I was talking to a few people on the team recently about where we struggled this season, we came to the conclusion that by end of week 4 all of our sub systems were at ~80% cad complete, besides drive train which was finished. We were definitely behind schedule. Having multiple unfinished subsystems, all with a high percentage of completion, may sound like a good thing, but it is not. Its a failure of prioritization, process control, and time management.

Going into kick off we created a lose schedule, with the goal to have the robot manufactured by the end of week 4. It was bag night when the robot was finally assembled for the first time. We even had additional robot assembly work planned for our unbag time, finishing the level three climber. We had hoped to use our unbag time fort drive practice. This caused us to go into our week one event, with less than 3 hours of drive time/robot run time, and our robot performance at our week one event suffered from the lack of run time. We had multiple break downs, and had to redesign a few things between matches. There were many kinks that needed to be worked out on our robot, and unfortunately we had to do it at our first event.

At the end of week four it should have been very clear to us that we were behind schedule, and needed something to change in order for us to get any significant run time on the robot. We could have recovered some ground if we had prioritized out work. Picking a sub system, doing the last 20% of the work required in order for it to be finished and sent to manufacturing. Instead we had three different, under staffed groups all work on finishing separate sub systems.

During our reflection conversation the failure of prioritization became very apparent. Focusing our efforts on one of the subsystems would have allowed testing and iteration to begin on that sub system sooner. Our problems with incomplete work is similar to the problems faced within manufacturing. There are many different processes focused around creating and enabling efficient manufacturing.

We hope to make use of lean manufacturing during the 2020 build season to get more done, and help build a better, more competitive robot. Lean manufacturing involves reducing waste from the manufacturing process. Waste is defined as anything that does not add value. Waste can be defective parts, time lost due to part shortage or even work in progress.

Our build season problems where mostly caused by having to much work in progress. We had invested the time split across 4 subsystems through out build season. Finishing final assembly of the robot left no time for any return on investment. Return on investment in FRC can take the form of driver practice, optimizations, finding design flaws, weak parts, and opportunity for iteration.

For an easy example lets take a look at a single sub team group, robot design(CAD) with four equally capable team members. Say a single person is capable of completing 5% of a task in a days worth of work. To make it easy lets also say we had 4 robot sub systems and they were at all at an estimated completion of 70% done(14 days of work for each team member). If each team member works on a individual sub system it will take a single person 6 days in order to finish the sub system. At the 5th day you will have 19 days of work in progress waiting on the final day of work to be finished. Now multiply this by 4 to account for each subsystem and you have a maximum work in progress of 76 days. This 76 days of work has not provided any value yet.

Now lets say we are trying to minimize work in progress through out the build season from the start. In order to minimize work in progress we will have our four person design team all working on the subsystem until completion. As a group of four the design team is able to complete 20% of a subsystem in a days worth of work. A subsystem will be finished in 5 days. On day four there will be a maximum work in progress of 16 days of work. Once the system has been finished on day five return on investment will start. It will still take the same amount of time to finish all four sub systems but you will have the first subsystem providing value in 5 days, the second in 10 days, the third in 15, and final in 20. Compared to all subsystems starting on day 20 in the first scenario.

This was a simple example that only looked at a single sub team. Its most likely that subsystems will depend on each other which will require analysis to determine the critical path. There are also other sub teams and coordination is required. It is not simply enough to CAD the robot, we must also manufacture it and assemble it. This is were another benefit of minimizing work in progress exists. Once a subsystem has been finalized it can be sent to the manufacturing team to be made. By minimizing work in progress, the manufacturing team will be able to start making parts for subsystems sooner. There will also be inefficiencies around more people working on a single task. Adding more people does not always mean a task will be finished sooner.

A FRC team can take the lessons learned from the business world and apply them to become an effective and more competitive team. Our team will be working towards minimizing work in progress during the build season. Look forward to future posts about our use of tools to minimize work in progress, and our progress as we grow as a team.