This is a design that I have been working on for a while now that we plan to put on a t-shirt cannon as a summer project. It’s driven by 2 mini cims with a ratio 16:56 18:48 with a final speed of 16 ft/sec on a 6in pneumatic wheel. Turning is done by two 775 pros with a ratio (12:48) (18:48) (24:84) so it has the same turn speed as drive speed. The caster has a 3.25 offset and is using an SFX 4 by 4.75 thin section bearing (JU040XP0). We are going to be running the robot off 24 volts so we will be running the motors in series.
4 Likes
Is there a reason to use the caster design over a normal swerve?
16 slots from the PDP…
I’m not even going to try to convince anyone why castered swerve is an oxymoron and a bad idea, it just is… sorry.
I mean, there are interesting things to think about…
FTFY. 
This seems like an overkill type of design for a promotional robot, not an FRC legal robot. You can do whatever you want in that case; more power = more fun! OP never even mentioned how many wheels they plan on using either, it might only be 3. Or 2.
Don’t know to much about the specifics but there are 3 wheels and I believe a triangular drive train.
Then why post?
Last I checked, no conclusion was reached in any of the CD castor swerve threads regarding the effectiveness of a castor swerve in an FRC context. Judging by your post history, you’ve not participated in any of those discussions, so you’ve either not read them, or decided that your opinions were already well stated by others.
In both cases, you really do need to explain why you think a castor swerve is a bad idea if you want to be taken seriously, and you most certainly should explain why you think it’s an oxymoron. (which I would reckon is only possible if one misinterprets the operation of a castor swerve or the definition of “oxymoron”)
As I replied to the most recent caster (not castor, which is vegetable oil) swerve drive posting, the purpose of rake and trail which results in caster is to make a following-wheel more stable while following, going straight.
The purpose of swerve is to be able to turn readily, at will. So making it harder to turn doesn’t make sense. To me.
If you can tell me why you want to caster it, I’m listening.
The purpose of offsetting a swerve wheel has nothing to do with stability. It is merely a side benefit. There is also no trail in such a module, which is evident by just looking at it, so your reference to rake and trail in this post isn’t actually relevant, as that’s just not how caster swerves are designed and built.
I won’t re-type the entirety of that thread, and I highly encourage that you reread it. Please take the time to understand the design and motivations before casting 
it away.
TL;DR: The purpose of offsetting the wheel is to allow fine adjustments through the steering of the module.
edit: Also, I hear others are partial to using olive oil for their drives, but once you go castor you never go back.
So yes, it is going to have three wheels. Also, the reason that we have chosen to overpower them so much is we know that the robot is likely to be over 200 pounds and we want it to have decent acceleration. The robot is also going to run on 24 volts instead of the normal 12, most of the standard FRC electronics can run on 24 volts but the motors can’t. So we are going to run the motors in parrel, for example, the two mini cims will be run of one Talon SRX and one PDP slot this also allows us to use one slip ring per module. So the drivetrain will only be using 6 PDP slots. The reason we are choosing to do a caster drive is that it hasn’t been seriously tested yet, we decided that this is a project that would allow us test casters without much risk.
-Owen Wilks
P.S. This design was never meant to be used in a real competition it is just a fun way to test the concept.
Wow. That looks like a very cool design. Let us know how it works out. I bet you learn alot about engineering swerves from it that you can take forward to competition. Just ignore the naysayers. There seems to be a lot of unnecessary negativity on CD as of late.
Nice work Owen. It looks great! One suggestion I have is to add a rib in the middle of the slot for the 775 vents. It shouldn’t block airflow, and will improve the motors’ mounting rigidity.
I’m really curious what your motivations were for making this post. Please enlighten me.
Owen touched on the motivation for this application, but in a more general sense, an offset wheel would be used on a swerve for the same reason that one would be used on an unpowered caster: So that the structure it’s mounted to can move immediately in any direction without skidding the wheel.
Cool design!
How did you go about determining the position of the idler pulleys?
I’ll offer what I do in Solidworks (it may be different in Inventor, which is what I believe 2471 uses):
Usually on the master sketch for the part, I’ll make a belt layout with the Input and output pulleys fully defined, but with the idler under-defined. Those pulleys are represented as circles with the pitch diameters (you can typically just look these up). I’d then create a path-length dimension set to the tooth count times the pitch (ex: 80t HTD 5mm -> 400mm length). I’d then drag around the idler until it’s in a position I like, then dimension it into place. When making the plate itself, the idler usually doesn’t get a hole but a slot, to make tensioning easier.
Feel free to pm me if you want more info or if I just don’t make sense.
Please verify the maximum input/battery voltage allowed on the current PDP manufactured by CTRE and the previous PDB to ensure that they will function properly in your proposed application.
It doesn’t look like they used a slot, so they’d have to know the exact positioning or have a plan to make custom sized idlers.
ClayTownR you are correct we do use Inventor (2017) and use sketches to calculate the theoretical perfect belt path, but we also build in a way to adjust it.
So for the idler pulley between the 775 pro’s the plan is a Delrin roller on a shoulder bolt that has been lathed down to provide the proper tension. It may take some adjusting to get perfect but will be a simple solution in the long run.
For the tensioning of the belt on the mini cim’s I have implemented a solution is very similar to what team 192 has done with their belt gearboxes in the past.
Philso we are planning to control it with a hero board that can be run up to 28 volts and the matching PCB boards (I believe these can be run at 24 volts but am not 100% sure). Talon SRK can also be run at 28 volts, the electrical system is one of the main parts of the robot that we are still trying to figure out any advice would be greatly appreciated.
-Owen Wilks
ClayTownR you are correct we do use Inventor (2017) and use sketches to calculate the theoretical perfect belt path, but we also build in a way to adjust it.
So for the idler pulley between the 775 pro’s the plan is a Delrin roller on a shoulder bolt that has been lathed down to provide the proper tension. It may take some adjusting to get perfect but will be a simple solution in the long run.
For the tensioning of the belt on the mini cim’s I have implemented a solution is very similar to what team 192 has done with their belt gearboxes in the past.
Philso we are planning to control it with a hero board that can be run up to 28 volts and the matching PCB boards (I believe these can be run at 24 volts but am not 100% sure). Talon SRK can also be run at 28 volts, the electrical system is one of the main parts of the robot that we are still trying to figure out any advice would be greatly appreciated.
-Owen Wilks
You will for sure want to have the pairs of motors in series, not in parallel. If you put them in parallel they’ll both run and 24V, which will likely kill them. If you put them in series they’ll drop roughly 12V each. You mentioned parallel in an earlier post and I’m not sure if that’s what you meant, so I just wanted to clarify.
Thanks, pkrishna3082
I work mostly on the mechanical side of the team and don’t know much about electrical, so this is great advice and the results we wanted to see.
-Owen Wilks