I´m at a project now for some time that may interest you. Inspired by the hangprinter and the thoughts about printing larger objects, triggered my interest in building one. The problem for me was that I didn´t had the space for the anchors and wanted the capability to print big on a small horizontal space.
I then evaluated different arangements and kinematics and ended up with a polar configuration for minimal part count and printvolume vs machine footprint. Alltough the polar kinematics also have disadvantages I build a prototype with a pretty old firmware from the R360 polar printer. It is mechanically build from very simple parts. The tubes are made from stainless steel and are pretty easy to source and most of the parts are “print at once” without the need for to much assembly.
The Z mechanics rely on 608 bearings just like the MPCNC does. As I discovered after long cnc testing that this system is pretty accurate and play free so I went with that. It is driven by steel reinforced HTD belts and they deliver pretty precise results. They are preloaded a little at the start to remove any possible problems with the first layer. The bed is a belt geared 7:1 drive with a nema 17 40 oz/inch for now witch is the minimum I had calculated with safety margin to move the bed dynamically without losing steps. The X carriage slides like Igus bearings but printed from PLA on a slightly oiled alumimum tube. Precision for it´s side is pretty good if you look at the layer deviation in XY. The Leveling is a bit tricky but there a several adjustable mechanics that can be tuned accordingly. Once its level it pretty much repeatable. The bed was a normal table plate from the baumarkt but was then replaced with a dibond plate later on. I could save almost 3 kg and improve the flatness for the same price.
The biggest hurdle to be taken was the transformation into polar coordinates from cartesian on the fly, as preprocessing created huge G-code files. I tried for some weeks to integrate the conversion into the klipper firmware and already had a working external python program that drawed some graphs for debug but was unable to integrate it into the klipper kinematics as they where a bit more complicated back then. Kevin O´ Connor than jumped in and wrote a demo code and I managed to get it all working based on klipper, which greatly improved the print quality as the processing wasn´t interrupting the printing and made it stutter in the marlin farmware.
I´m still working on this from time to time and try to polish it a bit more but there is still quite some things to do to make it easily usable and to open source the plans.
Are you interested? Got any questions? Let me know below to give this project a little love