Team 4253 Raid Zero - 2020 Build Thread

[itsYoshi]

Welcome to the official 4253 Raid Zero build thread. Located in Taipei, Taiwan. We are the first FRC team in East Asia, and this is our 9th season of competition. This is our first official build thread.

Introduction

We realised that it is probably unusual for a build thread this late into the season. However, due to the recent outbreak of the Wuhan Coronavirus in Taiwan, the government has suspended all school and school-related activities until the 25th of February. As such, we decided we would like to publicise our ideas and design processes to those who find themselves in our current situation.

We are deeply saddened by the impact of the Coronavirus around the world, and we send our heartfelt condolences to those affected by the pandemic.

If time permits, I will write a Mandarin version of this build thread for our Taiwanese and Chinese friends.

如果時間允許，我會寫中文版本給我們台灣與中國隊伍的朋友們。

I plan on sharing the CAD of the robot after the conclusion of the 2020 official season.

Season Info
Our lab hours are as follows (Tentatively):
Monday-Friday: 3:30~7:00 PM
Saturday: 9:00 AM~7:00 PM

In addition to the allotted hours, members of the team also work on the robot during their robotics or engineering classes.
We will be attending three regionals this year: Science Park Taichung (Taichung, Taiwan), Southern Cross Regional (Sydney, NSW, Australia) and Hawaii Regional (Honolulu, Hawaii, US)

With the regionals being back to back, we have decided to pack our robot in Pelican cases through checked-in luggage. We do not have enough time to ship our robot since we are an international team.

Kickoff
During kickoff night, our team gathered in our 3-story robotics facility dedicated to STEM education at 11 PM local time.

After the conclusion of the twitch stream, we split into small groups to analyse the game manual, each group led by one of the heads of the team. Each group was responsible for their respective section of the manual (EG: robot rules, game rules, etc.).

We then regrouped in our lounge area, and each head summarised their respective section.

Following the game manual analysis, we dived back into our groups to discuss game strategies.

At 3 AM, we decided to head back home for the night and rest for the following day.

This concluded our kickoff night.

Week 1
In the morning after kickoff, we decided to dedicate the start of week 1 to game strategies and overall season workflow.

Our overall goal was to develop a robust and consistent robot which could see hours of practice well in advance of our first regional. Thus, we have decided to stick with our regular 6-week build season, as seen in traditional bag&tag seasons.

Our design mindset is as follows: the ability to obtain all 4 Ranking Points (RPs) without the aid of others. This design philosophy allows us to potentially get all the RPs by winning the match, 3-stages RP and end game RP. In the worst-case scenario going against three god-tier teams, we would still like to obtain as many RP if the odds of winning were against us. Ultimately, we will design our robot based on our analysis instead of forming a game strategy based on our robot design.

Here were our crucial findings:

• The odds of completing the 3-stage RP were less favourable than endgame RP.
• To obtain the end game RP, it could be achieved by either three robots hanged or two robots plus a balanced switch.

We also prioritised the robot mechanisms into three categories: Need, Want, Wish. The “Need” category is what we absolutely must need to be competitive. “Want” category is what would greatly benefit us, but is not a priority. “Wish” is what would be a reach in designing, but we dream of having that on our robot. Below is the list we made

Need:

• Drive Train

• Shooter

• Intake Mechanism

• Climb Mechanism

• Wheel of Fortune (WOF) manipulator

Want:

• Turret

• Adjustable Shooter

• Switch-centring Climb

• Trench Run capabilities

Wish:

• Buddy Climb

Drive Train
After planning the priority level of subsystems, we moved on to the topic of the drive train, the most crucial subsystem of all. Before we begin brainstorming, our drive train has to meet the following conditions:

• Fit inside our largest suitcase of 65cm x 80cm x 30cm (25.5" x 31" x 12")
• Weigh under 15kg (33lbs) to obey airline regulations
• Easily detachable to allow ease in packaging other subsystems into multiple suitcases

Our first conclusion was to have a West Coast style cantilevering drive train. This style of drive train has proven to be effective ever since our first adaption of it back in 2017. With 2020 being a highly interactive game, we expect to encounter rough defence and more repairs than previous games.

