differential and trike steering?

rwgast_logicdesign · 2012-10-12 13:58

ok so ive been working with what i have which isnt alot and unfourtantely i cant find any scrap rc cars.

anyways i was trying to do a 4 wheel bot but do to differences in motors being a couple gears short i decided just to throw a caster wheel on for now. im not very happy just driving the two motors forward directly the thing wont go straight for longer than 20 seconds and the caster has problems turning.

so im thinking i have a servo i could use this to rotate a single wheel or maybe two wheels that roll freely in correraltion with the front diferential wheels. this seems like it would provide more accurate turning radiuses and lock the caster straight when its not turning.

i have never seens a bot with trike style and differential steering is there problem with this?

NWCCTV · 2012-10-12 17:20

Funny you mention this. We were just discussing yesterday. Yes it is doable and I am currently working on the same concept but with a stepper motor. http://forums.parallax.com/showthread.php?143017-Single-wheel-for-front-of-robot-to-turn

Ttailspin · 2012-10-12 20:34

i have never seens a bot with trike style and differential steering is there problem with this?

I don't think there is any problem...
>>>>Check out post #73 of this thread.<<<<

Casters are evil. I like Rudders...

It's not like I haven't tried to use casters...

Of course Rudders and Casters aren't allways needed...

"You will have to think outside of the box, especially if you don't own a box to begin with..."
Just use what you have on hand, and see what you can come up with.

-Tommy

erco · 2012-10-13 01:19

Ttailspin wrote: »

Of course Rudders and Casters aren't allways needed...

Sometimes outriggers are! http://www.youtube.com/watch?feature=player_detailpage&v=fuogP2h82B4#t=21s

You know how I love this alternate ending, Ttommy Ttailspin!

rwgast_logicdesign · 2012-10-15 00:59

Ok so this is what I came up with, I dont have any motor controlers yet, so I cant do any kind of real steering tests

but by moving the servo and swapping motor polarity everything seems to work alright, this is in no way final those batteries are getting moved and my circuitry is going there and im loading the inside of the frame with nicads. The tires might get changed too but this is the general concept im getting at with the motors the front two are direct drive while the back two spin free and interdependent of each other, controlled by a servo.

RobotWorkshop · 2012-10-15 11:02

rwgast_logicdesign wrote: »

i have never seens a bot with trike style and differential steering is there problem with this?

There is always going to be issues getting robots to go perfectly straight (wheel slipage, etc) no mater what drive method you choose.

The HERO 1 and HERO Jr robots use three wheels. Each one uses a single wheel for both the steering/drive and the other two wheels spin freely. It works but it doesn't seem to make it any easier to go straight than a robot with differential drive. Personally I think programming and navigation are easier on a robot using differential drive so that is what I would lean toward when making a new robot base. If you add encoders to the main drive wheels then the microcontroller can keep track of the speed of each wheel to make sure they are turning at the same speed. This can help keep it going where you want.

If you want to try something instead of castors you can always use a ball transfer instead. I'm using these on the last couple of robots that I built.

http://www.mcmaster.com/#standard-ball-transfers-for-conveyors/=jqfamf

rwgast_logicdesign · 2012-10-15 13:02

Well my issue is i have to make due alot of times, and my caster set up wasnt not going straight, it was doing doughnuts after about a foot of driving straight lol. As far as encoders I have some stuff that should work for the encoder wheel and some IR LEDs/Phototransistors once the bot is running with electronics I will start the process of adding wheel encoders, not trying to take on to much at once.

My biggest reason for 4 wheels besides the straightness issue is my cant seem to get over bumps at all when I use a caster and it falls alot more. The set up i have now should allow me o do tight 180s well enough to flip the bots oreintaion during travel. Im sure it will be harder to coordinate the back steering with the differential drive but im sure its also not going to be the hardest thing in the world.

erco · 2012-10-15 13:09

Agreed, the HERO robots had serious steering issues, they were marvelously inconsistent. Mainly because they lacked an accurate center position. Because I was aware of those limitations, I was pleasantly surprised at the repeatability of the little BS1 trike robot I made for the figure 8 challenge, see http://forums.parallax.com/showthread.php?138125-Erco-s-Figure-Eight-Challenge&p=1078311&viewfull=1#post1078311 and also see post #13 in that same thread. Properly done, I think a scaled-up trike chassis could be a winner.

As I've espoused many times, a good robot starts with a mechanically sound chassis and good solid motors & wheels. All the slick software in the can't make up for bad design or poor fabrication.

RobotWorkshop · 2012-10-15 13:20

Encoders aren't all that difficult to make. Alternately you can usually find old (no optical) computer mice for next to nothing. Most of those have a pair of quadrature encoders in them. Here are just a couple examples:

http://members.shaw.ca/swstuff/mouse.html

http://retrointerfacing.com/?tag=mouse-encoder

There are many different styles of castors, slides, ball transfers, etc. I've have the best luck using a castors (or ball transfers) front/rear and two differential drive motors. Usually balance the robot so it favors one of the castors (rear works well) and make sure there that both castors don't touch the ground at the same time. Otherwise it can get stuck. Robots with their differential drive wheels near the center can easily turn in place.

rwgast_logicdesign · 2012-10-15 14:11

Wow thanks for the mouse links, I never new that!! Its off to the computer junk pile for me

The thing is Im not really doing a trike design, my design is alot closer to the last picture Tailspin posted, but instead of having two roating platforms im only using one and differential drive in on the other side. If worst comes to worst I gues I could make the bot stop for a split second 180 the wheels on the servo then use the differential drive to turn.. but since the wheels attached to the servo are fee spinning I should be able to do differential turns while angling the servo at its max velocity.

I think this could work well as it will give me the tight handeling of a 3 wheel differential bot, with the stability of a 4 wheeler (most of the time) and the accuracy of a trike like the one tailspin posted in the figure 8 video..

Or this could all be a huge mess

but I guess I wont know until I get some motor control on the differential drive setup.

Ttailspin · 2012-10-15 19:44

That looks like a great start to a really fun Robot, rwgast. You have to start somewhere.:thumb:

I have found that I don't need to "180" the rudder wheel, and I only need four rudder positions,
Straight, 45(left/right), and 90.

Originally, I thought I needed more positions, but they are not really needed,
As you can see in the Figure-8 video, the Rudder servo is in roughly the 45 degree range.

When Spotbot is free roaming with IR sensors, I have about six positions per side(left/right).
The Parallax standard servo works just fine, It is plenty fast and strong enough to swing that rubber tire around.
I don't pause at all to move the rudder during a turn,
Only the wildest (rapid full right, forward, to full left turn) movements makes the wheel drag for a split second.

I think Rudder, or "Tiller" steering can add a little accuracy to turning a robot.
Oh, and I could mention that it just looks really cool too..

-Tommy

differential and trike steering?

Comments