Decision on swerve

Now is the time to begin the swerve project if your team is serious. Look at your bom history. What will purchasing modules do to it. Making them in house can save money at the cost of man power and time resources. Several of our swerve components are 3d printed saving significant money.

2 Likes

Ditto PatrickW and RC. We have broken only 1 module and that was kinda expected hitting the 2020 bump with a printed swerve which broke on a knit line. For peace if mind we always have a full set ready for replacement and the biggest reason for an R&R is wheels wearing out. Biggest replacement item has been RS775 drive motors and that issue appears to be resolved. If you are a team that has breakage issues now is the time (off season) to figure out how to drive up reliability. I will attest it is possible.

2 Likes

Thank you. In your experience, is there anything that drivers do that can damage the drivesystem, besides driving recklessly?

Lots of good info in this thread already. But also some info that was true a few years ago that is not as true lately. Here is the key information I have shared with other teams:

  1. Investing in COTS modules for your comp bot is the way to go these days. Although the upfront cost will be a chunk to bite off, the ability to re-use them will pay this back over time to your team’s finances. But probably more importantly, not needing to manufacture the modules during the build season will be a huge time saving for the team and allow you to focus on the other parts of the robot. We still use our custom designed modules, so we don’t have any direct experience with the options that are out there, but I think the recommendations above are good.

  2. You definitely want to do your development during the off season so that you have the confidence in your programming and driving going into the season. It will also allow you to start to adjust your mindset to how to take advantage of swerve in terms of the overall robot design (any side of the robot can be the front at any time, more nimble aligning and aiming, etc).

  3. Spare modules and heavy maintenance are really becoming less and less of a concern. although we always build one spare module each season, it is mostly for show and tell in the pits for teams that are interested in swerve. I can’t remember the last time we replaced an entire module on the bot. Occasionally, we will steal a motor assembly of the spare module because it is easier to swap the whole motor assembly than do something like making a wiring repair between matches. But we usually try to repair the motor assembly that was damaged at some point during the day and re-build it onto the spare module just in case we need it. I really don’t think you need to invest heavily in spare modules. If you want to have 2 sets of modules (one for your development bot and one for your comp bot) that is up to you. But, I would not by a complete 3rd set of modules. One or maybe two spares is probably more than you will ever need.

  4. Routine maintenance is important (replacing worn wheels) but generally is not needed during an event. Mostly, that is done between events. We went the entire 2018 season without replacing any wheels, but looking at the wheels after the season, we probably should have. I don’t think we replaced any wheels in 2019 either. Each design is different and a lot depends on your driving style in terms of how much wear you are going to see. But every design these days has made it easy to make such repairs. We have spare wheels built up and ready to install just in case we see some unexpected wear during an event.

  5. As @PatrickW and others have said, the issues that swerve teams had with their motors being damaged during 2019 was a unique issue having to do with the field. I saw several teams install guards on their bots to protect the motors from impacting the cargo ship. Our team did not see that aspect of the field until it was too late and we had broken the dust cover case off of 2 of our NEO motors at Worlds. Having spare motors is a good thing, but it really should not be a concern most of the time.

  6. Learning to drive swerve is not a big deal. In fact, most new drivers that have picked up the controls to one of our robots are able to pull of high speed maneuvers within the first half an hour of driving and are generally trying to fine tune their response after that. Experienced drivers that are used to driving tank drives (especially using the 2 stick controls) have the hardest time learning to drive swerve, because they have to unlearn their prior muscle memory. But, if those drivers are avid video game players, and play a number of different video games such that they are used to adapting to different controls, they can generally adapt to driving swerve quite quickly as well. Make sure that you have the ability to drive in field oriented mode as well as robot oriented mode. Many drivers prefer field oriented mode and it seems like most people have an easier time learning swerve if they start in field oriented mode. But each driver is different so having the option to switch quickly between the two modes is good.

  7. Transitioning from being a swerve driver to being a great swerve driver is a matter of confidence and experience. The more drive time you can give you your drivers the mode confident they will get. Some elements of that confidence will be specific to your robot design (for example, how quickly you can turn without tipping over will be a function of the CG of the robot) but other aspects can be learned with a generic practice bot. Swerve drivers love to drive and setting up challenging obstacle courses or pitting them against another bot to learn how to counter defense is a good way to help them build this confidence. While some people just have the knack for driving swerve well, unfortunately others do not (no different from video gaming or other fine motor skill activities). So, make sure you give everyone a chance to try.

2 Likes

Thank you. If you would be fine with sharing, which parts are you comfortable with 3d printing, and what filament do you use?

Make sure you keep track of your motor temperatures. We have found that students like driving so much that they don’t give the robot a chance to cool down. They wear out a battery and just replace it and keep going. We have never burned out a motor doing this, but many times they are very hot to the touch. So, I think there was some risk that we could have burned out the motors.

I will let other comment on any specific issues that you might need to be careful about with the COTS modules.

1 Like

Thank you. When our mechanical students were researching swerve modules, I believe they showed interest in yours. Would you have any issues with us trying to use a modified version of your design? Would there be any concerns if we tried to do that?

Thanks for the tip! We probably would have run into that issue, I along with the rest of our students aren’t exactly known for our self control :upside_down_face:

1 Like

Thank you for the advice! We aren’t too worried about driving, as everyone either has experience driving some form of holonomic drive, or just playing fps games. Our team is hesitant to try a COTS solution, since initial price is a major factor, and we are hoping that we can test a custom module off-season, and hopefully get some driver practice and work out any kinks with programming and driver preference.

