Disadvantages of swerve/crab

I love swerve drives, and by the looks of them a well implemented one SEEMS like the best drive train in FIRST. So my question is: What are the disadvantages of swerve? I don’t mean like it’s more difficult to build, or it’s heavier than other drives or harder to program, but disadvantages in performance.

Complexity of the hardware.

Simply, there is none. An independent coaxial swerve, programmed perfectly, will be the best drive out there. The one performance disadvantage, is that if one module brakes, your whole drive system is down(this applies to all drive systems, except wcd and variations). However, the resources needed to get to this point are tremendous, and without planning and time, the benefits are nullified.

Competitively, Swerve is, perhaps, the best ALL AROUND drive train in FIRST. All around meaning it follows the old saying “Jack of all trades, master of none.”

Swerve is so much desired because it offers the omni-directional movement of Mechanums, without relying on wheel slip and thus, is not as easily pushed. 4 Wheel Swerve will offer, more or less. similar power as a 4 wheel drive. Which is not particularly excessive, but well enough.

In addition, Swerve won’t outrun a lot of drive systems. It’s going to rely on that omni-directional movement to cut out the time used in turning to make up for the lack of excessive speed. Unless you’re 973 using Emperor Swerve, you’re not going to be moving at very fast speeds. That’s going to be the result

Average speed + average power + omni-directional movement = one very effective drive system. But it will not handle drive systems that excel in power or speed very well. That is, of course, based on an average Swerve. Some teams who have become very well versed in Swerve drive can overcome these shortfalls and create a drive system with great power and speed. But your average Swerve will boast a decent level of speed and power, not a wealth of either.

Edit: And, yes, as many are mentioning, Swerve is very prone to failure. And recovering from a failure in a Swerve module is no easy task.

Edit-Edit: I apologize for my bad terming. I’m being quite unclear about a lot of things. Apologies, again.

Disclaimer: I’m no expert. Don’t take my word as law.

Many many more possible points of failure.

It’s also likely to be devastated by a single part failure. If a module freezes up for any reason (even if it’s perfect, there’s a chance of damage) then the drive is almost useless.

Not that any other drivetrain is free of such problems.

[mecanum_troll] Mecanum wheels can drive in circles around swerve [/mecanum_troll]

While you say that swerve does not have the same “power” as a 6 wheel, I’m not sure what you are referring to, as they will have the same power, assuming they have the same traction.

These points aren’t necessarily 100% accurate. Both of these will vary with the implementation of the swerve and the drive it’s being compared to.

Pushing power is defined by the robots over all weight and wheel configuration, being a swerve doesn’t really change much here - other than the fact that a ‘unicorn swerve’ (or any swerve with independently powered modules like 111’s) has the issue of losing pushing power if one or more of the wheels lose the ground as the motor’s power is wasted and the weight on the driven wheels will now be less - this is an issue that any drive with independent wheels can encounter.

For examples of fast swerves other than 973’s Emperor Swerve, reference either 16 (In any year, but 2008 or 2011 would be best) or 148 (2008).

A crab drive will push with as much torque as you gear it to. Just like any other drivetrain. A crab drive will be just as fast as you gear it. Just like any other drivetrain.

Drawbacks to swerve include the interior space required, reliability, and serviceability. In addition, swerves can struggle to handle rough terrain as easily as 6 wheel drives.

Of course that can all be overcome by engineering. There were swerves that were quiet effective at going over the bump in '10, and swerves that have had lifting modules to go up a “stair”.

The big disadvantage of swerve is the amount of time you are going to take away from working on game specific facets in any given year. Anything other swerve disadvantage (again ignoring the complexity and difficulty in making it) is really just an engineering challenge that can be overcome.

I don’t think you can really ask what is better without taking complexity and manufacturing difficulty into account.

In addition to this- you (leland) keep saying “your standard swerve drive” but what are you defining that as?

The pushing thing can be whatever depending on the drive gearbox just as any other drivetrain can be. 2 speed solves all problems you mentioned.

How many of you guys have built and tested a swerve and compared it to a six wheel?

*Most swerves have one motor per wheel assembly. This means that when they are being pushed and are forced up on 2 wheels they have half the power to push back.

*118 had some wicked swerves specifically 2005 and 2007 are well known for the most powerful wicked fast drives ever. However, it could not rotate so to compensate they put their entire manipulator assembly on a turret.

Apologies, I’m being quite unclear about these things right now.

By “Standard swerve”, I’m not referring to anything particular about the Swerve itself. I’m referring more to a Swerve from a team who hasn’t really mastered the Swerve yet (i.e. 16, 111, 118, etc.). Those teams have worked up good implementations to Swerve, and have made steps in overcoming the common short comings that go with it.

Again, I apologies. I’m being very unclear.

Not a problem, your unclear term usage inspired quite a bit of good discussion.

Also, it’s worth mentioning that a team could build a ‘fast’ (Relative term), ‘powerful’ (Relative term), swerve using proven COTS components from AM and Team221 LLC - they’ve done the hard mechanical work for a team looking for swerve performance without swerve machining resources.

Doesn’t address the software though…

2079 has kind of done this. We didn’t really test any quantitative data, but we have run both in comparison during driver tryouts. The general sentiment from our drivers was that swerve is not as intuitive to drive (it does depend on the programming, but the programming is something that is really hard to get right).

And just from experience of doing both, 2079’s 2nd gen swerve broke with much higher frequency than our 1st gen 6WD. Swerve requires a lot more fine tuning before you get it “right”.

All 6 wheel drives have a zero turn radius (or at least most). Only well designed swerve drives can do this.

Both designs are good if implemented well, swerve just seems to require more effort to make it work well.

I’d say one big disadvantage is that it isn’t very practical to implement a two speed drivetrain with swerve. You’d need to pick one multi purpose speed and gear for that, making your robot prone to being pushed around by a 2-speed 6WD bot.

Well, if we’re ignoring the weight and manufacturing factors as instructed in the first post… :rolleyes:

What is impractical about making a multi-speed swerve?

They eat a larger chunk of the weight budget than a normal tank drive.

Only effective if you have a good driver who can capitalize on the directional freedom that a swerve offers.