FRC 5987 Open Alliance Build Thread

Right now our plan is only to reach the 3rd rung, not to go for traversal. Once we latch onto the rung 3, we’ll drive the arm in the other direction to lift up off of rung 2 and then engage a brake to hold ourselves in place there.

In theory it would be possible to get to the traversal rung with this design, but it would need a different way of releasing from rung 2 in order to swing under rung 3 instead of over, and it would need the ability to unlatch itself from rung 3 once it’s hooked into rung 4. If everything works well that might be something we work on after the first comp, but we decided early in the season that climbing to traversal wasn’t a high priority for us.

It looks like it can fairly easily be modified to go to traveral. Instead of running 2 hook, run 3 with the two hooks on opposite ends facing up and one on the end facing down. Grab onto bar 2 with the side with 2 hooks facing up, then you can spin it 3 times to get traversal.

Excuse the bad ms paint drawing

1 Like

Yeah, that’s the basic idea I was trying to describe before (sorry if it wasn’t clear). It would require switching the hook for rung 2 to a latch, and developing a way of opening the latches automatically. Should be possible to do, but not a priority for us at the moment.

Since y’all are Open Alliance and so are we, I will tell you that I think with very little tweaking (just figuring out how to open and close the grabbers) you have probably got among the fastest level 4 climb types in the game. Probably possible to get it below 5 seconds. I wouldn’t give up on that as a priority

12 Likes

We started checking speed vs distance yesterday, and found that with the hood angle we had we wouldn’t be able to shoot from as close as we’d like. We remade the hood plates for a slightly steeper angle, and we are now able to make shots from up against the hub (with the intake down). In doing so however, our adjustable hood now interferes with the shot trajectory when retracted and will have to be remade. Two steps forward one step backwards.

ezgif-1-82c09d88a1

We are having some problems with backspin causing the balls to bounce out, but a combination of playing with the speeds of the flywheel and feeding conveyor, aiming for the side of the hub rather than the center, and changing flywheel speeds between balls has helped mitigate that. Our job for the next few days will be to find the right combination of speeds and angles to be able to shoot accurately from our entire range.

ezgif.com-gif-maker (4)


At the end of the meeting today, we also tested our climber again after fixing our model hangar. We weren’t lined up quite right at first and this happened

After we recovered from that mishap and put the hangar back together, we got another shot

You can see that, although they are connected by an axle running through the robot, the arms had some angle difference causing only one to latch onto rung 3. After that latch was attached manually, it failed under the weight of the robot. We’ll be looking into what we can do to improve the rigidity between the arms, the strength of the latches, and the required alignment accuracy of the hooks.

5 Likes

In the past few days we finished the (preliminary) tuning of flywheel speeds for our range of distances, and we’re shooting pretty accurately.
ezgif.com-gif-maker (5)

Somewhat unintuitively, we are most accurate shooting from the outer edge of the tarmac; closer shots tend to bounce out more often and further shots are necessarily less accurate. We are attempting to compensate for the bounce-out by purposefully aiming close shots off-center of the hub so the ball hits the side slope and bounces sideways rather than directly upwards. This is showing some promise when done manually, the next step will be to automate it.

We also made some mechanical improvements to the intake and conveyor systems to better control the cargo. We accidentally had an intake belt that wasn’t properly spaced, causing it to skip when under tension. We extended the front belts of the conveyor and added flex wheels to help the balls transfer from the intake. We also increased the ratio of the conveyor to slow it and avoid marring the cargo.

At the end of the day today, we got to do some good driver practice. We had wanted to practice intaking and shooting, but unfortunately our targeting LEDs died right before we started, so we just practiced intaking. Now that the belt on the intake no longer skips, we see that we can increase the intake motor power to be able to intake two cargo at once. Here’s a short clip of our practice.

6 Likes

I don’t usually post two days in a row, but I want to share that we had our first successful climb today without human intervention

We were able to climb a number of times in a row without issue. We still haven’t gotten around to widening the rung 2 hooks, so it takes a long time to line up before the climb begins. That will happen in the next few days. Once the robot was up and locked on rung 3, we saw that the polycarbonate latch was bending and slipping under the robot’s weight. We made those parts out of PC for ease of manufacturing, but we’ll be switching them with aluminum for better strength. Overall though, we’re very happy that the climb is looking promising.


As an aside, we’ve had a problem in the past week or so having accidentally ordered endmills from the wrong company on Alibaba. The two products look identical and have the same description, but perform vastly differently. These need a much slower feed rate, DOC, etc and still have problems with chip welding.

Who needs chip clearance when you can make aluminum puddles?

