1339 Open Alliance Build Thread

I noticed in your CAD that the intake uses the Thrifty Bot bushings at the linkage joints. Would you mind sharing how these joints are built up? Are you putting small stubs of Thunderhex at these locations? Thanks for sharing all this, very cool to see how much work was put into this model.

End of Week 2 Strategy, Design and Build Update

Strategy

Hi folks, I hope that you’ve been having a positive build season thus far. This past week has continued to evolve our team’s understanding of the game, how we think it is likely to be played, and what role we see ourselves having in that game.

Commonalities with games from the past are always fun to discover, and many others have pointed out that the end-game climb is similar to 2013, that the center-field goal looks a little like that of 2017, and that the mechanisms required to play have a lot in common with 2020/21 and 2016. In terms of game play (autonomous, offensive scoring, defensive play, and scoring positions) I think that the game with the most touchpoints is 2014. Limited numbers of scoring objects that are team specific, but that can be interacted with by opponents, is likely to mean increased effectiveness of defensive play. Starvation strategy and keep-away will both be effective from low-level qualifiers to highest-level eliminations. The near-lack of protected space to take guaranteed shots from, and the apparent bounciness of balls in the high goal, are going to make the low goal and close-in high goal shots very effective. Blooping the balls in to either the high or low with bumpers touching the Fender very much reminds me of the use of the low-goal cube as a hard stop in the 2014 game. At all levels of play, one of the quickest and most effective things that a team could do was to slam into the wall of the cube and shoot either high or low.

I very much think that will happen in this game too, and I urge teams to worry less about (and spend less time on) creating long-range firing cannons, and more on dialing in one or two Fender shots that are more difficult to defend, and are less likely to bounce out.

With that in mind, our team is very focused on creating a ball shooter that can do just that: score from the Fender, in both the high and low. To help alleviate the possibility that backspin from a hooded flywheel will cause balls to bounce out, we are adding a top wheel to our hood. It’s possible that once we are good at the Fender shots we will try shots from a distance, but frankly that may not be feasible or even necessary. I’d rather take a guaranteed score than a risky one in this game, especially since there is a Ranking Point tied to the number of balls scored.

Design


The “S” curved ball path in continuing to evolve and will be the way the robot is built. It reduces the number of motors required to create a path, smooths out the transitions, gives us a clearer shot at the high goal from the Fender, and makes it slightly more difficult to defend a shot by blocking. We also feel strongly that intake/output on opposite ends is important for autonomous scoring as well as quickly snagging and scoring loose balls in teleop. Because climbing is a major focus of our strategy, we want to have a CoM that allows a quick movement from Mid to High and High to Traverse with as little swing as possible, and this arrangement helps with that as well.

A detail we added to this robot that isn’t super obvious is that we made our bumper rails across the sides double-wide this year, and are building them out of 1/8" thick 2x1. this gives a very solid second platform for attaching not only our bumpers, but our climber and intake 4-bar.

To answer the question above from nhos, the intake will be deployed only once, as it is on two 15-pound gas springs. The bushing is (we think) fine, since it will never see any kind of RPM and will only be moving in and out a few times. I believe the gap between the pivot mounts will be filled with polycarbonate tube standoffs, but I’m sure Thunderhex would work just as well. Thanks for the question!


One thing that we are carrying over from our design in 2020 is the extensive use of polycarbonate panels and long “columns” of polycarbonate tube that uses these extensively to create rigid 3 dimensional structures that are lightweight and easy to disassemble. I don’t think I’ve seen many other robots built this way. It worked great for us a couple years ago, I guess we will see how it works for this game! It should make it easy to create a ball path that is lightweight without tons of time spent on the router to create lightening patterns, as cool as those things are.

We are also experimenting with 3D printed (TPU) wheel treads for our aluminum 4" wheels. One of the more annoying things about running aluminum wheels is the constant drilling and screwing (or riveting) of nitrile tread. We don’t know if this will work out, but we are creating a few different tread patterns that we are stretching over the wheels with flathead screwdrivers, like bicycle tires and tire irons. This pattern ended up being too slick to use on carpet, but we are designing and printing others.

