Here are some updates on EasyMaker:
1. I’ve backed a new set of electronics on KickStarter to evaluate — the RA control panel. I’ve found that the pin header/connector system in use on RAMPS isn’t durable enough, and over time, I have to make a lot of adjustments to the electronics to keep everything working within the tolerances I like. RA supports 6 axis of motion ( RAMPS supports 5 ), can drive 3 extruders ( Ramps can do 2 ), has screw terminal throughout ( RAMPS uses pin header ), and supports an optional control panel that is already designed ( allowing for no-computer prints, providing GCode has been generated ). It costs a little more than RAMPS, but overall, seems to be a better solution.
2. My friend Grover has joined the project! He has designed a new extruder mount, and I’m working on refining it. The new mount solves a handful of issues that the old mount has — the new mount is a little harder to make, but returns almost 40mm of Z as printable area. We’re still working the bugs out of this new mount right now. Grover will be doing all build-outs/packing, etc… This way, I can focus more on design as we get moving.
3. I’ve slipped all my dates out by 3 months. Although my robot is design complete, I’ve got 2 reasons for this. First, I’ve got some durability problems in the bed and electronics I’d like to iron out. These problems aren’t major, but they require maintenance by the end user. All 3d printers and CNC mills have high maintenance requirements — but the lower I can make my maintenance, the better. The second reason is that I’m seeing things on KickStarter that address the first problem — New hot ends, new electronics, etc… Almost all of these will be ready around Sept/Nov, and I want to evaluate them as part of this design. The RA is a good example — if it ships to me on time, it may decrease the amount of maintenance work people who get a robot will have to do.
4. Bed flatness is a problem. The heating/cooling cycles are slowly causing the aluminum bed to warp. It’s still very flat, but has a few divots of around .1 — .3 mm at certain stress points. These stress points are random — caused by internal stresses during the rolling process for making plate aluminum. I could buy cast aluminum — this will have no internal stresses — but at 10x the cost. So, I’m thinking of ways to counter this. I had planned on using CNC mill mode to precision grind the Aluminum plate down to perfect flatness — but that’s a time consuming, labor-intensive step. Doing it on one makes a ton of sense. Doing it on 100 — not so much. So, I’ve got a few ideas in mind :
* use borosilicate glass instead of Al. This is known for it flatness and good price. Downside is weight, fragility, tendency to crack under stress conditions. It would make EM far less durable.
* Map the Aluminum bed and correct via a kinematic algorithm. This requires some new code, probably to be run against the GCode. It also doesn’t scale well — each bed has to have this run, and a custom file created. It’s a lot of time to each machine — I’d have to run heat/cool cycles on each bed, wait for the internal stress points to reveal themselves, then map them. During use, new stress points will emerge over time, requiring a re-map every once in a while.
* Some sort of conductive “bondo” filler. The divots are very small — less than 1mm. I should be able to find a thermally stable compound that can fill those.
I might just live with the .1 — .3mm error. I’ve found that this is within tolerance of the machine, but right at the edge. It makes large parts a bit harder to print, as you may not get cohesion in a divot, causing a “pile” of plastic to build up in those spots. That’s random, though — try again, and it all works. I’d like a hair more consistency, and hair less maintenance, before I feel the design is ready for production.