5 Likes

The focus for the past two days has been on getting an autonomous routine programmed. We started with a two ball auto and then moved up to three balls.

Unless we have extra time, we’ll probably stay with this as our primary auto routine for ISR2. Hopefully we’ll be able to add at least one more two/three ball routine in order to play well with partners. Unfortunately, our 6-ball auto isn’t working quite yet.


 
We also machined the hooks for our climber again, this time from aluminum. We got to test them right at the end of the meeting today, and they worked perfectly. We still need to work on automating the climbing sequence and easing the alignment, which will be work for the next few days.

 

We are going to XO 7039’s practice field to test our robot tomorrow. Big thanks to them for inviting us. Hopefully we’ll break stuff in an interesting way :slight_smile:

Edit: We also watched 179’s release video in the shop today and are really impressed with their climber idea. It seems like it would be a great way to modify our climber to get to the traverse rung without adding any releasing latches. It looks like we’ll have to deal with frame perimeter issues to make it work, but it’s quickly become the front-runner idea.

6 Likes

We’ve been putting the finishing touches on the robot in preparation for our first competition at ISR2 next week. I’m cautiously optimistic about the robot we’ve built. Our reveal video isn’t quite ready yet so I’ll share an uncut example of one of the practice matches we’ve been running in preparation.

Away from the robot, we’ve been working on prototyping potential upgrades for future competitions. Since our main priority this year was modularity, we have the ability to switch out pretty much any mechanism on the robot for a different version without affecting the others. So we’re looking at different possibilities for intakes (adding/repositioning rollers), shooters (adding a top roller), and climbers (modifications to get to traverse). We should have some more information on those fronts after the competition.

7 Likes

We are proud to present our robot for 2022, Stardust

Reveal thread

Check us out starting tomorrow at Israel District 2!

9 Likes

I’ll have a post coming soon with more details about the competition and what we’ve been working on for the past week, but I want to update here that we ended up ranking 1st at ISDE2 and taking home the blue banner.

I’m extremely proud of our students and the robot they built this year. Their hard work and dedication definitely shows. Thanks to our great alliance partners 3339 and 1580!

9 Likes

You can see here the robot book that we present to judges, which describes each mechanism on the robot and some of the cool programming stuff we do.

Robot Book.pdf (5.6 MB)
 

We were happy to win the Innovation in Control award at ISDE2 so I want to highlight the things that the judges commented on:

  • In order to check that the robot is functioning properly before each match, we use a BIT (Built-In Test) to activate all of the mechanisms and ensure that their responses match our expectations. Every motor on the robot has an associated encoder, and by checking that the encoder value changes when the motor is activated we can confirm that both the motor and encoder are functioning properly. For mechanisms with position control, the mechanism ends at its zero position both to confirm the encoder hasn’t slipped and the PID controller is working correctly.
  • We use swerve odometry to track the robot’s position through the entirety of the match. While the odometry on its own gets less accurate as the match goes on, a pose estimation from our vision system is used to reset the position whenever the target is in sight. This gives us a good estimate of the robot’s location and orientation throughout the match, allowing us to automatically turn to the target and shoot even when the vision system doesn’t see the target.

  • Beam-break sensors and a REV color sensor are used to detect the position and color of cargo once it’s inside the robot. This allows the robot to keep count of how many cargo it’s holding at any point (and present that information to the drivers) and automatically reject cargo of the wrong color.

  • We use a custom web application to help the drivers understand the robot’s state as quickly as possible. Included in this dashboard are two cameras (one for intaking and one for aligning with the target), a visualization of the cargo held by the robot, an estimate of the robot’s position on the field based on our odometry, and boolean outputs showing when the shooter is ready to fire.


Since the competition, we’ve been working on improving certain mechanical aspects of the robot. Thanks to our initial prioritization of modularity, it is fairly simple to switch out any mechanism on the robot without affecting the rest. We are experimenting with switching to a hooded shooter with top roller to help us shoot from close to the hub without the cargo bouncing out. We are also looking at implementing a climb similar to 179 and 1657 in order to reach the traversal rung.

Throughout the competition, our intake took a hard beating and the front profile had to be replaced multiple times. When that profile would bend it would be forced into the intake wheels, seizing the system. By the finals, we decided to run without the profile altogether in order to make sure our intake wouldn’t jam. Our biggest priority for ISDE4 was to redo the intake to be more sturdy. From our initial testing, it seems that the intake works just as well without the front roller. This makes it smaller, so it’s less likely to get hit, less likely to draw G204 penalties, and extends/retracts much quicker. Ideally we’d like to have the intake retract when hit, but that might not be possible without some major changes. We’re finishing testing with a 1:1 wooden prototype before we make the real thing from aluminum and polycarbonate.

