Corexy belt path question

I am working on modding/creating a Corexy. I have read about the corexy kinematcs, but I am not sure I really understand all of it. Anyway, I am thinking to modify the belt path to look like the drawing below. The standard belt path is blue and red. I am wanting to modify the red path and the blue path both to move just the motors. They would then follow the purple path. There would still be 2 belts, it is just the positioning of the motors I am wanting to move. The plane of the belt path would not change, just the x,y positioning of the motors. Since the belts new red and blue belts would be the same length would the corexy motion still function properly?

Sorry for the crude diagram, I just did this quickly in paint.

If I understand correctly you want to move where the steppers that are in the diagram to what you have drawn in purple. if that's the case then no it will not function properly. The belts coming from the Y gantry back to the steppers needs to be exactly parallel to the Y axis. You could put an idler pulley where the existing stepper pulley is and then move your motors where you like.

Edited 1 time(s). Last edit at 02/24/2017 12:20PM by Ryan_M.

For CoreXY, everything in the grey zone is up to your imagination.

Edited 1 time(s). Last edit at 02/24/2017 09:34PM by prot0typ1cal.

What the others have already said is true and in addition I'd recommend against twisting the belts. It is very hard to keep the belt from riding the flanges on the pulleys, and wears the belts faster. If you must twist, your drawing shows the right place to put it.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Thanks for the help, so a few more questions.

1) I am just taking some/most of this from FB2020 corexy design. It shows a twist in the belts. On an I3 and on a delta I ran the ribbed belts on 625 bearings, and never had a problem with the ribs on the bearings. Never needed to twist a belt to get the flat side of the belt on the pulley. So can the same be done with corexy, even though there are more pulleys? If so then a twist would not be necessary. I also like the idea of NOT twisting the belt. So any opinions on if the diagram without a twist would work properly?

2) I am certain you are right about the motor placement, but I can not reconcile why? Just moving the motors to the purple position, if the motor moves the belt 1mm, then the carriage would still move 1mm - sure the belt is longer and you have to ensure your Y rods and Y carriages do not get pulled to the side, but if that is designed correctly, then why would this not work. I am sure I am missing something, and I am NOT an engineer, only a tinkerer LOL. I am just trying to understand this point better.

It's easy to understand why you need parallel belt to movement direction. In your picture move the X carriage up and compute the belt length. Move it down and compute the belt length again.

Another way to look at it:
- the movement axis is a side of a triangle S1
- the belt is the hypotenuse of the triangle I
- the distance from the motor shaft to the movement path is the other side of the triangle and it is constant. S2
- the belt angle between S2 and I (at motor shaft) is A

The length of the Y axis should be S1 = I but in your drawing it is S1 = I * sin(A). So, to have S1 = I, sin(A) should be 1, therefore A should be 90 degrees.

Sigxcpu, thank you, I think I understand now, and even if I calibrated the steps differently it would not work as it is not constant, it is the sine, which would vary so you would be unable to calibrate the number of steps as a constant. I hope I understand it properly now, thanks again

It would be doable, if one of the firmware gurus would implement a stepper position offset correction in firmware. Much like probe offset.
32bit controllers would be able to handle the extra CPU load.

The main issue with that belt arrangement is that the belt path length varies with the Y position. The belts would be very loose when the carriage as at the top of your diagram, and very tight when the carriage is at the bottom. This is because the belt path follows the hypotenuse of the right angled triangle whose other sides are S1 and S2 in the diagram 2 posts above, but as the carriage movsd down, the length of the hypotenuse doesn't decrease as fast as the length of the section of the same belt on the other side increases.

I have been using twists in the X and Y belts in my Cartesian printer and it has never caused problems. The important thing is that the length of the twisted section is constant, otherwise the twist will cause the belt tension to vary. Belt teeth running over plain idlers cause the belt to vibrate.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

@o_lampe: I would really like a software feature that creates belt lengths out of thin air

Quote
sigxcpu
@o_lampe: I would really like a software feature that creates belt lengths out of thin air

Jaaa, I was focused on the sin/cos problem too much..

I saw an older project by RichRap that has a similar belt path. I think he used cables to drive the motion system. I would look into his project.

Edited 1 time(s). Last edit at 02/18/2021 08:02PM by shanehooper.

---

[www.thingiverse.com]
[hackaday.io]

You can use cables, but you still have to obey the rules about the routing- the variable length segments of the cable have to remain parallel to the guide rails or the tension and steps/mm will vary. That's actually harder than it first seems because cables are friction drive and require multiple turns around the drive pulleys to ensure sufficient friction that it won't slip. As the drive pulleys turn the cable will "crawl" along the length of the pulley shifting the angle of the cable segment that goes to the pulley on the Y axis bearing block. There are tricks to maintaining constant angle but it gets complicated and dealing with cables turns out to be a real PITA.

I built a cable drive corexy mechanism for the first generation of my sand table and eventually gave up on the cable and went to stacked belt drive.

Here's a motor mount that maintains a constant angle winding and unwinding from the drive pulley:



Video of the operation here: [vimeo.com]

More details here: [drmrehorst.blogspot.com]

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]