Wheel Placement on Robot

I would assume you would burn the higher speed motors as they will try to drag the lower speed motors forward.

I would assume you would burn the higher speed motors as they will try to drag the lower speed motors forward.

I was able to get 4 of the same motors. It took some time as the motors are not officially for sale, but parallax worked with me and sold me 4 new ones. I had 2 older ones, but they are slightly slower and you said I could not use two fast motors and two slow motors. So it took some time for me to convince them that I need 4 sets as they don't have many ready.

Think about it a bit: We are talking 4 wheels here. If you moved the wheels in the front back, and the ones in back forward, until they were near the center (but still spaced opposite each other), then yes, it would be easier to turn. Heck - if the wheels "merged", then you have the classic differential steered robot (add some casters on the front and back to balance).

I was thinking about putting one wheel/motor on each corner of the robot but you explained how that would screw up the turning.
What if I slid the wheels from the corner down towards the middle of the robot so that they were almost touching and then put a caster wheel on each end of the robot. It would be like putting two, three wheel robots back to back. That way the driver wheels are like .5in apart making the turning easy yet the caster wheels will help support the front/end.

Would that work? I think it would yet I have never seen a robot like that.

It wouldn't screw up the turning, you would just have some (maybe a lot) of wheel drag; I also don't know how well the center would remain "centered". Is there a reason to keep the center from moving much?

You could try it; it might work - but I fail to see how that would be better than just a single pair of wheels in the center, and a caster on the front and back?

A little belated response, but for a smallish robot, like shown in reply #7, I would imagine that motor/wheel arrangement will be fairly reliable. After all, Lynxmotion has been selling those things for years.

OTOH, if you're talking about 6" wheels and a 2' x 3' or so base, that's pretty heavy-duty, and you really need to go with strong motors and probably wheels mounted on shafts riding on bearings, and connected to the motors via pulleys or chains.
http://damencnc.com/en/components/mechanical-parts/ballbearings/279

Skid steering works, but as mentioned, the carriage needs to be extremely robust for something that size and weight, and they certainly work best on surfaces where the wheels can, in fact, "skid" easily (eg, not carpets).

It's very common for people to use casters as a 3rd wheel [tripod arrangement] for differential-steered bots, but you'll notice this thing is "tiny" compared to what's being talked about,

A large, heavy frame will tip over if not properly balanced.

Is there a reason to keep the center from moving much?

I wold like to keep the turning as accurate as possible.

you would just have some (maybe a lot) of wheel drag

that would make my turned less exact and decrease the life of the motors. (I'm mostly concerned with the turning) Right?

I fail to see how that would be better than just a single pair of wheels in the center, and a caster on the front and back?

That was my original plan, but two motors do not have enough torque to push the robot. It was cheeper to buy another two wheels/motors then to buy better ones. I assume that by adding two more motors/wheels I doubled my torque. Right?

wheels mounted on shafts riding on bearings

They are.

A large, heavy frame will tip over if not properly balanced.

If the wheels are towards the middle of the robot It will tip over, but the caster wheels will help balance it. Maybe tip is not the right word. Think of a seesaw. If the wheels are in the middle of the robot it will pivot and the back & front could hit the ground.

Thanks! Would my idea work? I can't test it out because I don't have the two caster wheels.

What I was referring to is, a heavy robot with tripod arrangement and only 1 caster can easily tip over.

What I was referring to is, a heavy robot with tripod arrangement and only 1 caster can easily tip over.

I agree! That's why I thought that a caster on each end would prevent it form tipping. Would my idea for a 6 wheel robot work?

That was my original plan, but two motors do not have enough torque to push the robot. It was cheeper to buy another two wheels/motors then to buy better ones. I assume that by adding two more motors/wheels I doubled my torque. Right?

You can add the torque indeed. But you are probably not so much interested in torque but more in "speed" and "acceleration speed"
You speed will certainly not double.
The acceleration speed will increase but it will not be double.
There are many things at play in a 4WD and engine loads in general. To be honest try and see is the best way to find out.

cr0sh his explanation about the square and powers for skit-steering is 100% correct if you only turn on the spot. If you also turn while moving forward (or backward) the square shape no longer has a benefit to the rectangle because in this case the centre of the turning circle is outside the ground plane of the robot.
The width of the contact surface of the wheel is very important. The wider the more friction the more axial forces the stronger the wheel bearing must be. In case of on the spot turning a square will make that the centre of the wheel is friction less and you have half of the friction on both sides of the wheel.

