Most robotics hobbyists are aware of what inexpensive and easy fabrication options they have available to them and tend to design as many components as possible in their projects to make use of them. A designer with access to a CNC mill will approach problems much differently than one with access to a laser cutter and 3D printer. However, both designers may find themselves stuck between using an inaccurate drill press or spending money to outsource if they can’t find a way to design out a large, stiff aluminum plate with precision holes.
While it is always nice to use the “right tool for the job” how can you know that you have an “adequate tool for every job” so that you are never stuck between dramatically decreasing quality or increasing cost? By breaking down all components into categories based on common manufacturing constraints you can be sure that you have a tool for every category, and therefore every component.
Typical Component Categories
As the length of the largest dimension in a custom component increases, the number of dimensions that can easily be customized during one step of manufacturing decreases. Because of this you can generally group a custom component into one of three distinct categories, each with their own ideal manufacturing processes:
- Small 3D Parts (largest dimension under 6-12 inches)
- 3D Printing
- CNC Mill, Lathe
- Medium 2D Parts (largest dimension around 4-8 feet)
- Laser/Plasma Cutter
- CNC Router
- Large 1D Parts (largest dimension over 4-8 feet)
- Bolted/Welded Extrusions
- Tube Bending
There are definitely exceptions to this, especially when you consider easy two-step manufacturing methods (bent sheet metal), large format machines (expensive industrial equipment), or capital-intense manufacturing methods (casting). However, for the typical robotics hobbyist, the above rule of thumb generally applies.
Additionally, most structural parts can be categorized as either requiring high strength and/or stiffness (metals), or not requiring high strength and/or stiffness (plastics, wood). Again, there are definitely exceptions to this generalization, but it seems like another good rule of thumb for the robotics hobbyist use case.
My Gap in Capability
My 3D printer is great at covering the “Small 3D” category for low strength/stiffness plastic parts. Most people have the “Large 1D” category covered using lumber or pipe (low strength/stiffness) and metal extrusion (high strength/stiffness) with no special tools. This just leaves the “Medium 2D” category for me to fill with a new tool, capable of fabricating both low (plastic/wood) and high (mostly aluminum) strength/stiffness parts.
There are plenty of machines that are good at making “Medium 2D” parts (waterjets, plasma/laser cutters, CNC routers). In order to determine which will best fill that role for me, the first thing to consider is what components I am likely to make based on my Project Roadmap.
- Large wooden items seem very likely. Although I do not have a good way to haul full 4×8 foot sheets to my garage, 2×4 foot “project panels” seem like they will be a common starting material for many builds.
- Medium sized plastic sheets cut with tabs and plastic-welded or glued together is a great way to make quick, cheap, light, watertight structures for robotics projects.
- “Small 3D” high strength/stiffness parts are currently unavailable to me with the equipment I have. While a desktop CNC mill might be the best compliment to 3D printers for fabrication of higher strength/stiffness “Small 3D” components, being able to offload some of that responsibility to the “Medium 2D” tool would definitely help fill the gap.
- Sheet metal (especially when bent and welded) can be a great way to make lightweight stiff, strong structures. However, forming those structures requires secondary processes and their associated equipment (breaks, welders).
- Plotting/rastering/engraving on wood/plastic/metal is another “Medium 2D” process that I will want for my upcoming projects. Cutting cardboard or posterboard might fall in this category as well.
- Axial-symmetric parts (like table legs or cups) can also be considered “Medium 2D” parts and can definitely be made with similar equipment. This type of part is less likely to come up in my projects, but equipment like a 4th axis/rotisserie on a mill or laser cutter/engraver would solve the problem.
While it is rarely the perfect tool for the job, a CNC router is technically capable of doing all of the above. For any pure 2D cutting a CNC router is significantly less time efficient than laser/plasma/waterjet approach, however time is not a constraint that I am worried about for an automated process building a hobby project.
Even though all of these approaches are basically just 3 axis systems with a special tool head, the biggest downside of the CNC mill is that it is the most sensitive to tool load, which will drive extra cost into the structure. The upside to this constraint is that adding an extra head later (perhaps a laser cutter/engraver?) will not require any major structural changes because the load case of the router head is enveloping.
Final Specifications
Based on the above analysis of projected projects and existing fabrication capability I believe that the machine must meet the following requirements:
- Tool Heads
- Mill spindle capable of cutting aluminum slowly and facing wood or plastic quickly
- (Upgrade) Vacuum system for collecting chips and dust
- (Upgrade) Laser head capable of engraving or rastering wood, plastic, and glass
- Size
- 2 x 4 foot working area (driven by project panel size) for all heads
- Upgradable to 8 x 4 foot working area with a similar architecture (in case full sheets are needed)
- (Upgrade) 3 foot (axial) potentially removable rotisserie axis (driven by length of typical table leg)
- 8.5″ vertical travel (driven by rotisserie axis holding 6″ square table leg)
- 2 x 4 foot working area (driven by project panel size) for all heads
- Speed
- Faster is better, but not worth dramatically overspending on. This tool is for one-off projects not cranking out products 24/7
- Fixturing
- Access to clamp down the outer edge of large sheets of wood
- (Upgrade) Vacuum table (to hold 2×4 foot wood sheets, or even sheet metal)
- (Upgrade) T slot clamp table section in one corner only (for small aluminum parts)
- (Upgrade) Rotisserie axis tailstock (for long parts)
- Profile Accuracy
- Should be comparable to the “cheap” option if it were outsourced (Waterjet aluminum plate: 0.010″, laser cut plastic: 0.005″, protolabs aluminum milling on small parts in the corner: 0.005″)
- Cost
- Compared to outsourcing (waterjet aluminum, laser cut plastic, protolabs aluminum milling) the machine should pay for itself over its lifetime (5 years?)
- When possible, use parts that I happen to already have (like giant NEMA 34 motors and external drivers)