The drive train will be constructed with 2" by 1" (2mm walls) box tubes, powered by 4 Falcons. The gearbox we decided to use was WCP DS Flipped gearbox.

There were multiple debates in regards to the ratios and wheel sizes. Ideally, we want a low gear for accelerating and point turning paired with a high gear to quickly traverse across the field in seconds. We also dived into the decision of using 4", 6", or 8" wheel. However, the verdict was quickly settled upon 6" wheel. The reasons being 4" wheels were not ideal to cross over the generator bumps. This was found through a quick CAD sketch along with proof of concept testing with our off-season bases. 8" wheel was discarded as the base will not physically fit inside our suitcase.

The wheel configuration was then debated upon. From 4WD to 6WD drop centred, corner omnis, etc. Our first concern was whether omni wheels could take the impact from going over the bumps. After abusement tests of our off-season base, we concluded omni wheels were prone to destruction over time. A 6WD was not ideal without omnis, so the configuration was down to 4WD or 6WD drop centred. We ultimately went with 4WD for the weight savings of the base and size constraints of the suitcase.

We then tested multiple types of wheels: Nitrile tread, Pneumatics, Colsons, etc. A 4WD with nitrile drew too much current, and to maintain the consistency of the robot’s performance throughout the match, this was out of the box. We ended up choosing WCP Pneumatic wheels. Those did not draw as much current during point turns and offered an extra level of protection going over the bump.

Attached is the image of the drive train CAD.

The drive train was fabricated in-house with our waterjet and welding station.

(More information will be written tomorrow, it is currently 2 AM in Taiwan)

[Andy5553]

That’s awesome that you decided to do a Build thread (we did our first one this year as well)! I hope you are all safe and that you will have a successful season despite the Coronavirus.
As you mentioned with the possibility of having a lot of defense this year having a 4WD decrease the maniability of the robot thus harder to avoid defending robots but I understand you choice due to the suitcase issue.
I see on the CAD upload that you made a lot of small wholes in the chassis is this a choice? If the’re there for reducing the weight of the robot why not chose a geometric shape that you can pattern all the way around?
Finally, which regional are you going to?
I hope we can meet one day!

[usymmij]

Those holes are not for weight relief, but for modularity. We do this so that as we have new ideas we can easily find ways to mount them without drilling new holes

[LouisDotA]

@itsYoshi

Science Park Taichung (Taichung, Taiwan), Southern Cross Regional (Sydney, NSW, Australia) and Hawaii Regional (Honolulu, Hawaii, US)

These 3 regionals

[itsYoshi]

The holes on the chassis are just for mounting. This allows our team to prototype on to the current chassis. In the end, only holes that will be used will stay. The rest will probably be solid or have weight reduction patterns.

To our surprise, the 4WD with pneumatic wheels turn pretty smoothly. We are able to point turn without browning out with current ratios. I will be posting those info later today

[waialua359]

Our State DOH just cleared us to continue as planned for the Taichung event today. We have a few members not going, but most of us still are. I was a little worried with 18k worth of airfares we were told were unrefundable.

Robot pickup is Feb. 20, so we’re scrambling, its crunchtime.
3 Hawaii teams signed up and are intending to still go.

Look forward to seeing all of the Taiwan teams, especially 4253. And, visiting the Tech Cube again with our team.

[Rick_McThunder]

I am very glad to see you all helping other teams with this thread, and hopefully your local teams will use it. The virus came at a very bad time for First teams in China, but i hope to see you all prevail and put out some killer robots this year. I am wondering how you plan to bring the robot in Pelican cases. How big are the cases, and how much did they cost?

[itsYoshi]

Week 1 Continued

Drive Train Continued
Our final dimensions for the drive train is 25" by 27" in a “fat” configuration.

Our initial ratios were 7.45:1 and 9.5:1, 18.29fps and 14.34fps respectively. Once the base was fabricated, we tossed steel plates to load our drive train to 154lbs, and gave it a stress test. We found out that the motors were not as efficient as we predicted. This was our first test with the new Falcon 500 motors.