Best regards
Jantje

If you mean 4WD with the drive wheels close together, and 2 casters on the ends, that sounds doable, but I don't see a situation where that design will be at all useful. IE, why even bother with 4WD?

I think a lot depends upon where this robot will operate most of the time. If outdoors on earthen terrain, I don't think having casters is a good idea at all, as they will probably dig into the dirt a lot. For that, I would go with a basic design with 4 drive wheels on the corners of the robot, and relie on a very robust motor/wheel system to keep it from breaking down.

Or else, use 2WD with 2 extra "large size" idler wheels [much larger than casters] on the opposite end from the drives. This guy has some ideas,
http://www.amazon.com/Build-All-Terrain-Robot-Robotics-ebook/dp/B001UQ5HW4

For indoor use on smooth surfaces, a caster system will work, You might check this out,

If you mean 4WD with the drive wheels close together, and 2 casters on the ends, that sounds doable, but I don't see a situation where that design will be at all useful. IE, why even bother with 4WD?

Yes that is what I mean. The only reason I'm doing it that way is that I have 4 of the same motors. If I understand torque correctly using 4 drive wheels will double the torque the robot would have if It was using only 2 drive wheels. To push the robot up the needed incline (I don't remember what it was off the top of my head) two of the motors I have will not have enough torque. That is why I need four drive wheels.

I think a lot depends upon where this robot will operate most of the time. If outdoors on earthen terrain, I don't think having casters is a good idea at all, as they will probably dig into the dirt a lot. For that, I would go with a basic design with 4 drive wheels on the corners of the robot, and relie on a very robust motor/wheel system to keep it from breaking down.

The caster wheels will be fine (link below). Also, I was going to put the drive wheels towards the center to improve the turning. I don't believe the motors will break. I need the turns to be exact when its driving autonomously and I think this design will give me the needed torque and turning capability.

http://www.trossenrobotics.com/store/p/5881-Parallax-Caster-Wheel-Kit.aspx

This is more what I had in mind for idler wheels, although smaller than 10" :-),

Something like that would be fine. The price is also much cheaper :slight_smile: Like I said I don't have the caster wheels.

This one might work. http://www.amazon.com/Swivel-Caster-Pneumatic-Wheel-Brake/dp/B005UZ4FQA/ref=sr_1_14?ie=UTF8&qid=1379185406&sr=8-14&keywords=Pneumatic+Wheels+Casters

Before I buy them I need to know if this design work. Am I right that when calculating torque you add up the torque of all the motors thus 4 drive wheels will have 2x the torque of a 2 drive wheels?

Thanks!

For identical motors, that sounds about right. However, one thing you probably haven't totally conceptualized is that having a single idler wheel front and back is really back to the "tripod" geometry that I mentioned, and on both front and back. If the idler takes any substantial amount of weight, say in turning while going downhill, don't be surprised with a possible tumble.

95% of the weight of the robot will be in the center so I don't think it will tip. Thanks for your help. I will post photos once the robot is built as well as the results.

Build it and see how it works, and refine as necessay. That's how this stuff works. Bootstrapping and iterative development. What you're doing is basically a Bobcat loader with outlier wheels. You can also do a search on "shrimp wheeled robot", but that is for really rough terrain.

oric_dan:
...and they certainly work best on surfaces where the wheels can, in fact, "skid" easily (eg, not carpets).

Well - they'd work ok on carpet - the carpet just might not survive all that well...

oric_dan:
If outdoors on earthen terrain, I don't think having casters is a good idea at all, as they will probably dig into the dirt a lot.

Depending on the size of the robot, you could go with a couple of larger "off road" casters; they make these larger casters that have knobby tires for such purposes. Not ideal, but they do exist...

oric_dan:
This is more what I had in mind for idler wheels, although smaller than 10" :-),
http://www.harborfreight.com/10-inch-pneumatic-swivel-caster-38944.html

Doh! Just saw this comment - you already know about 'em! Heh! :stuck_out_tongue:

Before I buy them I need to know if this design work.

You probably need to be specific with the size and use of the bot. About all I've seen is it is to be 2'x3' in size, goes up an incline, and maybe make zero turns. Kind of vague description for anybody to tell you if anything will actually work.

If only life were so simple that we would know ahead of time that our designs would always work. Rod Brooks, head of the MIT Robotics Lab, was fond of saying "the problem with simulations is, they're always doomed to succeed". IOW, things usually tend to work in unexpected ways in the 'real' world. A priori designs and simulations are necessarily always incomplete.