Building a printable steppe...

• to increase the percentage of a Reprap machine that it could build
• to get around the pig-headedness of the Chinese engineers who were refusing to sell me lead screws for their very inexpensive (~$2.50) linear stepper motors in the lengths I needed

I bought the original linear stepper motors for Tommelise 2.0 from Haydon. While their product massively simplifies the design and construction of accurate positioning robots, their pricing structure obviously factors that ease in to the point where a Haydon linear stepper and lead screw costs about $80/axis. The Chinese were much more attractive, pricewise. Unfortunately, they had had some customer complaints about the tendency of their small diameter lead screws to bend under load. While I tried to explain that I wasn't using their linear steppers to anything like their design loads, they still flatly refused to sell me the lead screws in the 12-18 inch lengths that I needed.

Having lived in China and liking the Chinese people immensely, I was altogether shocked and finally angered at this situation. In the months since July this annoyance began to inform my efforts at building a printible tin can linear stepper. The big problem with making a tin can linear stepper lies in the rotor. If you go for a one for one match you have to have an aluminum spindle and coat it with either a polymer or epoxy layer with an admixture of magnetic media. While this isn't technically a really difficult problem, you then have to create a special gaussing device to create the alternating magnetic stripes of opposite polarity on the rotor.

I came up with an alternative design which used thin steel strips and a ring magnet to simplify this process. There were two problems with my solution. First, the retail price of the ring magnets was on the order of several dollars. Second, and more importantly in terms of my annoyance with Chinese suppliers, was the fact that the Chinese these days pretty much make ALL of the rare earth magnets in the world. This situation has got to the point that the GPS guided bombs that the US uses to pound terrorists use Chinese magnets. Finally, I discovered that the nickle plating that the Chinese put on their rare earth magnets perishes after a few years leaving you with a really nasty mess where your magnet used to be.

All this kept me from continuing the effort I'd started to build a linear tin can stepper. A few days ago, however, I ran across what began to look like a much more suitable alternative. While, I'd read about variable reluctance steppers in a variety of stepper references, I hadn't really run across one in practice. Recently, however, I was nosing around in the Reprap forums "Things to Print" section, a place that I rarely visit. Back in May, Tim Atwood came up with a printable variable reluctance stepper, although he doesn't quite call it that.

While Tim was aiming at producing a stepper with a very small step angle, the Tommelise approach needed nothing this complicated. What really got me moving, however, what this little project.

The motor you see is a one phase variable reluctance motor.  As you can see in the video it is a bit of a pig to get started and keep going.  Watching its inventor work with it, however, I was reminded of  an illustration I'd seen some time ago.

While this design had a 60 degree angle step, I could get it to give me 30 degrees by either half-stepping it or using a cruciform configuration rotor rather than a simple bar rotor.

Further, if I used the general design of the single phase variable reluctance project that twigged my imagination, I could reduce the complexity of the fabrication dramatically.


With all that in mind, I dug out Rick Hoadley's Excel app for designing electromagnetic coils and had a go at designing a coil for a first cut at such a stepper out of stuff that I had at hand.


This was a nice little coil. I got considerably better gauss ratings than my linear steppers gave me and kept the amperage down to something managable. The only worry I had was that I was generating 2.91 watts/square inch of heat to dissipate while Rick suggested that 0.5 was ideal. I had no idea what "ideal" meant when Rick used the term.

So... I built the coil. The spool for it was made from two nylon sleves epoxied together.

To give you an idea of scale, that's a piece of 3/8ths inch threaded rod running through it. I pulled 5.4 and 12.2 volt power off of my ATX power supply. When I ran it at 12.2 v, it got quite hot within a few moments. I monitored the coil's surface temperature and shut it down when it nosed over 100 C.

I then applied 5.4 volt power. At that level the coil stabilised at about 71-72 C, hot, but not outrageously so. Apparently Rick's "ideal" dissipation rate heats the coil to only a few degrees above ambient.

Rick also mentioned another little game you could play with your electromagnet. The game was to see how many steel BBs (4.5 mm copper washed steel pellets used in American air guns) your electromagnet could pick up. Although Rick didn't mention it, it struck me that the number of BBs a magnet could pick up may be directly related to it's gauss rating. I had a go with this.

The only trick to this experiment is that to lift the maximum number of BBs you have to be very careful. Also, if you lose power before you get your BBs from the dish to another container things can get a little messy. In fact, though, I discovered that the pickup power of the electromagnet related to the gauss level predicted by Rick's Excel sheet such that...

1 gauss = 1.8 BBs

One end of my electromagnet would pick up 42 grams of BBs at 5.4 v.

I'm quite excited about going forward with my variable reluctance stepper motor project. Freescale described this class of motor in these terms.

The reluctance motor, commonly referred to as either a variable or switch reluctance (VR or SR), offers the simplest, lowest cost motor available to date. This motor consists of a shaped, highly permeable material for the rotor and two- or more-phase windings in the stator. The shape of the rotor concentrates magnetic flux lines into "poles" on the rotor which then interact with the rotating field being developed in the stator windings. The VR and SR motors must be electronically controlled to start turning and to produce the torque needed to continue rotation. Torque pulsations generated by this design tend to produce undesirable audible noise, leaving much research still to be performed. The potential for its low-cost application into many areas, however, makes it an extremely desirable motor to use in a broad range of products.

This looks promising.