After this discovery, part of our design & mechanical team quickly tested other ratios. We kept the spread the same but adjusted the second stage reductions (Thanks to the brilliant design of the WCP DS Flipped gearbox). Changing the reductions only required 4 screws to be taken off. One small mistake was made in the process. We accidently flipped the outer support plate of the gearbox which looked symmetrical, however it wasn’t. Our 8T pinion destroyed the first spur gear. We learnt this lesson and our team now labelled the correct orientiation on the outer plates.

In the end, we reverted back to the 12T pinion once the remaining Falcon order came in. Our final ratios are as follows:

We saw a sudden increase in efficency of the motors after switching to the 12T pinions, along with increasing our ratios.

Props
Part of our mechanical team went to prop building. We prioritised the prop list to not bottleneck our prototyping capabilities. The shooting goal zones were constructed first, then the switch, and lastly the wheel of fortune and the human loading station.

Prototype concepts
The shooter design was of our next priority. Given most of the points can be achieved from scoring, we placed a strong emphasis in a robust and consistent shooter. Another part being in the history of our team, our weakest points are shooting games. However, throughout the prototyping phases, we began to realise shooting was not as bad as we thought so.

In the first week, we pumped out multiple iterations of shooters, from double wheel, single wheel, catapult. We also tested different compressions, wheel types, angles, and motors. Below are some photos of our prototypes

[itsYoshi]

Looking forward to seeing you all at the Taiwan regional and Hawaii regional again!

[ProPain37]

Are you sure it was the pinions fault? I don’t see how a pinion would make the motors less efficient. Also, what do you mean by less efficient? From my understanding the gear ratio is what dictates efficiency.

[usymmij]

yea the situation is pretty scary in Taiwan right now, but not as bad as they have it in China

We plan on splitting up the bot into multiple components, ex. shooter, base, intake, and put them in seperate cases. We did this in 2018 and could rebuild and rewire the entire bot in 45-75 mins
The cases vary, although the biggest one is around 25" x 31" x 14"
They are pretty expensive, not sure on the exact price and model we have but they can range from 200-400$ but they will last. Our 2018 cases were cheap, but they could barely last through the season and we ended shipping the bot from hawaii to detroit in a wooden crate (as a full system, not split up)

[itsYoshi]

Yea. Part of the fault came from flipping the outer support plate which messed the shaft alignment up between the 2 bearings. We switched to 12T pinions because we were more comfortable with standard pinions as we have 0 experience with using involute gears.

Sorry for the confusion, what I meant was the motors were not being utilised correctly. We were not able to accelerate to top speed smoothly due to the relatively high initial ratios.

[ProbablyWont]

The sharp teeth of the steel 8T pinion absolutely shredded the only two 62T aluminum gears we had due to the misalignment of the plates, the belly pan was full of shavings and the 62T gear teeth were razor sharp. The 12T pinions have a much wider top land and probably wouldn’t have wrecked the gears as much before we noticed a problem.

[itsYoshi]

Week 2
As a recap, week 1 was mostly about game strategy and the mechanisms that we need. At the start of week 2, we then prioritised the order in which we would like to develop. Ultimately, by the end of week 2, we will have a robot that can play the game, but not perform well. This will be good feedback for our next iteration.

The list is as follows

Tier 1 Importance

• Shooter
• Climb

Tier 2 Importance

• Intake and Funnelling
• Wheel of fortune

We placed our focus on the shooter and climb because that is where the bulk of the points came from. Without a working shooter, it is pointless to have an intake and funnel system if the robot was not able to shoot. For the wheel of fortune, it was less of a priority. The wheel of fortune is only activated when we score a certain amount of balls, however, this requires a well-built shooter. The climb is independent of the shooter, but it was just as important as the shooter pointwise.

Shooter Progress
Below is one of our most successful prototypes.

Our initial plan was to test out with 775s at 3:1. The results were acceptable, but we also wanted to test out different motors. We attached a directly driven neo and falcon. The falcon appears to be the best of all 3 in terms of performance. We then decided to go with a falcon driven shooter through our experimental data.

Climb Progress
Some of our members also worked on climbing. They began by proving different concepts. The first climb idea was using a winch and a detachable hook, similar to our 2018 climb. Below is an example of our 2018 climb.

Final 2 - 2018 Hawaii Regional

Our first prototype was a 64:1 winch driven by 2 775s.

