Spartatroniks 3512 - Build Blog 2022

[Nick.kremer]

Hi Chief Delphi!

Team 3512 is officially joining the openalliance for the 2022 FRC Season!

Pre-Season Activities

Our team was pretty lucky in that with safety precautions in place we have been able to meet in person regularly since Summer of 2021. We had a really busy preseason!

Tidal Tumble

In October we attended the all the new SoCal offseason Tidal Tumble; shout outs JJ and the 4414 crew for a suburb inaugural event, although we we did lose our head during a match o.O

FLL

We hosting our regular FLL qualifier half in person and half virtual.

Swerve

We worked on a swerve chassis with MK4 SDS Modules.

Swerve was a big topic of discussion for us over the fall. While its never been easier to run swerve thanks to the amazing quality of COTS modules and publicly available Swerve Code, we didn’t finish making our swerve chassis until December. This did not leave us with enough time to confidently make the decision to run swerve for 2022. We subscribe to the mentality that it’s never a good idea to run a new-to-us drivetrain for the first time during build season.

2022 Drivetrain

With decision to not run swerve in 2022, we decided to fall back on WCD which we are very familiar with. We designed a custom version of WCP’s excellent SS Flipped Gearboxes to circumnavigate any potential backorder issues:

Prototyping

The 2020 season was a transformative season for us from a prototyping perspective. After having a less than stellar 2019 season due to not prototyping thoroughly, we ran head first into prototyping and documenting our efforts for 2020 and it paid off in spades. You can see our complied 2020 - 2021 documentation here.

For 2022 we wanted to continue this trend but also improve. One issue we had in 2020 was getting our conveyor prototype off the ground took the better part of 2 weeks using wood construction.

To address this we decided to leverage our team’s 3D printing resources and adopt a 3D printed prototyping system; namely TTB Blocks and HypeBlocks..

We made a 3D printing signup sheet with imbedded 3MF files which allows mentors and students to print at home, thus multiplying our 3D printing throughput 400%.

================================================================================

Well, that’s probably enough for this post, next post (coming later today) will be about Kickoff 2022 and our Week 1 Progress!

[Nick.kremer]

Kickoff and Week 1 Update

Kickoff

We were pleasantly surprised by how similar the 2022 game ended up being to the 2020 in terms of the game scoring element (balls) and the end game (climb). we believe this game will play nicely to our strengths after everything we learned during 2020 and 2021.

Our kickoff is structured after the one described in this video from Team 971 with some modifications and simplifications. In general we like the flow of:

• Breaking into groups to identify and present key parts of the manual
  • Robot restrictions (height, weight, extension limits, ect)
  • Scoring breakdown including Ranking Points
  • Field Zones, Field Elements, and Game Piece locations
  • Fouls
• Identify Robot Actions (anything a robot can do in a match)
• Combine Robot Actions into Cycles and map out possible strategies
• Look at the viability of different strategies, aided by this DOPE spreadsheet made by @cadandcookies
• Decide on team priorities
• Break into groups and brainstorm ideas for subsystems to address game tasks

We ended up with this list of priorities (again stolen from Nick A.):

Kickoff Retrospective

After any team event such as Kickoff, Competitions, Fundraisers, ect. we like to do a reflective identifying what went well, what we learned, and what we can improve on.

Newer kids on the team and/or kids on the business oriented half of the team identified during our retrospective that they felt as if they didn’t have enough knowledge of common FRC mechanisms to feel comfortable presenting ideas. We identified a few possible ideas to address this:

• Utilize resources such as Spectrum’s Training Curriculum
• Have all team members participate in 1 - 2 week prototyping crash course over Summer/Fall so everyone gets practical, hands-on experience

Prototyping Approach

Prototyping began in earnest on Tuesday. Most team members were split into one of 4 prototyping groups. To help keep everyone on the same page and provide an outlet to record the “Magic Numbers” such as ball compression, rpm, wheel sizes, ect. for any given subsystem, we created a Master Prototyping Spreadsheet. The spreadsheet features a tab for each prototyping group to record the results of their tests.

We encourage our members to record absolutely everything they test and upload it to our team Slack. To help keep things manageable we’ve created a temporary channel for each prototype group on Slack to be archived after the season is over.

