After browsing and consulting with the folks on RCGroups more, I decided to (once again) modify the (supposedly) final design for the wheelmotor. The tweaks increase the virtual gearing inside the motor by about 20% and also makes winding much easier.

I have to lock myself into a design eventually, otherwise nothing will ever get built as I continually update little things here and there. Therefore, I took the opportunity of a gap in the work scheduling to hop on the Media Lab waterjet with some thin steel sheets and erode myself some new 30T stator plates.

These thin steel formations are probably the most elegant things I have ever cut out of anything. The waterjet is officially my new favorite manufacturing implement.

With the change in number of stator poles comes a change in the required magnet dimensions. I emailed the people at Supermagnetman, my usual source for neo magnets, to see if they could assist me with manufacturing a set of custom magnet segments that will fill up the available magnet ring space 100% with no finicky spacing, trimming, or selective application of epoxy. We’ll see how that goes, but 40 5mm wide, 3mm thick magnets also fit well.

I split the 3D model for the “motor can” into three parts, from two. The left and right hub plates/bearing holders are now identical and symmetrical, so two could be popped off quickly on a CNC lathe. The magnet ring mount, which used to be integrated with one of side plates, is now its own structure. It is quite literally a ring.

Meanwhile, I have also been fiddling around with “internal logistics”. That is, how to squeeze all the necessary parts into a space about 40% of what I had to work with while building the first scooter. Here’s a picture of the new “conversion base”, a Razor A3 I ganked off EBay, as compared to Snuffles 1.

The wheelbases are the same, but the aluminum extrusion that forms the chassis of the A3 is half the height of the same on Snuffles 1. This means no more shoving a stack of big NiCd cells – they’re simply too big.

The hope for this build lies in lithium polymer cells, of which 6 4AH cells fit just right in the back two-thirds of the channel such that enough space is left for the controller and wiring.

I’ll get to the juicy parts later.

So, in the end, I decided to pursue the “normal” iron core motor design. Sure, I’ll lose some efficiency points, but it’s a design that I’ve worked with before and know the details of. On top of that, there is a huge hobbyist-level userbase of that type of motor which I can consult if something doesn’t work.

The “übermotor” design just had too many variables for me to juggle and too many tolerances to hold. I’ll tuck the design away for another day (or after the regular motor works and I can go “if you give me more money I can make a REALLY COOL one!”

• 
• 

33 pole stator chips, cut on an abrasive waterjet out of regular mild steel. I wanted to shove as many poles as I physically could manage (while maintaining wire clearance) to get the highest virtual gearing possible. It’s not electrical steel, but believe me – if I had a big sheet of electrical steel, I would have used it.

I’ll be optimizing the design some more over and seeing what else I can get done soon.