Build


My earlier complaints about the lack of phone-based customer service at Carbide3D turned out to be moot, as their email responses were fast, and got a new control panel into my hands within a day. A 20 minute swap of parts resulted in a fully functional CNC router, and by the end of yesterday nearly all the gussets for the drive base had been cut. We look forward to assembling them to the rails that the machinist team cut and milled on Wednesday, hopefully resulting in a complete drive base before the end of Saturday. It remains to be seen if we will take the extra day or two to have our powder coat sponsor paint the frame flake magenta, to go with our theme for the year.

Thanks again for staying with us. It’s been a very rewarding season so far, and a privilege to share it with y’all.

Mr. N

18 Likes

I appreciate the response, the robot looks great!

The hyperlink that you posted goes to a page that shows clip on nuts and tube-connecting nuts, you’re referring to the tube-connecting nuts right? Do you have any issues with these loosening?

2 Likes

Please :pleading_face::pray:

4 Likes

It is amusing to me that you included a picture of 125 - that bot was originally designed as a long range shooter before an early QF exit under heavy D at Groton left us a little irked. Some iteration on the release point (which may have resulted in the untimely demise of a few ceiling tiles) and we had the fender shot that was successful the remainder of the season. It’s really a story partly about iteration and designing in options.

2 Likes

Hey there, sorry for the slow reply. Yes, I was referring to the tube nuts. We’ve never had them come loose to my memory.

Just a quick update as I’m super tired after a long work day. The team finished cutting, milling and finishing all the metal parts for the robot, and it is now fully assembled. Wow, just two weeks. That’s a big deal for us.

18 Likes

Odds and Ends

Howdy! Yesterday was a marathon but it was worth it; having a robot frame fully constructed at the end of Week 2 means we are ahead of schedule and will have more time to build, test, fail, and redesign the things that make the robot do it’s thing. It also means we will get it painted in the next couple days, so yay for Imagery!

I wanted to spend today’s update on smaller things that might give additional insight into our mindset, or help for teams out there.


Plain Tubing
One comment we got recently on the Open Alliance Discord was that this robot appears to have a lot of (potential) performance for it’s simplicity. We definitely didn’t design this machine with a lot of custom machined stuff, and that was on purpose. We always (since 2014) use 1x1, 2x1, and 2x2 tubing, and 1/8" thick Aluminum sheet that we get from OnlineMetals or from local supply houses. Our CAD designs are basically drilling round holes in these tubes, then making gussets to fit the holes. We don’t do lightening patterns and haven’t for years, mostly because they take a lot of time to design and machine, and with current tech such as Neo and Falcon motors it is easier than ever to keep the robot weight under the maximum, so much less need to take material off to save weight. In summary, it saves a lot of design and build time and gives us the result we are looking for, which is a robot that performs. This robot, by the way, is looking like it will come in about 10 pounds underweight, including wires and connectors.

Polycarbonate Gearboxes
When we first put polycarb into our gearboxes three years ago we were pretty sure that they wouldn’t work. We saw other teams that we admired (like Team 33) were using nylon in some gearbox and swerve assemblies and we thought we would experiment. Our 2020 drivetrain boxes were made of 1/4" thick clear polycarbonate and they ran through practice, a full competition up to finals, off-season Covid driving, and an off-season competition, and they are still perfect. I know there are some things to be careful of with PC as compared to Aluminum in terms of flex, cracking, locktite (don’t), etc. but the advantages seem to be worth the drawbacks for us, at least. I would encourage teams that are wanting to design their own boxes for flexibility of design, or a new challenge, or to save money, to consider trying this route.

GT3 Belts, in the Drivetrain
Did you know that the GT2 and GT3 belts that are typically used in FRC applications are not the only size profile available? They also come in 5mm pitch/15mm wide, meaning that they can take up the same space, and use the same design calculations, as the much more common HTD belts. HTD are no longer under patent to Gates Rubber and as such are plentiful and cheap, but are significantly less strong than either the GT2 or GT3 belts. The cost of buying a drivebase worth of belts, plus some spares, was about 3X what we would have paid for HTD, but we want to experiment a little and see if it is worth it. We designed our own pulleys (with 20 dp gear inserts for the hex axle) and had them printed using Onyx by our friends at The Thrifty Bot for less than the cost of buying a metal HTD pulley of the same size.


We have run 24 tooth HTD pulley drives inside of 2x2 tubes on five different robots now without ever having a belt break or even show any wear. I understand that the guidance on belted drives would normally steer a team toward a larger pulley size, and that is wise, but I’m sharing our experience and we haven’t had any issues. Even so, we want our drive train to be basically worry-free, and the GT3 belts make that more likely.

