New Mobile Base Design Theory

This task hasn't been easy. Of course, it wasn't easy when we designed the old base either. Fancy that. Anyway, the ideas for the new base started to foment during Summer 2001 (the summer of my discontent) when the then-current renovation started to really go awry. By January 2002 it became really apparent that these ideas were going to be important since Brak, as we knew it, was history. My main idea was to make a modular chassis; make small, individual housings for each steering shaft so that it would be easier to align the bearing holes,

Top-down view of prototype chassis design

then fasten these together within an articulated frame which would allow the three shafts to "contract inward" for attaching the pulley belt, then "expand outward" to tension the belts in the final position. Sometime around September 2002 (I think; yeah, let's go with that date) I finally figured out how to the frame should work. It then took me another three or four months to find the time to draw it and send it off for someone in the Mechanical Engineering Dept to look over, give me feedback, and then get with our machine shop wizard to start making parts. As expected, the design isn't going to look like the drawing but should functionally do the same thing.

The prototype drawing shows the three housing and the articulated frame. Here's how it supposed to work: the left frame member is rigidly attached to the top and bottom left housings (the four white circles represent bolt holes). The right frame member is rigidly bolted to the bottom right housing but attached at a single "pivot point" to the top housing. The bottom frame member (between the bottom left and bottom right housings) is actually a two-piece link. It is hinged in the center and each end is attached to the housing at a single pivot point. When this member is "folded out" (shown by the two arrows as the bottom), the right frame member pivots inward so that the bottom housing are drawn closer together. This decreases the distance around the pulleys so that the belts can be attached. When the bottom frame member is straight (as shown) the three steering shafts are at their maximum separation and shafts are at the proper location.

Proper location is the key here; the base relies on an equilateral triangle with (almost) exactly 12.5" between the center of each shaft, and with the shafts being parallel. I can (and must) live with a little imprecision but the goal is to make it as simple as possible to minimize this imprecision through our manufacturing process.

Now, at this stage, the alert reader may be asking themselves what keeps the pivoting frame members from pivoting once everything is in place. Others may be wondering how "beefy" these pieces need to be to assure the chassis will be sturdy. Our plan is to attach a single aluminum plate to the top of the three housings (or the three frame members, or some combination thereof) which will not only hold the chassis in its proper place but which will add to the stiffness. This plate will (probably) also support the drive and steering motors, as well as being the foundation for the other structural pieces of the robot; for example, there will be at least one bottom plate, similar to the mbase1a plate on the old mobile base, to provide support for batteries and act as a "bumper". There will be one top plate to support the ultrasonic sensors, probably identical to the old mobile base's mbase7a plate.

I also want to change the drive and steering systems. The old mobile base used identical gearing for both (except for the miter/bevel gears on the drive part), but as previously mentioned we retrofitted the steering system with extra gear reduction to increase torque and reduce speed. I've since been introduced to the concept of worm and helical gears, which I'm beginning to like since they are comparably priced, take less space, and (in the case of worm gears) allow for higher gear reduction. I've found that for less than $110 I can get a 30:1 worm gear set and adapter hub; the comparable parts on the old steering system cost about $75. The new drive system would replace the each miter/bevel gears (costing about $200) with two universal joints, two helical gears, and some additional shafts and bearings (costing about $180). The advantages of each new system is that the steering's gear reduction is 30:1 versus approximately 9.5:1 on the old base and the drive system allow the each drive wheel to be placed directly under the steering axis of rotation (eliminating the old mobile base's "Steer Drive phenomenon"). At this point, however, there is a new phenomenon; whenever the robot steers with the drive motor is off, the wheels will slip in place. This can be corrected by connecting the steering and drive motors thru a differential to the main steering shaft (I built a model using Legos, so it has to work, right?), but this will probably need to be after the gear reduction from each motor to avoid having one motor accidentally turn the other.