So I’ve got a question about deltas for those who are more experienced on the subject matter than I. Is there an optimum ratio of arm length to machine size (height, or diameter?).
I’m noticing that the end effector on the machine I’m working on tends to stay level near the center of the print area, but actually tilts (ever so slightly) when near the edges. I’ve measured the rod arms, they’re completely parallel to one another, equidistant around the perimeter of the effector, and the same length.
Note: This is also on a machine that I have guaranteed that the linear rails are also equidistant and aligned perfectly. What else could it be?
Gotta have some give somewhere @ThantiK . If there’s just a tiny bit of play in any of the joints, as the one or more arms get closer to horizontal there less bracing action. Linger arms are better for that, but, then there’s the flex from longer arms vs shorter. What a conundrum eh? What kind of universal me hanism are you using? Maybe some type of backlash spring/rubber band to pull the arms together
@J_O11 the ends we’re using are very similar to the male-threaded one in this picture: http://3.imimg.com/data3/ON/LS/MY-7898592/rod-end-bearing-500x500.jpg – it is rather close to their extents. The arms are made out of 8mm OD aluminum rod, 6mm ID. They’re epoxied into place so that the threads don’t change their length any.
The ball joints are custom-toleranced joints from IGUS made for us so they don’t have any slop, and just enough preload.
Those are nice ends. Check if there’s any clicking when you wiggle them. It’s an angular thing. If there .001" play in them, it shows more at the platform. It’s more an issue with the connections to the platform than the carriages if you get any platform “rock”.
What if I’m reaching close to their extents? (left/right) They do tend to stiffen up a little, but there’s no deflection in the rods that I can detect.
@ThantiK check the spacing between the pivot points of the joints at the effector and compare to the same at the slide end. These must be identical. If not you will experience this behavior.
Ive had some problems like these, but have had different arm lenghts than specified in fw. It was not much but enough to make problems. Have milled miself a jig so I can make sure the arms are really what they shoul be, not just all the same lenght but also the right lenght.
Aslo, like @J_O11 suggests, make sure you have some clearence in joints even after reaching your max travel. These can also make this happen.
On the other note, have you found any solution on the other problems you encoutered (unstraight printing in vertical)?
@Matej_Rozman , yep, solved those. Was an easy fix too. I had just overlooked it at the time because it never caused much problem on my cartesians and the machine I’m working with is pretty closely toleranced. The wobble was actually belts/pulleys not being completely planar with the makerslide and belt tension changing as the 3 carriages slid up and down the bot. Now, I’ve figured out that the auto-leveling problem I’m having is a result of the end effector not staying completely level. Our hot end and probe are both mounted off-center from the carriage, which is why we get a lift on one end, and a dip in the other. I’m trying to figure out a simple fix so that people can print to the far-extents of the delta. And those who have their machines already are likely to be offered some sort of fix-it kit if it comes down to that…or we may end up “fixing” it in software. All depends. Really not trying to do anything other than make sure people are getting a machine that isn’t going to frustrate them.
Good news! We actually ended up tracking down our auto-leveling mishaps to the bushings just being too damn tight at the extents of the bed! (Due to the rocking of the carriage I was talking about) On a whim, I asked if we had some old prototype units from when different types were being tested, and we did. I swapped out the arms for the delta printer and redid the auto-leveling procedure. It ended up turning out perfectly even at the far extents of the bed. There’s a little wobble in the end effector, which is causing some backlash-looking issues, but we’ve narrowed down the issue.
So now, on the stainless steel bearings we have, I’m running an overnight test. 300mm/s circles at the extent of the bed, in hopes that doing this will “break-in” the bearings we’re already using. So hopefully a software fix for a hardware problem.