Intake

The spare 3 roller intake from 2020 has served as a great test rig for intaking the 2022 balls. So far our findings have been:

• TTB compliant wheels and rubber sleeves stretched over versarollers both have great grip on these balls, with complaint wheels being better for initial contact (lowest roller) as they help deaden bouncing balls better
  • Here’s a video from 1678 explaining how to float the rubber over the tube
• ViW as the last roller work good for vectoring if needed, although initial tests show passive guide fins may also work for indexing
• Bottom kicker bar like 2910 2020/2021 helps with getting the ball off the ground
• 1.5 Compression is just about right for intake

Here’s some videos of the different configurations we tried:

versaroller + rubber bottom roller w/ kicker ramp

versaroller + rubber bottom roller w/ kicker bar

compliant bottom roller w/ kicker ramp

Next up affixing a kicker bar solidly to the intake plates to see if we can solve the deadzone issue we are seeing when the balls get captured between the first roller, bumper and ground. We did get a fun video of the ball “pushing” a robot though!

Conveyor

The conveyor team has made an incredible amount of progress aided by our investment in using 3D Printed Prototyping blocks to quickly mock up an adjustable test rigs

Our conveyor setup and some videos:

• 2" ball compression
• 1" between belt centers
• Gates 655-5M-15 Belts (131 teeth lol @Boomie)
• 18T HTD Pulleys
• 1"x1" frame + 3D printed TTB Block, and HypeBlock construction

2 Sided Conveyor, 2 Ball Feed Test

2 Sided Conveyor, 2 Ball Start Up Test

2 Sided Conveyor, 1 Ball Resistive Test

1 Sided Conveyor, 2 Ball Feed Test

We are really happy with the performance of timing belts for conveying balls, and we think we can get away with a one sided conveyor. Next up to test is replacing the plywood on our 1 sided conveyor with polycarb to dial in the compression (the though being polycarb may require more since it will have less friction). We also want to test if we can index the ball with passive guide fins inside the conveyor.

Climber

We pulled the trigger on buying several TTB telescoping kits the day after kickoff after seeing the stock dwindle overnight, and after team 1986’s 2013 Reveal Video was posted on the our Slack.

The plan right now is to replicate the climbing sequence shown in @s_forbes Lego video. We like this climb for a few reasons:

• Minimal degrees of freedom
• No need to articulate the entire robot frame with an arm\pneumatic
• Easy to revert back to L2 climb in the event where L3 and L4 climbs don’t pan out

While we wait for the TTB kits to arrive, we made progress by prototyping a sprung hook and mapping out the climbing sequence (thanks 2357 for the idea) in SolidWorks using a series of layered sketches:

Protip: you can create simple stop motion animations using Google Photos!

Shooter

We are making an adjustable shooter prototype modeled after the Greyt Shooter with CNC’d side plates which allows us to move standoffs to adjust release angle and compression:

We used this Desmos Calc made by @dydx to help plan our trajectory. Plan at the moment is to shoot both high and low when pressed up against the hub using an adjustable hooded shooter.

More to come this week!

2022 Drivetrain

Production of the 2022 drivetrain is in full swing!

• Gearbox plates are machined

• Exploded Assembly Views are done

• Drivetrain encoder mounts and hex spacers are being printed

Overall Robot Layout

We threw together a few possible robot layouts with simplified geometry to get a sense of how to package the various different subsystems together:

================================================================================

Thats about everything for Week 1! We’ll be back at it again on Monday.

[wolly_efendi]

Very excited to see this climber in action! Any thoughts on how you’ll power the telescoping and pivoting?

[Nick.kremer]

Yep!

We’ll use the new REV MaxPlanetarys to power the winches and pnuematic cylinders for the pivots.

[jakeamusprime]

What ratio have y’all decided to use on the drive gearboxes?

[Nick.kremer]

We designed for ~16ft/sec w/ 11T pinions with the KLib Design Calculator, but swapping out the pinions gives us a nice range:

• 10T pinion 14.5ft/s
• 11T pinion 15.96ft/s
• 12T pinion 17.41ft/s

[AdamHeard]

How did you arrive at these ratios? What did you run the prior season for reference?

