This is an idea for a drivetrain that I’ve been kicking around for a little while now. I was surprised to not find anything like it on Chief or anywhere else. The Versablocks are all oriented upwards, so there is no center drop. The omni wheels should allow easy turning and also increase traction when moving straight back and forth. No drop = no rock from acceleration, so bumpers on this drivetrain will always be lower than an opposing robot’s, allowing it to “steal” traction in a pushing match using the other robot’s weight. All in all, I think this design would perform exceptionally well in playing defence and harassing opposing robots with it’s pushing strength. I’m interested to hear any feedback available.
That looks great! How heavy is it, and about how much would it cost?
Weight is about 38lb without chain (I haven’t decided weather this would need #35 chain or it 25 would be sufficient). I think cost would be around $1000 with the 3 CIM gearboxes (haven’t done a precise calculation).
Great, now I’m thinking about lawyering the rules to use differently-sized pool noodles which make wedged bumpers…
2016, R21.C
Blue box from that rule:
All pool noodles used on a ROBOT must be the same in order to
maintain the desired interaction between ROBOTs in the cases of
BUMPER-to-BUMPER contact. BUMPERS containing pool noodles
of vastly different construction may cause a “ramp” effect when
interacting with other BUMPERS.
Does that 38 lbs include the battery? 38 sounds a little heavy, but that may just be the 8 wheels and the relatively heavy gearboxes and motors. I would suggest you look at JVN’s design calculator to figure out your gear ratios. You’re designing this drivetrain to win pushing matches, so you’ll need to make sure your low gear is low enough that your motors don’t burn out.
Also, you may want to push the Colsons on each side out to the front and back of the chassis. Your inner wheels are quite close together, and while the Omni’s will cause some turning scrub, you may find this still turns too quickly for your driver’s comfort.
If you want to drop your price point on this, you could look at a different (custom?) gearbox. You could also consider switching to smaller wheels for the purpose of using the 3.25" Omni’s that already have the hex bore. This will save about $12 per wheel since you don’t need VersaHubs. There are other small things here and there you can do to lower the cost without increasing the need for machining much.
Also, I see no reason not to use #25 chain.
You might consider looking at this post, it’s for VEX but concepts apply.
Not to discourage you, it’s great to see people looking and figuring things out for themselves, but most combinations of “what happens if I put X wheel in Y position” have been pretty thoroughly looked at for tank drivetrains. You are just left with a bunch of trade-offs, and the magic is in selecting the right drivetrain for what you need it to do, not necessarily trying to find a magic drivetrain that is far superior in all cases.
In this case, let’s compare to a 4 wheel omni (basically take out your center wheels), an 8 wheel omni, and an 8 wheel tank.
Increasing the number of wheels is usually a factor of what you need to traverse. More wheels reduces the amount of space between the wheels, allowing it to claw over. Additionally, by staggering the drop of the wheels, such as in a 6 wheel drop center (or 8… 10… 12… whatever), you can trade off stability (rocking) for reducing turning scrub. Adding more wheels can also give you a little more “middle stability” in the neutral state.
An omni wheel already has pretty good traction in the forward direction, and can push fairly well. What it doesn’t have is any resistance to lateral movement. If you look at a 4 wheel omni drive (like the 2014 JVN buildBlitz, or any number of other ones), it almost drives like a drifting car. With drive practice, you can do some interesting things to throw it around… but you are also at a higher risk of being spun.
So to your drivetrain, you have a risk if the omni wheels are preferentially weighted (design tolerances, weight shift under acceleration) of acting more like a 4 wheel omni drive. You probably don’t have substantially more pushing power than an 8 wheel omni drive, but you have a little more resistance to being pushed sideways… while still more at risk of being spun. So basically, you have a set of tradeoffs that gives you some of the benefits of a 4 wheel omni and some of the benefits of an 8 wheel drop center tank, some of the disadvantages of either, etc. At the risk of beating this horse dead, I just want to emphasize that you don’t get the advantages of both as well as dodging the disadvantages of both, it is just a blend.
