Arc sides

[Generic Default]

Hi everyone,

I've been helping a friend make a delta printer for a while now, and just a few minutes ago we got the first successful (completed) print. It is a 20mm cube.

The sides were supposed to be flat, but the X and Y plane sides had noticeable arc to them. The square was slightly undersized (0.3mm) in X and Y dimensions and perfect to a thousandth in Z. The radius of the printer is well calibrated so all towers and the center are on the same Z level.

The best way to describe it is with a picture. I can easily change dimensions to fix X and Y undersizing but the arced sides are bugging me because I really haven't seen this type of defect before.

Can anyone explain why? My hypothesis is that the frame of the printer (non T-slot alum extrusions) is not consistent; the length of extrusions between the towers are not exactly the same. They should each be 12 inches but they might be off slightly since they weren't precision cut. Would this cause arcs?

Thanks!

[Polygonhell]

I think you'd get that if either the delta radius or the arm length were significantly incorrect.
The usually primary issue with home made deltas is arms that are not exactly the same length, it's a very critical dimension, but there are so many dimensions that can be miles off if your not careful. Did he cut the frame pieces by hand? Did he use a jig of some sort?

[Generic Default]

I just assumed that the radius would not affect straight sides as long as it was flat across the plate.

The frame piece were rough band saw cut by online metals, so they could be off a bit. The arms are more than 12 inches long so I can't measure them with my calipers. The magnetic arms on my own delta used to be a huge problem until I got them accurate to 0.04mm; now they're very good.

I'll check into the arm lengths and frame consistency more. It's not much of a hassle for me to mill stuff to a few thousandths as long as I can measure it and get consistent results.

[bot]

Is your effector very heavy? I bet it is ringing with a very low frequency due to the inertia of the hot end.

[Polygonhell]

My experience from Online metals is the rough cuts are pretty rough, usually slightly oversize and off by up to a 1/4 inch, so it might be worth at least putting a ruler on the sides.
Having said that I wouldn't think the frame would cause that, it would more likely be bowed outwards in some places and inwards elsewhere.
If it's low frequency ringing, or just plane slop in the effector motion printing something larger should give you a less consistent bow, i.e. you'd expect it to wobble as it "rings", wort a try printing say a 100x20x10 object.
The reason that radius can affect straight lines is that the overall curvature is affected by the arm length as well as the radius, so if the arm length is wrong (by a significant amount) and you correct it with radius you can introduce errors in Z on parts of the bed and X/Y all over it. Having sad that it would certainly affect dimensional accuracy, so 20mm wouldn't be 20mm.

[Generic Default]

The effector itself is very light since it's nylon with my Tri hotend on it, the weight of the effector is miniscule. However the arms are made out of stainless steel tubes so they are heavier than most. We did the test cube at a decent speed and acceleration of 2500 and the corners didn't show much overshoot ringing, so I don't think that's the problem. There is no slop since it is a magnetic arm system and the Z layers look very consistent.

I'll do a larger print when I have time today, the printer uses my old RAMBO board with a few blown inputs so we don't have a heated bed yet (large parts may warp).


I'm considering making a bunch of the printer's moving parts out of aluminum with tight tolerances so that future printed parts will be guaranteed accurate.