Live Axle vs Dead Axle

[OWilliamson]

Our team is planning to build an off-season drivetrain, but we are not entirely sure on what we want to do yet.

We were wondering what teams did, and if there were different advantages for either live axles or dead axles. Or does it just come down to what type of drivetrain you are creating (WCD, Sheet Metal, KOP etc).

[MechEng83]

Our decision to use live or dead axles usually depends on if there is a sprocket driving the wheels.

Our 2013 robot was mecanum which we direct drove each wheel out of the gear box, so it was a live axle.
Our 2012 robot was octocanum (4 mec, 4 traction, switch between) and those were dead axles, because each wheel was driven by a chain and sprocket.
Our 2011 was 6 wheel drive w/ 2 center traction and 4 corner Omni's. That robot had the middle wheels direct driven live axles, with chain going to the corner wheels that were dead axles.

[pfreivald]

The past several years, we've used direct-drive live axles. This year is the first that we used hex drives (and hex bearings, which we actually happened to get enough of), and we're very happy with it.

[nathan_hui]

It really depends on the final gearing that you want. Live axle usually means faster, less torque, and less backlash. Dead axle usually means slower, more torque, and more backlash. The type of drivetrain usually doesn't directly drive the selection of live axle vs. dead axle - the type of drivetrain combined with the chassis design combined with the final expected performance does.

If you could tell us a bit more about the drivetrain you want to build, we can give you more advice. Otherwise, there's really not much we can tell you. Each design is specific to its application.

[OWilliamson]

Although my team hasn't discussed what we want in too much detail, we were planning on making it 6WD with a dropped center that was fast and fairly agile.

Originally Posted by nathan_hui

Live axle usually means faster, less torque, and less backlash. Dead axle usually means slower, more torque, and more backlash.

Based off of what you are saying, in this instance would it be a better idea to create it using live axles?

[orangemoore]

I have to ask, what other type of axle is there?

[OWilliamson]

The only reason I put other is because I remember seeing a picture of a team (191?) who had, although a very long time ago, used balls as the back wheels for a test robot. I can't find the picture right now, but it appeared that there weren't any axles that the ball was attached to. I figured that I shouldn't leave some team out if they went with something that was a little unusual like that

[Koko Ed]

Originally Posted by OWilliamson

The only reason I put other is because I remember seeing a picture of a team (191?) who had, although a very long time ago, used balls as the back wheels for a test robot. I can't find the picture right now, but it appeared that there weren't any axles that the ball was attached to. I figured that I shouldn't leave some team out if they went with something that was a little unusual like that

That was team 45 the Technokats.

[the.miler]

It depends a lot on what sort of resources you have. We have easy access to a Tormach pCNC 1100 in our room on campus, and less easy access to a sheet metal shop with laser cutting capability. The choice is simple; we do a cantilevered style live axle drive train with chain tensioning that works by swapping in different sized idler sprockets between our wheels. We CNC the side members of our drivetrain and the bumper standoffs out of tube (2"x1" generally, sometimes 1.5"x1" if we can get away with it).



Since reliability is, above all else, most important to us, we run direct drive to the wheels. One more chain run is one more chain run to tension, and one additional point of failure that could cripple an entire side of the drivetrain. We achieved our desired reduction entirely through the design of custom gearboxes and machining the side plates on our CNC (not a difficult endeavor with an in house machine). 16 ft/sec served us well this year.

If you have access to sheet metal resources, I would consider making something sweet and lightweight like Team 33's WASP drive; the use of the bumper rules to allow a lighter weight frame but heavier duty bumpers is very cool.

[iyermihir]

Originally Posted by nathan_hui

It really depends on the final gearing that you want. Live axle usually means faster, less torque, and less backlash. Dead axle usually means slower, more torque, and more backlash. The type of drivetrain usually doesn't directly drive the selection of live axle vs. dead axle - the type of drivetrain combined with the chassis design combined with the final expected performance does.

How can a live vs. a dead axle change your speed, torque, or backlash?
The power transmission method would be the same. The only difference would be what the sprocket/timing belt pulley is connected to which does not affect any of those things.

-Mihir Iyer

[Nathan Streeter]

I don't think there's a general "one is better than the other."

I typically prefer dead axles, as it typically allows you to use a smaller shaft (since it doesn't need to carry the torque), avoids the use of keystock, and generally makes it easier to avoid set screws.

