I recently sent out a link to our CNC team Titan quality #29, and when I first saw it I thought ok I get why he puts so much emphasis on this but all those forms, isn't that a bit overkill? I dismissed it at first as too much red tape, but then I have been thinking about it and remembering all of our failures in previous years. In this article I wish to present what was successful and how important it is we have some kind of inspection on critical structures as well as identifying what structures need the most attention.
In 2012 we attempted to make a tier 1 type of robot. We outsourced the CNC tasks to m2 global to cut parts, and we had two of the best CAD designer students/mentors from 148 on our team to run the show. We gave 148 a bit of a scare that year, the design was pretty good, but setting aside the "put magic here" missing cad. There was a fundamental problem with the alignment of the shooter. Since it had to rev up to fast RPM's the misalignment caused serious vibrations, and as a result not only did this shatter the encoder casings (giving our programming team a headache of noise to solve) but the planetary gears would eventually fail and we would have to replace them.
In 2013 it was a similar story, but this time we worked along side Texas Torque (Who won the championship that year)... to reiterate, we had the same drive design as a championship team, what happened! Over time the gears to the drive would mesh too tightly and grind away and robot would then no longer be able to drive... how did this happen? Well, some say the robot had too thin of sheet metal on the drive frame and when it got hit it knocked the drive out of alignment, but I believe it was a combination of possible misalignment either in the parts or assembly. While the cause may never be never truly discovered, the point remains as seen in next entry.
In 2014, due to many factors, we abandoned this workflow of custom sheet metal to start using COTS, which really had a better fit for our capabilities at the time.
In 2015 I remember making a printed part that mounted on the robot, while where this was... wasn't a critical part like with other years, it reminded me of the imperfections and loss of precision we had on the Curivator project. The outtake from this is that there are parts that are not as critical and so, given the nature of getting things done quickly. I think I can live with this as long as we know which parts don't have to be as precise. The long term results from these had no consequence.
In 2016 we used different technique where most of the parts were already COTS, so the quality of the parts were already confirmed among the vendors. This time around was like a redemption from 2012 where the concentric alignment was there and we didn't have the same problems as before.
In 2019 we are going to turn over CNC manufacturing to the students, so my goal is that we take the time to check our work. The CNC machine we have is a bit harder to use than a professional grade, but we can still achieve greater precision with it than doing it by hand, we just need to inspect the parts mostly the alignment of the holes, with a caliper. For CAD teams be aware of what structures require precision, by looking at concentric mates and the load requirements on them... drives and shooters come to mind.