For fun, I started reverse engineering the file to see what I can see. Here's a snippet of the file displayed as hex:
(exported from frhed, and trimming off the first 0x14 bytes on a hunch (I think it's the header))
Code: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 see two different kinds of record here. The first 4 bytes looks like a timestamp to me (reverse the byteorder (little-endian))
018ab94b
019da94c
01b0a94d
I've also been looking for floating point numbers (put a 4 byte likely candidate into an
IEEE-754 calculator to see if it's a meaningful number).
My guess is that it's recording low resolution data quickly, or it's higher resolution data very quickly and they are multiplexing the current sensing.
I'm still trying to make sense of it. It would be super helpful to have a file from a robot under known conditions. Something like
Code:
power up
disabled for 30s
start teleop, but do nothing for 10s
run a motor on channel 1 forward and back in 5s intervals for 30s
run a motor on channel 2 forward for 5s
ch3 for 5s, etc
disabled
power off
I'll poke for a little while longer and maybe try to get a file off the robot at comp, but I probably won't have it decoded by the time a viewer is released.
I know the OP wasn't looking for a research project, but this is an interesting type of problem that can be solved with a little testing.