DISCLAIMER: I am speaking from my team’s experience here. I’m not saying I’m right or this is the best way to go. Thank you have a great day

For a custom swerve, I would agree. However, after my team’s experience using the SDS swerve modules for the first time this competition season (no offseason experience at all), I don’t think it’s as difficult as people make it seem. We had few issues with the modules themselves, nothing that wasn’t fixable, and one issue being a simple mistake when wiring and nothing to do with the module itself. For anything else we had issues on, we contacted @PatrickW and were able to get quick and accurate help with.

If you start from scratch, yes. But (in reference to the SDS modules) there is a ton of documentation on the programming. Our programmer had never previously programmed swerve, but having someone who can understand the code is key. He was able to understand the code and fix something if needed.

I’m not promoting jumping into swerve blind during competition season, but if it happens, i do think it is very possible to succeed, even for a lower resource team like us. We never intended for our team to try it for the first time during the season, but given the wonderful resources that the FIRST community provides, we were able to successfully get through it.

4 Likes

Would definitely at least entertain the idea of custom swerve. This year we decided to go custom and with out motors, we came to a price of around 140 per module. In the long run cots is generally more cost effective because they can be reused but if you can stomach rebuilding every year, being able to make changes to the module to tailor it to each game is a nice advantage.

1 Like

Thank you for the advice! We saw you guys at Palmetto, and by the end of the comp you guys had a very consistent shooter, and imo one of the best bots at the comp. It’s a shame you guys didn’t make it to playoffs. Also congratulations on your Chairman’s win this season!

1 Like

Thank you! Do you guys have any experience with custom modules?

Depending on your machining resources (or access to a sponsor that can machine parts for you), you may find that the cost of a custom solution is not that much lower than COTS. Our BOM costs for the COTS components for our custom modules is $250 per module when you exclude the motors, motor controller and encoder, and $500 per module when you include NEO motors, SparkMax motor controllers and a Lamprey encoder. We bought a lot of that stuff on sale and are able to re-use a lot of the gearbox components from year to year, so our actual budget costs are lower than that. Our BOM costs for the raw material are about $50 per module for the aluminum plate and filament for the 3DP parts (the aluminum is in fact donated by a sponsor, so our budget cost for that is 0). So, you may want to run the math on your custom design (for both BOM and budget) and see how much, if anything, you are saving over the COTS modules.

Like I said, the value in the COTS modules is that you only have to buy it once and you can re-use it year after year. It is also extremely valuable when you realize that you can build it in the off season and use it during the season so you are not wasting valuable build season time manufacturing your custom solution.

As a team that continues to use our custom solution, I can tell you that custom modules are great and an excellent way to give the students a technology that they can own, modify and become experts in (which is a huge part of why we do FIRST, right). So if this is what you are hoping to get out of the switch to swerve, then by all means, go the custom route. But, if you are just looking to do so for cost reasons, you will probably find that it really doesn’t save you as much cost in the long run as you think it might.

2 Likes

Alright, thanks for the info!

+1 for pretty much everything here. An additional angle to consider: strategy.

Definitely get the whole team in the mindset that “we’re doing this to learn and grow the team, at the possible cost of max performance on the field in 2021”. Even if executed perfectly, swerve may not be the best option (think giant bumps or similar). This is ok, as long as everyone has the mindset of learning.

You’ll want to make sure everyone (including new students) understand this, so that in future years you can say “we have mastered swerve, it is a tool amongst many in our toolbox, and we can choose between our options after kickoff to build the best performing robot”. Ultimately, this is how building a swerve drive this year helps you build better robots down the road.

2 Likes

Thank you! That’s actually one of the main reasons we want to do swerve. Honestly, we don’t expect our custom design to function very well during our first season, but we want to try. (Also we have all been obsessed with swerve for a while :rofl:)

1 Like

That’s excellent! Every team has a different way of making their inspiration printer go “brbrbrbrbrbr”!

Have fun! Sharing any tips or tricks (especially “lessons learned by a first-timer”) is well appreciated!

4 Likes

Will do!

I’ll also throw some questions out there that I think every team thinking about swerve should ask themselves -

  • Did your team seed high this year or in previous seasons?

  • Is your team currently good at traversing the field with a standard 6 or 8 wheel drive, especially in autonomous mode?

  • Has your team won an event as an alliance captain or high pick in recent seasons?

  • Do you have more than $2000 to dedicate to the drive train from a budget standpoint?

  • Do you feel like your team played the 2019 / 2020 game well from a robot standpoint and that the swerve drive is the thing you need to give yourself an edge over the competition?

Swerve is really cool. We made one based on 2767’s in the 2017 off-season and went forward with it during the 2018 build season. It absolutely did not help us win matches and actually hurt our robot performance overall because we sunk so much time, effort and resources into it. In 2019 we did a complete 180 and went back to using the kit drive and did the same in 2020. 1720 won our first blue banner since 2016 this year.

I know it’s not all about banners and swerve is a fantastic learning experience, don’t let me stop you. But definitely frame what your goals are and decide whether it makes sense to spend time on this instead of what goes on top of your robot. It’s fun to fly across the field in any direction you want but if you can’t pickup a power cell (or insert 2021 game object here) then what good is it?

5 Likes