Hope this is helpful, or at least gives y’all something to think about. There are many things our team could do better, and we are open to suggestions and criticism. We also know what has been working for us, and think it might help more teams do better for themselves. Thanks again for your time and interest.

Mr. N

9 Likes

Have you considered using 0.1" or 0.09" aluminum plate in place of 1/8"? We’ve found that the weight savings can be pretty significant on unpocketed plates, especially if your robot is primarily constructed out of tube and gusset.

1 Like

Yes, we’ve also used 1/10" plate and it’s fine, nothing wrong with making that choice. This year we had 1/8" on hand and the price of new metal made it the easiest choice for us.

3 Likes

Mid-/week Update, Week 3

We have had to make a few minor adaptations to the kinds of setbacks that typically happen in build season, but we are still doing well schedule-wise, and hope to have a fully assembled robot by the end of the day Saturday.

Setbacks
Our powder coating sponsor is in the middle of a major move, from a 10,000 sf facility to one more than double in size. This means they are backed up on jobs and couldn’t do their normal one-day turnaround, but rather would take two weeks or more. We appreciate their continued support but needed to make a quicker choice, so chose to paint it ourselves with automotive paints. We used acid-etch primer, regular primer, paint and a matte topcoat. We will definitely need to do touchup, as we have had to redrill some holes already. That’s okay!
The gearboxes seem to have been mis-cut. They are the same design we ran successfully on our fall drive base, but this time there is a definite “clocking” when the motor pinions interact with the first stage. It might be that the new CNC didn’t cut them right, or it might have been the speed of cut was too high, or it could be something else. In any case, we need to rebuild both drive boxes before finishing the install, which is annoying but not a huge deal (yet).
The weather was uncooperative yesterday, so we lost a day of work, as after-school activities were canceled. We also struggled with productivity on Monday. These things happen.

In any case, we are still pretty close to on target, and we believe we should have most of the robot functional soon. Here are some pictures of today’s progress for you. Thanks folks, hope your build season is going well.

Mr. N

14 Likes

Beautiful! Have you guys considered a sponsorship from Pepto-bismol?

CAD looks solid!

13 Likes

1726 built a nifty sort of long range shooter for 2009, spent a lot of time playing with vision and it worked great in our shop, but not on the field at our first regional. Near the end of the event, we finally figured out that if we just added a bit to our hood, and waited till we were on the opponents’ trailer to shoot, we could dump a load of balls in quickly, every time. Too late for that regional, but we qualified #1 and won our second regional. So yeah, going for the short, quick, reliable shot can be a good thing in some games.

Also…the robot is beautiful! thanks for sharing your design as you go. I just started reading, I’ll be sure to follow along now.

I love the lack of lightening holes…it’s the sensible way to build a robot.

2 Likes

Weekend Update, Week 3


Upcoming FUN Video Update

We are hoping that when we film with FUN on Tuesday night this week, we are able to give y’all information that you want and need from us. If you have any questions you’d like us to prepare for (including photographs, videos, programming questions, etc.) please let us know here or via PM before our meeting Tuesday afternoon, so we can prepare. Looking forward to seeing you then!

Twinsies!

If you haven’t yet looked at the 7492 Build Blog, you should. This team is doing a super job in documenting their process, and the robot looks great. I keep seeing ways in which their process and their product are running parallel to ours this year. For instance, both of our teams are choosing to build many parts out of polycarbonate sheet rather than pocketed sheet aluminum. It makes machine time much, much less and is less expensive and lighter. We started doing this intentionally in 2020 (though we have often used a lot of polycarb before then) and honestly I don’t know yet what the disadvantages are, if they exist. In any case, 7492 is doing a fantastic job and their work will be helpful to any team wanting to learn stuff about this game.

What I Wish We’d Done Differently, Part 1

  1. Thin Wall Drive Rails. Using 1/8" stock has meant that our internal belt pulleys are harder to fit inside than we thought they would be, which was frustrating this week and may cause problems if and when the GT3 drive belts ever break. If we had used 1/16" stock I think this problem would have not occurred.
  2. Found a Practice Field in the Off Season. This is always the problem, isn’t it. I wish that we had found a place where, when this robot is mechanically and electrically complete (later this week, according to the schedule), we would be able to send the programmers and drive team to get in all those needed hours of practice. As it is we will have to put in precious hours trying to figure out a solution while we are in the midst of the build season.
  3. Not Written Off Pneumatics. Not that I want to use an entire other system that could fail catastrophically. It’s just that we’ve had to come up with some gear-and-sprocket workarounds that probably could have been lighter, faster, and more reliable if we just hadn’t said “no” so quickly. This choice might also affect our intake, which will deploy via servo and gas springs, and will stay down throughout the match. I understand there are risks involved here, that could have been mitigated with some pneumatics.