[Nick.kremer]

@AdamHeard

What did you run the prior season for reference?

We ran 14.5ft/sec in 2020.

@AdamHeard

How did you arrive at these ratios?

We almost always run around 14-15ft/sec as a baseline, as we feel this is suitable for pretty much any modern FRC game. Because we designed this gearbox before the season, we wanted the ability to go faster than that with the option to use the 10T pinions to stay under 15ft/sec depending on the game.

Designing the gearboxes before we know the game inherently means our gearing may not be fully optimal, but the tradeoff is less work during the season and we can get the drivetrain done faster, and in the hands of our drivers.

I’m sure some people may disagree with this approach.

[Nick.kremer]

Design → Machining → Assembly

I wanted to detail how 3512 produces the parts for our robots, and the process and tools that make that job more manageable.

Part Numbers

Like many teams, we use unique part numbers to identify every part our team designs. I’ve written a detailed guide on part naming schemes on our CAD Training site, but here are a few examples:

21-000-000

• 21 – The first two digits represent the last two digits of the year in which the assembly was created.
• 000 – The next three digits are the higher level assembly identifier. When this is 000 it specifies the highest level assembly (which is the entire robot assembly).
• 000 – All assemblies end in “000” to denote it as an assembly.

In this example, the part number denotes an assembly created in 2021 and is the highest level assembly containing the entire robot.

21-DRT-102

• 21 – The first two digits represent the last two digits of the year in which the assembly was created.
• DRT – The next three digits are the higher level assembly identifier. DRT in this case is an abbreviation of drivetrain.
• 1 – **The next digit is the lower level assembly identifier. This denotes a sub assembly within the drivetrain such as a gearbox.
• 02 – The final two digits indicate the Part Identifier of the part. This is the second part in the DRT-100 sub assembly.

In this example, the part number denotes a part created in 2021, the part belongs to a lower level sub assembly within the high level drivetrain assembly. An example would be a drivetrain gearbox plate.

It doesn’t really matter what naming scheme you use, but be consistent and use one!

Part Tracking Spreadsheet

We use a part tracking spreadsheet called the PBS (Product Breakdown Structure) to track the custody of a part from design all the way through manufacturing and assembly. Our spreadsheet is modeled off the one described in this Wintergatan video.

Our PBS tracks custody of any given part by capturing 2 pieces of information (among other things) to establish which subteam “owns” the part:

• Build Type: Mill, Lathe, CNC, Subassembly, ect.

• Status: Designed, Printed, Assmebled, ect.

Different combinations of Build Type and Status are entered on the DATA tab allowing for parts to have different workflows throughout the design process. As an example, here’s some workflows for different types of parts/assemblies:

• Mill Part
  

• Plate CNC’d part
  

• Subassembly
  

Intuitively, It makes sense that a CNC part would have a different workflow than a Lathed part; the former needs CAM, while the later needs a printed drawing.

Here are some other useful features of the PBS:

• Build que tabs to see a snap shot of what parts any given subteam needs to work on itemized by machine
• Material
  • List of all common materials we use, entered on the DATA tab
• Assigned Team Member
  • List of all students on the team, entered on the DATA tab
• Automatic Cell colors to represent higher and lower level subassemblies
  • Achieved via conditional formatting
• Flagging parts with issues causes the entire row to highlight red
  • Achieved via conditional formatting
• Deprecated parts automatically crossed out
  • Achieved via conditional formatting

Kitting Assemblies

While we wait for parts to produced for any given subassembly, we kit out that assembly by gathering all the necessary COTS and Custom parts into a staging area. Some tips to make this easier:

• Labeled Plastic totes to store all parts for a given subassembly

• Where possible, keep COTS parts that come in individual labeled bags (like gears and sprockets) in their original packaging

• Set aside a place in your build room to place COTS orders as they come in, and add parts to their relevant assembly totes as soon as possible

• Part Bags to keep multiples of the same part together

• Masking tape + sharpie, again to keep multiples of the same part together
  

  

Exploded View Assembly Drawings + BOM

We mandate that the CAD team produce Exploded Assembly View Drawings with Balloons and BOMs to assist our mechanical team during the assembly process.

