Annirak Wrote:
-------------------------------------------------------
> Hah. Couldn't stop thinking about it so I did the
> math. You may have some trouble with the speed of
> the controller, because the coordinate
> transformation takes some heavy lifting as far as
> the math goes.
>
> ...
I haven't gone through to check your math but the geometry that I picked has the ability to greatly simplify the equations.
![]()
![]()
![]()
Note: 2b2 is a constant so it can be hard coded. I don't know what the firmware can do yet but you could even store 1/2b2 instead to avoid division all together.
That equation for theta will work for all the arms. You have to just add or subtract constants from the x and y coordinate so that (0,0,0) coincides with the middle of each swivel.
By my math that comes out to...
12 multiplies
12 add/sub
3 acos
...for 1 position operation.
I would like to use a RasberryPI's GPIOs to interface straight to pololus and hall effect sensors. (I think this will get the total under $100 for the electronics.) Proof.
If that proves too problematic due to interrupt requests causing the steppers to stutter, can this be done on the printrboard or a mega? Which one would be better?
-------------------------------------------------------
> Hah. Couldn't stop thinking about it so I did the
> math. You may have some trouble with the speed of
> the controller, because the coordinate
> transformation takes some heavy lifting as far as
> the math goes.
>
> ...
I haven't gone through to check your math but the geometry that I picked has the ability to greatly simplify the equations.
Note: 2b2 is a constant so it can be hard coded. I don't know what the firmware can do yet but you could even store 1/2b2 instead to avoid division all together.
That equation for theta will work for all the arms. You have to just add or subtract constants from the x and y coordinate so that (0,0,0) coincides with the middle of each swivel.
By my math that comes out to...
12 multiplies
12 add/sub
3 acos
...for 1 position operation.
I would like to use a RasberryPI's GPIOs to interface straight to pololus and hall effect sensors. (I think this will get the total under $100 for the electronics.) Proof.
If that proves too problematic due to interrupt requests causing the steppers to stutter, can this be done on the printrboard or a mega? Which one would be better?