This past Saturday we got the prototype rolling around on the ground. We ran into some issues.

It’s all fixable stuff. We’ll have to reinforce and redesign the shell.

For the long form from Mike, look below:

1. The shell is buckling severely. This is occuring in the spans between the welds (rather than at the welds.) And it has a shape more consistant with distributed forces parallel to the surface (sideways buckling) rather than crushing at the point of impact (bending over.) Since the structure maintains its integrity in any position while stationary, this must be related to dynamic forces encountered while rolling.

2. The effective “facets” created by the grid of ribs are too large for even rolling and steering, especially in the critical +/- 20 degrees or so around the equator. I believe this is the primary cause of the dynamic loads which result in buckling. Essentially, the sphere is repeatedly rising and falling a portion of an inch, resulting in impact loads perhaps an order of magnitude larger than those of the static load. These facets also make steering very difficult.

3. Currently, there is a bias or high point pushing us away from the equator. One reason for this is built in to the geometry- the equator rings (or annuli) are uniformly at the perimeter of the shell, and whereas other “latitudes” have peaks on the perimeter, and lower troughs in the middle of the “facets”, resulting in a smaller average diameter. In addition, it may be the case that due to fabrication issues, the sphere is somewhat ovular (we should be able to measure this, if it is the case.) This effect pushes us away from straight line travel, resulting in instability.

In general, we can improve the stiffness against this sort of buckling for the same weight be trading off rib thickness with width. Doubling the thickness of the sheet metal will result in an 8 times increase in bending moment, whereas halving the width of the rib only cuts its bending moment in half- this is a good tradeoff. It also makes it easier to weld, and possibly easier to cut.

However, I don’t think this entirely solves our problems. We need to make the facets smaller for better control, and a tighter network of ribs has a much better resistance to buckling (and reduces the impact
shock.) Therefore, I am considering adding an additional set of short ribs between each of the existing ribs. This scheme is nice because the full-sized ribs can be laid in as they currently are, and then the short ribs can be laid in at the end. It is also possible to retrofit our current sphere this way.

This is illustrated in the attached jpeg. It also has an elliptical cut out at the equator, so that it is a quarter inch closer to the center on the equator, and gradually eases out to the full diameter four inches out. This should provide a stabilizing effect which nudges the sphere to default to straight-line travel.

I’m still tweasing these ideas out, and reformulating operating theories to improve the design. If some of this rings funny to you, please let know- especially if you’ve got a good way to test which model or solution is better.

Also, if you’re interested in this stuff, but not really clear about what this sea of text meant, please corner me or give me a call and we can talk it over. Later,

-mike

Comments are closed.