All exploded view drawings are printed on 11x17 paper, and the BOM may or may not be on the same sheet or a different sheet depending on the size of the physical assembly.

This is absolutely my favorite picture from the 2020 season of a team member assembling our drivetrain gearboxes:

Maintaining a CAD Binder

In addition to assembly drawings, we also create drawings for Mill and Lathe parts. All of these drawings are added to a 3 ring binder to keep everything organized. Here are some tips we’ve picked up along the way:

• Every drawing gets a sleeve, when mechanical is referencing a drawing keep it in the sleeve; nobody likes dirty drawings

• Use dividers to separate different subassemblies

• Inevitably drawings may need to change, or parts might change. We ask the mechanical team to use a red pen to mark up any changes on the drawing itself, and then leave a comment for us for that part on our PBS

• Bring your CAD binder to comps to wow the Judges!

================================================================================

In the next post I want to detail some of the CAD tips we use to work smarter not harder when it comes to designing an FRC robot.

[Nick.kremer]

Conveyor Prototype Update 1/18/2022

We are now confident that 2022 cargo balls can be conveyed reliably with a timing belt on 1 side and solid surface on the other, and have made the decision to move forward with this design on our competition robot.

Why Single Sided Conveyor?

Single sided conveyor has a few distinct advantages:

• 2022 Cargo are much more forgiving when it comes to task of intaking, indexing, and conveying compared to 2020 Power Cells
• Less parts and less complexity
• Cheaper and easier to make

Conveyor Prototype Overview:

• Timing belts on 1 side
• Polycarb on the other
• Single 2" TTB Squish Compliant Wheels axel under polycarb to assit with feeding

• Bottom angled cardboard ramp
• Gates 655-5M-15 Belts (131 teeth lol @Boomie)
• 18T HTD Pulleys
• 1"x1" frame + 3D printed TTB Block, and HypeBlock construction

The additions of the single complaint wheel roller under the polycarb and bottom cardboard ramp means we can reliably feed a ball into the conveyor from the side.

The Magic Numbers

Testing Videos

Over the course of Monday and Tuesday this week, our conveyor prototype group has been hard at work testing, iterating, and recording different conveyor configurations. Here is a recap of our 1 sided conveyor testing:

• Initial testing

• Dead zone noticed @ slow roller speed and slow feed rate

• Added bottom ramp to address dead zone

• 2 ball feed test

• Adjusted ramp position closer to compliant wheels

• Final testing

Conveyor Next Steps

We have a few action items left:

• Add 2" compliant wheels on the other bottom shaft between the pulleys to further eliminate the dead zone we see at the bottom with low RPM or slow feed rate

• Test running the conveyor in reverse for outaking balls

• Adjust overall conveyor width to directly interface with our soon to be assembled hooded shooter prototype

• Begin conveyor mastersketch in CAD

Prototyping Google Drive

You can access all of our prototyping videos here, organized by subsystem and date.

================================================================================

That’s it for this post, but don’t forget to pet your cargo!

[Nick.kremer]

Week 2 Update

Intake

We were pretty happy with our intake prototype coming out of last weekend, so this week we focused on intake Design and CAD.

Intake Prototype Overview

• Spare 2020 Intake as a test rig
• 3 roller over the bumper
  • Bottom roller: 2" TTB Squish Wheels
  • Mid and Top rollers: 1.25" O.D. Versaroller w/ Rubber Sleeve

Intake Magic Numbers

Intake Design Considerations

We had a few issues with our 2020 intake that we are hoping to address with the 2022 intake:

• Bearing flanges oriented on outside of intake plate, flange retained with #10-32 button heads
• Using oil embedded bushings w/ shoulder bolts for pivots
• Safeguarding the pneumatic cylinder by orienting it to be retracted when the intake is extended, and extended when the intake is retracted.

• Tapped Thunderhex standoffs for rigidity
• Uneven 4 Bar linkage to allow intake fold into smaller horizontal space

Intake CAD Progress

The intake mastersketch and subassembly are coming along nicely:

Conveyor

I made a blog post earlier this week that contains most of our progress, but I do have a few updates to share.

