pic: 4WD Concept



I’ve been messing around with this idea since I read about Winnovation using a similar concept drive-train. My team doesn’t have a lot of sheet metal resources or knowledge to fabricate sheet metal drive-trains. I used cantilevered 4WD in FTC with 2 skid wheels and 2 omni wheels before, but each wheel was connected to it’s own motor.

Using the limited knowledge I’ve gained about West Coast Drive this summer. I designed a drive base that is essentially a 6WD chassis without the middle wheels. It’s a 28"x28" frame but the wheel configuration could easily suit a long base frame as well. What I’d like some feedback on is how I’ve placed my 6CIM gearbox in relation to the wheels. I’m using a 221 systems style tensioner blocks to do chain tensioning. The picture is missing a few elements like bearing plates, but the overall concept is there. What do you think?

It would make sense to direct drive the traction wheels with the gearbox.

Personally, I would run a 6WD here. The 4WD worked for Winnovation because of the game and the fact that their driver is magic. For most drivers, this kind of drive can be harder to maneuver and score with, and if the game were something more precise (like 2010-2012), it might not have worked as effectively.

Regardless, running a WCD, you should probably direct drive one of the wheels (probably the traction one), then chain/belt in the omnis.

First thing I would be tempted to suggest is direct driving the back wheel. You do have to make an asymmetrical gearbox to get everything to package nicely but it saves you a few hex bearings and a shaft in weight. The other benefit of this is that your traction wheel will always be powered even in the event of chain failure, and more of your robot’s weight is over the traction wheels.

The solid bearing blocks inside the tube like that really shorten the distance between your two bearings, resulting in a less well supported shaft. I would consider a tensioner free belt drive (no bearing blocks) or some other kind of bearing block like the VersaBlock for adequate shaft support.

If you’re willing to step up to 3" tube and shim your bearings out a little you could pop a 24T 15mm wide pulley in the tube to power your omni. This is how we ran a 4WD in 2013. A 27T 9mm wide belt might also work.

Looks good, these are fairly nitpicky suggestions and it largely comes down to preference.

4WDs like this are indeed an option, just know what the trade off is. You’re sacrificing a good deal of traction and a centered turning point for better performance against T-bones and a turning point about one end of the robot. T-Bones were a big concern this year with all the defense but they are not always a big problem. It’s not inherently harder to drive than a 6WD, just different.

This drivetrain was brilliant pretty much for the sole reason that Winnovation’s driver is magical. I don’t know if any other team could pull it off.

I’ve driven a 4WD myself (just like this) and it just takes a bit of practice. It is different, but it doesn’t give a disadvantage in my mind. I’ve programmed plenty a successful autonomous program for this style drive. Magic drivers help, but this drive doesn’t require magic.

I’ll look into the possibility of remaking the gearbox to direct drive the traction wheel, that makes a good bit of sense.

My current mentor for the summer is interested in making a tensioner-free drive as a prototype to see how well it performs and how easy the maintenance would be on it. I’ve also never run chain-in-tube before either and I’d be interested to explore that option as well.

If the open field disappears and I need something more accurate I would obviously consider the potential of another style of drive. I just want to experiment with the simplest possible drives for a given problem/scenario. Say another 2013 or 2014 style game returns in the near future.

Additionally, if the game requires you to sit a few seconds to score with defense (such as 06* and 07), you can pretty easily be spun around. We tried to take advantage of that against 1625 at Midwest but it takes them just a split second to get the truss pass off.

*I believe 1625 ran a similar drive in 2006 but added a drop down breaks for the championship to prevent this.

We ran a tensioner free belt drive this year - indeed , we didn’t even have bearing blocks - just sat flanged bearings in the tube wall. It worked fine, but we did have to add idlers to the back belts to compensate for a belt-sizing issue.

In the future we’ll probably run ansi-25 chain and sliding idlers (sprockets with shafts that sit in a milled slot).

As the main mentor who helped with the design of last years drive train, here’s what I have to say.

I really like how easy this thing is to build and it isn’t super hard to drive in my opinion (although I have driven some difficult robots in my day). I would recommend trying to keep your center of gravity shifted a little bit back, this is another good reason to direct drive the rear axle.

As far as pushing goes, I was really surprised at how much grip the omnis actually have. Also, because most 6wd only have 4 wheels on the ground at a time, the 4wd can go toe-to-toe with many 6wd’s in a pushing match.

I really like 6 CIMs on the drive. I have looked a lot at the numbers, which seem to say it doesn’t make a huge difference. But I think it’s worth it.

These are just my opinions, feel free to prove me wrong.

The bearings are spaced plenty far apart. Most teams that run 2 piece blocks (like 254) are counter boring the bearings anyway, yielding the same bearing spacing (or only .125" more).

The rigidity and concentrically advantage over most 2 piece blocks (as most 2 piece blocks don’t adequately align to each other) is nice as well.

