Team 8013 the Boston Lions is a small team that started in 2020. We got our robot build done in time and just completed our 3rd in person competition ever at the NE District Granite State Event in Salem NH this weekend. Thanks to everyone for making it a great event. We had a lot of teams interested in our turret design so we wanted to post some images in case they are helpful. This is our first post on Chief Delphi so hoping we are going about this the right way.
5 Likes
This year our team tried developing our initial build in CAD which has worked out well. Our team decided early on that our main goal should be to build a high shooter. We noticed in 2020 that a lot teams also used a Limelight vision/tracking system so we were keen on developing a rotating turret that could really take advantage of that. We had a lot of success with a pneumatic climber in 2020 (thanks to another team for helping us out on that) so decided we’d use a similar design and try to be a reliable middle rung climber (we only missed one climb at this weekends event) and not focus effort past that. We used the standard AndyMark drive base and felt that an over the bumper intake system would be effective and fairly straight forward.
Our initial CAD model was done in Fusion 360. Here is a rendering of it out of Fusion 360. It’s quite remarkable how realistic their renders are with just minimal work on applying materials.
5 Likes
This was our 2020 model before we took it apart for parts we compared the ball size and reminded ourselves what we liked and disliked about the design. We tried a variety of materials on that and in particular were impressed with the durability of Polycarbonate so decided to rely more heavily on that this year. We also have limited experience cutting aluminum on a CNC but have done a lot of wood, so felt more comfortable machining Polycarb.
2 Likes
With a CAD model done we cut all our Polycarb parts. We used 1/8" polycarb for less structural components, 1/4" for the sides and 3/8" for the main turret plate (will post more detail on that latter). Parts were cut on a ShopBot Desktop MAX CNC with a single flute 1/8" Carbide bit (and we only broke 1 bit due to bad tool pathing!). Here’s an action shot of cutting parts:
One technique we particularly liked was using blue painters tape to fix our parts to be cut to the table since it eliminates the risk of running into a hard metal fastener. With a tool path printout near by it was easy to see where on the table attachment would be needed and blue tape was stuck in those locations. We would then lay the Polycarb sheet on the table and sharpie where to put matching tape on the Polycarb. With matching tape on the table and Polycarb we would spray the Polycarb side with superglue (cyanoacrylate) accelerant (make sure it dries completely in a few seconds) we would but a small bead of medium thickness superglue on the table side blue tape. We would then carefully place the Polycarb on the table for a nearly instant tight hold down. Once cut we two glued together tape layers were easy to remove and clean off the table and parts.
Here’s a photo prior to glue-down:
1 Like
Some photos from our initial build weekend:
Base assembly:
Intake Assembly:
Unwired but largely assembled robot:
1 Like
We started this thread mentioning our turret. Despite seeming like a fairly straight forward thing it took a significant amount of effort to figure out a solution. As they say a picture is worth a thousand words, so we’ll jump to that first.
Zoom out of the finished assembly:
Top down view of the middle plate which is fixed (does not turn) but has the circle that the movable part rides on. This is made of 3/8" Polycarb. We used a standard motor that is available on AndyMark which has an indexer built in so that we know where the turret is pointing.
This is the assembly that turns. It is made from 1/4" Polycarb and cut entirely with a 1/8" flat endmill (see below but we chose a belt pitch that was roughly compatible with the dimensions that bit could cut).
We used a 200 Tooth 1/4" wide, .2" pitch drive belt from McMasterCarr. We found that this particular pitch is wide enough that the large circular gear can be cut with a 1/8" end mill and be close enough to get a grip. Link below to the particular belt.
Here is a closer view of the turret when assembled. The 8 bolts are comprised of a few large washers that help hold the belt in place, and flanged bearings with a small washer in the middle to get to the 3/8" width we need to ride nicely on the fixed plate circle.
Here’s a view of just the portion that turns.
Figuring our dimensions/spacing etc. for belts can be tricky. We found this video to be very helpful as a tutorial finding spacing when working with drive belts.
You can see we are fans of using T-slot (aka 8020 to bad it’s not 8013 
) channel particularly for this where we were not 100% sure how the turret and the base needed to lineup with out field testing.
1 Like
We just returned from our 2nd regional event in Hartford CT. Thanks to all the site organizers for throwing an amazing event! We’ve learned a tremendous about this season and figured we’d share some notes here in hopes they are useful down the road.
1 Like
Some tips and notes we thought were helpful from the day. Please excuse that these are somewhat raw, many are clearly easy to learn online about we are just trying to provide an overview:
- Try mounting a limelight if you have one vertically for a larger field of view
- Wire management for a turret can be done a lot of ways one of the slicker solutions involved having a channel that the wire could wrap around the turret on and then have a constant force spring which pulls the slack down and holds it in place
- Many teams used a system from Nvidea called Jetson for image processing which apparently they sometimes used for field awareness (sounds way more advanced than anything we tackled)
NVIDIA Jetson Nano Developer Kit | NVIDIA Developer - On this Jetson platform one team referenced using Tensor RT (tensor flow real time is what we understood this to be, which we think is here:
TensorFlow | NVIDIA NGC - Our understanding was this can also be combined with YOLO (you only look once) framework for some powerful object detection
https://github.com/ultralytics/yolov5 - This Pigeon 2.0 accelerometer seemed very popular on swerve drive robots and some mecum drive
Pigeon 2.0 - We heard several teams mention they had printed plastic parts out of “Onyx” with a little Googling we’re thinking they are referring to this not sure if it’s just a mark forge product but seemed very tough:
Onyx One Desktop 3D Printer | Markforged
It took just the right set of robots to make this triple hang from one bar possible. We were impressed by the pro move of the last robot pushing the rest of us over to make room!
Just finished watching the Hartford district. You guys were very impressive for such a young team, and I’m very excited to see how you do going forward!
1 Like
This photo is a good example of using a pressure relief valve to help soft close pneumatics. In our rookie year we didn’t know these existed and all our pneumatics would simply slam full force up and down which is quite rough on things. We figured out these relief valved exist and are a great way to soften pneumatics
The team next to us also had this same problem with the self-tapping screws that came with the standard Andy-Mark drive base backing out and basically milling a section of the drive base away. We fortunately caught it before it compromised the integrity of the bearing housing etc… we learned others were not so lucky. We found that with a 1/4" wrench we could tighten them between matches, also tried some red thread-lock where the metal screw met the metal plate and dried it with a heat gun, that seemed to help hold them a bit better. If we were going to run a lot more we’d want a better solution.
Our poly carbonate intake was remarkably tough. We took numerous bit direct hits and hits from behind into the wall with it down. The 1/2" churro would bend and the bearings got mangled but both were pretty easy field replacements. Here we are swapping the front one for the 3rd to 4th time.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.