That being said, depending on the year, it might all be a valid set of trade-offs and lead to a pretty solid drive train. Build it and see what you think 
Edit: Also as a note, I’m not sure how significant the “lower bumper” comment is. Most drop drive trains (outside of pneumatic wheel years like last year) are on the order of 1/16" to 1/8". When you are talking about a several inch bumper contact patch, I think the drop center difference will be lost in the noise when compared to other variables (bumper construction, driver practice) when you talk about defensive play. (just imho)
I think you will have a hard time trying to find a 3.25inch Colson, let alone a 1/2 inch hex bore one.
This isn’t exactly new; this wasn’t an uncommon layout in 2010 and later when 8WDs started gaining traction.
In short these have similar results to a 2 traction 4 omni setup vs a 6WD. Most of the problems “solved” by this drive are minuscule / imaginary. Rock shouldnt be big enough to be the difference maker in a bumper to bumper pushing match; we are literally talking about a 1/16th inch difference in bumper height here. The omni wheels lower resistance to being spun more than the 8WD all traction designs do given the same wheel layout. Omnis also don’t match the lateral traction of Colsons or traded traction wheels, so you’re giving up a little there.
Basically, you need a really good reason to not have a drop center for this to be the right call. Given that an unequal spaced 8WD already doesn’t do much rocking at all, it’s hard to find a niche for this drive.
This isn’t exactly new; this wasn’t an uncommon layout in 2010 and later when 8WDs started gaining traction.
Do hou have any pictures of 8 wheel drives like this? I haven’t been able to find any.
Regarding your concern about lateral traction and being pushed or spun, I don’t see how this would be any more at risk. An 8 colson wheel drivetrain with center drop will only ever have 4 colsons giving traction at any one time, just like this drivetrain. So why is this more susceptible to spins?
A “proper” setup for an 8-wheel with center drop will often have the outer 4 skimming the carpet. There’s considerable sideways “traction” available if you’re hit–more particularly if the impact rocks you “onto” the end being hit. With omnis, you have zero sideways traction available in that situation. It’s not that they aren’t skimming the carpet, it’s that they’re omnis. They’re designed to allow free sideways motion.
Of course, maybe you shouldn’t just trust my word for it. Weight up a proto-frame set up for 8WD (you can just have the 8 wheels on a frame) in one of your favorite configurations–you want 150 lb or some considerable fraction of that weight. Try to turn it by pushing it, hitting it, etc. Swap wheels and repeat. Quantitative data may be harder to get than qualitative, but you should be able to get an idea of how the robot will try to act.
1986 used a drive like this in 2014. A video of just the drivetrain is located here. There are certainly many other examples, but this is the one that comes to my mind.
Mounting your bumpers as low as possible and limited rock is an excellent way to gain traction in a pushing match, and something 558 does as well. Where we differ is that if we are designing for maximum pushing force we believe that any wheel that touches the ground should be a traction wheel. Understand that any wheel that is in contact with the ground is providing traction, and increasing the number of contact points with the ground reduces the normal force on each wheel. Typically 558 will design an 8wd with a larger center to center distance between the middle wheels to increase scrub and make the robot both more stable at speed, and more resistant to spinning.
In basic terms, an 8wd robot with drop center will provide more pushing force than an 8wd with omnis on the outside all other things equal.
If accurate, 38 pounds is actually on the lighter side for a 6 CIM shifting drivetrain with wide traction wheels.
By placing omni wheels in the front and rear, you are creating a drive train with what amounts to a very short wheelbase and very wide track.
This will make turning very easy, almost too easy. It may turn out to be a bit squirly to drive.
You may want to consider adding a gyro to assist with driving straight.
It may be beneficial to consider what the normal force on each wheel would be and the behaviour of your chassis as it drives over a floor that is not perfectly flat. What happens if the four high-traction wheels in the middle have a slightly smaller diameter than the omni-wheels?
Today (https://twitter.com/MechaDojo/status/784923996345077761) on Twitter which basically looks a 6-wheel version of OP’s model.
