Sent from my iPhone using Tapatalk
Nice design, keep on iterating and improving it! Swerve is a great design challenge.
That being said, please please please please please do not use it in-season. Making your own swerve is not a wise choice for a majority of teams. I know a lot of people saw 114 and 192 at SVR this year and were impressed by their driving, but unless you’re extremely confident in controls code and have high performing robot design down to a science, it is not a good idea to use your own swerve. If you’re interested in swerve, consider using a COTS module that you can develop your code on in the offseason first, and then use that module in-season. Making your own swerve is exhausting.
Not saying you’re planning on using it, but I’m just gonna put that out there just in case anyone else is considering it. Be it 775drive or knockoff Aren Hill swerves, CD is really good at bandwagoning onto an idea because of its “coolness-factor” and it can often lead teams down the dangerous path of reaching too far outside of their capabilities.
Very compact looking. How much does it weigh and how fast does it go?
Do you have an absolute encoder in this design?
I might be a little concerned about the CIM mount plate. With only two mounting points, I see the possibility of it flexing.
What’s your bearing setup like? It’s a bit tough to tell from the picture.
Each module weighs around 6 pounds and they are geared for 15 ft/s. I haven’t tackled the encoder/sensor placement yet. The bearing setup is a slimline bearing press fitted into the main plate.
Sent from my iPhone using Tapatalk
Could you post a STEP file or something so people can actually look at the design and give feedback that would actually help you improve it and iterate it? These screenshots aren’t too helpful…
Sure I can! Here’s the link the the grabCAD Page: https://grabcad.com/library/team-100-swerve-v1-beta-1
Let me know if it doesn’t work!
I’ll admit I didn’t look too closely, but 3/4" plate is wayyy overkill. 1/4" plate will be just fine.
I agree the CIM pointing plate needs to be better constrained. 2 bolts is not going to be enough.
Also, what’s taking up the thrust load on the module? It looks like the module is pushing on the bearing, but I don’t see what’s keeping the bearing from moving.
Would Loctite work for holding the bearing in place and taking the thrust? Then manufacturing tolerances on the bore wouldn’t have to be as tight as they would for a press fit. What would you recommend?
Sent from my iPhone using Tapatalk
The other option would be to capture the outer bearing race with some screws near the edge.
Sent from my iPhone using Tapatalk
I would not trust loctite to hold the entire weight of the robot. Screws to hold the bearing in will probably work if you have a bunch. You still need to make sure the spinning part of the module isn’t rubbing on the rest of the bearing.
ie. A thrust bearing. A perfect application for one.
I’d add in some lightening on that cim mounting plate, and it should be really easy to add holes that align with one of your preexisting holes or some new ones so you have 3 or 4 to properly hold down that cim. Your 3/4 plate is really really massive for this type of application.
Definitely take a look at the VP integrated encoder for the turning but also be sure to add solid mounting/wire runs for your wheel rotation encoder.
Overall, a cool design with lots of solid concepts. I’m by no means a swerve expert, but I’ll definitely echo Andrew’s thoughts on the matter. Unless you already have this made in the offseason with code running on it successfully and have drop tested and everything, I would seriously caution using swerve for the first time in season, especially a custom swerve.
Good luck in your testing and pursuit of a final design!
Also I’d check the gear ratio on that. IIRC single speed swerves usually have ~5:1 reduction. That looks closer to 3:1 (I didn’t actually look at the part numbers or anything, just speculating). What are your free and adjusted speeds?
The ratio is 4.89:1. The free speed is 18.89 ft/s. Adjusted is 15.4. This may be a tad fast, but I guess that’s what off season testing is for.
Sent from my iPhone using Tapatalk
To take in the thrust forces, I have added in a lip on the top end of the main plate to capture the outer bearing race. The bottom will be held with washers and screws. The inside piece that holds the wheel forks has a lip on the bottom and captures the inner bearing race with washers also.
Sent from my iPhone using Tapatalk
Definitely a unique take on the bevel-beside wheel. The 3/4" plate may be overkill, but it serves an interesting purpose in completely capturing the bearing and elevating certain components. That being said, 3/8" could probably do the job just as well. Even thinner can work but 3/8" is quite safe.
Good work using the 61916 bearing. 6816 bearings might work better due to their smaller cross section. A few teams (1323 comes to mind) have used a single 6800-series bearing to take both thrust and side loads, albeit with a much smaller wheel. The lip.flange option is the right choice there. If you want to use this style of swerve you’d be better off with a much smaller wheel like a colson to reduce the torqueing on the bearing though.
19fps is pretty fast for almost every year. If you want to practice for the season, consider using a more realistic speed like 15-17fps free. Reducing wheel size also helps with this.
I really like the way you’ve set up the pulley sandwich. Looks like a fair amount of machining, but it’s all one-sided anyway. What purpose does the long tube capturing the bearings serve?
CIMcoders might work for this for drive encoders. VP integrated encoder is easiest for rotation, but you’ll need some way to zero the sensor.