Plan for the Week

  1. Robot v1 should be mechanically complete on Tuesday, and fully wired on Wednesday. This means installing the climber subsystem and the intake, both of which are waiting on parts delivery (hopefully coming in on Monday, fingers crossed).
  2. Robot in the hands of the programmers for testing all systems on Saturday.
  3. Build Hangar and improve the Hub, with accurate placement of the Fenders.
  4. Procure some donated carpet, larger and more accurate than the piece we have rolled up in my classroom closet.
  5. Build the bumpers.
  6. Finish the team shirts, and continue to work on the sweatbands we are giving away at competition. Make the bumper numbers, the vanity bumper logos, the banners, and other Imagery-related things.
  7. Create the various Autonomous paths from each starting position.
  8. Get a Thrifty sponsorship.

    Have a great week, y’all.

Mr. N

9 Likes

Gorgeous work. Keep it up!

1 Like

Snow Day Update

School was canceled today due to an overnight snow storm that left about a foot of soft fluff all over Denver. We were not sure what our team meeting status would look like, but by noon the streets were pretty clear and we decided to have the shop open for anyone who was able to come in. Six students and a mentor were able to meet me, and together we got a lot done. The wiring all works and passes the tug test, though it still needs to be made more neat. The main breaker was wired up as well. The intake was assembled and installed, and we found that the tolerances were just a bit too tight, so we made a slight redesign and cut two new panels. The intake now fits beautifully, and it looks like we won’t even need the servo-initiated drop on the gas shock springs to start the match, as the assembly stays in place nicely on its own.

We were able to cut the plywood for the bumpers , and have now got the blue set bolted together using angle brackets, and the noodles are neatly taped down. The whole assembly fits, which is great. We’ll finish assembling the red set and the vanities on Saturday.

We’ve now finished setting the CAN ID’s, tested the intake for the first time, and are calling it a night.

What I Like
The new Rev PDH was very clean and easy to set up, and we really appreciate the extra 40A slots this year, since we decided not to run any pneumatics, and everything is built around motors.

We are also super impressed with the Rev Max Planetary gearboxes, and we are using them exclusively when we need reductions greater than a belt and pulley can provide.

Students Leading Each Other
This team has two lead mentors (including me), one professional programming mentor for ULA, and a couple of recent college grads who were team members back in the day. We’ve all experienced the brain drain that Covid has caused, and it has been amazing to me to see how our students have stepped up to train each other while simultaneously doing their own work on this robot build. We have probably the strongest group of Seniors in my memory, Including Isa our Imagery lead, Diego who leads Programming, Helen who is in charge of Manufacturing, Nick our Driver and head of Design and Strategy, and Julia the team Captain. Because of their dedication and hard work not only are we close to having a working robot mid-week 4, but we have an engaged group of younger students who are really learning the ropes and making major contributions. Getting a little choked up here. So much has gone wrong with the world over the last couple of years, and I’m really amazed at these young folks, who are some of the brightest of bright spots in these hard times.

I look forward to testing the full robot on Saturday, when we next meet. There’s certainly a lot that can go wrong, and I look forward to failing early so we can succeed later. Hope you are all having a safe and productive time.

Mr. N

11 Likes

This robot is totally radical! 1339 AngelBotics showcases their initial assembly, provides a CAD overview and details a PathWeaver Demo for Rapid React. https://youtu.be/PkwMCldNpLU

13398toa4BTB-YT-Thumbnail-new

3 Likes

End of Week 4 Update

Climber
Much of our effort this weekend was focused on finalizing the assembly of our climber. As a result of much internal debate, we ended up painting the first stage of the extension element a new shade of green. After doing so we fully re-assembled the climbers and optimized their chain tension with half-links.

We cut out the first iteration of our hook designs for the climber and mounted them to both the static and dynamic climb elements. After a bit of sanding and filing, the carabineer hooks operate smoothly and only need to have a spring mounted to them before being complete.

