Dimensional accuracy problem

General Discussion
22

Well, I guess all arguments pro and con have been exchanged in this thread:

  • If you feel like doing it, then do it. It’s not like that you end in hell for being a rotation_distance heretic
  • If you want to trust the experienced users here, then do not bother with it. Likewise you likely won’t automatically enter the 3D printer’s paradise
1 Like
23

You seem to misunderstand the difference between absolute accuracy and relative accuracy, or the periodic error to be more specific in the case of 3D printer kinematics. If you are using 2GT (or GT2) belts and pulleys, then the nominal motion system “gain” is 2 mm per tooth. Your actual belt and pulley will have local inaccuracies, but as the travelled distance approaches infinity the tooth pitch will asymptotically approach 2.0000[…0] mm per tooth.

Correcting kinematic system periodic error or any printer geometry errors by adjusting the rotation gain is outright wrong. It will never achieve what some lore makes people believe they are achieving with such “calibrations”. The only place where step gain calibrations are applicable is the extrusion system, but that too is subject to periodic errors (a.k.a. wood-grain effects, etc.)

24

As I already said in the previous post, everything is clear and I don’t dispute the correctness of such statements, but like what you say “but as the traveled distance approaches infinity the tooth pitch will asymptotically approach 2.0000[…0] mm per tooth.” it is precisely inapplicable in my case where we are talking about about 20 cm, normally the error that is found is between 0.1 and 0.3, an error that can be easily compensated with a negligible rotation adjustment, almost all printer manufacturers they give the possibility to modify those values through the stock firmware.
However, I would like to thank all those who replied to my post, providing me with very useful explanations.

25

would you mind sharing where you purchased the parts (belt, pin, wheel and possibly the sprocket that attaches to the stepper motor) .
I cannot find a vendor for Gates

26

Hi. If you are in the US then PrintedSolid carries genuine Gates belts and pulleys. Otherwise you can source Gates belt from reputable vendors on Ali such as Powge or Fysetc and Mellow has good quality pulleys. The wheels came from OpenBuilds.

You can use anything you have on hand for the retention pins. I used a stainless steel rod that I cut to length. You can also use a small diameter (~2mm) nail or even non-metallic material since the pins are in radial compression. Just make sure that it’s not too long and has smooth ends as they can rub on the inside surfaces of the extrusion.

27

Thanks for the info.
Is 2mm to 3mm diameter enough to make the belt horizontal?
Did you do it also for the Y-axis?

28

I am going from memory here, but I think about 2mm dia will give you a quite parallel run. I think I used around 2.5mm and it was a tiny bit too much. Although that little difference should have virtually no effect compared to the original design.

The Y axis belt run was parallel on my printer so a similar approach is not required. However, I used the same fundamental pin approach to retain the Gates belt because the original Creality belt was crimped at the ends and I did not want to crimp the Gates belt. Instead of pins though I used M2 screws so that the screw heads keep everything in place since the Y belt is vertical.

1 Like