Re: Arm length calculation help on own-design delta needed

Your post contains a few errors:

Quote
allempty
I built a delta earlier this year and found that the most important point is to make sure the angle of the push rods from the ball joint centres is exactly 60 degrees to the vertical slides.

If you mean angle between a diagonal rod and the corresponding tower when effector is at position (x,y)=(0,0) then it is typically 30° but can be more or less. This number is not that important and has nothing to do with SIN_60/COS_60 in firmware.

Quote
allempty
Using Marlin as control sw :- got the angle wrong on my first attempt (58.62 degrees), I modified the angle value in config.h expecting the calculations to then be corrected.
This approach did NOT work so after very careful measurements, recalculated and extended my rod lengths to correct the angle to 60 degrees.

Of course it did not help. The SIN_60/COS_60 constants in firmware are related to computation of virtual tower positions and not to diagonal rod. They do not have anything to do with diagonal rods at all.

Quote
allempty
// Effective X/Y positions of the three vertical towers.
#define SIN_60 0.8660254037844386 // 0.8660254037844386
#define COS_60 0.5 // 0.5

These are for my machine it will not be correct for you, notice the number of decimal places for SIN 60 !!! rounding errors will cause problems.

Not really. You can round SIN_60/COS_60 to about 4 decimal places after dot and there will not be a noticeable difference.

Quote
allempty
I have errors in the Z plane when the head moves, the bed appears to be dished but with in 0.15 mm, a sort of doughnut shape ,but after the first layer is layed down the errors are compensated for.

No. You are just printing your parts distorted by about 0.15 mm. (That is provided your bed is planar.)
You should calibrate your printer better if you care about precision.
Either try Rich Cattell's firmware: [groups.google.com]
or try this approach: [groups.google.com]