MakerBot, Past and Future

Present

Yesterday Makerbot Industries announced that they were going to have an open house today to discuss what everyone thinks they should focus on.

Since many, er, most of those interested won’t be able to make it to NYC with 24-hours notice, a wiki page was started so that everyone could share their thoughts remotely.

I like most of those ideas. I agree with most of them. But, if the question of the day is to determine the roadmap, I think that simply citing specific bot upgrades or hardware/software changes is too specific. Not worthless, but we need a little bit of Big Picture too.

I read an article on the Technocratic Anarchist where Forrest Higgs muses about how (as I interpret it) people are getting caught up on things like infill, heated build platforms, and support material when they really should design to the strengths and weaknesses of the medium instead of constantly fighting it. It made me think about hearing Michael Curry talk about his Gothic Cathedral design that is very popular on Thingiverse. He designed a model of a cathedral for printing on a Cupcake CNC. The cathedral that he chose was designed to overcome a common problem of the time where horizontal overhangs were hard to make with stones stuck together with mortar. This is a great fit for the Cupcake because it has the same problem of horizontal overhangs with no supports, and demonstrates how the old solution is perfectly applicable to a modern problem.

Past

As I understand, the Cupcake was conceived as a RepStrap. And I think it has been wildly successful. Many Mendels have been made with Cupcakes. There has even been a Cupcake made by another Cupcake!

Then it became more than that. The popularity and lowering of the barriers of entry made it, in some ways, better than a RepRap. It was becoming a viable commercial product. Over time and batch after batch, in order to meet demand, the parts that were part of the cupcake kits and were printed on cupcakes, in the spirit of the reprap project, were replaced with off-the-shelf parts made in traditional factories. It became apparent that these were no longer just RepRaps. They were becoming serious competition for $20,000 rapid prototyping machines.

Present (again)

I’ve been months getting my Cupcake (Wintermute) in working condition enough to start printing it’s first child. It’s not just fiddling with temperatures and tightening bolts, but getting the dizzying array of software parameters figured out and tuning them. Most of this is in Skeinforge.

There are a few things that the Cupcake is missing. It’s not closed loop, meaning that it doesn’t really know where it’s parts are. You just have to start the build platform and extruder in the right places so that it won’t drive one into the other or the build platform into the side of the bot, etc. It’s very good at driving blind, but a little perception would go a long way. But then you have to support it in software. Once you support that in firmware, you need to hack on Skeinforge.

Skeinforge is a hack with hacks added to it. It’s a great hack. It’s very capable. But it’s still a hack. And it’s the single largest point-of-failure for anybody operating a Cupcake. No amount of hardware changes will make that right.

I was at the KC Mini-Maker Faire answering questions about MakerBots, and kept hearing myself referring to it as a hobby, “basically like hot-rodding a car.” “You spend a lot of time with parts all over the floor and messing with this and that to get it just right. It’s definitely not for XYZ yet, because you would be spending most of your time fiddling and fixing.” I attribute 90% of that to Skeinforge and 9% to stuff that was just fixed by the MK5 or about to be fixed by Gen4 electronics. The remaining 1% is yet to be determined.

Future

I think that MakerBot needs to have a goal. That goal shouldn’t mention the how, and maybe not even the why, just the what. I think a simple goal of making a 3D-printer that is affordable, of reasonable quality and capability, and that has a pleasant, approachable user experience.

Of those three, the first two are complete but the last one is totally missing. The community surrounding MakerBot is excellent, and it the only thing that makes that missing element excusable so far. But to move forward there needs to be a lot of cleanup. I don’t think this will require anything drastic, really, with the possible exception of replacing Skeinforge.

One major assumption that is made with the Cupcake kit is that the user has a technical background, but that’s not necessarily a very clear assumption. Does that mean they can solder and know what a capacitor does? Or does that mean they are comfortable with a command line? Does that mean they know how to make a Bézier do their bidding or even what a Bézier is?

Any of those assumptions may be wrong, and worse, requiring that knowledge means that using the bot will be more hassle than it needs to be. Since it’s a kit, the only real requirement should be the ability to follow instructions and what they entail, and how to download and print 3D files from Thingiverse or elsewhere. Anything outside of that should not be required for making objects out of a loop of plastic. Anything else should be left as an exercise for us hobbyists.

-Rob

Posted via email from tinkerin.gs' posterous