RC Truck Differential MkII

For the RC Truck project, one of the areas that was in need of a big update was the rear differential. While it was generally smooth running using a helical ring and pinion and nice tolerances on the spider gears, there was a bit too much compliance that allowed for quite a bit of gear skipping. The previous differential was relatively non-modular when it came to the type of suspension and didn’t have any provisioning to accommodate wheel speed sensors or space to package mechanical brakes.

Current Updates

Suspension Mounting

There are a number of different suspensions used on class 8 trucks. In order to more easily provision for mounting a variety of suspensions, I added a few generic mounting points. These will allow for mounting a number of different types of suspensions. In addition, the nut-bolt fastener scheme has been replaced with integrated, 3D printed threads.

In order to allow the springs some rotational freedom, the mounting eye is sized a bit larger than M3 cross bolt. While this will lead to some wear over time, it’s an acceptable compromise for the time being. The current spring design is most similar to camelback suspension. Decades ago, this was a commonly used design for on-highway, long-haul applications. The new suspension mounting points will provision for future experimentation with some other types of truck suspensions.

Camel back suspension

Walking beam suspension

Air ride suspension

The last major update for the suspension is to incorporate some kinematic restraints in order to restrict the movement of suspension and axles with respect to the chassis. The previous designs were more or less rigid connections with no real suspension. This was fine initially, but as the design has expanded, the constraints become a necessary component for the axles and drive train to function properly.

Spider Gears

The previous differential essential had no central case in order to more easily get a few of how the mechanics. There were some very small stand-off arms to place the spider gears appropriately. Unfortunately they allowed for quite a bit of flex - under high rates of change of torque, there was a strong tendency to skip gear teeth.

Previous design with simplified spider gear constraints

In order to fix this, a full central case was added to keep the spider gears more firmly in place and maintain their tolerance. As a consequence of having less floating and variable tolerance on the backlash, there was significantly more friction trying to spin the spider gears.

New design with a more rigid central case

Generally there are two methods for setting the backlash in a gear train - spacing the gears further apart than would be called for based on the nominal pitch circle diameter or grinding tooth root slightly narrower than design principles would call for. The second method is the more popular choice though until now I have been adjusting the mounting distance.

Ideal gear meshing with no backlash

While both methods can produce acceptable backlash tolerances, when adjusting the mounting distance many more components need to be resized and reprinted. This has slowed the process overall for dialing in the backlash quite a bit. Given the small modulus of these spider gears relative to the printing resolution, it was also difficult to make meaningful steps in the backlash settings. A major sticking point here was the small artifacts in printing small gears that would lead them to bind easily. I tried using a Dremel tool and wire brush, but the results were very inconsistent

Backlash via adjusted mounting distance

In order to reduce the iteration time, I switch to reducing the gear tooth width and root depth, while keeping mounting distance fixed. This allowed for several fast iterations finally dialing in the backlash so that the differential could run smoothly. The relatively coarse surface finish and larger backlash tend to make the gears a bit louder, but they’re generally acceptable.