In order to test the overall functionality of the extending climber, we set up the rigging and ran the spool. Aside from a bit of stickiness in the second climber stage, which was remedied with some lubrication, everything seemed to work well.

Shooter

After receiving some falcon spline-to-hex adapters on Friday, we were able to fully belt up and assemble the power transmission for our shooter. We then cut, drilled, and ziptied some 1/8" polycarbonate sheet to the tubes comprising our shooter’s back wall, giving the balls a consistent, even surface to accelerate against.

Preliminary testing of the shooter seems to indicate that we will be able to make the upper and lower shots from against the fender quite consistently. The back wheel of our shooter offers ~20 degrees of adjustability in projectile angle. I’m not sure if this has already been stated, but we’re currently running 1.5" of compression.

Ball Path

Our singulation system has remained largely unchanged, except for a strip of nitrile tread which was added down its center near the front of the robot. This was largely a response to the intake geometry change, which necessitated greater compression directly behind the last intake roller. The nitrile strip was a quick and dirty proof of concept, and we’ll improve upon it in the next few days.

So far we’ve identified one dead zone in our robot, which is located in the center of the indexer. We think that the ball is getting stuck mostly because of the aforementioned nitrile tread, and the issue should resolve itself when we replace the tread with a better solution. If this does not prove sufficient, we’ll add some kind of tensioner to hold the belts inwards against the cargo.

Just as with the shooter, we added a polycarbonate backing to the indexer, ensuring constant compression as the ball moves through the S-Curve.

Bumpers

Assembly for the plywood components of all three sets of bumpers started today. One set is further along than the others, and will probably be complete in the next few days.

Work Space Update

A parent of one of our students volunteered to construct the hangar for our team and offered us a space to set up some field elements. Today a few members of our team traveled to the new space and set up a carpet. While we initially planned to test our climb there today, we weren’t yet confident in it, and instead opted to stay at the school and test the shooter.

In the next week or so we will tape out different areas of the field and build some improved versions of field elements to practice with.

It was great to see the robot come to life today for the first time. Although I’ve been through a few seasons of robotics now, the joy of seeing something you designed on a computer fulfill its purpose in real life never seems to get old. I hope everyone’s build season is going well, and am more than happy to answer any questions that you might have!

Nick

11 Likes

A few more images to share with y’all to go with Nick’s post.

We received a donation of carpet this weekend which was awesome.

A team parent is kindly allowing us to use their shop space as a practice field. We drove up to drop off the carpet and saw the narrow Hangar that he had built for us. Fantastic, very sturdy. Guaranteed to help us climb!

We spent some time this week with 8283, a team we value as the newest member of the Denver Public Schools crew. They are building a cool version of the Everybot, powered by all Neo motors. I am personally jealous of their teacher’s classroom size and machinery. Awesome stuff.

ezgif.com-gif-maker

Here is a short video showing the ball path for a test fire. Ball was released at around 25% on both the top and main shooter wheels. Playing with the percentage with either results in very different flight trajectories.

Finally, just a pretty picture with the bumpers.

Mr. N

12 Likes

A Part of Our Success

An important but often overlooked part of building robots is the actual time spent building the robots. We have an awesome team, but I think one of the keys to our success this far in the season is our meeting times.

Pre-Covid, we used to meet 6 days a week, as long as possible after school, and full workdays on Saturdays. But this year, our schedule was cut down to 4 days. Now upon hearing this I was quite apprehensive. How would we get a robot done in 2/3s the time? It seems counterintuitive, but I think fewer workdays have contributed to our success. Now there is ample rest time, time to just take a break and collect our thoughts before embarking upon another week.

And with fewer meeting times, our meetings themselves have become more productive. Compared to years before, meeting 6 days a week, it seemed like we had all the time in the world, there was time to goof off a bit. But this year, with only 4 days, there is a sense of limited time, a bit of urgency that we need to keep moving and complete our goals. And so far it has worked. We are ahead of where we normally are and that is due to an awesome CAD team and more productive meetings. Might not work for all teams, just something to keep in mind.

And we even still find the time to have a bit of fun.

Julia

14 Likes

Planning outside of meetings and going hard for a couple hours with everyone there anecdotally looks like it works a lot better than “labs open 6-7 days, come thru!”

The secret ingredient is planning outside of meetings, IMO.

5 Likes