Snowpocalypse Now

Snow and ice - not at the lab.

I personally hibernated at my house, and missed 2 days in the lab this week. Fortunately, Slack works, and all members of the team were able to keep up to date and make progress in their respective areas. This week was about refinement, starting "real" CAD, and getting the first chassis finished and into software's hands.

Prototypes Evolve

Our V2 prototypes are coming together. These are 2-d sketch CAD'ed wood mock-ups that are approximately the right size, approximately the right intake/output, and will be the first prototypes we measure against their respective sub-system requirements. We will start using real motors for power, and start tuning for precise control.

Proofs of concept are done - the next version will be cut to specific dimensions on the router.

One of our not-so-secret weapons is the use of (real) 3D printers. Previously, we had one printer that could try to print prototypes and/or small parts. Most of the time, they didn't work well. Now though, we have three Mark-Forged 3D printers - running almost full-time, printing prototype parts, real parts, template to assemble parts, and more. It has taken me some time, but I now see the magic of 3D printing. This year, our prototypes are higher fidelity than in the past, and turning out faster than in the past. We're recording prototype data the same as the past. Which is good.

We've 3D printed so many things! Except the hammer... and the file...

Here's a fun story. We have some PVC rollers in a prototype. Hoping they go on the real robot. Last year, we made 3 elevators, had a ton of bearings. We had a ton of belts from other systems in previous years. And we have those new 3D printers. Turns out - we can print pulleys that press into the PVC and are powered by all these old belts. Prototype made in 2 days with one $8 Home Depot run.

The last note on prototypes that I can share here, is that this year (last year, the year before that...) details matter. Oftentimes I peruse Chief Delphi, looking for technical notes, programming updates, and of course, all the build blogs. So many build blogs. It's great to see! A lot of teams showing their strategy notes, prototype progress, and sharing how they run their manufacturing. I find this to be a good thing. One thing I have noticed, is that so many teams have started converging on paths for shooters, intakes, and fuel cell storage. I'm pretty sure I could grab a number of shared sketches from Chief Delphi and mock our planned robot. :-)

Here's what our head coach rendered for our 2020 robot:

It.... Looks a little familiar... Maybe it had blue bumpers..?

Anyway - with so many teams running down the same path, what's going to be the competitive deciding factor among teams? Details. And well, drive team. But - all drive teams being equal (they won't be), the devil will really be in the details this year. Like 2015, this is a very optimization heavy game. Over half the teams in the world will probably build a shooter (er - launcher) of some kind. Probably only a hand-full will be able to shoot 5 fuel cells into the inner goal in less than 1 second whilst driving toward the next fuel cell. Those engineers are so preoccupied with whether or not they could, they didn't stop to think if they should! ... Well, in this case I suppose they should!

Am I doing this right?

As your teams finish your prototypes and begin cloning the critical dimensions into full robot CAD and final mechanical systems, keep the details in mind. Your prototype should meet your robot requirements - if not, how can you tweak it to meet or exceed? What are the critical dimensions of the prototype? How can we score faster? How can we improve our accuracy? How do we ensure that these stupid balls don't stick together and squish into some compartment that jams the whole system? (I'm waiting for this to happen to ours. Both because it will be spectacular, and to see what ingenuity it brings about!)

Evolution of the intake -we should have solid contact on the ball for the full duration.

Hey look! A flywheel shooter! <gasp!>
(Its's only a model...)

Sustainability

Imagine this - you're just hanging out at the lab, watching some stuff get put together, and head coach comes in and says "If you're a sub-system lead, you get to build robot 1. Also, you don't get to build robot 2. You'll be doing other things." Erm - what? None of our student leads will be able to build, assemble, or wire the second robot? How is that going to work. (This was new to most CyberKnights, I do know that some other teams already do what I'm about to describe!) For a team with ~35 students, we actually have plenty of hands to help out, and a pretty even spread across the age range. For every lead, there's 1-2 students in training or just learning the foundations of mechanical, electrical, software, etc. The purpose behind this decision is to solidify the pipeline (while giving us a competitive advantage!) As leads progress through the process of design, manufacture, and assemble for robot 1 - they will document and teach the next group of students. Those students will then go on to build robot 2. Next year, those students, already having built a robot, will build a another robot - while they teach the next group of students, and so on. If we pull off this trick - it will create a completely sustainable pipeline of trained, prepared students and leads.

These CyberKnights are learning how to build the modules.
On robot 2, they will own them from the start.

Plus, the existing leads (for this year) make up the drive team. So, with no need for robot 2, they get to spend that entire time... driving robot 1. Yeah, this can work.

Outside of creating a sustainable pipeline, this year we also created a new team wiki. Each subsystem has a page, and it is part of the exit criteria that the subsystem design process is documented. As we build more subsystems and more robots, we will start to build a library of easily replicable pick-and-choose pieces. If we continue to keep documentation entwined within process, and make small process improvements every year, this can soon become second nature to record our technical history.

Speaking of technical history - here's our current chassis.

Chris... poor guy, still hasn't gotten any sleep since last build season...

Just keep swimming.

-B