Justin Dunham

Parts for the Z Stage Assembly.

The completed X and Y Stage Assemblies, together with parts to install them in the body

Mechanically complete Makerbot!

For tomorrow...

Today I tackled 3 separate stages of the Makerbot build sequence: The Z Stage Assembly, the XY Stage Installation, and the Z Stage Installation. The end result is an (almost)  mechanically complete desktop fabricator.

First, let me explain briefly what each of these stages means. The Z Stage is the part of the machine that controls the movement of the extruder up and down. So today I basically built and then installed the parts of the machine that enable that to happen. The X and Y Stages are the parts of the machine that allow moving the extruder (actually the build platform) left and right and back and forth. I built these parts yesterday, and installed them today.

The way that the Makerbot extruder moves up and down is actually pretty ingenious. (Much of the machine is, but this was a particularly interesting piece). Basically, there is a platform on which the extruder will sit. The goal is to get it to move up and down, for reasons similar to the reasons why you lift a pen from the paper when you’re done writing.

The way you do this with a Makerbot, is that a very long bolt runs through each corner of the platform. Bearings and nuts are put onto the bolts also, in such a way that you essentially move the platform by screwing and unscrewing the bolt. So think of the way the head of a screw goes down when you drill it into wood. The Z-stage motion of the Makerbot takes advantage of this principle, with the result that you end up with an easy-to-build and elegant mechanism for this motion.

Anyway, other than that finishing up the mechanics of the Makerbot pretty uneventful – a lot of the same old nuts and bolts (literally) that I’ve been doing for the past two days. One minor problem was that the tolerance between the pulleys and the shafts of the motors is way too tight, which means that you have to do a fair amount of sanding to get everything to fit together. The worst part about this is when you put one of the pulleys on a driveshaft to test the fit, and then it won’t go into position, but you can’t remove it anymore, either.

However, we then move on to hooking up the electronics and software. This is an area where I had significantly more trouble, and I was actually surprised at how much the quality of the documentation seemed to rather suddenly deteriorate. “Now you’ll just need to hook up the electronics and Plastruder”, indeed!

My major issue was with the instructions for assembling ribbon cables. Makerbot doesn’t provide any of their own, though they do point you to the instructions that come along with the RepRap, which is the design on which the Makerbot is based. The main problem is that these are somewhat difficult to follow – it was hard to get a good sense from the pictures of what goes where, how everything should be aligned, and how much force should be used.

The most annoying part about this was that preassembled ribbon cables aren’t that expensive – does it really make sense to build these from scratch? (This must be a fundamental decision that has to be made with all the parts Makerbot provides, and it’s one I recognize from lots of other projects – build or buy, or this case, supply parts or supply part?). Anyway, it seemed like there wasn’t as much support on this step for people who don’t already know their way around this stuff.

After hooking that up, I went on to check out the instructions for building the plastic extruder. For some reason, I thought these would be relatively easy to complete. They… are not. I think it will go OK, but this is the first time I’ve felt somewhat intimidated by this project.

The kit for X Stage Assembly

Completed X Stage Assembly. The build platform will slide along those rods.

The kit for body assembly.

Completed body!

Today I completed two more steps in the building of my desktop fabricator – the building of the X Stage Assembly, which will allow the build platform to move along the X axis, and the building of the Makerbot body.

This was a pretty major couple of build steps; I’d say I’m about halfway done now. Overall, the build was successful, but I ran into a couple of problems. The first was that I managed to put in the aluminum drive pulley in upside-down. The aluminum drive pulley will pull the belt that allows the Makerbot build platform to move along the x-axis – so, it allows movement from left to right. I’m not sure exactly how this happened, but I think I was confused by the instruction to put the pulley onto the motor shaft “set screw end first”.

This was definitely my mistake, but I guess I thought that meant “set screw end up“, and for some reason I ignored the picture. I only managed to fix this when my fiancee’s dad pointed it out to me – otherwise I would happily have allowed the belt to sit slanted, which would have happened because it wouldn’t have been vertically aligned with the idler pulley on the other end of the belt.

