Quad- 1/2 track drive system

[Ben Mitchell]

Dr. Johnson and everyone else,

What do you think of this drive system:

Robot oriented such that the 36" side is front, and 30" is side. This is to dampen effects of scrub.

A drive system consisting of 4 small tank treads. (perhaps 4-5 inches each), spaced apart to spread weight and traction evenly.


Each tread being drive by a motor (2 chiaphua, 2 drill),
Each small tread has a driven sprocket, and three idlers on the ground (resembling a triangle)

Each small tread has a center idler which is slightly lower than the other two, creating a "digging in" effect on each tread, gripping the carpet.

The treads themselves could be either brecoflex timing belts (high-end double sided with gripping pattern) , or custom tracks (a la Technokats).

4 motors could be replaced by the 2 chiaphua's.

Programming could be used to account for high amps, if we were rammed, and dull the ampage spike. (True or false??)

The pros of this drive: high traction, lower scrub (than full treads) Increased speed (than full treads), whuile still getting the traction that treads would produce.

What does anyone here think of this? I'm interested in getting opinions from the folks who know.

Would 4 motors cause a major battery/breaker problem?? Any ways to compensate??

Don't spare the criticism, or foresight to any problems you might expect to occur, I'm thinking about this design for next year (its a little late now). I thought I'd ask the folks who know what they're doing.

Thanks for your imput,

--Ben Mitchell

[ggoldman]

you can get jsut as much traction from four wheels as tread....this is because the weight distribution...if you distribute weight ona smaller area it has more traction....but f you spread it out..like on a tank tread...you decrease your traction...

I think i might be confusing with friction..but traction is friction...

Can an engineer check me on that..or am I way off:confused:

[jasoni]

I don't think you'll gain anything other than complexity by going with 4 smaller tracks instead of 2 tracks running the length of the machine. Most of the turn-slip in a tank-style track occurs at the corners, which is just where you have your tracks.

We had a somewhat similar concept, with one caveat. Instead of empty space in the middle, we had two wheels. The wheels were slightly lower than the treads at the corners. It might be hard to visualize, but if you were looking at a side-profile of the machine, you'd see a short track at the front, a wheel at the center, slightly lower than the track, and a short track at the rear. We were pretty confident in this design as a power machine, but we've gone the way of speed so we ditched the treads and went with center-drive only.

[Joe Johnson]

Weight in the corners is the problem not the length of the track, as others have pointed out.

You will gain by having your machine wide & short rather than narrow and long, but will it be enough? Difficult to say. Having 4 big watt motors rather than 2 will help muscle through the scrubbing but again, it is difficult to say how much you are giving up.

As to voltage spikes when rammed, they are so far down my list of problems, I don't even think about them. Your machine breaking is more of a threat than electical damage.

Sorry I can't be of more help.

Joe J.

[Ben Mitchell]

4 wheels have the same traction as 4 treads??? No way!

Would'nt having 4 tracks dampen the scrub issue, or would it be so little a gain as to render the entire concept useless?

What if the drive was modified to a simpler 1/2 track deisign, 1/2 tracks in back, wheels in front? Would that be better in terms of scrub/traction.

Also, would treads and wheels be much faster than 4 1/4 tracks, or would having treads in any combination cut out the speed factor??

What if the quarter-tracks were shortend to 2-3 inches, making them basically the size of wheels, but with the traction of treads.

Thanks for the help.

--Ben Mitchell

[Mike Norton]

We have run with track systems for the past 8 years. The best way we found out is having the track the 36" long. Putting a center wheel in that can move up or down depending on the traction. Get different track that have different friction factors.

We have had fast robots and power robots. we always used two drill motors to run the system. Blowing breaks is a problem if your driver does not know how to drive. With a good driver you can go 2 mins under big loads without breaking the breakers.

This year is great because we added more motors to the drivetrain.

with track system done right nobody will pull you sidewards. once you get what you want you just need to be 90 degrees to your puller.

pulling three goals with 3 130lb kids in the middle is easy done with this track system.


I still think track system done right can not be touched with Wheelchair wheels.

[Joe Johnson]

Wheelchair wheels are not the only legal wheels a team can have.

Traction is the thing in my mind, not tracks vs. wheels.

Joe J.

[AdamT]

Everything has to be wheels or treads, wheels or treads.....

So what about the other guys? Maybe there's something else out there...*wink wink*

[David Kelly]

i think there's a good change we may see something totally different this year. it could be us, i just can't tell. you'll have to wait and see:D :p

[AdamT]

Shoot, I'll go on and tell you a little right now....

That is, look for team 401!!!! Yeah, we don't have a "traditional" drive train.

[Ben Mitchell]

So would 4 1/4 tracks be a waste of time?? Would 2 full-length tracks be better in terms power/effort convervation.

