Re: [CAD_CAM_EDM_DRO] Re: A "simple" Hexapod project - envelope

Graham,

To decide on the envelope needed, there may be some good, closed form way
to do it, but I use a mangled version of my kinematics code. I put in
estimated geometry information (joint locations, leg lengths, joint angle
ranges), and then try moving the tool through the work
envelope. Violations indicate that I don't have enough range.

The output of this process is a very crude (very) ascii plot that shows how
much Z-axis above the reference 0 I have at each x-y position, for a
particular range of roll and pitch. I make a plot at 0, 15, 30, 45, and 60
degrees of pitch and roll. This makes for an interesting, highly
non-linear plot.

With this testbed, I can try different joint arrangements for the fixed
joints as well as the variablle joints. I can see the effects of leg range.

I also have (or have had) a mode where it records the angular range and leg
length range needed.

In short, I don't have any wonderful mathematics that drops out a design
for a given envelope, but I do have tools that let me experiment before
building.

Again, these are written in C and run on Linux in a purely text mode.

-- Carl