moving a limit switch

Ok peeps I have run into a block wall, its called a budget. and i need a few ideas. i have a project that i need to move a limit switches on a liner bearings.

there are 4 of them that are used for back stops on a shear.a robot picks a sheat metal blank and put the blank on the shear table and pull it back till it hits 3 of the 4 limit switchs then the shear cuts the metal.

depending on the job there are in deferent places. now the operator has to go in the cell and move them. I would like to automate them so when they select a job they will move in to postion. my fisrt thouth was to use a ball screw and servo motor, but that is to much money. so my next idea is a sevor or stepper motor drive a timing belt. i am controling this setup with a plc-5 with a profibus card and control net installed.

What i am looking for it a cheap way of doing this. not asking for any one to do this just looking for ideas and products

thanks for the help

if it would help i will post some pics so you all can get an idea of the project

thanks agian

How accurate is the positioning on the robot?
From what I know, most seem to be in the 1/10 mm accuracy and repeatability range, which also seems to be a bit better than what you get with a limit switch.
Why not reprogram the robot to use this inbuilt accuracy for the job?

If you need to use limit switches, why not set them up as a reference point, then have the robot move from that to a position dependant on the product setup? You could just have product A off set 30mm in the X direction and 200mm in the Y. Product B can go to a different position etc.

This gets rid of the extra hardware and would more fully utilise the robot's inbuilt functionality.

Another option, have the robot move the limit switches......

Idea No. 1,

Forget the limit switches all togeather.

Mount a sliding backstop on the shear. When the material hits the backstop it will push the backstop back away from the blade. The sliding backstop would ride on two tracks (one at each end of the shear). Each end of the backstop would be tensioned with a spirator (coil spring gizmo with a pullout cable).

Now mount one of thoes pullout cable type encoders on the shear and attach the pullout cable to the sliding backstop. The operator would use a small HMI to set the cut lenth. The material will push the backstop untill the encoder matches the cutlenght. The shear cuts the material, then the back stop will be pulled back into it's origonal position and be ready for the next feed.

Note: this may also require some sort of shock arrestor to keep the backstop from slamming into it's home position.

This shouldnt be a budget buster. If you have a shear in your shop then your likely to have whatever else you need to fab the backstop.

Spirator: http://www.springcompany.com/Spring%20Reels.htm

Encoder: http://www.encoder.com/modellce.html

elevmike the shear is not set to use the backstop idea but i like the encoder might be able to use it

doug_adam the parts are not alway square they are deferent angles
so it would be hard to set the part in the same place went over this idea with a coworker and he dosen't think we can keep the part the right shape

so i guess that leaves me with 4 stepper motors if i can find a resonable one with a controlor that i can talk to with profibus
i was trying to upload some pics but they are to big
but thanks for the ideas

I was thinking along the same lines a Mike until the last post.

OK, if I understand you, you have multiple switches because the parts have different profiles, and the switches are adjusted accordingly.

Well, now I'm still thinking that you might just want to ditch the discrete point limit switches and go with someting that that measure actual position. What about four long stroke LVDTs? How far do these limit switches move between setups anyways?

they move on a 4 ft guide with micro adjustments
here are some pics

pdelno,

Two encoders, one on each end of the table, will give you the ability to measure orintation and horiziontal position. (aka angle & feed). Requires some fancy math (ever do trig on a PLC?), but a lot less hardware.

After looking at the pictures:

How hard would it be to work up several bars with the switches already correctly positioned on each? Bring the connectors out to a common point. When the part changes, install the bar for that part.

.02

The rail idea would be ok but we have 10 jobs on the shear.i would need 20 bars for that. what i would like it open up the shear to run just about anything on it with a minamum change hasel

Hey, if the robot is picking up the material from a stack of sheets, there should be benchmark 3D positioning information available to properly position the sheet for cutting by using the various positioning sensors on the robot. Surly the robot has some way of determining X, Y, Z & rotation. I mean like how are you moving the material from stock to the shear?? All you have to do is take it a step further, code wise & get rid of the switches all togeather.

I would like to but i can't guarantee the postion of ever part in the stack.

I would like to but i can't guarantee the postion of every part in the stack.

Click to expand...

Ahhh, but you can...because you can always determine the exact center of the robot's pickup-thingamigig. The center of the robots pickup is your benchmark used to determine the position & orintation of the stock as it relates to the shear. If after the cut the robot is left holding the product, and not the waste, then you need to measure from the center of the robot's head, and not the edge of the stock.

If the robot is left holding the waste, and the product is dropped in the shear bin, then you have three choices:

1) use a fitted stack. Presumeably all the stock sheets would be the same size and shape, so they need fixed guide posts, (a jig of sorts), that an operator would load the sheets into.

2) provide a homing station on the table. 3 switches (fixed) would would do it. The robot would set the stock material on the then move the stock into the limits and stop on the limits, and zero for home; then push it into the shear for cutting.

2) Have the robot do a trimming precut. Let's say the sheets are slightly irregular or out of place, but the robot will always pick in the same position (center from stock). Before cutting the final product, have the robot position and cut the sheet at a fixed location just outside the final cut, but would be known to be inside the outer edge of the stock.

ok elevmike you got my attention
The productn stays on the robot tooling.I am still trying to learn robots. This is an ABB robot and they are a little harder for me then fanucs. but the way it sounds is i have a lot more to learn about them. but i guess you always have more to learn.

so what i what i don't understand is how to I tell the robot where the part is with out a vision system and with out prequlifing the part.
it would be a lot easier that way.

Ok pdelno,

The robot always knows Z (vertical where it picked up the material), but maybe not exactly X & Y & Orintation. So your stock stack has guide posts right?? So you pick up the stock and pull it stright up to the top of the posts. Facing the stock from the robot, on the two left posts you mount one prox each. As soon as the robot reaches the height of the prox's then move the material to the left till you hit one of the two proxes, then rotate till you hit the second one.

Now the material is square and you know where the left edge is. Thus you have determined X & Orintation/rotation or whatever you call it.

Now move the material stright back twards the robot till you hit prox switch for Y. Now you have X, Y, Z, & orintation. That's your starting benchmark for moving the stock into the shear.

Knowing or having a benchmark position will eleminate the need for the moving limits, or having to manually setup for each job. Additionally the Robot/Shear station could switch jobs on the fly, (if you use the same stock for each job). The three limits are now used for homing only and should never have to be moved. Your process will fly, and your boss will get great credit for not firing you, and you may even get a pat on the back; but he'll be sure to get the raise.

Good luck with the robot code. I would consider this a multi-axis motion control, wich Peter would know much more about then I. Maybe you'll get lucky and he'll chime in...