My idea is basically replace wheels wityh treads, gaining the traction of treads, and with less material on the ground, have less scrub and more speed (as with wheel)

I am seeking a medium between a tank and a humvee,

Am I wasting my time, or is the 1/4 half-track idea worth pursuing (or at least trying out) ??

As always, opinions on this idea are welcome.

--Ben Mitchell

[Matt Attallah]

Wouldnt' tracks be better if you are dealing with some high-forces on wheels. Where you would eather sink-in too much (than you are carying too much weight!) or wouldn't your coefficent of friction become so much that wheels won't do it? Wheels just place the weight on a point, while tracks place it over a distance. So, in my opnion (sp?) if you some how can "tilt" the goals to get more traction, i say go with tracks. They turn on a dime, and i personally like the experience with them. My team (#5) got feather weight in the finals 3 (?) at GLR, when everyone thought we broke our bar hanging-mechanisim and we didn't hahahaha :), we weighed like 105 or something lbs, and we where pushing bigger, taller, very much heavier robots than us. I say go w/ tanks cuz' of my personal experience with it, and i think the motors will like it better when going straight, especially with a gear-box!!

[AdamT]

Word of warning. Tread are very hard to engineer properly. If built wrong, very bad things can happen. Stalling, breaking curcuits. Fun stuff. Like I said before, ours aren't exactly traditional treads or tracks. We like to think, "ouside the box."

[Matt Attallah]

Yes, 401Mentor, if you don't design the tracks correctly, you will slip the belts, eating them up and they are very, very expensive. And we did have a prob last year with our heavier robot with the treads slipping and poping breakers while turning in high gear...

[Ken Leung]

Basically, with treads, you have more surface area of contact.

So what would that do to you? Well, you have more surface area to take advantage of frictional force, and you won't be pushing as hard on each square inch of carpet your robot drive on.

I don't think using treads will necessary give you more traction. I think this because I believe if you wrap the same tread material around wheels, you will get the same traction.

Say team 45's chain tread with metal pieces digging into the ground, and team 159's wheels with similar metal pieces attach all around the wheel digging into the ground. I don't think there is a lot of difference.

As for being pushed on the side, you do have more side area on your treads to resist against the push, but the resistance should be the same with wheels with similar material wrap around them. With the same pushing force from another robot, that pushing force will spread through out the tread, or concentrate on the 4 wheels. Tread have more surface area to resist, but less force per surface area... Wheels have fewer surfaces to resist, but more force per surface area... They should both come out to be the same...

So, do treads give out a false sense of more traction, because in general the material for treads has better coefficient of friction than normal wheels?

Or, is there some special physics behind treads that help it gain traction over wheels with the same material wrap around it?


By the way, I also believe that it is more important to think about traction and friction of the material you use, rather than thinking about wheels vs. treads.

[ajlapp]

you need to really think about your objective and how you want to accomplish it rather than get wrapped up with wheels or treads. i've built both....treads are definately harder to build, and not always the answer. i've taken treaded robots against The Chief with two 8" wheels and lost horribly. it's all a matter of weight distribution, time, and desires. i know it's cryptic, but there are die-hard tread people, and there are cavemen(wheels), there is no clear cut winner. by the way, we definately don't have a standard drivetrain either...it should be a FIRST first!

Anthony Lapp
team 221 ----> now 857

[Chris Hibner]

Apparently team 401 and 857 are doing a totally new drivetrain concept. I want to take a guess at it:

First, we know what it isn't:

Typical drivetrains:

- wheels
- treads

Not-so-typical motive device:

- walking

However, I believe either 401 or 857 are considering something a bit different. I think it will be: A SCREW DRIVE SYSTEM!!!!!

Instead of wheels or treads, how about making two auger screws that are about 33 inches long (one for each side of the robot), and just spin the screws like crazy? You get a a lot of traction and it is a VERY simple design (except for making the screws, but that might not even be that difficult). I'm kind of surprised no one has tried this in the past (or did someone try it and I just missed it?).

Anyway, is that the secret system? Are you planning on using screws to move the robot? I guess you don't have to answer if you don't want to.

-Chris

[kevinw]

Why not coat the bottom of your base with some extremely low coefficient of friction material (available from SPI, of course), and use the motors to turn fans and propel you? Of course, you'd need to drop part of your chassis down to prevent people from pushing you around when you didn't want to move...

[LittleDave]

The formula for the force of Friction is F = u N NOTICE THAT THERE IS ABSOLUTLY NO CONCIDERATION FOR SURFACE AREA!! The thing thats going to make the biggest difference in traction is the material you are using and the coefecent of friction it provides with the carpet. Timng belts and other track materials are made of hard rubber and that means they have a low coeffecent of friction! I suggest that some of you re-think your driving mechs.



-Im sorry for my * spelling errors. * means i couldent spell terrible or atroshus

[Matt Leese]

Originally posted by LittleDave
The formula for the force of Friction is F = u N NOTICE THAT THERE IS ABSOLUTLY NO CONCIDERATION FOR SURFACE AREA!! The thing thats going to make the biggest difference in traction is the material you are using and the coefecent of friction it provides with the carpet. Timng belts and other track materials are made of hard rubber and that means they have a low coeffecent of friction! I suggest that some of you re-think your driving mechs.



