Spectrum 3847 - Build Blog 2021

With the bulk of our FRC at Home challenge work completed, we have begun work on designing a new robot for 2021 off-season events: Gamma Ray 2021. It is still in the early stages of its design, but we wanted to allow people to follow its progress.

OnShape Public Link

Photo & Slide Gallery


  1. In the fall of 2020, we decided to register a 2nd FRC team at our schools, Photon FRC#8515. This team will be the “Development Team” for Spectrum. The general plan is new students will be on Photon but will also be members of Spectrum at the same time.
  2. The plan is for this new robot to compete under the banner of Photon at The Texas Cup event.
  3. We have been kicking around the ideas for this robot since March 2020, but the actual CAD didn’t get started until last week. We didn’t build or modify our 2020 robot for the skills challenges; we just used it as is.
  4. Spectrum has switched from Solidworks to OnShape, so it will be easier for anyone to view the robot’s progress.
  5. Most of the build blog updates for this will be the design slides our team uses to organize our thoughts and keep everyone updated. The full design slide set will be uploaded to our photo gallery.
  6. One of the major goals of this robot was to design and build a full swerve robot before the 2022 season. If you haven’t heard, there will likely be a lot of new swerve teams in 2022.

  • Spectrum

I really appreciate the work y’all do when sharing with the community! I can’t wait to see this come together and to compete with Spectrum and Photon at The Texas Cup!


What are you using to couple the servo to the shaft it drives?

I assume you are talking about the image on the launcher slide, correct?

That isn’t our CAD model that is an image of 2102’s launcher. I can answer the question though, they have a small 3D printed adapter from the servo horn to the hex shaft.


Week 2:


Spectrum was recognized as one of the FIRST in Texas District Chairman’s Award teams. We had one of our students from 3847 and one of our students from 8515 advance to the District Champions Dean’s List round.

Congratulations to all the teams

Subsystem Updates

OnShape Feature Scripts

Here are some of the feature scripts we have been using on this robot to help us generate parts faster.

Laser Joint

  • Creates the finger joints that we often use with laser cut parts
  • Video

T-Slot Joint

Advanced Gear

  • Creates Spur, Helix, and Double Helix gears

Shaped Flange

  • Allows you to make a flange from a sketch profile, and can also offset the flange easily so you can make “Bend Outside” parts nicely.

Week 3 and 4

This has been a much slower build. We are only meeting in person on weekends and CAD development has slowed as students prepare for AP tests.

In Week 3 we were able to get our new router setup to drill hole patterns in tubes and used our laser cutter to cut our belly pan.

With rails cut and a belly pan, we were able to assemble the chassis for our practice drive train.

In Week 4 we were able to add electronics and mounting to our practice drive train and get it wired up.

We also mounted some temporary bumper corners to the chassis so we can easily use this for programming and driver testing.

The competition chassis was also advanced to the belly pan and rails. We have one of the swerve modules for the competition robot assembled and should get more of the other 3 done this week.


Now that it’s been some time since the Texas cup, did the robot live up to your expectations? Did you change anything between the last update and the event?

The build process for this robot was definitely tougher than most seasons. We were only able to finish the mechanical and electrical systems the day before we had to travel to the Texas Cup.

We’ll have more detailed updates in the future but some of the issues we experienced at the Texas Cup that we will be solving for future off-season events.

  • Inadquiet quality control and Loctite application during swerve module construction led to loosening hardware and caused multiple problems.
  • Insufficient time to tune automated systems, we didn’t have automated aiming working so we were having to aim by hand. We did have reasonable velocity control on the launcher and tower and field-oriented drive working on the swerve (thanks to the 364 example swerve code).
  • As everyone saw we have some major issues with the climber that we are still diagnosing. We were relying on the Falcon brake mode to hold the robot up which worked in testing and on the practice field. It also worked for Tuesday at the event but Wednesday and Thursday we had constant issues on the field. Our climber gearbox plates are made from laser-cut acetal and we believe our flexing in a way that is causing issues during the climb, we will likely be replacing them with aluminum plates going forward.
  • Our indexer mechanism was able to destroy two 3mm timing belts during the event from ratcheting and removing teeth on the belt. We were able to 3D print a gear to replace the belt and pulley at the event and the problem was fixed for Thursday.

