Spectrum 3847 - Build Blog 2020

Personally I’m waiting for the spectacle of a robot that tries to run full speed under it, gets the two POWER CELLS underneath wedged under its bumper, and crushes the robot into the underside of the CONTROL PANEL.

7 Likes

The bottom of every bumper has to be at least 1" below the midplane of a power cell. Unless you’re talking about a team with an split bumper for intake, I don’t see this happening.

Easily >30% of teams. And you’re not even taking into account how sticky this year’s pieces are.

This is totally going to happen and I hope a GIF makes the rounds on social media when it does.

Bonus points for leaving “unnecessary” parts of the robot behind in the process.

2 Likes

Also even without a bumper gap, teams can get stuck on the balls. Last year teams got stuck on cargo, even without a bumper gap.

Trajectory Calculators and the “Sweet Spot” Parabola

For those of you that weren’t involved in FRC back in 2014, you may not be aware that Aren Hill, a friend of Spectrum, “invented parabolas.” During VEX’s Build Blitz that season Aren detailed a method of deciding on your shot trajectory by defining a wide “sweet spot”. The same principle should work this year.

During tonight’s meeting we did some analysis using two tools published on chiefdelphi to calculate the trajectory of the ball from a shooter. Thanks to @cmarley, and @sgeckler for posting these calculators.

The goal is to find a trajectory to limit the amount of variations you need in your shooting mechanism, while maximizing the distance range from goal for which you can still put the ball in the goal.

Basic idea, have the top of your arc at the top of the goal and a long flat tail on each side of the apex that still allows your ball to go into the goal.

Here are a few of the graphs we made tonight.

A wide range of distance that make it in the outer goal.

A similar shot in the other calculator

A shot that focuses more on the inner 3pt goal

If you can tune your shooter to fire every ball at a similar trajectory you may not need to adjust your velocity or angle as much as you think.

Robots That Inspired Us Today

2013 - 1114 (Dingus), 118 (Frisbee Angle Deflector)

2016 - 2056 (Simple adjustable rigid hood)

2017 - 33 (shooter/feeder), 1986 (Single Big Shooter Wheel)

2018 - 2877 (vision)

2019 - 148 (Climber Legs)

Spectrum

23 Likes

Day 6: Advancing the complete idea

Robot Design Plan Updates

A little more thought, some more white board drawings and conversations, and a few bench tests have led us to a few design changes for our prototypes.

  1. Shooter - reduce some complexity with fewer rollers and feed wheels than planed. Shooter and tower moved towards the front of the robot, to allow a much easier ball path while keep the intake behind it.

  2. Shooter Hood - Two positions, a long wide shot that will make a ball into the 2 pt goal from behind mid court and up to the initiation line and a shallower angle that is specialized for 3 pt goal from a smaller range on the field. We are currently choosing to not shoot from the target zone.

  3. Shooter Tower - Stop balls from entering the shooter by decompressing the top of the tower by moving one of the belts away from the ball.

  4. Climber - doesn’t need to be directly over our CG it’s okay if we curl a little while we solo climb as long as our climber is rigid. Climber has also moved to a standard single stage 2x1 elevator as it now longer needs to stay below our shooter path as the shooter is turned around on our robot.

  5. Buddy Climb - The buddy climb forks are now powered down instead of being sprung down. We believe this will let us make it easier to keep the forks on the floor even if our CG isn’t directly under the hooks.

Drive Train Update

CAD has been underway on our drive train since Sunday. There is still work to do but it’s coming along nice. We have had issues with a large sheet metal front rail on multiple instances so this year we are planning to combine sheet metal drive rails with an extrusion cross bar that are put into compression with threaded rod and some 3D printed inserts on the side of the tube ends…

Robots that inspired us today

2013 - 610 (drive construction)

2019 - 148 (elevator & climbing legs), 971 (climbing legs and bumper mounting)

Sketch of the Day

Spectrum

15 Likes

Oh boy are you going to be surprised.

8 Likes

As a driver in 2012 when the diameter of the balls were around 1/2" larger than these, I can tell you they will go under a surprisingly low bumper. They’re like rats, they’ll sneak in any crack you give them. This includes inside your conveyor in your robot, not just the bumpers on the outside.

Day 7: Making a few shots and a lot of misses.

Shooter Testing

We modified the shooter from our 2017 robot to be able to launch power cells. It’s mounted to our protocube (just a cube made of 2x4s but we find lots of uses for it).

In 2017 it was one 775pro geared 4:1, it’s now 2x 775pros geared 2:1. The hood has 1/8" rubber sheet to ensure the ball is rolling and doesn’t slip on the polycarbonate panel. This shooter was based on the FRC#33 2017 shooter when be built it that season.

It shoots with 4 solid Fairlane nitrile rollers (RR-754-35W-RP). It also has been setup with two smaller diameter accelerator motors to attempt to have the effects of hand feeding the ball some what mitigated.

We are running it currently without velocity control, we have a small tachometer and it shows that the wheels are spinning at 5700 RPM at full voltage. This shows that the 775pros motors are having trouble getting the wheels anywhere near their free speed.

Video of some of the initial test can be seen on our photo gallery

This shot was about 25ft.

The compression at the tightest was around 2.5" (4.5" from wheel to hood).

