Uh Oh - How to Fix Driving Configuration?

Hey guys!

So a group of us is working on a project to build an all-terrain cart. We had one guy doing a lot of the CAD and manufacturing and we unfortunately found out that our robot cannot turn left or right (kinda frustrating because we mentioned the robot needed to be designed with the front and back wheels closer to turn better).

My question to you all is if you can think of any fixes we can make in a short amount of time (less than a week). The main issue is that we are using ebike wheels with tires that have a lot of friction and make it difficult for steering.

Some thoughts that come to mind would just be having some sort of cover over the tires to create less friction, maybe drilling new holes to mount the front and back wheels closer to each other, or possibly adding two more wheels to the assembly in the middle to have a “drop down” wheel to get a smaller turning radius.

Here is a photo of the current drivetrain (also attached):
https://imgur.com/a/0nzsCkJ

Any feedback would be appreciated, you guys are always so helpful!





How about turning one of the wheel assemblies around? Put the post where the shock/spring is, and the shock/spring where the post is. That’ll move the wheels much closer together, making it easier to turn, and probably wouldn’t be too much work. You would have to keep an eye on load-balancing to make sure it doesn’t tip over, though.

The better, though more time-intensive, solution would be to convert one set of wheels to casters. That’ll make it much easier to turn, working like a shopping cart does.

Can you elaborate on what your requirements are for “all-terrain”? Because a cover over the wheels idea could work on concrete or carpet, but the wheels will likely sink into the ground on grass or dirt and render and friction mods (And possibly extra wheel options) far less effective.

You could always put a cover on just 2 wheels if turning on a dime isn’t a requirement. Still get friction from the non-covered wheels that way.

If turning on a dime is a requirement, however; you may be able to do something similar to an Omni wheel and arrange strips of (insert covering of choice) around the diameter of the wheels that way you have cover and no cover at the same time for traction purposes.

You describe an “all terrain” cart - what is the actual terrain involved? You might find that on the actual terrain, you have fewer wheels on the ground or break friction more easily (gravel or sand), or conversely that the wheels sink in and you need to reorient the wheels to roll instead of skid - such as replacing powered wheels with casters like Jon mentioned.

This sentence is really concerning. It sounds like you’re blaming the fabricator for not listening to something that was “mentioned” and then speculating on a fix “if only they listened”, rather than looking for the real root cause in the system design process.

Why should the fabricator have to remember something that was “mentioned” - did you have any kind of design document or specification that they could refer back to check that the design met all the criteria? Was wheel spacing in the specification?

As for your speculative fix - is “turn better” the same as “able to turn”? How did the team check that it would be possible to skid-steer on these wheels, before completing and building the design? Is there a specification (“X times longer wide than long”, “maximum Y inches apart”, something like that) that the fabricator could meet to make sure that the design works?
What if changing the wheel spacing to help it turn makes the cart tip over?

I get that it’s frustrating that the design doesn’t work, that’s normal - what you need to learn and grow forward from is to take that team frustration, take ownership of the “miss” as a team, and fix both symptom & root cause together, rather than claim “oH if HE just LIstENeD*, we wouldn’t be in this mess!”

Which is super hard to do, actual real-life leaders don’t always successfully do this. Do it! I believe in you! :slight_smile:

*(I’ve translated your paranthetical from Business English to Internet English here. Yes, you sound this bratty.)

That’s definitely an idea, and yeah we probably would have to watch out for tipping.

And it will just be going on some flat dirt road/gravel most likely, it’s also more of a demo robot for a project and doesn’t have to be perfect.

Have you tried turning it on a dirt/gravel road? or only in your shop, on concrete?

Can you get away with radiused turns instead of turning in place? That usually requires less energy.

These are some great points, really appreciate the feedback. To clarify we can’t really turn in place lol. Our group has been operating where we each have been just working on subsystems independently (CAD, programming, electronics, etc) and meeting every once in a while to check in on each other. I have been doing mostly electronics, but from what I’m aware of there were no turning or spacing specifications :frowning: . I think for big projects it’s always a bit hard to consider everything but maybe having the specifications written out in the future for a design could help.

No, we need to do that still haha. I imagine we would be able to turn a little better if we tested it on a dirt/gravel road.

We may be able to, it was being super grippy and difficult to do this on the tiles in our workshop though

What do you plan having in the cart? Even with closer wheel spacing, having a lot of weight in there could still make it really hard to turn.

It’s more of like a self-driving thing so won’t be carrying much. Main load will probably be lead-acid batteries, motor controllers, and a few small electronics.

The classic FRC approach would be to run a zip tie around the rubber of the wheel, between the spokes. Idea being that the ziptie has more physical interaction in the direction the wheel is rotating and less in the direction you need it to slip. Its a pretty quick thing to try at least.

But from how you make it sound, the motors may simply be geared poorly for this application and it wont make enough of a difference. In which case you may need to look into more drastic changes.

OP’s problem aside, I hope you aren’t doing this on your FRC robots. R06 says:

Ziptie heads sounds to me like hard plastic studs/cleats.

Heads didn’t go towards the carpet. Just the smooth side.
And thankfully havent had to do this on our team ever, just to get older kitbots to turn when their first time on carpet was at a competition. Think pre toughbox days.

+1.

If you want it to work, you really need to fix it right. Castering one pair of wheels is the right answer.

It actually does not look like it would be that hard to modify what you have. It looks like the wheel “fork” is made up of two side plates that are bolted on to the swing arm and shock mount plate. You would need to remove that and probably modify the square tubing on the swing arm to make it a little shorter and then add a new plate that ties the swing arm to the shock.

Then you will need a new plate or channel to bolt the side plates to rebuilding the fork as a separate assembly.

Then you will need to attach the new fork assembly to the swing arm assembly with a thrust bearing and a bolt in the center. You could probably find a suitable thrust bearing at a big box hardware store (you may need to be willing to cannibalize a COTS caster wheel assembly).

The idea of turning one module around for shorter wheel base should work - Less than 2:1 is a rough rule of thumb (wheelbase Width to length) and then to mitigate tipping over put a set of small diameter (6 inch or 4) omni wheels - unpowered - out front (or back) - the 2009 plastic wheels can works as well of you have some of them.
or you need to have an actuator to steer the front wheels - (Ackerman style) but more complicated.

This looks fast with 4 e bike wheels! (cool).