FRC 4481 Team Rembrandts 2022 Build Thread

WEDNESDAY - WEEK 1

We took a day or two to prepare our prototypes but we finally got around testing a bunch tonight to get a first feel for some different materials vs. CARGO on intakes and shooters.

Intake Prototype 1 - Carwash #frc95 style

Carwash style intake with non-slip drawer liner. These were mounted in a 3D printed adapter > 4 flaps.

Looks quite promising but we need to improve compression and maybe 2 flaps instead of 4.

Intake Prototype 2 - Mecanums + Nitrile

VEX 2" mecanum wheels on left and right, a 3D printed hub with nitrile tread in the middle.
Worked quite nicely but we’re probably not going to be using mecanums since the robot would be small and we’ll serialize into the shooter. We rather have a full width grippy roller than vectored intake wheels.

Intake Prototype 3 - Dual nitrile roller

PVC pipe with nitrile tread on there, dual roller.
Spacing of rollers is far from optimized but main goal was to get a feeling for materials and general concepts.

In the end we added a sheet of the non-slip liner and that worked quite nicely. A combination of a carwash intake + nitrile seems to be working quite well.

Intake Prototype 4 - Nitrile with velcro straps

Everyone figured velcro would be a magnet to these CARGO balls right?
Well… turns out it does :wink:

They do quite some harm though so teams should be careful with velcro intakes. We love the function of velcro vs. unsettled bouncing CARGO. Maybe these flaps were a little wide and smaller strips would work to + damage the ball a little less.

This concept vs. the non-slip liner seems to be our top two.

Intake Prototype 5 - 3D Printing Magic

Using every 3D printing trick in the book one of our team members came up with an interesting intake wheel concept.

It uses 0.8mm wide layers of TPU-> shore 95A (flexible).
It has Gyroid infill so it will flex in any direction. We could play with the infill density + number of shells to change the flexibility of the wheel.

Using “fuzzy skin” settings the printer will generate a random squiggly outer shell that gives it a more grip like texture. If you make the fuzzy skin very big it will start make plastic loops mid air: if we cut these (like they do with velcro) it could become kind of a velcro surface

Video down below! Haven’t tested them on a full roller yet.

WHAT ABOUT THE SHOOTER?

We build a flexible test setup out of some flat wooden sheets, polycarbonate and threaded ends.
Hooked up a NEO on a 1:1 ratio on there and tested various compressions, wheels/treads, speeds and masses.

We’ve been trying LOWER and UPPER shots.

There are 6 different configurations shown in the videos below. Variables of each configurations is displayed in the video if you’re wondering while watching.

Main take aways of our shooter prototypes:

  • The low goal is very do-able and the upper goal is very tempting. We’d still highly advice teams to first get low goal shots working consistently and then upgrade from there. Being able to contribute to your alliance by 100% consistently scoring your pre-scored ball will be very valuable!

  • 1:1 with a NEO seems like a pretty good ratio to land on. We’re debating options now on adding an extra NEO vs more flywheel mass. This was tested with a PID or any control. End of week 1 we’ll have a upgraded shooter with control and better consistency.

  • Colsons seem to shoot okay. We’ve had better consistency with a bigger drum consisting out of 5 x a 4" Andymark Stealth wheel.

Shooter Prototype 1 - Colsons 4" x 2 [LOWER]

Shooter Prototype 2 - AM Stealth Blue 4" x 2 [LOWER]

Shooter Prototype 3 - Black Fairlane Teardrop Style 4" x 2 [LOWER]

Shooter Prototype 4 - AM Stealth Blue 4" x 5 [LOWER]

Shooter Prototype 5 - AM Stealth Blue 4" x5 + Extra Mass [LOWER]

Shooter Prototype 6 - AM Stealth Blue 4" x5 + Extra Mass + Tilted Shooter [UPPER]

WHAT ABOUT THE CLIMBER?

We thought about a flop-out single stage pneumatic climb that could unfold fast. While building a prototype we succeed in doing something we had in mind.

Still we went with something we knew that will work 100% for sure. We’re going by the philosophy this year of first building simple things that work and have a completely functioning robot in 4 weeks.

We can upgrade after we have a consistent LOWER cargo bot with a MID RUNG climb. At least by the end of week 4 but hopefully faster.

The climber subteam has started their sketches, CAD and calculations for a single stage telescope with a pneumatic stroke.

If you have any questions you know where to find us :wink:

19 Likes

Would adding a strip of the “hook” portion of hook and loop onto the bumper help intake the balls a bit easier?
We would try this but both cargo balls were missing from our kit of parts black tote :frowning_face:

That 3d-printed intake wheel looks really interesting, would you care to share the CAD/ G-Code?

1 Like

That’s an interesting concept! You mean instead of on the rollers or both?

