Hello, chief! I’m next year’s head captain for FRC 5438 in Jersey City, and for the first time in our team’s history, we are thinking of purchasing swerve modules (mk4is). However, we are unsure which gear ratio to get (we’re thinking L2 or L3). If needed, we can always change it later, but we were just wondering what the best one would be that would fit most all scenarios. L3 is definitely stupid fast, but I’m not sure if the potential drawbacks would outweigh the speed. Any thoughts?
Thanks!!
1 Like
Another thing is to consider whether you are using NEOs or Falcons for you drive. NEOs will have higher torque, but lower free speed.
I think it will depend on the game, mostly. I think this year, L3 was the better way to go, but I don’t have the most experience with choosing gear ratios. (Literally no experience)
We will likely be using falcons (if we can get our hands on them) Thanks for the advice!
L2 is very fast without sacrificing much acceleration and is suitable for most games, especially for full-weight robots or when rapidly changing directions is important. L2 ratio is pretty close, but not quite traction limited meaning it gets close to the maximum possible acceleration while still being fast. Having driven L2 modules in 2022, it is still extremely fast and it accelerates almost instantly. I would generally suggest going for L2 as many games favor the increased acceleration over maximum free speed. There isn’t a game where L2 is a bad ratio, and in many games, it is the best ratio.
L3 has an even higher free speed, but you will lose acceleration due to increased current draw unless you can build extremely light robots (< 100lbs). L3 is better suited for games like this year where sprint distances are long, but will perform worse when acceleration is more important.
4 Likes
Is that true? NEOs have lower stall torque to a pretty significant degree.
At 60a, NEOs have a stall torque of 1.18 Nm, while Falcons have a stall torque of 1.06 Nm
(at least according to ReCalc)
Going to lower current limits, from some math I’ve done on ReCalc, at 30a, NEOs have ~2% more stall torque than Falcons when geared to the same free speed. However, I could be completely wrong.
We use L3s.
In software we basically limited down the speed. ITs most likely ‘too fast’. If I had to rebuy I would probably think about getting L2s. We didn’t need to sacrifice some of that Torque for Speed considering we then limited out some of that speed. I think you can buy other gear ratios if you wanted to move to a L2 or L3 from your original gearing.
That being said… I’m sure this is just the nature of having 4 Falcons directly geared to a wheel (this is partly why I roll my eyes at these 3 motor swerve systems, its just throwing money at something trying to solve a problem that does not exist). We never got pushed around. I never felt like we needed pushing power we gave up from going L2 to L3. Or the additional speed improvements of a second motor
So having that improved speed might have future proofed us a little bit, but i highly doubt there will be a game where you sustainably travel at 18 ft/s.
1 Like
Aren’t teams such as 4414 utilizing three motor swerve to go faster? It doesn’t seem like they’re trying to get more pushing power, so their money does seem very well spent since it gives them a competitive advantage.
This was our first year using mk4i and we used L2 and they worked very well for us.
In my opinion its overkill. I don’t think you ever are going to sustainably get to travel at that top speed. I’d be curious to see what their average speed was and how often it was realized. I’m sure they did it more for an acceleration boost but even then I’m not sure.
Not to mention now you have 4 extra motors on the same battery and need to worry about brown outs and other considerations such as can bus utilizations and extra weight. The two latter things of course are pretty negligible though.
I’d speculate that robot would have been just as elite on an 8 motor swerve system. I have no way of proving this. I just don’t like this trend in FIRST where we try to brute force solutions. I’ll die on that hill.
1 Like
it’s not a brute force solution . it has to do with efficiency. this was heavily discussed and data was provided leading to the conclusion that “They were really whipin”
my understanding is It has to do with duty cycle and efficiencies of the falcons and torque per amp at different duty cycles. they found a way to significantly increases efficiency of the drive system. it starts to makes sense when you look at falcon motor curves at lower voltages, my understanding is 50% duty cycle corresponds to 50% of vbus when looking at motor efficiency curves . same total current draw for there drive system accelerates faster.
2 Likes
12v 1500 rpm ~3.6 NM ~200 amp draw .018 NM PER AMP
6v 1500 rpm ~1.5 nm ~36 amp draw .0416 NM per AMP
1 Like
I’m not here to argue. Noones going to agree with me. In my opinion you are chasing efficiency you do not need. I guess brute force isn’t the best term as it clearly improves the efficiency of their swerve systems… but why. You can already drive at such a high rate of speed is my only point.
Apologies for high jacking the MK4i Gearing thread! I’d go L2 but L3 is fine too. Wouldn’t recommend L1.
1 Like
If you can do it semi cheaply, REV MAXSwerve with 2 neos driving per wheel is cheaper than SDS MK4is with one falcon driving and one steering.
However, it’s a significantly greater design challenge than putting another falcon on a SDS module.
So maybe it’s not just throwing money at the problem in some scenarios?
We used L2 and I feel we looked fairly zippy out on the field. Data is a few years old now and was half field cycles but unless you are pushing for top .1% on field performance the L2 is probably the most forgiving.
You Probably Geared Your Drivetrain Wrong: Definitive Proof has some interesting discussion around those speeds but it was pre swerve proliferation. I think the general consensus was 12ft/s for half field and 14ish for full and then the next priority was spending less time standing still manipulating (gets and sets) of game pieces. If you find you nailed that last part you can always easily swap L2 for L3 if you feel you need it and can get the parts
We were in a similar situation last summer when purchasing our own swerve modules. We went with L2s due to the availability of each ratio with the understanding that we’d likely be under-geared if the game required full field cycling…and what do you know, that’s exactly what happened. Although we would have preferred using the L3s this past year, we still felt like the robot was quite agile at our districts (Escanaba and Troy 1) and we didn’t start to feel uncompetitive with our sprint speed until the Championship.
You’ll have an easier time acquiring L2 gears and that ratio will likely be closer to optimized for next year’s game than the L3s, but if the GDC happens to give us back to back full field cycling games, you may have opportunities to upgrade to the L3s later in the year.
sure…
FWIW we are on NEOs with a free speed gearing of 17.2 on our “mk2.5” swerve, but it does take some current draw (we run a 60A current limit).
I’d run L3 (16ft/s) if using NEOs for sure for a game like Charged Up.
If we were running Falcon’s we would have also chosen to run L3s or something close to that gearing, but L2 is a very good trade-off though at 16.3ft/s on the Falcon. It will be a very good “all-around” speed.
1 Like
Offseason 2022 we went with L2s and they were fine for offseason 2022 comps, but we regret not having L3s this game due to the crossfield cycles that were required. L2s are a good middle ground but it all comes down to your sprint distance. L2s are great if you’re going short distances because they have a faster acceleration than L3s. L2s were great for going ball to ball in 2022 because there was rarely a time where you had to cross the whole field. L3s are more optimal for long sprint distances like this year where you had to cross the whole field to get to your substation because they are higher geared for higher top speed, but slower acceleration
If you are planning to build an offseason bot for THIS year’s game, go for L3s. If not L2s are safe. IMO L2s and L3s are equally good options, they just specialize a little differently.
We used the L1’s as it still had a higher top speed and acceleration that the KOP drive train and it gave us a solid base to practice and figure out swerve drive. We are looking into either the L2 mk4i’s or MaxSwerve as our programmers have more experience with the Rev API.
1 Like