I’ve made more changes to EasyMaker. I’ve changed the shaft mounts to SHF-12 mounts, which have exposed a problem I’m now trying to fix. The SHF couplers do a good job at making the machine more space efficient, and making the machine easier to understand. SHF12 couplers keep more of the Y axis in the frame, and reduce overall parts a little for mounting a spoilboard. They also make alignment a lot harder. I can still make an EasyMaker and make it work — but it makes calibration time a lot longer. It violates my design ethos — a machine that’s easy to build and easy to run. The alignment process is too complex. I need to design some jigs to help make mounting/aligning easier for new builders. Either that, or redesign the bearing mounts for the gantry.
The new mounts create squarenes problems — linear motion is still good. But I now have 1mm error every 100mm travelled. It can be compensated for with a touch probe and some manifold math. But that requires some complex code, and only solves the problem in R2. In R3, a new problem arises — one I don’t even know the word for. I’ll call it eccentricity. In R3, the angle of attack in drilling a hole matters — a lot. Most drill presses, for example, can’t drill perfectly square holes in long bores. If you ever try, you’ll find noticeable error. I’d say all holes I’ve ever tried to drill on a drill press have been off by a degree or so, from bottom to top. This forces you to make hulls where you wanted holes, and introduces error in things you make. 1mm/100 is great — .01 mm square error is considered very good. But I hate it. It took me 6 hours of calibration to get the error that low, and 2 rebuild attempts. I can’t imagine anyone being happy with a machine they build in 3 hours, and then rebuild and recalibrate for 12.
So, I’m now off to work on this problem. Luckily, I’ve already mentioned earlier that I’m in a holding pattern on this project. I’m waiting for some kickstarter projects that I’m backing to ship me items. I want to evaluate them for use in EasyMaker. Espeically the RA controller board. So, while I wait, I figure I’ll work on this problem as well.