We slapped the winch onto our 2017 base and loaded steel plates up to 150lbs. The winch was able to smoothly lift the robot up without stalling or producing weird smells.

Base Updates
The first iteration of the base was finally produced. There were some manufacturing errors, such as the waterjet taper enlarging the holes, however, those will be accounted for in the final version of the base.

To allow members of the team to prototype with ease, we made 2 wooden plates to move all the electronics to the side, which will allow all the free space within the robot to be prototyped on.

Bumpers
To allow fast and easy bumper mounting solutions, we went with sticking a pin through the 1 by 1 box channels.

Intake
We also started out working on the intake. We made this adjustable intake to determine the approximate placement of our rollers.

We wanted to do be able to intake balls but also quickly spit it into the low goal. We slapped on 2 drills and spun them to test it out.

Funelling
For our funnel system, we tested out different methods and approach to organise the balls and prevent jams.
Below is a proof of concept of transporting the balls to the shooter

Conclusion
Week 2 was mainly about proving concepts before prototyping and prioritising our approaches.

[itsYoshi]

Apologies for the long gap since the last post. Our school switched to a virtual learning platform during this period of quarantine.

Week 3
This was our last official build week before our Chinese New Year holiday and the following Wuhan coronavirus outbreak. Our original schedule was set to resume right after the conclusion of Chinese New Year. This unforunately did not happen.

Our main focus for this week was to solve the obvious problems we have in our first bot, and fine tune it to a point where it is competition ready. This allows us to figure out the changes we need to make for our next iteration of the robot. Once we are done testing the robot we handed the robot to the programmers to tune pathing and limelight features.

Shooter
Our next goal for the shooter was to change its launch angle depending on where we are shooting from. We set up a quick manual adjustable hood to determine the angles that favoured us. Below is how we set it up. The adjustable hood was able to be adjusted at small increments by locking it with a screw.

Adjustable Hood

Through our testing, the two angles that favoured us the most was 35 degrees and 65 degrees, for trench shots and close up shots respectively.
Here are some video clips of us shooting from the trench

Kobe Bryant x5

Intake
Two intakes were made

The first intake was then made with 1/8 bent polycarbonate and powered by a 775 with a 5:1 reduction. The piston was mounted as it is to act as a 1st class lever. When the intake retracts, it is able to spit the balls into the low goal.

The second intake was made with box channels. The only difference was the piston acutating was a third class lever.

In the end, we went with the polycarbonate version. The polycarb absorbs contacts better than box channel and was much lighter. Although the box channel version provided more rigidity, we went with polycarbonate ultimately.
Funnelling
From our testing earlier, we discovered polycord was the best method to transport the balls into a funnelling system into the shooter. It is powered with a falcon with a 3:1 reduction. The pulleys were custom printed with our onyx printer. In addition to the conveyor, we added a passive funnelling system which proved to work well. We attached latex to the side cage which allowed the balls to slide right into the conveyor without any jams.

WOF
With the amount of real estate left, we did not want to introduce a complex mechanism for the WOF. We slapped a 180-degree versaplanetary from Armabot and some compliant wheel. The WOF is actuated by a 90 degree rotary piston. In addition, we attached a lifecam as our software team were working on machine learning for the wheel.

Below is the WOF spinner in action

you spin me right round

Climb
We tested a few concepts for buddy climbing, such as bumper biting. We used the kitbase along with our 2019 Deep Space robot.

V1 Robot
Below is the first iteration of our robot. Sorry for the blurry quality. We did not plan for a robot build thread beforehand so we did not take showcase photos.

Merch
Designs were submitted and finalised for our hoodie and uniforms this year. Below is what our apparel will look like.

The colour will be Navy Blue although in the samples it is black/white

Hoodie

Uniform

Photo Dump

[itsYoshi]

Weeks 4,5, and 6
During the following weeks, we were asked to stay at home and work remotely. We analysed the current robot we have and started CAD models for re-iterating. Our mentors were able to have access to school were able to fabricate the parts for us to build as we return to school.

Intake
We began to realise the cantilevered 3/8 shaft that the intake pivoted on was torquing downwards. Another feature we wanted was to move the belt and the chain inside the intake as a precaution against defence robots. We printed a standoff layer for the 775 with our Onyx printer. We also added an additional box channel bearing to support the other end of the pivot shaft.