I think “doing something with your bumpers” might be an interesting concept. In practice I think it will be little more difficult to incorporate into your design. I rather look into a stiffer/rougher bumper material and leave it as is.

The bumpers in our videos are like truck vinyl and rather smooth. In one of the intake videos we added the anti-slip drawer liner on the front bumper and that helped for taking in cargo over the bumper.

If you’re trying to center a CARGO with vectored intake wheels then have a slick bumper material might be better.

Adding the strip of hook onto your actual bumper would be correct. It can also allow teams to leave the bumper be slick close to the bumper corners if teams would like to try and vector the cargo towards the center and then add the hook strip towards the center of the bumper.

WEEK 1 - THURSDAY

Today wasn’t as busy at the shop. Many worked remote from home as help desk, designing new subsystem iterations or playing the xRC Simulator.

Shooter Version 2 - Kicker wheel + Dual motor setup

As you can see below, the concept of the shooter construction is relatively similar. The ball path changed a little and they added some extra options.
Hopefully we’ll have this up and running tomorrow and we can change a couple things.

  • There will be a separate driven 2" kicker wheel

  • A 4" flywheel drum based with either 1 or 2 NEOs 1:1 ratio

  • We can add an extra 4" wheel if we like in front of our shooter

  • Controls department is working on a PID tuner to control the shooters better

  • If we like, we can add extra mass on the shooter axles to see the difference between # motors vs amount of added mass

Intake Prototype 6 - Week 1

We worked tested a similar concept as before but with a gap in the bumper. We are currently liking the fact of having a gap in there since it might simplify other things such as:

  • Indexing/singulation inside, towards the shooter
  • We’ll probably only need 1 roller, so deployment of the overall intake concept is easier
  • The overall intake can be pretty lightweight and flexible
  • The gap wouldn’t be huge, probably, since the intake will be mostly deployed out during the match anyways.

Improvements for this intake:

  • We’ll be wiring up the KIT BOT, so we can drive the chassis around and into the CARGO for better representation of testing
  • The middle rollers are 3D printed out of TPU but are still quite stiff. We’re making more flexy-ones.
  • The overall mounting of it can be more sturdy and give equal spacing in between the mecanums
  • Having no gap between the mecanums will increase intake speed from what it seems. We need to confirm this by driving in the CARGO.
  • Maybe incorporate some velcro straps in between the mecanums to see what it does vs. unsettled CARGO.

Storage Prototype 1 - Week 1

They’ve been working on a mock up subsystem to test compression and spacing on the CARGO with timing belts.
The overall compression on the hubs vs cargo is around 35mm.
Putting the hubs tighter next to each other increased consistency.
In the video below, we ended up on 70mm between the two hubs.

Drivetrain Update

We’re leaning towards 7.6:1 with 4" wheels. [Gearbox: 12t → 50t, Chains: 12t → 22t]
Free speed 14.6 ft/sec (4,45m/s), adjusted speed 13.2 (4.02m/s)
Drivetrain base: 550mm wide, 600mm long
In inches that is: 21.65" wide, 23.62" long
Wheels 50mm wide nitrile tread, custom made aluminum milled wheels.
Below our the wheels shown of our test rig, these are little smaller in OD than 4".


3DM UPDATE

Our 3DM department has been working hard on some impressive stuff. Everybody thinks that with the limited amount of game pieces and the feedback time of the field a starvation of game pieces could happen. But what are those odds and can we predict a starvation of game pieces. Should we be worried and how would it determine our own cycles?

They been modeling the game with potential outcomes, cycle times and average CARGO availability throughout the game on the field amongst other things.

They’re wrapping up their analysis and we’ll be sharing our approach and results in The Open Alliance Show hosted by @Greg_Needel and @Tyler_Olds on FIRST Updates Now.

After the show we’ll upload our findings and conclusions on CD as well.

If you have any questions, you know where to find us!

11 Likes

Hey, I’m curious as to how you landed on 4" wheels for the shooter rather than any other size. We were testing with 6" and 4", but ended up liking the extra surface speed the 6" gives. Since this year the size constraint isn’t too bad, we’re thinking of the 6". Is there anything we might not be considering other than mechanism size?

Good question! Whenever there is a game piece we are trying to launch with a flywheel our starting point for prototypes has been a 4” wheel 1:1 with a CIM and adjust compression. More recently we have switched to a NEO because of the ability to control it from a laptop and get accurate RPM feedback. If this can hit the desired results we can then choose different size wheels and gearing to achieve the same release velocity. In theory a larger diameter wheel has a longer distance to transfer energy to the ball for the same release angle which can result in more consistent energy transfer, you can see this taken to the extreme in FRC2481s 2020 robot. Hopefully this answers your question, if not please let me know!