-Im sorry for my * spelling errors. * means i couldent spell terrible or atroshus

While surface area doesn't matter, weight distribution does. The larger area in contact with the ground means that more of the weight will be positioned over the drive wheels (or treads,etc....therefore increasing N). Treads (as in with teeth) work a bit differently because they don't work on the properties of friction. They instead apply a contact force to the carpet parallel to the carpet (ie, they dig in and grip the carpet). In this case, larger contact area does help. Also, hard rubber has a satisfactory coefficient of friction (meaning not particularly high nor low).

Matt

[ajlapp]

we do have wheels.......but the drivetrain is untraditional, think the first time you saw the chief. walking would be cool though, and i seen screws used before, but speed is good for us this year!

[Ben Mitchell]

I've seen robots with treads perform fantastically (team 365 is a great example)

My whole purpose is to combine the high points of treads and wheels, while minimizing the negative qualities of both.

Treads would make a robot much harder to push or move, simply becuase of the friction spread over the surface area.

Treads also bear the weight of the robot over a larger area, making less stress on the drive.

As for 2 auger screws, think about it: they'd only be able to go forwards and backwards. They'd need 4, and thats the same as wheels (:rolleyes: ). duhh.

Getting back to the subject of treads versus wheels, I think that is you want a muscle robot, you'll probably want treads. I compare such robots to bulldozers: they may be slower, but they don't stop!

The advantage of wheels is that they are much faster, but they are also more vulnerable to being shoved across the room, are prone to lack of traction, and bear the weight of the robot on 4 smaller surfaces.

Having 4 wheel drive conpensates for some problems, but |||nothing||| compares with some nice solid treads.

Anyone beg to differ? Also, how have teasm improved wheel traction in the past.

And how about them pnuematic wheels (the ones w/ air inside them) Would'nt robots bounce around with them on??

Hmmm..

--Ben Mitchell

[Chris Hibner]

Originally posted by LittleDave
The formula for the force of Friction is F = u N NOTICE THAT THERE IS ABSOLUTLY NO CONCIDERATION FOR SURFACE AREA!! The thing thats going to make the biggest difference in traction is the material you are using and the coefecent of friction it provides with the carpet. Timng belts and other track materials are made of hard rubber and that means they have a low coeffecent of friction! I suggest that some of you re-think your driving mechs.



You are not quite correct in your assumptions.

Actually, the friction formula does not account for surface area, so you are right in that regard. However, what people are forgetting is that we are not dealing with pure friction here.

The carpet acts more like snow or sand rather than a smooth floor surface. When travelling on snow or sand, most of the pushing force of a good wheel design does not come from friction - it comes from a normal (normal meaning orthogonal) pushing force. That is why there are treads on snow tires (or paddle ribs on dune buggy sand tires). The treads or ribs sink into the snow or sand and push off of the snow or sand much like you pushing off a wall using your arms.

A simliar idea is the track design on a snowmobile. There are paddle ribs that protrude from the track (now most snowmobiles come with cleats for ice, but if you've seen an older snowmobile the track was made of formed rubber). Once again, the ribs sink into the snow to push off of the snow. The friction between the rubber and the snow accounts for a very small fraction of the total motive force. The snow can only hold so much stress before breaking away, so more track area means more paddle area, which means less stress on on the snow, which means more traction.

The carpet acts quite a lot like the snow or sand since it is quite deformable. Good tank treads act like the paddle ribs on the snowmobile track. The teeth of the treads sink into the carpet between the fibers and push off of the carpet with a normal force. Friction plays a much smaller role in this design.

If you want proof of this, look at either our wheels of the past few years or the treads of the TechnoKats from last year. The coefficient of friction between the carpet and lexan (our wheels) or smooth metal (the TechnoKats treads) is quite low compared to any rubber. However, we both got traction that was equal to or superior to rubber because we dug into the carpet and used "normal force" rather than friction.

Just like the snow, if you put too much stress on the carpet fibers, they will deflect enough to let go of your tread and you will begin to slip. Just like the snowmobile track, in this design the more cleats you can dig into the carpet, the less stress you put on the carpet, which means you are less likely to have the carpet deflect enough to cause slippage, which means the more traction you'll get.

[Chris Hibner]

Originally posted by Ben Mitchell

As for 2 auger screws, think about it: they'd only be able to go forwards and backwards. They'd need 4, and thats the same as wheels (:rolleyes: ). duhh.



You can turn using the auger screws. Turn one forward and the other in reverse and you turn just like the tank steer system with treads. Should I say "duhh" now?

[AdamT]

I'll spill the beans in a week or two, but here's another hint. It may not be WAY out there. There is a heavy machine that uses them already. Yeah, that's the ticket.