The inverse belt box hub motor is a brushless, fully enclosed transmission system with a 10:1 reduction. By rotating a center pinion, an inverted timing belt transmits torque through a ring pulley that is attached to the wheel hub creating a quieter, weather-resistant, and more compact reduction over a traditional spur gearbox. The timing belt is retained using an array of bearings and acetal sliders. This system was developed for the drivetrain for my Major Qualifying Project (MQP) at WPI.

Initial testing for the generated tooth profile and the design concept was all done with 3d printed components. Since this was the first time I have designed something like this, I needed to check to make sure I could setup my design constraints correctly, 3d printing was a cheap and fast way to prototype the design.

After confirming that the tension from the belt was correct, I printed a full version to see how the system behaved. 

This version had nothing retaining the components in place so the ring gear slipped off a few times. Testing this system was difficult but it did show that the ring gear would rotate.

The first motorized spin up showed me that the belt naturally wants to pull tangent to the lead in bearing that interfaces with the ring gear. If you look closely at the powered spin up, you can see the belt loses tension on the opposing side, eventually leading to the belt slipping. This realization drove the design decision to remove the inner tensioner bearings and instead pull the leading bearings in closer to tension the belt.

In addition to the tensioning problem, the belt also seemed to skip from the 3d printed ring gear. Since the ring was designed to be made out of aluminum and be compact, the plastic ring gear that was printed to scale was not holding up to the testing. 

This version worked much better with belt tension but it highlighted a new flaw in the design. The bearings that were making contact with the ring gear only allowed for 1-2 teeth of engagement at the time which was allowing the small GT-2mm belt to slip. The idea to fix this problem was to use a short section of low-friction slides that would hold a longer length of belt contacting the ring gear.

The new sliders did help with keeping the belt in place but there was too much friction in the assembly for the wheel to spin freely. This was because the first sliding surface is not a rolling surface and although low friction, the belt pulling on that surface just increases the total friction force, causing it to bind. The solution I thought of was to combine the two previous methods of belt retainment such that the bearing is the first point of contact, but the belt is pinched in place by the acetal slider so that the belt can't jump.

This system required a ton of manufacturing involved. Lots of cnc machining and a fun challenge of EDMing a thin ring pulley took place. The ring gear was made using a reamed out fixture pate and dowel pins to locate a blank in place. Since the wire EDM is a zero force cutting machine, it didn't need any clamping force to keep the part constrained. Eventually, I stacked 3 blanks at once to minimize the cutting time by 75%!

This system was created for my senior capstone project: a delivery robot. Its intended purpose is to traverse the WPI campus to deliver various items to our community members. The hub motor provides a weather sealed transmission system to the robot so that it can operate in wet conditions.