Shortly after receiving the sun gear holders, I received the first iterations of the planet output and front housing.

Both of these seem to have actually adhered to the tolerances I requested, so thankfully it won’t be too hard to fit everything together.  However, getting everything together for the first time did involve a comedy of errors — a lack of planning for assembly order, a lack of foresight into how things would be *dis-assembled*, a stubbornly stuck shaft, and plenty of broken parts.

Sun gear holder assembly

The sun gear needed to be bolted into the rotor, and have a bearing installed that interfaces between the rotor and the back housing.  That bearing interface had a similarly poor tolerance as the interior interface from last time, and thus the bearing needed to be pressed on.  I had pressed one on as part of verifying that it would work at all, but then didn’t realize that the bolts to installed the holder to the rotor wouldn’t actually fit down.  I ended up grinding a flat on each bolt so that it could fit around the bearing.

Front housing

The front housing is slightly different than my original plan.  I was initially going to machine the outer housing included into the front housing as one piece.  However, having a shop do that turned out to be prohibitively expensive, both because of the need for a 4th axis, and also because of the depth of the part.  It was looking to be something like $500 per piece at ~20 quantity.  Thus, I simplified it into something that wasn’t very deep, didn’t need more than 3 axes, and had just the attachments necessary to maintain a rigid drive train.  I figured I could for now continue 3d printing the outer housing.

That said, the front housing still has a lot attached to it.  The output bearing fits in the central hole, the internal gear rests on the interior of the cylindrical protrusion, the stator rests on the outer surface of the cylindrical protrusion, and the outer housing bolts to outer recessed holes.  Additionally, in the future, the position sensing board will fit into the end of the cylindrical protrusion.

The internal gear ended up closer to a slip fit, which meant I needed to use retaining compound if it was going to serve any practical purpose.  That would also mean it would be really annoying to remove.  Fortunately, I did manage to install the output bearing before using retaining compound on the internal gear, as it does need to be installed first.  However, I didn’t manage to install retaining compound on the output bearing, which wasn’t exactly a slip fit, but did press out easily with my hands.  That would come back to bite me a bit later on.

My next, relatively minor problem, occurred when I was test fitting the stator in place.  That surface was within tolerance, however, the front housing still required cooling with an upside down duster before the stator could be installed easily.  I got overly excited and installed it before having the outer housing installed, so had to take it off again.  Then, after I had installed it again the second time, I realized I had not aligned the stator so that the phase wires lined up in the proper location on the outer housing.  That, I just let slip, which turned out to be fine, since I had to tear it all apart later anyhow.

Gear train

Next up, I installed the planet output shaft bearing.  It was a very loose fit, so retaining compound was used, and possibly applied too sloppily as evidenced later.  Also up were the planet gear shafts, bearings, planet gears with 3d printed bushing adapters, and the planet input with input bearing installed.  Those all came together fine.

Next I installed the rotor which had the sun gear holder attached.  It seemed to mostly fit, and looked like the rotor and stator would clear, so I got excited and super-glued the position sensing magnet onto the end of the sun gear holder.  That way I’d be all set to power up the controller.  Yeah, right, if only it were that easy!

Next up, my challenges in getting the rotor and stator aligned…