Conveyor
The conveyor did its job so no major changes were introduced. One change we added was to place the battery under the conveyor. We added a pivot to the very end of the conveyor so we can flip it up and down to change the battery.

Climb
Perhaps this is the only part that I did not go fully in detailed with. Our climb folds horizontally within the cage and is deployed by a servo and a spring. The video should explain for itself.
After the claw latches on to the centre bar or the tube, the robot tilts sideways with a winch powered by a falcon with a ratio of 64:1 on a cim sport gearbox. We metal printed 8mm keyd shafts for falcons.

Climb Deployment

WOF
Our previous WOF was actuated with a 90 degree circular pneumatic. However we have concerns the mechanism going under repeated impacts when driving on to the table. We replaced it with a 180 degree circular pneumatic with it geared 14:26. This allows the WOF mechanism to have some backlash to absorb the impact. The 14T pinion was able to interface nicely with the D-Bore shaft on the pneumatic device.

Overall Robot
The current weight of the robot is 97lbs.

[itsYoshi]

Updates
Hey all, It has been a week since the last post, and a lot of progress has been made to our robot, along with news on the island

First of all, our first regional, Science Park Taichung has been postponed to a later date due to the ongoing coronavirus, presumably after the World Championships. This decision was made by the Taiwan Government’s Centre of Diseases control.

We feel deeply saddened to our international friends from 359 (Hawaii), 3008 (Hawaii), 4817 (Singapore), who had to adjust to the cancellation of the event in short notice.

Besides, we plan on supporting our local teams in the upcoming months to prepare for the next season or the eventual Taichung Regional. The Taichung regional was the only regional for many of the local Taiwanese teams.

Intake updates
In our previous intake iteration, we discovered the location of the motor was not optimal. It caused clogs and jams as the balls attempt to compress under the motor. We moved the motor to the outside of the ball’s path and added carbon fibre reinforcement. The motor was moved just enough to stop the balls from jamming. Additionally, we added poly belts to assist with moving the balls in our hopper and compliant wheels to give the balls a little bit more speed when intaking.

Conveyor
We added a home-made poly belt with poly cord flaps to assist with funnelling. Here is what the belt looks like.

This new homemade belt improved the consistency of feeding the balls into the shooter.

Autonomous
Our programming team was finally able to work on autonomous. We started first by getting the control system set up for the drive team, then move on to tuning the drive train, shooter, and other subsystems. The turret was programmed to auto-aim in tandem with the adjustable hood by using the limelight.

After getting every subsystem tuned and set, the programming team then began running autonomous paths for our ideal starting positions. Here is an example of a path we were able to master consistently.

autonomous dank

Reveal
We have a robot reveal planned coming soon! Stay tuned for it

[itsYoshi]

Updates Part 2
A lot of FRC events have been cancelled or postponed due to the Wuhan Coronavirus. Our second regional, Southern Cross Regional has been postponed as well. Although it is a big disappointment for the team, especially for the graduating seniors, the health and safety of everyone take priority over anything else.

Lab Invitations
Last week, we invited 6191, RoboKryptonite to our lab for them to tune their robot with our field. In addition, we were able to assist them when troubleshooting problems. We plan on sending out more invitations to local Taiwanese teams to invite them to our lab.

Climb
For some reason, we did not have a video of the climb in progress until yesterday.

Monkey bar climb

Shooter
We ordered 6 inch fairlane wheels and replaced the Andymark Stealth wheels with it. We found improvements as the exit velocity of each ball was more consistent with each other. The increased mass of the wheel also acted as a fly wheel.

After watching some reveals with a top kicker wheel, we quickly made a proof-of-concept and attached it to the current shooter to see the results

Reveal
We are in the process of filming a reveal video

CAD Release
The CAD of the current robot will be released sometime next week. The CAD will be based on the robot that was intended to be brought to Southern Cross Regional.

[jack42405]

What kind of spring did you use for the climber deployment? We were working on a similar spring deployment method for our second competition before everything went crazy

[itsYoshi]

We used a double hook spring that will naturally want to actuate the hanging mechanism.

The spring is right under the hanging arm