At DRP, I was given loose design concepts and expected to refine them into manufacturable products. It's often a daunting proposition, but I found it can be broken down into loose steps:

• Establish the product requirements with the customer.

• Consolidated ideas into a concise functional specification. 

• Make designs fitting the functional specification and model them in CAD. If there are existing CAD models or prototypes, they are modified to fit the functional specification.

• Manufacturing methods are evaluated, and production tooling designs are drafted.

• Product undergoes another set of revisions to ensure its manufacturability, tooling is adjusted accordingly.

Rinse and repeat.

Maintaining product quality while minimizing production costs was the greatest set of challenges I faced on this project. High-quality standards, leading to supreme durability are the product's primary selling points; it was critical those attributes were maintained. Given DRPs machining capabilities, I turned to vertical integration and adjusted the shovel's design accordingly. All shovel components and their associated production tooling were designed and made in-house: Uniform quality is assured through every stage of development and production.

When the design for the shovel was finalized, I was tasked with developing a process to manufacture the shovel. 

The shovel handle assembly consisted entirely of cylindrical bodies with axial symmetry. The geometry of these components made efficient use of bar stock, and are machined with few operations.

The blade was much more challenging, its final geometry was ill-suited to any machining center (excluding 5+ axis grinding). The shovel has a profile easily cut and bent from sheet metal; however, it also required sharpening along complex, 3-dimensional curves. These motions are easily made by a human arm, and I figured they could be replicated by a 6-axis industrial robot the shop had available. This ancient robot, a Fanuc LR Mate 200iB, was designed in the early 90s, produced in the early 2000s (older than me!), sat in a corner unused. It was too small and old to effectively serve the CNC machining centers. A prime candidate to make shovels!

Mimicking the motions of a human arm proved to be very difficult. The robot was programmed entirely by teach pendant, and writing then transforming functions representing blade curvature was beyond my mathematical capabilities. Ultimately I opted for a more unconventional solution, developing a 3D-printed "programming tool". It is a fixture holding angle sensors and represents the curvature of the final ground shovel. The programming tool is brought to a point, adjusted to meet the correct angle reading, then the point and angles are recorded with their respective approach/retreat cases into the robot's memory. This method allowed the robot to be taught complex routines, using only a teach pendant, avoiding the expensive Fanuc ROBOGUIDE programming and simulation software.