One evening last year the glue holding on the rod-ends on the original Kitchissippi-Delta prototype slipped due to a head crash skewing the mechanism and causing a subsequent further crash of the effector into the build platform.
Not only did this shatter the glass bed, it irreparably damaged the mechanical assembly. It was a rather heartbreaking event which hilighted some of the key flaws to the machine’s build.
Rather than re-print the smashed components and re-assemble the machine, I realized it was an opportunity to refine and fix some flaws in the initial design and build.
I noticed a few mistakes that I had made – particularly some of the hardened steel pins were not the same size as the majority of them, I must have grabbed them from a set of loose bicycle parts or accidentally scavenged a few of them from a different length of bicycle chain.
Further, there was definitely a flaw in the swivel design which allowed for too much play in the mechanism.
The 25mm fans on the effector were not providing adequate cooling for PLA prints, they either need to be moved closer to the nozzle or replaced altogether with something with a higher CFM.
I have been of the mindset that this machine should be focused on ABS printing since it excelled at that PLA cooling can be easily provided using a large externally mounted fan. I have a Hexagon all-metal hotend with a 0.4mm nozzle that I wanted to mount onto this machine with that in mind, it’s not particularly good at PLA but works wonderfully on hotter plastics like ABS. Given the long bowden system in use on this printer, ABS tends to slide more readily and not jam along the length. Fitting in a Hexagon instead of a J-Head will require a new effector, so that also entails some redesign of the parts.
Initially I had started to do some testing of redesigning it in SketchUp but was more focused on the Gemini 3D Printer project I was already working on at the time things went south on the Delta.
Some initial steps were taken to use OpenSCAD to recreate the effector but I did not get very far at the time, being fairly unfamiliar with OpenSCAD. A few aborted attempts at coming up with refined parts in SketchUp have been tried but left incomplete.
Ultimately the partially disassembled machine has been lording it’s hulking 1M tall frame over my little Prusa Mendel i2 mockingly for the past year whilst I was occupied with other things.
While cleaning up the other day I was absently fiddling with one of the swivel parts and had a sudden inspiration as to what the core problem was with the previous design. I realized that if I embedded the pins deeper into the swivel arm, fixing them on one side (rather than expecting them to rotate freely on both ends) it would greatly stabilize the mechanism as well as make it easier to 3D print.
I’ve learned much with OpenSCAD in the past several months by working on the Gemini 3D printer’s effector and Z axis redesign. Using this newfound knowledge I set about building a new project framework for the Kitchissippi-Delta and implementing some test parts.
The resulting swivel arm worked out exactly as I theorized – the pins are now adequately stable within their sockets – but what about the host component they needed to swivel on? I spent a few hours one evening recreating (and refining) the carriage adapter piece in OpenSCAD and ran a few test prints and am also satisfied with the results.
The pins sit tightly and swivel well in their constraints and the assembly mechanism is much easier to work with, no longer requiring heating or partial disassembly of the carriage to attach. Now they snap fit in with minimal pressure and a low risk of snapping the parts.
All that is left to do now is to script some new rod-end parts that will snap onto the pins and then a new effector piece. This is one thing that appeals to be about delta style printers – once you’ve made one axis, the other two are just a matter of duplicating that exact sub-assembly.
Recent Comments