The idler pulley was my second mistake, and I don’t just mean the vertical alignment. Instead, it was simply mounted in a wobbly fashion. The reason for this is as follows: the instructions say to “set aside the small idler pulley”, but it wasn’t until I tried explaining this to someone else that I realize there were actually different pulley sizes. As a result, I spent a fair amount of time using a pulley that was unstable because it was far too big for the screw that was anchoring it.

Finally, it also turns out that during yesterday’s build, I had actually used the wrong drive belt for a certain task. The instructions specify e.g. “the 196-tooth belt”; I just guessed at which belt was which, as I didn’t realize the tooth numbers are actually printed on the sides. I realized this when today’s instructions called for the “second-shortest drive belt” (the one I had used), so I was able to fix the error. I wish the previous day’s instructions had been as specific!

I ran into a few manufacturing defects as well. As usual, some of the parts don’t fit together perfectly as they aren’t machined exactly enough; this is totally understandable. Also, one of the bolts, which happens to occupy a critical position, is a little bit too long. This means that the Makerbot is unable to move the build platform in a particular direction. I had to go through a fairly long process of trial and error to figure out how to shorten this bolt with the equipment on hand (I ended up using a hacksaw to shave off the extra length). I was slightly surprised at this oversight, though it is explicitly mentioned in the instructions.

Other than that, the build was pretty uneventful. I built the body as well, which basically meant a long process of tediously bolting together the sides, back and front.

This is the sort of thing I would pay money to have done for me – it takes about an hour to get all 60-odd bolts and nuts in the right place, but there is no real thinking involved. I also thought it might be interesting to paint and otherwise decorate Makerbots. Since most of the structural pieces are just wood, this would be pretty easy, and the instructions even explicitly mention that you should do any decorating before you do the assembly. Anyway, at the end of the day I had a complete Makerbot body.

In the course of the built, I was also asked the very interesting question of what the “killer app” is for the Makerbot. That is, what is the application that results in millions of people buying a desktop fabricator and putting it in their homes, the way they bought the telephone, dishwasher, refrigerator, television, etc. – all the standard equipment that had to be introduced and marketed at some point in the past.

I answered that I don’t know. Right now, the Makerbot is really targeted at hobbyists and other people who are curious about the technology. It’s not yet particularly usable “for” anything. As far as what it could be sold as today, I think it’s mainly useful for artists who want to incorporate printed objects into their work, whether that’s individual sculptures, or mass-market pieces as might have been created with Bakelite ninety years ago. There are some other random applications as well, such as certain technical tasks in e.g. jewelry making. But no killer app yet, until the technology can be dramatically improved.

The next key step is to figure out a way to generate cash flow so that more can be invested into R&D on desktop fabricators – specifically, improving their ease of use and capabilities. This will require lots of market research, and identifying at least the interim “killer app” for certain niche markets.

The Makerbot box.

Workstation.

The parts for the Y Stage Assembly - structural pieces, bolts, a belt, etc.

Completed Y Stage Assembly, with the Build Platform in the upper left.

• Note 1: This entry is the first in a series.
• Note 2: I am looking for likeminded people who want to work together on building some kind of desktop fabricator business. If you read this and agree with what I have to say, will you drop me a note? For more about me, check out my LinkedIn profile.

Today (12/31/10), I began building a Makerbot, which is a special type of machine called a desktop fabricator. A desktop fabricator allows you to print. But you can print objects instead of documents. Remember the replicator, from Star Trek? This the first step toward that technology. I’ll document the build process on this blog.