10 Likes

The students have off school most of this week for Purim, so that gives us a nice block of time to work on improvements between district competitions.

Yesterday we finished designing the new intake, and it was manufactured and installed by the end of the day today. Without the front roller, the redesigned intake is significantly lighter, faster and stronger. And from our testing, it works just as well as the original.

We experimented with adding a top roller to our shooter to decrease the backspin and hopefully keep cargo from bouncing out of the high goal, but we found that it didn’t make a significant difference. Instead, we remade our shooter hood with a steeper release angle for close shots and a stronger piston holding the adjustable hood in place. This should help us when shooting from the tarmac areas, and make our far shots more consistent. While upgrading our shooter, we also added a second Falcon motor to improve spin-up times and rewired our custom vision system, which failed a number of times during our first competition.

We are still working on our final improvement, upgrading our current climber to reach the traversal rung. We were able to retrofit our original climber prototype with the flexible hooks, and have used it to successfully climb in medium-scale (full scale in size, small scale in weight). We are still playing around with the final geometry and how to constrain the system for the start of the match. Hopefully that will be done in the next day or two so we can put the mechanism through its paces before District 4.


We were happy to be featured on FUN’s Top 25 this week as rank 23. We were a little less happy to hear what the hosts had to say about us.

“…they did not score very many from what I saw. Missing a lot in auto, and in teleop as well. And I think they only have a mid climber as well.*”

“I’m not sure that they belong here.”

“Top 40 team, yes. Top 25, I would not have them in that position.”

* Not true, we climb to the high rung

I’m sure it wasn’t personal, but I guess we’ll just have to show them that we deserve our ranking in competition next week :wink: @Tyler_Olds @Mike_Starke

13 Likes

In two days, thanks to the long weekend and much dedication from our students, we were able to design, manufacture, assemble, install, and begin testing the new climber. And somewhat to my surprise, everything seems to be going smoothly. The assembly and installation only took a few hours (thanks to our focus on modularity), and the robot climbed all the way to the traversal rung on its first attempt.

(The people on the sides were only there to catch the robot if it fell, it was completely supporting its own weight.)

After a bit of programming tuning and driver practice, we were able to go from not touching the hangar to full traversal climb in around 17 seconds, and that might get even quicker with some more practice or automation.

In addition to working on the climber, the programming team has been hard at work tuning the new shooter and vision system. They’ve also been working on speeding up our autonomous routine, and we’re now able to get four cargo.

6 Likes

Buckle in, this is a review of our time at ISR#4 and it’s a bit of an emotional rollercoaster.

We came into the event with very high expectations, being the only top 10 OPR team in the district attending the event. We started off day 1 well, winning our first four matches. Then at the end of the day our luck began to turn, losing the next two and tying the third based on a string of bad luck (getting tangled all match in an alliance partner, some partners missing their climbs, etc). We ended the day ranked 6th, but were confident teams that were scouting well would recognize our ability. Overall our new climber was working very well to get to the traversal rung, and while our shooting efficiency had probably gone down a bit since ISR#2 we were still one of the top scoring robots at the event.

At the start of day 2 we started off well again, winning our first two matches of the day. Then once again our luck turned, and the swerve drive that had been so reliable for us until now started freaking out without rhyme or reason. After a long troubleshooting session, and three bad losses where we barely moved, we finally pinned the problem down to a broken rotation encoder gear. We brought spares and the fix was fairly simple, but at that point we were ranked 10th and no longer had any more matches to show that were were working properly again before alliance selection.

Now due to both not being ranked very high and not working in our last matches, we were passed over by the top four alliances. We ended up being accepting the invitation to join the fifth alliance with 2230 General Angels, and together we picked 6168 alzahrawi as a third. Spirits at this point were not very high, believing that we should have performed better and assuming we would be defeated by the higher ranked fourth alliance in quarterfinals.

As the playoffs started, it became pretty clear that the lack of super high-scoring robots at the event meant all of the alliances were pretty evenly matched. Every matchup in quarter- and semi-finals (other than the ones we played in) ended up going to a tiebreaker, and more often than not the lower ranked alliance came out ahead. Now that our swerve was fixed, we were performing at the highest level we had all competition, scoring a ton, effectively evading defense, and consistently climbing to the traversal rung. Our alliance plowed through quarters and semis pretty convincingly and found ourselves in the finals against alliance six.

Finals 1 arrives, and seconds into teleop hard defense from our partner 6168 ends up with them climbing 5990’s bumpers and getting stuck on top of their chassis. Most of the match goes by with both teams trying and failing to get unstuck. After a long deliberation by the referees, our alliance was given a red card for the match. It’s extremely unfortunate since it wasn’t intentional, but ultimately the right call as 5990 was effectively disabled the entire match. Now it’s finals 2 and we’re down 1-0. Our alliance plays a hard but clean match, everything comes together well, and we pull out the win forcing a tiebreaker.

