I decided to create a CoreXY using a Duet Maestro and a bunch of spare parts that I had lying around. I have gotten Delta's down to a science, and while they are easy to build, and space stingy, they are hard to tune when you want a large build plate. I didn't want a big bed-slinger, so CoreXY.
Unlike the Cartesian printers, which are hard to build, but easy to tune, I feel that the CoreXY is both hard to build and hard to tune. Or maybe, just really hard to build.
By build, I mean build well.
I have now read DD's site: [
drmrehorst.blogspot.com]
and it is giving me information on why the "really hard" is really hard, especially for someone that isn't a Mechanical Engineer. I wish that I'd read it more carefully a while ago!
My CoreXY works, and it works great, until it doesn't.
Every once in a while I experience a gods-awful "gronching" sound and a layer shift occurs. Reading through Mark's site, it must be because my belts are not perfectly aligned to the X and Y paths. They look "really close", but I get the feeling that "really close" is not good enough and my thumb-in-the-air mechanical design must be off by enough that in certain locations the belts gets too tight and the system binds. This bind always occurs in the X axis, and always when there is a long traverse. If I orient my model so that the long traverse is in the Y axis, no problem. The head is carried on the Y-axis extruded aluminum bar which is carried on the X-axis slides. This is opposite from every other build that I have seen, but there is no reason that one way is better than the other, IMO. My carriers and corners are printed ABS.
What is odd is that it seems like the bind happens more often when the head is moving faster. As I slow the speeds down, at some point the bind stops happening. Or happens at extremely long intervals, which I do not reach often enough to notice.
I am using Exoslide mechanisms for my sliders, which determined how I configured my axes as I did, and they are not forgiving during a binding condition.
My system has the "crossed belts" because I do not stack my pulleys. I did that because I printed my corners and slider mounts rather than having metal ones and I did not want to over-stress the ABS and cause the spindles to bend over under belt tension. I am beginning to wonder if that was a mistake since it means more locations to be out of line than when using stacked pulleys.
I have checked everything else in my system, all motion is smooth, my chassis is square, by belts are good, plenty of current, drivers are cooled and nothing interferes with anything else. I will post a couple of pictures as soon as I can.
Does it make sense that my belts are not aligned well enough, which then causes a bind condition during a long traverse at higher speeds along one axis, and not the other?
DLC
Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.