6 Likes

Awesome, I think we’ll stick with the 6" then (ordering some more “official” testing parts this weekend). We also landed on ~1.5"/40mm compression (possibly more if we expect deflated cargo in competitions) based on similar testing, which was reassuring to see in your post lol. Thanks so much!

2 Likes

Hi! Your wheels are seems really cool and effective. Are you planning to publish print or design files?

SPIRAL INTAKE WHEEL
Link to STL: https://drive.google.com/file/d/1fFsVHWEGH7i9qONw-rHE0DelWVYZJpum/view?usp=sharing

Printer Settings for Spiral:
Material = TPU 95A
0,8mm nozzle
0,2mm layer height
4 Shells or 100% infill

FUZZY INTAKE WHEEL

Link to STL: https://drive.google.com/file/d/1EW6iyxyHyTwzikRBIcqdReDWcUh_ryiC/view?usp=sharing

When importing it to CURA it shows like a normal hub with a hex hole. You’ll get the compliance and fuzzy skin through the settings below.

Material: TPU 95A
0,8mm nozzle
0,2mm layer height.
No top and bottom layers
Gyriod infill → play with density for stiffer or more compliant (5%-15%)
3 shells → play with number of shells to stiffen or weaken the rim
Fuzzy skin thickness 1.5mm-4mm
Fuzzy skin density 1 mm^-1-0,5 mm^-1

7 Likes

thank you for your kindness and charitableness. I wish you a enjoyable season :slight_smile:

3 Likes

Oh note that on the spiral wheel we’ve added our own nitrile on there. So it’s not just 3D printed. You can add different type op material treads that you’d like to try

2 Likes

hmmm… the nitrile’s shape that you mentioned is like a tape? (to cover the wheel)

It’s very similar to this: Blue Nitrile Roughtop Tread 1 in. Wide 10 ft. Long - AndyMark, Inc

Only the one we’re using has has non of the backing on there. We sourced it locally at a industrial vendor of one of our partners.

@EricLeifermann linked a plain rough-top belting type a time ago on McMaster: McMaster-Carr

Having no backing on there makes it easier to spin/wrap around smaller diameters.

1 Like

Week 1 - FRIDAY

We’re back at it. AGAIN! We’re loving the non-stop action. As said earlier in the thread, we split up the team in two and we’re alternating days at the shop. Team members that are home for the day work mostly remote if they’re available and people in the shop can join google meets to ask questions if they need more help.

It’s been really effective so far since everyone can continue working on prototypes from the day before ór prepare decent revisions of the prototypes a day in advance.

Here’s our progression from today!

INTAKE PROTOTYPE 7 - WEEK 1

Yessir another concept in the category: 3D printing.
A #125 - 2016 inspired corckscrew, helix style, intake.

Models used in the video were rigid plastics with 30mm fin height or 10mm. It worked until a certain extent but definitely lots of room for improvement.

To do’s for this concept for tomorrow:

  • Change fin height to 20mm
  • Print in TPU and make sure that the stiffness of the screw is printed in the correct direction
  • Test single start, double start and triple start thread.

image (24)
Tuning helix pitch on a 1 start thread will also change the direction of the force vector on the cargo. multi start threads will let you get more fin contact per surface area while keeping the same >45* angle on the force vector. (Thanks for the tip @Jeremy_Germita )

UPGRADED SHOOTER PROTOTYPE - WEEK 1

This is the revision of the shooter based on the input from last tests. We’ve been changing and testing quite a lot of different things.

See pictures of the test setup below.

Testing different variables with flywheel mass

*Testing different variables without flywheel mass

Narrowing down our choices:

As you can see in the first video we tested mostly LOWER shots since that our first shooting priority.

We came to a flew conclusions:

  • With 4" wheels and 2x NEO on a 1:1 you’ll most likely NOT need a flywheel mass.

  • We’ve been shooting with 5 compliant or stealth wheels and there hasn’t been much difference. If you have the extra wheels we recommend creating a wide shooter drum since that will increase overall consistency.

  • Adding a feeder wheel increases consistently by a lot. (Duh)

  • Overall feedthrough/total shooting time of 2 CARGO is ~0.39 seconds which meets our requirement. (Shoot 2 CARGO <1 second)

We’ve changed the hood angle by 1 hole. I don’t know the exact change in distance but for us it’s not possible to shoot for the LOWER and the UPPER with the same angle. Now knowing the needed angle for UPPER and LOWER finalizes our shooter geometry.

What are we going to be testing over the weekend:

  • The shots in the video were without PID. They’re still working on that so that should increase overall consistency across and entire day of testing.

  • Backspinnnnnn. We’re going to add some nitrile, rougher, anti-slip materials on the hood to see by how much we can increase the backspin. Current thought is that if we can stay around 80% power on the shooter but lower the total height of the ball path that would be huge. *
    Lots of teams are going to be shooting high, CARGO is going to be bouncing and hitting in the air. If we can narrow down our shot to JUST shoot over the rim in the UPPER we’ll be hitting hardly any other CARGO in TELE-OP but most importantly in AUTO.

xRC Simulator

Some team members have been playing lots of games on the xRC Rapid React Simulator. They’ve written down their analysis and findings. One of them will be sharing their analysis tomorrow on here.

