Week 2 - Drafting

Meet dummybot.
Dummybot helps test all of our ptototypes.

In what seems just another blink of the eye, Week 2 has come to a close. In pre-season, we set a rather aggressive schedule. Week 2 is normally where we start to find discomfort with our aggression. We're still meeting our goals, but still feeling some discomfort. Or tiredness? Let's see if I can recall the major beats of Week 2. Grab your magnifying glass, and let's take a peak!

The major goals of Week 2 involve building on the prototypes and learnings of Week 1. We built a decently high fidelity output for handling Coral last week - this week we took a shallow look into deep climbing. Electrical continued supporting software, adding more components to our test board for initial testing and familiarity. Software themselves started skeletoning classes and functions for the currently planned subsystems. Design continued taking the prototype measurements and building subsystems on their laptops using Computer-Aided-Drafting (CAD).

We generally inform and warn students and new mentors about the "CAD Crash" that occurs during weeks 3-4. The first 90% of the design is together, but once we get into checking every nut and bolt, everything slows down. The number of parts we can manufacture slows to a trickle, and nothing new gets into the Electrical/Software team's hands. The quiet calm of the shop juxtaposes the thunderous stroke of laptop keys coming from the design table.

For uncertain reasons, we started the CAD crash early, before the end of Week 2. This is fine? (First time I've said that all year!)

Student and mentor actively engaging in the CAD Crash.
If we didn't do CAD, we wouldn't have a CAD Crash. <taps side of head>

Our mentor musings this week once again skipped ahead to future steps - we're getting into the devils and details of the design, but then we actually have to manufacture, fabricate, and assemble this thing. Our manufacturing resources have proven to be a limiting factor in years past, but once again this year we are determined to leverage more of what we have available to us in the shop.

Previously we fought through a process of center punch drilling using paper stencils and elbow grease. This was low precision, and led to a not insignificant amount of match drilling and snowman-ed holes. It did mostly work though, and we never had any actual parts failures. But we can do better. We have a Computer Numerical Control (CNC) mill that can produce much more accurate hole and pocket patterns, however it has only about 13 inches of linear travel space. Most of our parts are longer, and need hole patterns at both ends. An experiment we're thinking about utilizes 3-D printing - create re-usable jigs that slip onto our chosen tube material size with pre-centered holes to match drill. For smaller parts, we have started working with test pieces on the CNC mill, cutting simple hole patterns and checking tolerances. Each machine has its own little quirks, but our test patterns look promising.

Simple hole patterns, but we all need to start with the foundation.

Design:

Kicking off with design this week - we have lots to manufacture, and we have a set of robot priorities, so our focus has been implementing systems for our priority 1 and 2 items so we can get them on the robot as fast as possible. Does this mean we're more than 90% complete with bits and bobs involving cycling game pieces? Yes, indeed! Does this mean we still have a rough space claim and no real definition for an endgame mechanism? Also yes, indeed! It feels weird, but the space claims should prevent any issues even as we add final details to our "primary" subsystems.

As with many things, the final 10% of the design takes 90% of the time. We're on that side of that fence, getting into the details - resolving forces and moments, designing gearboxes, checking final actuations and clearances. We will hopefully be cutting metal soon! (Especially if we want to hit our week 3 schedule targets!)

Mechanical:

Mechanical slowed a little bit this week. We have the two critical mock/practice field elements (Reef and a Human Player station) complete with April Tags. We did end up prototyping more with the climber and Algae - once again testing concepts, then refining and gathering measured data that we can feed into the design team once they're ready.

We always said building with wood is faster.
Yet again the field elements are finished before the robot. Slackers.

Electrical:

Once more, our Electrical team was very much in service to our Software team. With the basic chassis together, the next items to place on the test board were a suite of sensors. This year we are hoping to incorporate vision tracking for increased accuracy in autonomous and in scoring commands. Last year we made use of a Time of Flight sensor as a beam break, noting when we had acquired or ejected a game piece, and will do so again this year. Finally we added an LED strip for increased horsepower. It will also help visualize changes in robot state for our drive team during matches and our software team during debugging, but really, it’s all about that +10 horsepower.

Software:

Last year was our first year with swerve, and specifically third party swerve libraries. It took 3 weeks to get running through much trial and error. (Mostly error.) With last year’s lessons in our back pocket and new versions of the libraries for 2025 (and one bad motor controller swapped out), our chassis is moving! With sensors now attached to our test board, Software team started testing additional libraries and incorporating feedback into our eventual robot commands.

Lights! Camera! Time of Flight!
This is the earliest we've ever given a sensor to software.

Week 2 Wisdom: CAD

Once again I feel bad putting 'wisdom' in the name of this section, cause I have such limited knowledge of CAD. I've never created a primary geometry, I've never imported models, and I've never taken a sketch and converted it into an assembly of 3D components. With that disclaimer out of the way, what wisdom can we find in CAD?

Like many things about FIRST and STEM in general, CAD is a tool. If we think about the "simple" act of cutting metal, we've developed several tools like endmills, lasers, the power of friendship, or high pressure jets of sand-infused water that enable us to cut faster, more accurately, and/or more efficiently. CAD is our tool to create and verify designs faster and more accurately.

Caveman 1: "I find stick"

Caveman 2: "Stick too long"

Caveman 1: "I go find new stick"

These poor cave people should have used CAD to see what length stick they needed before lollygagging through the forest.

(While proof-reading, Kayla suggested the cavemen break the stick to get it shorter. She’s very good at thinking outside the box.)

Like most tools, there is a learning curve to using them properly. Worse still, a tool used improperly can slow things down, or produce its own failures. In my first absolute humdinger of wisdom - one should learn how to use CAD. Learn the tools available for mates and splines and patterns, learn the keyboard shortcuts for when you gotta go fast (tm), and learn how to order and organize all the sketches, geometry, parts and assemblies. (Dear Bearcats - an assembly called "Celebrated too Soon" is not very descriptive. And we are still not cut ready.)

Once you've learned how to use CAD, the next remark from Captain Obvious is to <use CAD>. This is more about the thought process behind CAD - Did you use the data measured from the prototype? Can your part be manufactured or purchased? (Just where does one acquire a 23 ½ tooth gear?) If it's multiple parts, can it be assembled? Can tools fit where they need to? If it's a moving part, does it collide with other static or dynamic parts? It may seem like a tedious task, however adding the bolts and rivets, planning where wires and pneumatic tubing will run, will only increase the correctness of your final pieces. Fortunately, most CAD review items can be distilled down into reusable checklists - yet another tool in your belt.

We leave ourselves helpful little notes in CAD.
(Half the time it just says "Hi Bobby!")

Here’s a few sample checklist items we’ve got on our list:

• Belts - Loose, taught, tight? Add or subtract 0.005 depending on the belt application.

• Chains - Will always stretch. Always add a tensioner. <Always.>

• Bearings and Axles - Arrest all lateral movement. We generally use 3D printed “doritos” to capture bearings, and bolt+washer to capture axles.

• Standardize hardware - All bolts in the 10-32 family (where applicable), all rivets in the 3/16” family (where applicable).

Finally, even when the CAD is "done", make sure you continue to use the CAD. In my experience, it's not always the designer who ends up assembling a component. The builders should learn some CAD basics to navigate through a software assembly to help with their physical assembly. Just tell them it's like the Lego's, and CAD is the little instruction booklet.

Quote of the Week:

Mentor 1: "That begins with a 'g'."

Mentor 2: "Yeah, we don't do 'jussets' here!"

<Everything on the robot is now called a Jusset>

-B