FRC 95 The Grasshoppers 2023 Build Thread

Welcome to The Grasshoppers 2023 build thread!

As public documents get created I will post the links here. These will include OnShape, GitHub, imgur, and YouTube.

I have a few general things to talk about before the build thread really takes off on the technical material.

  1. We are feeling pretty good about this season. We are on #teamRev for 2023, and are going all-in on some REV staples and new products. More on these details later.
  2. Just one student graduated last year, so we are returning at almost-full-strength of 2022 veteran students, plus our recruiting efforts have paid off with around ten new faces. It is looking like the largest Grasshopper team in quite some time! Now we just need to keep all these kids busy…
  3. The kids have decided on a bit of a philosophical evolution in our team’s goals. We are going to weight learning, experimentation, and wild ideas more heavily and competitiveness less heavily than we have in the past few years. Look forward to, hopefully, more and wilder content this year!
  4. The Grasshoppers do not meet year-round. We do not have a super well-defined organizational structure. We play things a bit weird and chaotic most of the time, and sometimes are willing to tolerate more risk than others. Thus our example may not be the one your team wants to follow. Pick and choose what works for you from what we have to share, and remember that we do not have the optimal solution for many things.
  5. If there are experiments or tests that you may be interested in having us run, please make a request. I cannot promise we’ll get to it, but we will give it consideration. Bribery with pizza is strongly encouraged.
  6. We will have two students, @Wesley and @JWC_95, contributing more regularly to the build thread.

Good luck teams! I hope 2023 is good one for you!

50 Likes

A ‘B-Story’ of this year’s build thread will be our pit setup. We won a grant to buy road cases in 2022, but COVID/supply chain/(pick your favorite nightmare from the last 16 months) drastically slowed our acquisition of them. But, this fall they arrived! We will be outfitting and customizing these cases to suit our needs.

For now, here is a mock layout in a 10x10 pit, and a close-up of the battery pockets.

34 Likes

My favorite thread of the season.

21 Likes

Let’s talk about crimping and critical thinking. Crimpical thinking. No wait… I had something for this…

As many of us know: crimpin’ ain’t easy. We are going to try and fix that. This post will cover stripping, die selection, and what to look for to indicate a good crimp.

I always recommend doing test crimps and cutting/polishing the crimp to inspect the cross section and doing load testing, but these tests are destructive. Destructive tests are good to do before and after a production run, but not so useful for the production components. In these example pictures I’m using 2/0 gauge copper HV wire and lugs, but the same lessons apply to any other hex die crimping.

Preparation

Hold the terminal next to the wire
Mark a strip line slightly longer than the terminal barrel length

Carefully strip the insulation by your favorite method. Real strippers are the best bet, but a talented and careful operator can do it with a knife. Just make sure you don’t nick any strands.

Fit the wire and terminal together, push hard, and ensure a small amount of conductor in still visible. This means you have stripped enough insulation to completely fill the terminal.

Ta da! With no strands poking out or cut and confirmation that the whole terminal is filled with wire strands we are ready to crimp it.

Die Selection

So, you might think ‘that’s a 2/0 wire in a 2/0 terminal, he should probably use a 2/0 die.’

Well, that’s mistake 1. We need to think critically and evaluate what we are seeing and what it means to our final product. I am going to test the 2/0 die first.

When I crimp the terminal with this die I notice a key problem: the die hit each other, and thus a portion (in this case, most) of the crimping force resolves through the die and not the crimp. Sad times.

This crimp is so bad I actually pulled it apart with my squishy caveman hands.

Next, we’ll go down two sizes, to show what a die that is too small looks like.

In this case we have a LARGE standoff between the die and a large portion of terminal material squishing between the die. This is better than the too-big die, but still not great because a large portion of crimping forces are being resolved through the excessively large flashing.

Now the goldilocks die, the right size, allllmost bottoms out the die, but not quite.

It forms an extremely thin and small flashing. This does take some crimping force away, but much less than either being too big or too small.

Here are the three crimps together. Left is too small, center is good, right is too big.

Here is what the flash looks like from the side.

Now, why didn’t the ‘right’ sized die work? Well, could be many reasons. Might be sized for a lug of different wall thickness or it could be the lug isn’t quite the right flavor, or it could be that I did not spend $900 on a name-brand calibrated crimper and die. The important thing is that I tested the termination technique and found the right recipe for this job.

Test your gear, think about what you’re seeing (where is the force being resolved?), and how does it impact your goals (does this crimp have enough force to be made properly?) and you will likely have better results. Do not just follow ‘what looks good’ or ‘what seems right’ without thinking about it critically.

This is what the crimp test section from this effort looks like: solid copper.

The test section is from what I initially made this recipe for an EV project, but I redid the die sizing check with pictures for this writeup.

30 Likes

What’s your polishing method?
I’ve sectioned on a bandsaw, but the teeth definitely smeared the copper around.

1 Like

Progressively finer sandpaper does the trick. Some compressed air to blow particles out helps.