Part of goal is to see how we can make consistent shots with a variety of balls. We will continue to play with this shooter setup tomorrow to see how we can improve consistency.

Our next prototype will be moving up to a 6" wheel to get it closer to the correct RPM range for the free speed of NEOs at 1:1 or slightly over driven.

3D Printing Prototype Parts

All this week we have been printing parts for prototypes.

Today we started making some larger parts. Wheel extension on the left, intake ramps on the left.

We have used countless Protopipe parts as well we printed up a bunch of HYPEblocks for the next iteration of our shooter prototype.

Robots that inspired us today

2017 - 33 (shooter)
2019 - 148 (elevator chain attachment)

WCP - Greyt Elevator V2

Sketch of the Day

Figuring out some of our climber geometry

Spectrum

13 Likes

Allen,

We wonder what is the weight of the wheels that you have used in your shooter? We are currently trying to figure out what kind of “wheel” to use in our shooter in terms of toughness and weight. So, can you please point out a few key-points that we can look and implement into our shooter design?

Thanks a lot

1 Like

I wish we were that far a long in our protyoping to be able to tell you that.

Our current plan is to use 6" wheel as main shooter wheel. The larger diameter allows for lower rpms for the same surface speed and it increase the contact distance with the ball around the hood at the same surface speed (ball traves a large diameter arc).

We are currently looking at 6" VEX pro traction wheels or 6" colson wheels.

There will be teams that can dial in an amazing shooter with just about any wheel. In almost every FRC season I can think of, at least one team has shot objects really well with some version of what used to be the skyway wheel, kit wheel, and now the AM smoothgrip is a close approximation of that wheel type.

The point is tuning a system is more important than just choosing the right wheel.

8 Likes

Thank you very much Allen, we are so pleasured to get some advice from you! I think, (at least for now) we will certainly worry less about the wheel. As opposed, we were so worried about the wheel selection and choosing the “right” wheel as it will be our first shooter ever. I will definitely have my software team spare some time on fine-tuning our hopefully working shooter this year.

Thanks a lot and have a great build season !

PS: I just want you to know that you guys are inspiring so many of us here !

3 Likes

Power Cell Shooter Testing

Quick shooter testing video from today, we are now starting to believe we can make a more consistent shooter. Our plan is to have a wide 6" wheel and concentric flat hood. That will be were are next shooter starts.

Block CAD of the Robot

As subsystems start figuring out how much room they need we start laying things out in quick block sketches to make sure we won’t have systems needing the same space and see how we can leverage nearby subsystem together.

This is all very rough and almost none of this other than the drive train will actually be on our robot. It’s also missing an intake at the moment as we are still figuring out exactly what that will look like.

Spectrum

8 Likes

Have you given up on creating a short robot that fits under the tunnel?
Or is this just a prototype Power Cell shooter?

We never tried/planned to build a short robot that fits under the tunnel.

5 Likes

We are headed in a similar direction, so thanks for all of the updates. After living through 2016 low bar, we aren’t going low.

I like the path the ball is taking through your robot, have you done much testing on how to hold and transfer the ball through the robot?

Not nearly as much as we would like, we just got more balls Thursday, so know that we can feel how they interact we are working on more of the options

2 Likes

Ok thanks, we just our extras Friday, so on Saturday we set up our first real rig to test the metering. Still determining compression and we are setting up guides on the side out of HDPE, hopefully it’s slick enough.

Day 9 & 10

Everybot and Week 6 Robots

This past Sunday, Spectrum was able to attend an event hosted by the Robonauts were Everybot and Week 6 demonstrated game play together. Both robots were very well done for only needing one week to complete.

Takeaways (Some are just confirmations of what we already knew)

  • The outer goal is very large. Week 6 was able to easily manually aim their robot to score consistent shots.
  • Fast low bots may have value in getting up to the RP level of balls but you need to be very quick in getting balls dumped into the goal and quick at intaking balls. Everybot does both very well.
  • The balance even on the wooden switch didn’t seem hard to accomplish. A reliable and consistent climb will be incredibly important.
  • Driving across the Dance Floor (rendezvous points) will test how robustly made a lot of robots are. Make sure you test your robot driving over 1" bumps (the kit 2x1 is a good approximation) before coming to your first event. We are going to see a lot of radios, roboRIOs, batteries and other electronics and connectors come loose if they are only zip tied loosely to robots.


More photos and videos are in our photo gallery.

Shooter Prototype V2

Hopefully will have test shots tomorrow after we make it a little more rigid.

Intake/Indexing

Intake and indexing plans are still up in the air. We are re-investigating an intake similar to 971’s from 2012. At first we didn’t think it would package nicely but after some other changes to the design we think we can make it fit, but we need to build a better model to make sure it will index the balls fast enough for our robot.

Indexer.PNG

Robots that inspired us today

2012 - 971 (intake & ball funneling), 2056 (ball funneling)
2013 - 180 (shooting game objects inside of your robot)
2017 - 125 ( 4-bar intake)

Spectrum

14 Likes

Day 11: Design Review 2

Tuesday = design review day

Here are some of the slides from today’s review put together by each of the subsystem lead groups.

Photo of V2 Prototype using HYPEblocks, hopefully we will have some testing tomorrow. We have to refactor the hood a bit.

Spectrum

12 Likes