The Nature of Technology
That’s a book whose review I just read in Science (2/19/10) by W. Brian Arthur and the reviewer’s comments connected with the work I’ve been doing to revise the azimuth drive assembly for the telescope. Michael Alexander, the reviewer, describes one section of the book that “analyzes the easiest form of innovation to comprehend, standard engineering, which solves problems by combining well-accepted, usually well-known, technologies. The new technologies created through these combinations may be important and complex (jet aircraft and bridges, for example) even though the underlying inventive principles may not be novel.” Arthur’s point is that “What is common to orginators [i.e., inventors] is …the possession of a very large quiver of functionalities and principles.” While I don’t have the illusion that this design is “important,” it is certainly fun and the trip is much of the fun.
The early iterations of the azimuth drive started with my simple ideas of how a wheel could be attached to a shaft. I knew the word “keyway” but I hadn’t yet associated it with the “key.” (chuckle) So the design depended on locking shafts together or locking the drive wheel to the shaft with set screws driven into the conveniently placed keyway. Mechanical engineers and practical mechanics may now cringe. For an instrument whose value will be depend heavily on a repeatable control of movement (so you can find things without thinking about the telescope!) this approach potentially leaves the connection between shaft and drive wheel so loose that even the backlash would be unpredictable.
I didn’t know it at the time, but I needed more arrows in my quiver. I also needed to know how to use an arrow. So I started reading about woodruff keys (half circle thingies) and square keys, and, oh, those keys sit in the keyway of the shaft and stick out of it into a matching keyway on the drive wheel, and then you tighten the set screw on top of the key! I was feeling pretty proud of myself until I realized that the wheels we had planned on all had no keyway. They seemed mostly to be intended as cart wheels or conveyor wheels that would spin freely on their axles, not to be driven by the shaft, like our application.
The quest for a keyway began in earnest now. It was the holy grail of locking the drive wheel to the shaft. Google is a great tool for finding arrows. And since this problem has been around since Og made a wheel, if you figure out what to search for you get 10,000 results with arrows for this problem (if you don’t know what to search for, you get 10,000,000). So I started adding arrows and whole quivers to my arsenal. QD bushings have keyways but don’t really fit the hub of the wheel. The wheels themselves don’t have keyways but do have bearings (optional) so maybe we can use a pulley wheel, cut down and fixed to the hub somehow (it has a keyway and it’s cheap). And as it always seems to happen with such searches, I found by accident a manufacturer called Fenner Drives that had an intriguing idea – no keyway to attach wheels to shafts with no backlash.
Well, I’d been having my doubts about keyways anyway. I mean, how tightly could that key fit in the keyway? And if it wasn’t very snug, in both keyways, the shaft would rattle back and forth as it turned back and forth between the two limits – backlash. And Fenner knew how to get to me – they had drawings of how keys and keyways wear and make backlash even worse! Oh no! Luckily, they had a solution, the Trantorque Mini (whew!).
This ingenious product is arranged so that when the big nut is tightened, both the inner bore on the shaft and the outer surface in the wheel bore expand, creating a pressure fit between the shaft and the wheel bore – with no backlash. Made by Fenner Drives (http://www.fennerdrives.com), they specialize in these devices that come in different sizes. This one is on the small end of their line, and while I got the right size for the shaft, which is 0.5-inch diameter, the outer diameter is looking for a 7/8-inch wheel bore and the closest I could come with the wheels from Service Caster had 13/16-inch bores. No bushing I could find had those two diameters, so the search was on again, now for a custom machined bushing.
Google found me a site that I had actually found in a past search for a different context. It turns out that quite a few hardware manufacturers now offer CAD (Computer Aided Design) drawings in several formats of their products. These can be downloaded and inserted into a drawing, like the one at the left. Fenner Drives provided a CAD drawing of the Trantorque Mini, and in previous searches for simple things like bolts and nuts, I’d found emachineshop.com. Since I was looking for drawings at the time, I hadn’t realized exactly what this company did. As a free download from the site, they offer a simple CAD system that you can use to design a part from scratch. And part of the application allows you to check the part for manufacturability, and then to place an order for as many of them as you’d like, custom made. I specified a bushing with, you guessed it, a 13/16-inch outer diameter and 7/8-inch inner diameter and a slot cut on one side of the cylinder. That allows me to fit the Trantorque mini inside it, tighten it to push out on the bushing, and for the bushing to expand and push against the wheel bore.
It’s hard to describe the feeling of designing something that will be made to order, even for such a simple design as this. It’s one of the things that has always attracted me to making and building things. Once I got through some frustrating moments making the software work, and then learning about the limitations of the machines that would make my part, requiring some compromises in the design, I placed the order. The bushings are scheduled to be delivered on Tuesday next.
The connection between the drive motor 0.5-inch shaft and the extended shaft that this drive wheel is attached to was the next target. My first design had a simple shaft coupling — a solid piece of steel with a 0.5-inch bore and two set screws in a keyway. With the wondrous Trantorque Mini, I could hardly settle for this potential backlash.
Good old MartinSprockets had something that was close – a flexible shaft coupling that was a three-piece affair. But it still locked to the shaft with set screws on a keyway.
My search ended on the applied.com site with a device manufactured by Ruland (thanks to them for the CAD drawing) that has the same flexible coupling feature — the central piece is a plastic that allows the hubs to connect slightly misaligned shafts — and it has the bonus of a bolt that can tighten the coupling over the shaft with a pressure fit. Voila! No backlash!
Well, you’re getting the idea by now. The nature of technology is to build your new product from the technologies of other products, and in as simple an engineering task as this design has been, I’m still able to enjoy the thrill of finding a way to put existing devices together to create something new that will have utility and beauty (at least to the eye of some beholders!). You’ll remember that this wheel will be the drive wheel for both the azimuth (the lazy susan rotation) and the altitude axes of the Project 40 telescope. One of them (pictured below) will drive that azimuth by itself, while a pair will drive a balanced set of altitude bearing surfaces to move the telescope direction up and down. It gave me great pleasure to find a near zero backlash solution to the drive design problem, and it will give me even greater pleasure to come when we put the drives together with the mount that is growing in Larry’s garage and see it actually move.