That said, which one we use on 1519 generally comes down to what our wheels are and how they're designed to mount to an axle. Most of the AndyMark wheels are designed with that nice standardized 1.875" bolt circle, so it's easy to bolt on a sprocket and have it be dead axle. The Skyway wheels which we used before usually only had a keyway, so we typically used live axles. To further drive the decision one way or another, we've recently used primarily timing belts... our pulleys are primarily for live axles though.

So really, I prefer dead axles for reasons like it often giving you more freedom on the shaft you use (a shoulder bolt can even be a great way to add structure to your drive channel), but 1519's typically made the choice more based off wheel and sprocket/pulley availability. With more machine access you generally have a bit more freedom in dead vs live axle.

[mman1506]

I think one advantage of a live axle is they usually make a more rigid chassis . A sheet metal dead axle is usually a lighter setup.

[EricH]

Originally Posted by iyermihir

How can a live vs. a dead axle change your speed, torque, or backlash?
The power transmission method would be the same. The only difference would be what the sprocket/timing belt pulley is connected to which does not affect any of those things.

-Mihir Iyer

Backlash COULD be increased by an extra chain/sprocket or belt run--most live-axle setups have a direct-drive off the gearbox. Dead axle cannot do that, and so must have a chain/belt run. If some stretching happens, then there will be some backlash--overall, however, it probably won't be noticed.

Speed and torque can be adjusted to be identical between live-axle and dead-axle. I think nathan was confusing axle types with direct drive and chain drive, which are by no means the same as live-axle and dead-axle--though it's rather difficult to have a direct-drive without a live-axle somewhere in the system, it's actually pretty trivial to have a chain drive go either live-axle or dead-axle depending on what you attach the sprocket to.



Oh, and the other drivetrain, with the balls: that would be team 45, TechnoKats, from sometime around 2003. I think that system saw use by another team, or two, at various points, but memory fails me. The balls had no axles; the front wheels were omni wheels if I recall correctly (and casters if I don't). All power went through the ball drive; each ball was driven by a pair of rollers directly mounted each to a motor. Vary the inputs to the motors, vary the direction of the robot.

[DampRobot]

The correct answer to these types of questions, of course, is that it depends. Here's my take on it, and the thinking that got me there.

Direct drive live axle is in my mind the best way to go a majority of the time. It's best suited to a drive like a WCD, where you direct drive one of the wheels, and transmit power 1:1 between them. Live axle typically lets you cantilever the wheels, which introduces some frame design benefits and allows you to swap out wheels more easily (I'm speaking generally here). It also means that you have your bearings somewhere other than the wheel, which can allow you to decrease the size of your wheel. In my mind, a live axle system is easier to repair because the wheel and the chain or belt path isn't between two frame members.

Generally speaking, live axle drives are better when the wheels are smaller (so you don't need that extra reduction in the chain/belt from the gearbox to the wheel). Dead axles work well where you are going to have the wheel inside the frame anyway, or where you have large wheels that need that extra reduction. There's also a perception that it's "easier" to build a dead axle system, as usually it doesn't require a team to machine bearing holes themselves (they're already in the wheels).

I recognize that my beliefs are probably neither totally correct nor totally typical, and in all honesty, this type of thing really comes down to "to each their own." I'd encourage you to build/design a WCD, as it's widely considered to be a very high quality DT (generally). Most important, though, is that you consider these types of design tradeoffs yourself, and make informed decisions accordingly, even if you don't happen to end up with the type of drivetrain like I described.

[MichaelBick]

It largely comes down to what kind of frames you build and what your final power transmission stage is. However, there are ways of getting around the disadvantages of either.

Dead axle is nice because the axle becomes a structural member. This is really helpful on sheet drives which can easily get crushed. In non cantilevered drives dead axle is usually easier to service. Dead axles also can be easier to machine, because you don't need to precisely machine bearing holes. However dead axles can be harder to tension without losing efficiency somewhere. A nice way of getting around this is by using C-C belts, though this may be above the capabilities of some teams.

Live axles tend to be better to use in cantilevered wheel drives. Live axles also have an easier time direct driving wheels, which makes it easier to make live axles lighter, more efficient, and more reliable. Live axles are harder to machine if you don't have the ability to precisely machine bearing holes, however are easier to efficiently tension.

Regardless, speed and torque have nothing to do with which axle type you choose and everything to do with your gearing. Dead axles don't have to have more backlash then live axles either. For example 971 uses gears to direct drive their center wheel from their 1 stage transmissions which allows them to have the same efficiency and backlash as any live axle team.