The 3DM department is currently comparing their findings towards the simulations they’ve ran themselves. Don’t forget to tune into The Open Alliance show tomorrow!

REV MAX PLANETARY GEARBOXES

We’re just going to put it out there.

These
Gearboxes
Are
THE
BOMB!

It’s been a couple days we have them in the shop but these are truly life changing. Swapping out gear ratios on the feeder wheel has been so quick and easy today. We’ve went through 4-5 different ratios and changed them literally in a minute.

Yeah really, that’s it. You don’t need an written instruction.

If you’re thinking of needing some more planetary gearboxes we highly recommend these!

If you have any questions you know where to find us

21 Likes

It’s game analysis time!
This week we have been playing the simulation game for quite some time with some team members and drivers. Playing this game gave us a good feeling about the overal game experience, but also strategies that might work and not work. Here is a detailed summary of our observations:

General observations:

  • Almost every ball lays against the wall.
  • Some balls will roll against the pillars of the hanger (and might be hard to pick up).
  • Balls scored in the upper hub are not worth waiting for to pick up. They bounce too much and it takes too long.
  • Balls from the terminal are not worth waiting for to pick up. They bounce too much.
  • From the driver 's perspective it is very hard to see and pick up balls on the opposite side of the field.

2 V 1 match:

  • After a while all your alliance-colored balls lay on the opponent’s side. When your shooting with 2 robots and there is a 50/50 chance of going back to your side of the field there are quickly no balls left.
  • Defense: while playing offence and gathering balls you can also play a little defence by ramming into your opponent just before they want to shoot. It is very simple and fast.

3 V 0 match:

  • It is very busy on your side. The hub and the hangar take a lot of space.
  • You struggle with getting balls since 3 robots are using them.
  • When trying to hang, 1 robot should already go hang BEFORE end-game. It is once again very busy if 3 robots are trying to hang. You easily push someone causing him to realign for hanging.

Autonomous:

  • 4 balls auto is possible, but it must be executed perfectly else you won’t make it in time. I managed to perform it 1 time.
  • More balls in auto is impossible i think, unless your robot has a turret. The simulation doesn’t have a robot with a turret so we were not able to test it.

Calculations:

I tested the reliability of the balls actually going in the hubs and not bouncing out. I did this assuming no driverskill issues. I tested it from two positions because they would have different launching angles: The fender and the tarmac line (I wanted to shoot from the launchpad but the simulation robot is not able to do it, so I had to choose a closer position).

Results:
Shooting 20x (so 40 balls) in HIGH HUB from fender:
33 scored (82.5% accuracy)

Shooting 20x (so 40 balls) in LOW HUB from fender:
39 scored (97.5% accuracy)

Shooting 20x (so 40 balls) in HIGH HUB from tarmac line:
32 scored (80% accuracy)

Shooting 20x (so 40 balls) in LOW HUB from tarmac line:
39 scored (97.5% accuracy)

During the test I saw the ball that bounced out was ALWAYS the second ball that was shot. The first ball knocks the second ball out.

I also notice that aiming slightly off centre changes the whole outcome. The first ball won’t interact with the second ball and they go much more gently inside. I quickly tested it and it seems that it does help for shooting from the tarmac, but not for the fender position. Probably because the fender position is shooting almost vertical and bounces the second ball out anyways.

Disclaimer:
Keep in mind that this is a simulation. Not everything is 100% realistic (for example: the balls bounce way too much on the ground (more than 10x before you can pick them up), gravity might be off).

24 Likes

Awesome stuff Rembrandt’s! Really appreciate you putting this all out there for us to learn from. I had a couple questions on your most recent shooter prototype. What is the amount of ball compression you are testing this with? Are your cargo inflated in a way that you believe is accurate to the 3.5 psi (like is said in other threads you need a more specialized pump to measure that low). And do you know how many degrees of wrap the cargo is engaged with the flywheel for your two shot heights?

3 Likes

The team is busy at the shop, but I can answer this for you:

  • 40mm of Compression
  • Between ~35-45 degrees of wrap with the current configuration, I don’t have the specifics of the release angle, but it will be different depending on the distance between the shooter axle and the fender and the height from the floor (this is still in discussion based on the CAD integration and the prototype testing
  • The most recent videos the balls were filled and verified with a gage to 3.5psi, the older videos the two balls were at different pressures (unknown, one soft, one hard)
  • On Sunday we will test with the balls at different pressures
7 Likes

What speed are you running your feeder wheel at relative to the main flywheel?

1 Like