You ought to be able to rip the crimp apart and see each strand formed into a ~hex under a microscope, but I don’t have the ability to photo that in my home shop.

1 Like

A GOOD CRIMPING DEVICE NEEDS TO BE ADJUSTED FOR THE BEST CRIMP. While James was testing crimped connections, I was making crimped connections to test. While its great to crimp, without a good squeeze on the wire and ferrules there isn’t a good connection between the wire and the ferrule. We occasionally need to adjust our crimping devices as they wear, increasing play in the linkages that give the mechanical advantage to crimp thus applying less force to whatever it is squeezing. Usually there is a cam screw that you can turn to compensate for these tolerances. It should take some force at the end of the crimping movement with nothing in the die. As James explained above when hex die hit each other a portion of the crimping force resolves through the die and not the crimp. On the other hand, manual ratcheting crimpers are designed to have the die touch, which is why we look for a little force at the end of the crimping cycle. The adjustment of many crimpers is similar to this one

I can’t wait for January seventh! It better be a water game.

11 Likes

Swerve dev chassis runs! First time it hit the ground too. Apparently simulation helps!

And a quick test of field centric mode looks decent too.

14 Likes

Limited to a little less than half speed in these tests. Didn’t want to smash anything too hard yet.

2 Likes

Content links below. These will be growing throughout build and competition seasons.

imgur album
YouTube Playlist
OnShape Main Robot Assembly

5 Likes

@0NA0 and @Wesley have been working on various programming efforts. The latest to show from this week’s bonus meetings are a swerve auto and absolute heading control with joysticks instead of a game pad (offers the driver more control resolution).


Thrustmaster T16000 if memory serves, old versions but they hold up nicely.

We have mostly assembled a new robot cart. This will be the first time in a long time we’ve had a ‘real’ robot cart. We don’t have the handle material yet, of course.

Model here.

4 Likes

To follow up on the topic of flashing/extrusion during crimping (it has come up on other threads) I emailed my coworker who used to work at Burndy designing tools (including crimpers) to make sure I was not losing my mind or forgetting something.

Big takeaways:

  • Flashing is normal with hex die
  • DO NOT rotate and re-crimp to smoosh the flashing, it will damage the crimp
  • Here is the link to a flash-cutting die, sadly not available in FRC-applicable sizes
  • Living and working remotely on your yacht and sailing to the Bahamas for the winter is a galaxy-brain move
13 Likes

Hmmmm… If only I had a yacht with room for a full practice field

2 Likes

New robot cart is together! This puts the robot at a reasonable working height for the pits and yet low enough to get through most doors. This is the first time in a while (20 years?) that we have made an all-custom cart.

8 Likes

We hope everyone had a good kickoff!

We are taking a little bit of a different approach to our initial season work this year. Instead of creating and evaluating specific robot archetypes we decided to identify robot functions (as normal) and then prioritize them based on our own criteria (not necessarily the best criteria for every team). This is what we arrived at yesterday, though nothing is written in stone yet.

And a quick (also not set in stone) pps analysis to determine the potentially highest-value activities during each portion of the game. Cycle times are just my best guess for now for what I think we can do, knowing we make faster-than-average drivetrains and encourage our drivers to run the wheels off. That is to say YCMV, don’t take our numbers as gospel.

Some of my own impressions as of now (subject to change):

  • Seems like a really good year for middling teams to focus on low/medium scoring opportunities given the relatively small points disparity between low/high, especially with links. Hi link = 20 pts, med link = 14 pts, low link = 11pts. Unlike numerous prior games the low goal is worth >50% of the high goal, effectively, and contributes equally to earning the ‘cycling’ RP.
  • Auto is deceptively important, though the difference is just 1 point, the faster cycles available in auto (going to midfield instead of full-field) will bring those link points closer
  • Engaging the charger during auto is worth ~3 mid cycles and 2.5 high cycles, likely worth 30-45s of teleop play, yet ought to be achievable with a minimum of programming and design effort. Very high value activity that many teams should focus developing on before worrying about auto-scoring game elements. Docking is even easier, and still worth 8pts. I hope every team will try to have a ‘drop item in low goal and dock’ auto move.
  • If your robot can engage in auto and teleop your partners only need to park at the endgame to get the charger RP. This also frees up robots on your alliance to score nodes until the end, increasing your alliance chances of getting the sustain RP
    I am likely wrong here based on the current manual wording
  • A chassis width of ~<25.5in will enable 3x teleop engage if everyone does it. This could be a valuable feature for playoffs and late-season stamina.
7 Likes

James, I’m totally missing something here. Pts 3.7, 4.7, 6.7??? Cycle times are pretty much what we came up with.

QFT.

Look great. We are using the same swerve with neo too. Are you using pigeon2? Because we have some problems with our field centric.

1 Like

We are currently using the old pigeons, but I think we have ordered some pigeon2s.

If your issue is that the robot skews in the direction of rotation while driving & spinning, then this is a natural consequence of the discretization of the first-order kinematics. More information about this here.

1 Like

Is your code shown on github?