Finals 3 is tight the whole time, but seems to be going our way thanks to our alliance being able to get more climb points. But bad luck strikes again, and our climb that had been extremely reliable throughout the competition gets snagged on itself and has to stop at rung 3. This doesn’t leave enough room for 2230 to climb onto rung 2 without contacting us, they don’t end up getting off the ground, and we both lose our climb points. Alliance 6 ends up taking it by 13 points, meaning if we would have climbed successfully it would have flipped the outcome. It was a heartbreaking loss for our team, who finally believed we could come back from the underwhelming qualification performance.

Overall though, other than some bad luck and matches lost to a faulty swerve, we performed well throughout the event. I’m incredibly proud of the work the team has done to get to this point, and it’s certainly paying off now. With the addition of our Engineering Excellence win, we came away with 52 district points, putting us in 3rd place in ISR. We’ll have a tight turnaround time this weekend, with the DCMP starting on Sunday, where we’ll face some of the stronger teams in the district that we haven’t yet played against. That being said, I’m looking forward to the challenge and confident that we can hold our own with the best of them.

Big thank you to 2230 General Angels and 6168 alzahrawi for being amazing alliance partners, and congrats to 6168 for winning Chairmans at the event! Congrats also to the event winners 5990, 3835, and 1943 for a hard-fought finals and a deserving win. Looking forward to playing with everyone in Jerusalem!

11 Likes

What a week we’ve had!

Less than three days after packing up at ISR#4, we get right back to it and start setting up at ISCMP. The students have barely had time to catch their breath after their difficult finals loss and now have to start fresh against a significantly higher level of competition. Though we did sometimes struggle during teleop, our 4/5 ball auto was fairly reliable and we had the fastest traversal climber at the event. We managed end as rank 7, going 9-3 with two losses of less than 2 points and a ranking score of 3.00. With some in-picking, we ended up as the 5th alliance captains, picking 2231 and 4590 as partners.

Unfortunately, we lost in quarter-finals to a strong 4th seed alliance. At that point, I was sure that was the end of our season. But as a result of winning ISR#2 and being finalists at ISR#4, we came into the event with a ton of district points (3rd in the district). As the playoffs continued, we realized that, combined with our good quals performance, we still had a chance to get enough points to advance to Houston. On the bus ride home, we learned that we ended ranked 6th in the district, with 3 teams below us qualifying on awards. With 9 teams from Israel going to Houston, this means we barely squeezed in by all of 2 district points.

Along with this great news came another challenge. Like everyone else, we haven’t flown to Champs in three years, and our crate was destroyed over that time. We were informed that if we wanted the robot shipped by FedEx, we would need to have everything packed by tomorrow morning. And in addition to the normal standards for crates, those being shipped international need to be built out of special wood treated with pesticides. Once again, we had to rally the troops to get the crate built, approved, packed, and sealed in time. And sure enough, the team came through. A week after loading in to our second district event, our robot was packed in its crate and ready to be picked up for Houston.

Before the season started, the team sat together and set their goals for what would make a successful season. The two big goals they set were winning a competition and advancing to Houston. We have now completed both of those goals, and anything else we do is a cherry on top of a very successful season. Other than a few meetings to debrief from ISCMP and prepare for CMP, we’ll take a bit of a break now to give the students a chance to recover from an extremely hectic two weeks. Everyone is looking forward to attending Champs for the team’s fourth time and getting a chance to play with the best in the world.

And some great pictures I just have to share:

10 Likes

We love you ari

2 Likes

It was a great pleasure playing with and against you guys this season. best of luck in Huston. Represent us well!

5 Likes

I’m going to bring this thread back with an update of our offseason work over the past few months. We decided to redesign and replace a number of mechanisms on the robot for the Israeli Off-Season as practice for the upcoming season. The intake was reverted to the wheel spacing from ISDE2, now powered by a Falcon and without the bumper bar in front. The conveyor system was split into two independent zones to better control the ball positions inside the robot. The shooter was completely replaced with a new mechanism, now featuring an infinitely variable angle hood and a top roller to get better control of the shot trajectory. We also managed to get our hands on a limelight, so the vision system is all new as well. The only systems on the robot that haven’t been changed are the chassis and climber.

We’ve selected our drivers for next season already, and they’ll be at the controls for the offseason competition. Here’s a clip of their driver practice from today:

And of course, the programmers are getting autonomous ready:

4 Likes

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.