XY Acceleration / Speed / Resolution on a Delta

bberger

As I know there are tons if smart people in here:

How can I calculate speed/acceleration/resolution for a delta on the XY plane?

As far as I understand in RRF I set the speed and acceleration, jerk etc values per axis (tower).

How can I translate those values to planar XY (and also take into account how far from centre this deviates)?

Same question about resolution (positioning accuracy)..

I'm trying to understand the effects and limits of the delta kinematic. I've always been fascinated by it and finally have built/modded a delta to my liking, but I'd like to actually understand those things and implications of print diameter, motors, reductions etc.

PS: is it just me or are the delta kinematics owly dying? All (and I mean all) the talk is about CoreXY now, you barely see new delta builds anymore which kind of is a shame..

Danal

@bberger said in XY Acceleration / Speed / Resolution on a Delta:

PS: is it just me or are the delta kinematics owly dying? All (and I mean all) the talk is about CoreXY now, you barely see new delta builds anymore which kind of is a shame..

Still plenty of Delta being built; they are mature enough not to need as much help. CoreXY fascination at this moment is primarily because of Toolchanging and specifically because of the E3D toolchanging printer release.

If you want to have fun thinking about delta kinematics, play with this for a while: Grab the extruder and move it around. Roll the mouse while in the graphic area, etc, etc, etc. Be sure and click "details" on the output side. That will begin to give you some hints on the topics you've asked about (it will not answer everything).

http://danalspub.com/DKcalc

Then, download the below file and open it in a text editor. It is the javascript for all the calculations in that tool/animator. Over 80% of it is dealing with Browser, and sliders, and graphics and... but the other parts are calculating the delta itself. In particular, see the functions that start with "calc...".

It is FAR from "optimized". It was specifically written to be readable. Long/meaningful variable names, calculations done on multiple lines when they could be condensed, things like that. Anyway, that could be a fun resource for the things you wish to calculate.

An astonishing amount of Delta kinematics can be calculated with right triangles, or simple planar trig. High school math. Again, there may be "more compact" ways, if we were really trying to write a motion planner or whatever... but to understand this kinematic, simpler math works very very well.

http://danalspub.com/DKcalc/js/deltacalc_de.js

bot

The accel, speed, jerk settings in RRF for deltas is based on the planar X/Y motion, not the towers.

bberger

@Danal thank you. That's really not a bad idea

Also thanks for pointing out that RRF values actually are for the XY plane. This just made it 'click' for me.

Delta kinematics are just so beautiful and fascinating to watch. It's and art form and I can spend hours sitting in front of it drawing up a stupid XYZ cube..

Still have to figure out the resolution part of it on the XY plane now.

AndreS

On a delta there is no fix resolution in XY plane. If one arm is nearly horizontally, the carriage has to move many steps to move the printhead a tiny bit, and is the arm nearly vertically then the printhead moves a 'huge' amount if the carriage move a little bit up or down.

bberger

@AndreS exactly. That's why I'm after calculating this for certain areas to get an idea of machine limits.

Danal

Your cases for calc are:

Center of platform.
Near a tower
Opposite edge from a tower. (this will also be the limiting case regarding not having a diagonal rod below 20 degrees).

This could be done in a spreadsheet...

bot

https://reprap.org/mediawiki/images/b/b5/Rostock_Delta_Kinematics_3.pdf

bberger

@bot said in XY Acceleration / Speed / Resolution on a Delta:

https://reprap.org/mediawiki/images/b/b5/Rostock_Delta_Kinematics_3.pdf

It's been a while, but thanks! That's exactly what I was looking for.

Let's see if I got this right:

if you care about XY accuracy don't print in the center but along the long sides. For Z go for the middle.
If you need equal accuracy on 3 parts either print them one by one or rotate each them by 120 degrees and place them at opposite towers.

Did I get that right?

Now that I know the expected max errors for different parts of the bed, is there something similar regarding resolution?