What do I mean? Let’s say you are looking for a bottle opener – as has happened to me many times. You have a bunch of beers, but you can’t find anything to open them with. If you own a desktop fabricator, you can print a bottle opener. Not a picture of a bottle opener, but a physical bottle opener that lets you open bottles. Or lets say a knob breaks on your over. Print a new one. Or if you want a copy of some tchotchke that someone’s designed. Download the schematic and print one yourself. The technology’s primitive right now, but it’s improving rapidly. And pretty soon 3d scanners will come into their own, and you’ll be able to copy things for real.

So wait, wait. How is this machine able to print things? Well, check out the Makerbot video page for some demonstrations. But the basic premise is that it extrudes ABS plastic, much like honey coming out of a squeeze bottle. This ABS plastic is extruded into layers that make up the object you’re interested in. The layers are pretty thin, so you can create a wide variety of objects.

There are a few limitations, obviously – plastic is the only material you can use right now (with the exception of frosting). The Makerbot is slow; it takes an hour or two to print a complex object, though this is improving rapidly. Lastly, the geometry you can produce is somewhat constrained. For example, you can’t do severe overhangs because the plastic will collapse before it has time to solidify.

But still, I strongly believe that this technology will change people’s lives in ways that we haven’t imagined yet, and this will happen soon, like in the next 5 – 15 years. My favorite way of talking about this is that the desktop fabricator will do to physical objects what the MP3 did to CDs.

Because I believe very strongly in the promise of this technology, I’m hoping to build a business on it at some point. Ideally, I would want to find a way to finance rapid improvements in the Makerbot, for example, a major new version every 6 – 9 months. The machine is an incredible achievement, but there’s lots of work to be done on improving the technology so that it can produce a broader range of usable objects.

Also, I think there could be a lot of work to do on making it easy to build (eventually buy) and use. Right now, you have to use 3 separate pieces of software to create a new printable object. One day, I would like this to be as easy as desktop publishing. I read an article recently that said that at one time, the idea that a home user could produce professional-looking documents on their personal computer was laughable. Today, kids use Microsoft Word, Powerpoint, etc. to create documents that simply wouldn’t have been possible 30 years ago. 20 (10? 5?) years from now, it should be possible for someone to publish an object just as easily.

Anyway, in order to better understand how a desktop fabricator works, how easy it is to use, and what can be done with it, I decided to build one. I ordered the Cupcake CNC kit from Makerbot – there are others, but this one seemed like the best buy and easiest to deal with. Over the next week, I hope to complete the kit and print at least 1 or 2 test objects. I will also document the build process on this blog.

Today, Day 1, I built the first stage, which is called the Y Stage Assembly. I wish I could explain why it’s called that. Y almost certainly refers to the Y (up-down) axis, and the Stage, which is the orange thing in the pictures, is what objects are printed on.

The MakerBot components in the kit are really well organized, and the documentation is extremely good – it even incorporates mistakes that past users have made, and videos are available to make the build process as straightforward as possible. With that said, I am going to use these entries to point out things that could be improved. There are a few things that were more difficult than they could have been today.

• Some of the parts are slightly different than the pictures in the wiki (for example, a couple of parts now have a Makerbot logo, and the flanges are white now, instead of black). This made it surprisingly difficult to find pieces, since otherwise I don’t know what a “flanged slide bearing” is. It seems like a really small detail, but it actually matters a lot for more naive users, like me.
• The build quality is not that precise. On several occasions, I had to sand a part down to get it to fit, or holes were slightly misaligned, so I had to work really hard to get a screw into place (while worrying about cracking the materials I was screwing together). Could one of the included tools be a metal rod with a rough surface that makes it easier to sand holes to an appropriate size, or to a large enough size to accommodate imperfections?
• Early on, you have to get a bunch of magnets into holes, and they all have to be aligned in a certain way. This is because two of the parts are held together by strong magnets, rather than by glue. I think this idea is ingenious, and I assume it comes in handy later although I don’t know when. However, this is an immense pain to gain right, since you’ll put a magnet in a hole, and then fill another hole with a magnet, and the magnets will sometimes jump out and stick to each other. I don’t know what the solution to this is. Scotch tape?