Conveyor Testing

We reconfigured our conveyor width to interface directly with the shooter. Below is a test video verifying the setup still works in a narrower configuration, as well as a test to confirm we can outtake balls:

• Narrow 1-Sided Conveyor Feed Test
• Narrow 1-Sided Conveyor Outtake Test

Conveyor Next Steps

Next up for conveyor is integration testing with the shooter prototype, testing to confirm that passive deflector shields like Team 3636 work for indexing balls as well as advancing the CAD model.

Short Conveyor?

As a final note, our current conveyor configuration can hold 2 balls vertically, however we are leaning towards only holding 1 ball in the conveyor, and 1 at the entrance. We know this will work as we did a similar thing in 2020. This change is advantageous for us for a few reasons:

• Entire conveyor + shooter tower is now 1 ball length shorter
  • Lower Center of Mass
  • Less parts/complexity
  • Solves potential interference issues with climber

Shooter

We finished assembling the shooter prototype right before the end of Saturday’s build session. I will be making a mid week update with testing videos and numbers.

Shooter Prototype Overview

• Parallel Plate + Standoff Construction Hooded Shooter

• Adjustable standoff locations to vary compression and release angle
  • .5" - 2.5" compression range in .5" increments
  • 90° release angle range in 10° increments
• 4" Colsons for flywheel and accelerator wheels
  • Main flywheel driven 18:12 w/ 2x NEO
  • Accelerator wheel driven 1:1 w/ NEO through single stage MAXPlanetary

Climber

We are still waiting on the TTB Telescoping kits to arrive, but that should be any day now! We’ve spent the week narrowing in on the necessary lengths for the passive and active stages of our climber, figuring out hook geometry and advancing the CAD Model.

We also made a pretty neat collapsing hook prototype, but most likely will not be moving forward with this design.

Climber Design Considerations and Overview

• L4 climb via pivoting telescoping hooks & passive sprung hooks
  • 2x TTB SS Telescoping Kits for extension
    • Driven with 2 stage MAXPlanetary
    • MarkForge Onyx 3D printed winch spools w/ WCP 3D Print Steel Hex Inserts
  • 2x Pneumatic Cylinders to pivot telescoping stage
    • Pivot made from 1.25" OD Versaroller, 1.25 I.D. Oil Imbedded Bushings, and Tube Nuts

Drivetrain

1 set of drivetrain frame rails and 2 sets of gearbox plates are machined. We expect to have a mechanically complete chassis some time this week once the drive shafts have been machined on the lathe.

Digital Animation Award

We will be submitting an entry for the Digital Animation Award this year.

• Here’s a sample video

Woody Flowers and Dean’s List

We conducted a series of interviews over the course of the week with our team’s Juniors and picked our 2 candidates for Dean’s List.

Meanwhile the students met and picked a mentor for the Woodie Flower Award by a blind show of hands.

================================================================================

That it for this post, but don’t forget to have fun at robotics!

Protip: 2020 power cells are excellent for a pickup game of team dodgeball!

[robofrenzie]

Hi, great blog. We are a rookie team and your blog is helping us learn a lot. We do not have CADders this season and hope to get the students ready with CAD skills for next season. We managed to do our drive train and intake designs but got stuck at the conveyor / shooter design. Any chance you can share the CAD files to help us this year. Also please let me know the dimensions of your Chassis.

[Nick.kremer]

@robofrenzie

Any chance you can share the CAD files to help us this year.

I can see about getting our current CAD published on GrabCAD for you tonight. I will encourage you to check out Team 3636. another OP team who has a very similar robot with released CAD.

@robofrenzie

Also please let me know the dimensions of your Chassis.

Length: 32"
Width: 27.5"

[Nick.kremer]

Week 3 & 4 Update:

Sorry for the posting delay everyone, it has been a busy few weeks for us. I am going to try splitting these posts into subsystem specific posts moving forward to make them more manageable. The first of these will be an update on our intake progress.

