Author Topic: Linear Track positioning system (Read 7180 times)

BANE · « **on:** January 26, 2008, 04:02:05 PM »

Hey I making a linear track for a robot to slid back and forth on. I have a modified a servo for continues rotation and mounted a gear on it. The gear makes contact with a 12" long gear purchesed for ServoCity and thats pretty much the machanics of it. (see pic) The only way to make the servo move is to send out a pulse for a certain amount of time. So i have to relie on time for the position of my robotic arm. the position of the arm is farly important and my question is, is there an "affordable <$20.00

" positioning device that i can get and can it work like a servo. (like a very long linear pot)

Any other ways to do this?

Bane

ed1380 · « **Reply #1 on:** January 26, 2008, 04:16:54 PM »

IR rangefinder mounted on base. and aim it at one end of the track.
w/ simple math you can figure out where it is

BANE · « **Reply #2 on:** January 26, 2008, 04:27:52 PM »

I considered a PING))) sensor but a IR sensor would probably work better because of short range. How accurate is the IR sensor thought? (+/- 1"?)

bane

airman00 · « **Reply #3 on:** January 26, 2008, 07:59:20 PM »

Quote from: BANE on January 26, 2008, 04:27:52 PM

I considered a PING))) sensor but a IR sensor would probably work better because of short range. How accurate is the IR sensor thought? (+/- 1"?)

bane

I would say IR sensors are accurate to +/ 1 inch IF you are using a proper surface for the IR to bounce off ( meaning NOT foam , NOT dark , etc.)
It all depends on the surface of the IR hitting it

OR you could make a linear encoder like in a printer .

Easiest: IR sensor , just be careful of the surface the IR sensor is hitting

dunk · « **Reply #4 on:** January 27, 2008, 04:56:42 PM »

i would say the easiest way would be to count how many revolutions of the drive servo.
just a simple encoder made from a push switch that something attached to the servo horn hits every revolution (or every fraction of a revolution if you require more accuracy).

dunk.

BANE · « **Reply #5 on:** January 27, 2008, 09:09:14 PM »

Dunk, your idea of the switch would probably be my best bet if i had the right materials that could have that type of endurance. (previous experience with Lego mindstorm robot, switch broke after a while) However, i have a Dell A904 inkjet all-in-one printer and scanner that I'm going to take apart. Is there anyway to get the linear encoder out and use it? (just a thought)

bane

airman00 · « **Reply #6 on:** January 27, 2008, 09:12:48 PM »

yea the encoder is the long strip of darkish plastic and there is a reader on it

I just looked inside a printer and it looks translucent , but I have no idea how to read the encoders, its the same color and see throughness throughout the entire strip

Search on google, it should give you some good results
Please post how you used the printer encoder ( if you will use it)

BANE · « **Reply #7 on:** January 29, 2008, 06:58:14 PM »

I can seem to find anything desirable

, but I'll keep looking. I can see the translucent thing in and I'm guessing that something in the type changes (transparency and maybe it will be linear) but not quite sure thought.

if anybody can help or have any opinions, I would be most grateful.

jareth frank

HDL_CinC_Dragon · « **Reply #8 on:** January 29, 2008, 07:04:55 PM »

This is pretty much exactly what you need:

Its called a Limit Switch (or in the pic there is 3

)

That would be for using Dunks idea which IMO would be the easiest and best option. But like Dunk said, using a radial encoder would be much much more accurate

airman00 · « **Reply #9 on:** January 29, 2008, 07:10:58 PM »

Quote from: BANE on January 29, 2008, 06:58:14 PM

I can seem to find anything desirable , but I'll keep looking. I can see the translucent thing in and I'm guessing that something in the type changes (transparency and maybe it will be linear) but not quite sure thought.

if anybody can help or have any opinions, I would be most grateful.

jareth frank

there might be lines very close together in the film like this http://www.goldmine-elec-products.com/prodinfo.asp?number=G15602
or maybe its transparency that changes

the material is mylar , so maybe that can help you somehow

gamefreak · « **Reply #10 on:** January 29, 2008, 07:51:46 PM »

back to the limit switch idea, it couldnt detect if the motors were forced backwards(

) it would still count it as one revolution in whatever direction you have it going.

HDL_CinC_Dragon · « **Reply #11 on:** January 29, 2008, 08:20:12 PM »

True but if you got creative and had one at the 0 degree position and one at the 45 degree position, you could see which one was hit first and that would tell you which direction its going in

BANE · « **Reply #12 on:** January 30, 2008, 12:07:25 PM »

Thanks guys for all the great ideas, but please note that I'm only trying to find away around depending on time for the positioning. Most of all, this robot can have about a 1/4" tolerance

and probably still work right.

About the cam and switch idea, i think that it would work better on a larger scale robot and it would have to take quite a few ports from the microcontroller. As far the direction, you would already know that direction its going (unless its bewitched or something) so the question would be how much? (on this robot that is)

the encoder from the printer would be an awesome idea (if we could figure it out) because its so fast and printers a so readily available now a days

bane

bukowski · « **Reply #13 on:** January 30, 2008, 01:11:11 PM »

Just out of curiosity...

How do you plan on "holding" the pinion gear on the rack?
It seems that unless you are running the bot itself on a track, or you have a flanged gear, you would have a tendancy to wander off the gear rack.

BANE · « **Reply #14 on:** January 30, 2008, 05:30:59 PM »

(Check out pics

)
I have the ball bearings to make a smooth and percise ride. However, they only pervent the servo sperocket from moving up and down. As far as side to side, I have nylon bushings that will make contact with other side (~+/- .01 $:-\$ ) of the rail and will be enougth to keep it on track.

bane

HDL_CinC_Dragon · « **Reply #15 on:** January 30, 2008, 05:53:28 PM »

Looks pretty nice

What program did you make that in BTW?

BANE · « **Reply #16 on:** January 30, 2008, 06:12:38 PM »

Solidworks

, is fast, simple, and pretty cheap for its functionality http://solidworks.com/pages/products/SWOPremium/SWOPremium_Features1.html?pid=916

bane

bukowski · « **Reply #17 on:** January 31, 2008, 11:35:49 AM »

Wow, nice design job on the trucks. It'll be a beaut when it's done.

BANE · « **Reply #18 on:** January 31, 2008, 12:06:35 PM »

when i complete it I'll make a tutorial of it. I think its pretty cool project because everything looks really hard and expensive, but I'm almost making everything in my garage and it out of scraped items and junk (cool)

.

bane

BANE · « **Reply #19 on:** January 31, 2008, 03:21:23 PM »

Hey, about 10 min ago UPS came with my stuff from Lynxmotion

. I ordered the sharp sensor so I'm going to build the track system this week then if time permits. I think the Sharp IR sensor is going to be my best bet. I know that there are alto of toterials about doing the math on the sensor, but could someone give me a link to a short range one ( < 12")

bane

Admin · « **Reply #20 on:** February 16, 2008, 04:57:03 PM »

a servo encoder, if your Sharp IR isn't accurate enough:
http://www.acroname.com/robotics/parts/R239-WW01.html

News:

Author Topic: Linear Track positioning system (Read 7180 times)