Quote
Dannydefe
Correct me if I'm wrong ,but this shouldn't change any of the geometry using the fork design.
Double shear joint reduces error:
There is the potential to print a more accurate part, and reduce the error with Lisa if you use an inline double shear arms (i.e. eliminate offset-shoulder),
I believe it will save you 12,000 CPU cycles, (4000 cycle hit per motor x 3 motors = 12,000 CPU cycles).
Pickle fork saves CPU cycles:
Using a pickle fork to constrain the nut from rotating eliminates the g-code preprocessor, and I think will speed things up as well.
To prevent the nut from rotating, substitute a rigid metal column for the PVC column.
Fabricate a bracket to carry 2 bearings to ride up and down the metal column.
Quote
uncle_bob
“Since the main impact is from the tan function,”
Quote
Hazer
“For LISA, the shoulders and the hub offset stays in line with the arm in the X-Y coordinates.
They can no longer be ignored like the Rostock does. By throwing them back into the equations,
you would need to perform two more squares and one more square-root per motor.
With the threaded rod offset, you would also need to add one SIN/COS equation to compensate the movement of the thread.
That unfortunately is a 4000 cycle hit per motor.
So, Marlin Rostock would be 3120 processor cycles per move at 16MHz would take 195 usec.
Adding LISA calculations to firmware would change to 18480 cycles, or 1.1 milliseconds.
I have no idea what impact that could have.”