The resistance your robot is going to be able to give is related to the amount of friction your robot has to resist that force. When it comes to lateral force, it’s obvious that an omni wheel is going to present negligible friction compared to a traction wheel. Even ignoring the wheel placement for the moment (the outboard wheels will have a greater lever arm to your center of mass for resisting the applied moment), you can determine the friction from your wheels with the equation f=μN. For our purposes here, the normal force is equivalent to the weight placed on each wheel.
When the quantity of wheels in contact with the ground is increased, the normal force on each wheel is decreased (less weight on each wheel). In the case of a 8WD with 4 “drop center” wheels, you’d have the weight of the robot on 4 wheels. 100% of the robots weight would then be placed onto those 4 traction wheels. In your corner omni 8WD, the weight of the robot is distributed among 4 traction wheels and 4 omni wheels. Assuming an even distribution of weight, you’d have 50% of the weight of the robot placed on wheels with a high coefficient of friction, and 50% of the weight of the robot placed on wheels with a low coefficient of friction. As a result, the total friction your robot generates to resist that lateral force would be less than a robot with all of its weight placed upon high traction wheels.
Now, we’ve made quite a few assumptions to reach this point, and many things will end up being far more complicated in reality than I’ve presented here. For instance, resisting a spinning moment is going to be very dependent on wheel placement, drop height, and frame interactions. But I wanted to illustrate a general point. To phrase that point differently, the advantages your design has in terms of turning itself easily also serve to make it easier for other outside forces to turn. None of this is to say your design is poor, just that it will behave differently than a drop center drive. In somecases, team’s have taken advantage of ultra-low resistance to turning and incorporated it into how they wanted their robot to behave.
Thanks for sharing your design! CD needs more of this.
Some quick comments. Unless you adjust the spacers on the Ballshifter third stage I don’t believe there is enough room for #35 chain. Remember you want to have clearance on both sides of the sprocket for the chain.
With the 8wd if you switch your chain routes it will allow you to sneak the outer chains in a little making the shafts shorter and save a little real estate in your bellypan.
It looks really solid. Nice work!
I’ve seen this done before, and it provides little to no real world advantage - ignoring the fact that this is illegal. Bumpers don’t seem to be rigid enough to act as an effective wedge, which seems to lead to minimal weight transfer.
As others have noted before, low (literally, the bottom of the bumper zone) bumpers are pretty much all you need if you want to build a drivetrain that’s effective at holding it’s ground and/or moving objects.
If you really want to have fun, you can start playing with the center wheel spacing relative to the outer wheels, a larger center wheel base will lead to a drive train that has more of a tendency to drive straight while a shorter wheel base will lead to better agility at the expense of straight tracking.
I personally am an advocate for minimal drop on 8 wheel drives. The last two I had direct involvement with had .090" drop (2013) and .060" drop (2014). The difference in drop were due to different strategic objectives and robot configurations but each was somewhat optimized for the role. The 2013 drive train was optimized to sit on the back 6 wheels when in shooting position to provide a stable platform and while in transit position it rocked forward to “tip” the whole robot back for more stable driving and a slightly angled leading edge, helping to get “under” other robots. 2014 was designed to “Squat” when pushing, digging all 8 wheels into the carpet while also providing good stability at speed.
A couple things to remember when chasing high traction performance:
-Wheels sink into the carpet, sometimes as much as 1/8". A drive with minimal drop may not ever have it’s wheels leave the carpet, nor does it truly rock, instead the force on each wheel changes depending on the conditions.
-Traction on carpet is a bit more complicated than F=CoF x Fn. There are a series of factors such as carpet wear, tread geometry and tread hardness to consider as some wheels actually “Dig” into the carpet, creating additional mechanical forces that appear to increase traction.
-When working with 6 CIM drive trains, you have to be extremely conscious of current draw, especially when pushing. Once you’re pushing another robot, if it’s weight begins to transfer to your machine, suddenly a traction limited machine may no longer be traction limited, which usually doesn’t end well.