PmWiki | Lab2 / L2G4Q4

Name: Chris Merrill (Red)

Purpose: To determine the function and abilities of the Create's battery-and-power-related sensors.

Here are graphs of all the data I collected, with the wheels moving at 100mm/s, no load:

Charge (blue) and Estimated Capacity (Green) in mAh vs. Time - Image.

I used the stream command to get my data, which sends packets every 15 ms. The data table for that would be too large for here, so you can download the datafile here. I used sensor packet 3, so it sends back:

[19] [a byte] [3] [Packet 3 data, 10 bytes] [Checksum byte]

The data was plotted with MATLAB, using this script file to generate the matrices. You'll have to remove the "txt" extension if you want to use it, though.

Voltage (blue, mV) and Current (Green, mAh) vs. Time - Image.

Results: As can be seen by the above graph, the estimated charge remaining for the battery is inaccurate when compared to the actual remaining charge. The charge also seems to go down at a constant rate, which is to be expected as the current (seen on the other graph) remains constant, ignoring noise, at a given current.

The voltage measurement, on the other hand, did not seem accurate at all. It followed a series of plateuas, first decreasing and then jumping back up to a higher level. Given that the current is constant, and there are no external factor affecting resistance, it is highly unlikely that the voltage would have any reason to spike that way.

The temperature sensors were giving roughly constant readings throughout the experiment, at around room temperature, so I thought it futile to include a graph of that data.

The only other data returned by packet 3 is the charging state, which for the duration of this experiment was 0, meaning normal use.