[Nick.kremer]

Week 3 & 4 Intake Update: Or How We Learned To Stop Ignoring Our Own Design Review

As of the last post, the intake was still under design and we just started working on the CAD assembly. A LOT has changed since then and we are currently working on V2 of the intake, but I would like to detail the issue we ran into with V1

Intake V1 CAD

We spent Week 3 refining the V1 intake design to a point where it was ready to begin manufacturing. Here are a few renders showing V1 in all its glory:

V1 Design Review

We did a design review for the intake before machining with senior mechanical students and our technical mentors. The following problems and solutions were discussed:

• Bearing retention screw holes too close to outer edge main intake plate. May lead to failure of plate at the leading edge.
  • Will add more material and relocate holes away from the edge. they don’t have to be spaced 180 degrees apart.

• Pneumatic pivot supported in single shear with loose cotter pin. Loose joint may cause extra wear and tare on the polycarbonate leading to binding.
  • Replace cotter pin w/ 1/4-20 bolt + washers for more secure joint

• Pneumatic placement does not have favorable geometry and may not have enough force to actuate the intake.
  • We decided to not change this as the cylinders had already been ordered and the CAM for these parts was already completed. Spoiler Alert: this was a boneheaded decision.

Intake V1 Assembly And Testing

All though Week 4 we machined, printed, ordered parts for the V1 intake culminating on having it 95% assembled (missing belts) on 2/5:

V1 Unpowered Movement Test

Everything was looking good until we plugged in the pneumatics to test the deployment and retraction and found that the cylinders just didn’t have enough force to reliably actuate… sigh.

We did not get a video showing this, however I should be able to get one tonight as V1 is still assembled.

Intake V2: The Search For More Favorable Pneumatic Geometry

As I type this we are still redesigning the intake. We’ll have to see if the 3/4" bore, 1.5" stroke cylinders we already have for V1 will be sufficient.

What’s the quote about there never being enough time or money to do it the right way the first time?

================================================================================

Next post will be about the conveyor.

[Nick.kremer]

Week 3 & 4 Conveyor Update:

Coming out of Week 2, the Conveyor was still in 2D land. We spent a while spinning our wheels trying every way we could to package the conveyor in as small a space as possible until it dawned on us that we really don’t need to store 2 balls vertically. We can store 1 ball vertically, and the other at the entrance of the conveyor. This revelation solved most of our packaging issues and allowed us to continue forward with the design.

Conveyor Design Considerations

Since we built a conveyor in 2020, we had a good idea of how to approach the design for the 2022 conveyor:

• 1x1 tube and gusset construction
  • We toyed around with the idea of doing large parallels plates for the entire conveyor shooter tower, however sticking with tube and gusset construction makes things more modular in that it provides easy mounting locations for the shooter, climber pivots, and electrical boards
• Tapped Thunder Hex axels require less machining and and are easy to service compared to regular hex
• Timing belts + pulleys give more consistent tension compared to poly cord
• We plan to use ziptied polycarbonate and churro for any angled ramps
• Main conveyor powertrain should be separate from forward conveyor roller so they can be independently controlled similar to our 2020 conveyor
• Mounting conveyor to drivetrain chassis with WCP 1x1 Tube Plugs w/ screws increases modularity. We try to attach whole subsystems to the chassis with screws rather than rivets.

Current Conveyor Design

Conveyor Next Steps

We have a few small action items before we can do a Design Review for the Conveyor and begin Manufacturing:

• Add diagonal cross brace

• Add climber pivot holes

• Finalize outer conveyor roller wheel spacing

• Finalize main conveyor plate, add sensor mounts, pocket plate

================================================================================
Next post will be about our shooter progress.

[KeshavNarasimma]

@Nick.kremer

Conveyor Prototype Overview:

Hey, what speed have you guys been running the conveyor at?

[Nick.kremer]

Around 80%

[Nick.kremer]

@Nick.kremer

Intake V2: The Search For More Favorable Pneumatic Geometry

As I type this we are still redesigning the intake. We’ll have to see if the 3/4" bore, 1.5" stroke cylinders we already have for V1 will be sufficient.

I actually have an update on this! We used some polycarbonate to act as spring and it gives the oomph we need to deploy the intake reliably. We will be moving forward with the current intake design w/ spring assist. We will reassess later on if this solution doesn’t hold up long term.

• V1 Intake Spring Assist Test

[Andrew_L]

Congrats on the Finalist at Ventura! Y’all were awesome!