You don’t even need bearing blocks-- The robonauts didn’t use any this year, they just pressed their bearings straight into the frame rail.

As time goes on, our drives get more and more interesting. We started with cantilevered wheels, then aluminum sprockets and shafts, and now, we’re starting to ditch the bearing blocks.

In my research of the 118 press-bearing system they still had some method of tensioning the chain if needed. I’m not a huge mechanical guy. My background is mainly in electrical and programming, but I’ve started learning this stuff to help my team design better. How would one make sure the chain is tensioned properly without the use of tensioner blocks…I would assume some pretty accurate calculations have to be made.

This past year we used this calculator produced by 1640 to space our wheels apart using 35 chain. We CNCed our drive plates, assembled, and haven’t touched the chain since. It has stretched a little over the course of a long season with 6 events but it still runs great. We had a plan for tensioning if needed using a round piece of delrin with an off center drilled hole for mounting so we could rotate the piece around until proper tension was re-achieved and tighten the mounting bolt.

However I would assume since your single chain run is nearly the entire length of the robot (assuming you direct drive the traction wheel) the chain slack would be more noticeable compared to teams running an 8wd with C-C around 8in as you have more links to stretch out. A bearing block system might be more appropriate but you can use the chain calculator to help in design so your bearing block starts where you need it and tension from there.

For 1:1 ratios, the math becomes dead simple. For Chain, the c-c must be a multiple of the chain pitch (.25 for #25, .375 for #35). It can be a multiple of the half pitch if you’re willing to use a half link.

For belt, it must be a multiple of the half pitch (BUT don’t assume all belts are sold!). Required toothcount on the belt (B) is;

B = T + 2*C-C/P

T = Pulley toothcount
C-C = Center to center (in same units as P!!!)
P = Pitch (in same units as C-C).

For both chain and belt there are arguments for adding a fudge factor to the c-c. I generally don’t for shorter distances, but will more commonly do so for longer distances (or for small pitch belts).

My team used 4WD set up like yours last year, and to quote our driver, “it turned like an old lady”. It was poor in situations with heavy defense (glad we didn’t do it this year). If you’re going to go 4WD, I’d suggest to go 4 omnis or even mecanum. Other than that, I’d recommend sticking with 6WD

I didn’t realize these were COTS bearing blocks when I made the post, which increases my confidence in them a lot.

That said, since the bearings are hex, I think there’s at least an argument to be made to be paranoid about how heavily the bearings are loaded. This year, our bearings were spaced 1.75" apart (just happened to be, we didn’t design around this problem) and we had no hex bearing failures. Other teams with bearings 1" apart or less had problems with some hex bearings exploding under regular drivetrain loads. Hopefully the manufacturing problems with hex bearings are cleared up for 2015 and we don’t have to think about out of spec COTS parts when designing, but if you want to be paranoid it’s something to consider.

An easier solution than going with a different bearing block would be to just turn that part of the shaft round and use round bearings (less prone to failure), or better yet use a dead axle bearing block (also available from 221) since your chain is on the same side as the wheels. But this might not even be a problem next year, so…

What Wheels were you using and where was you CG located?
We had no problems whatsoever with turning.

Our team ran something similar to this in 2010 (by accident…) but it actually turned out working very well. It allowed us to pivot more around the ball that we had possession of in the front of our robot. We had one of the best acquisition units that year at FLR and we could keep control of the ball in almost any situation.

I honestly don’t know why this would occur. The turning scrub of this drive-train would be noticeably less than that of a 6WD.

I didn’t realize these were COTS bearing blocks when I made the post, which increases my confidence in them a lot.

That said, since the bearings are hex, I think there’s at least an argument to be made to be paranoid about how heavily the bearings are loaded. This year, our bearings were spaced 1.75" apart (just happened to be, we didn’t design around this problem) and we had no hex bearing failures. Other teams with bearings 1" apart or less had problems with some hex bearings exploding under regular drivetrain loads. Hopefully the manufacturing problems with hex bearings are cleared up for 2015 and we don’t have to think about out of spec COTS parts when designing, but if you want to be paranoid it’s something to consider.

An easier solution than going with a different bearing block would be to just turn that part of the shaft round and use round bearings (less prone to failure), or better yet use a dead axle bearing block (also available from 221) since your chain is on the same side as the wheels. But this might not even be a problem next year, so…

These bearing blocks though designed to work similarly to 221 systems blocks are in fact somewhat custom. 4451 and I CAD up these versions ourselves and then CNC milled them ourselves. They used this version on their 2014 6WD drive-train without issue. When you say the bearings were spaced (distance) apart, are you referring to the distance between the two bearings inside the block?

I was referring to the distance between the bearings, as that determines how well the shaft is supported and how loaded the outer bearing is, but it seems like that’s not a very big issue, especially if you’ve successfully done it before.