Seam Tracking - Arc Welding

mattdavis

Does anyone have some experience using seam-tracking?

I have made a program that is doing some seam tracking whilst reorienting the torch.

so I have

ArcLStart (gun pointing down)

ArcL (about 75mm up and gun level)

ArcLEnd (200mm up gun still level)

I use touch sensing to locate the first point and then just offset the other points z location from the first one, so the second to points are not exactly on the seam.

What happens is that seam tracking will track pretty well upto the ArcL instruction and jumps to the programmed point and then starts tracking again. Is there anyway to have it just continue from the tracked point rather than the programmed point? You can see the point 75mm up the weld seam where the robot jumped about 5mm to the left and then worked its way back onto the seam.

Kleo

You have to consider the sensor's look-ahead.

The first measurement that gives a correction for the robot is at the position where the sensor is located when the robot TCP is at the starting point of the weld, which is the robtarget of ArcLStart.

mattdavis

Sorry, i am using AWC (thru arc tracking).

So the sensor is located at the tcp and gives correction signals with very little delay.

The robot has used the tracking to follow the seam up to the ArcL point and them jumped back to the program point once it reached the ArcL and then continued to track correctly. I really want it to ignore the programmed position of the ArcL point, its only there to reorient the torch, but i don't know to program this.

Kleo

Sorry, I was involved in Laser tracking questions minutes ago and did not even think about AWC when I saw your question.

It is difficult to say what the problem is without having seen the RAPID code (including the values for the robtargets and ev. workobject).

 

Did you try to activate the tracklog and look at the values you get from the sensor?

 

Just to be save: are you using IRC5 or S4C(+)?

mattdavis

I am using IRC5, and was not aware of how to use tracklog (other than having the laptop connected to the AWC unit during the weld).

This is the code for one of the vertical sections. The workobject is defined with z positive being up (but the weld seem is not always exactly vertical so the midpoint and the top point are not exactly on the seam). The position values of ver1straight and ver1top are overwritten once the search for the start position is complete, only the orientation (and external axis, yes its an irb140 so i need to use external axis for a weld 190mm long argggg) is used. 

LOCAL VAR robtarget ver1straight:=[[89.54,84.33,109.69],[0.251941,-0.670475,-0.308028,-0.626186],[1,-1,-1,5],[12600.7,9E+09,9E+09,1688.05,12600.7,1849.71]];

LOCAL VAR robtarget ver1top:=[[93.19,92.36,266.63],[0.251944,-0.670475,-0.308032,-0.626183],[1,-1,-1,5],[12600.7,9E+09,9E+09,1688.05,12600.7,1692.52]];

LOCAL VAR robtarget ver1start:=[[84.94,70.77,119.51],[0.191433,-0.770212,-0.343752,-0.501958],[1,0,-1,5],[12748,9E+09,9E+09,1562.09,12748,1682.68]];

LOCAL VAR robtarget ver1startweld:=[[68.44,82.91,199.59],[0.207696,-0.749445,-0.339514,-0.52908],[1,0,-1,5],[12577.5,9E+09,9E+09,1694.02,12577.5,1711.93]];

WeldSearch y, 50, 6, wobjdn600ps10, tWeldGun36;

WeldSearch x, 50, 6, wobjdn600ps10, tWeldGun36;

WeldSearch z,-100,2,wobjdn600ps10,tWeldGun36;

WaitUntilInPos, TRUE;

ver1startweld := Crobt();

ver1straight.trans := ver1startweld.trans;

ver1straight.trans.z := ver1straight.trans.z + 75;

ver1top.trans := ver1startweld.trans;

ver1top.trans.z := ver1top.trans.z + 190;

ArcLStart ver1startweld, v200, VER3F_1_1sm, VER3F_1_1wdWeave:=VER3F_1_1wv, fine, tWeldGun36WObj:=wobjdn600ps10Track:=VER3F_1_1tr;

ArcL ver1straight, v200, VER3F_1_1sm, VER3F_1_1wdWeave:=VER3F_1_1wv, fine, tWeldGun36WObj:=wobjdn600ps10Track:=VER3F_1_1tr;

ArcLEnd ver1top, v200, VER3F_1_1sm, VER3F_1_1wdWeave:=VER3F_1_1wv, fine, tWeldGun36WObj:=wobjdn600ps10Track:=VER3F_1_1tr;

Thanks for your help so far, I would appreciate any suggestions you have.

mattdavis2010-03-24 08:39:35

Kleo

I think your problem is the 'fine' point in the ArcL instruction. You should not use a 'fine' point in the middle of a seam, use z0 instead if you need. But in your case a larger zone would fit better (e.g. z5).

mattdavis

Makes sense... I will try it out.

Thanks for the suggestion.

Smith

I have a question has anyone integrated a laser search system on a old s4C 2400l M98 with miller delta weld? We are currently using nosel search and have a hard time catching the edge of some lap joints that have thin sheet metal we would like to use a laser. What would this all entail and how would I go about setting this up?