

LR, RF, and RR wheels' Encoder values set with +/-3 of each other

This shows three wheels' EncoderCounts to be as close to each other as possible, i.e., over a spread of 7 in this case (I imagine that the width of this spread will vary as I check in various parts of the PWM/sec. range):

Adjusting PWM to get the Encoders to match

This shows adjusting the PWM values very sightly, in order to get the Encoder outputs to be as similar as possible.
It shows adjusting the PWM/sec. values to get the EncoderCount/sec. values to match, and how a small PWM/sec. change causes a 8-10 change in the EncoderCount/sec. value:

Max and Min EncoderCountPerSec and Current values for each wheel

This shows the max Encoder values and current for each wheel (The left-front wheel looks a bit more macho than the rest. The right-rear wheel has always used less current than the others):

This shows the min Encoder values and current for each wheel (The odd value for the right-front wheel isn't necessarily reproducible):

This shows the Robot moving in a Clockwise-Diamond Test Pattern:

This shows the Robot moving in a Counter-Clockwise-Box Test Pattern:

Back to Main Page...