Some of the positives

  • The SDS MK3 modules when fully tightened and proper Loctite applied worked great. We are planning to move to MK4 modules for 2022 if swerve is a reasonable option for the game (i.e. not 2016 style field). We also need a lot more driver practice (30 mins is definitely not enough) before we think we will see the real advantages to swerve.
  • The indexer, tower, launcher worked well once the belt issue was solved and we were getting some very consistent shots. We think there is a lot more potential for ball scoring in this robot once we get things more automated and tuned.
  • The intake had some issues but the combination spring deploy and pneumatics is a promising combination that achieved most of its goals; being able to take up a different volume of space during the match than it does pre-match while still be able to be retracted while moving around the field and being able to flex when we drive into walls.
  • The bumper latches worked very well, we will tweak the designs a bit but it’s likely they see a return on our 2022 robot. Toolless mounting and not requiring any space above the 2*1 frame are both nice features.

Allen, can you elaborate some on the bumper latch system? How it works, etc.

1 Like

I think this post shows some of the latch system, I too would like more details

1 Like

Can you say what components of the SDS mark III modules required loctite to avoid failure? Other users might like to avoid similar issues.

Am I wrong, or did you change out the indexer from the new horizontal belt design to a v-hopper (or was that a different robot)?

I am still not clear on exactly how it works even with the picture in the post you linked. I think a picture of it on the robot would clear it all up for me.

1 Like

Bumper latch pictures and explanation

These two parts are on the robot in 4 locations, they were both riveted on for the competition robot, this photo is just of the test mounting.

Then the mating part is attached to the bumpers. We ended up making this set of bumpers a little too loose to the robot so we needed these parts to be about .4" thick instead of .25" so we added spacer parts.

Basically all of the bolts since none of them use locknuts, it was our mistake for rushing through the assembly process.

We didn’t change the indexer on Gamma Ray (8515). Ultraviolet (3847) used the same v-belt indexer it had for the 2020 season.


We’ve uploaded some more pictures to the photo gallery. - 2021 Swerve IR - Spectrum 3847


We really loved your 2020 and 2021 robots, especially the dead axle roller system. We implemented it on our own robot and it made our intake design much cleaner than having separate struts from our rollers to provide rigidity.

One issue we had, though, was that our clamp on pulleys would slip on the rollers to some extent when there was load on the intake (when intaking, for example). It wasn’t a big enough problem to prevent it from working, but it made some pretty awful noises and probably will end up being an issue at some point. We printed ours out of PETG. I think we copied the ID of the pulley relative to the roller from your CAD, but I’m not positive on that.

Have you seen this issue and what would you recommend to resolve it?

We didn’t use clamp on pulleys in 2020/21 only 2019. For 20/21 we had the pulley-hubs inside the polycarb and tapped them for bolts and drilled holes in the polycarb tube. This works because we were using large polycarb tubes (2.5" and 2").

When using clamp on pulleys you want to increase friction between the pulley and the tube to prevent slipping. Sanding the tube and/or inside of the pulley to rough up the surfaces, adding friction tape between the tube and pulley, increasing the clamping force by tightening the bolts even more, etc. Another option we haven’t tried yet is this stuff, I bought it to try a while ago for applications like this - https://smile.amazon.com/dp/B0012RIEM6/ref=cm_sw_r_apan_glt_fabc_TW2FNFRHXC93W9AE7CNT?_encoding=UTF8&psc=1


Thanks to @Tyler_Olds and FUN for filming a behind the bumpers segment on the team and this robot.