SearchL Function

  • Hello Everyone,


    Anyone have good knowledge on searchL function Because I wanna try. And It will really helpful If someone shares ideas about searchL. :help:


    My actual plan is robot needs to find deviation in a part location. I'm using the s4c+ controller and it having a searchL function but I wanna give DI to the tip which means 24V supply to search the part deviation and store the value, Can I use laser distance sensor(eg: Omron laser distance sensor E3Z-L61) instead for giving 24v to the contact tip?


    Please do share your ideas about it ??? ??? :help:

    sandeep

  • SearchL instruction can only be used with a digital IO input.
    For me makes no sense to use a distance sensor with SearchL. If you already can read the distance using sensor, why you need to perform a search?

  • Hi roulv,


    Thanks for the reply. Can you tell me how to use searchL? I'm trying it for the first time. I read the manual but didn't help much. The sensor is the second option but first I wanna use searchL. if you familiar with searchL please share your experience about it. :help:

    sandeep

  • Just move robot to a start position and run searchL. In SearchL instruction you must write the digital input signal, create a variable robtarget that will get information about where the robot found workpiece and another robtarget with the end point. This position must be taught on the very end of the search.


    CONST robtarget pStart:=...;
    CONST robtarget pEnd:=...;
    VAR robtarget pFound;


    PROC SearchWorkpiece()
    Movel pStart, v1000,fine,tGripper;
    SearchL\Stop, diSensor,pFound,pEnd, v100,tGripper;
    ENDPROC


  • SearchL instruction can only be used with a digital IO input.
    For me makes no sense to use a distance sensor with SearchL. If you already can read the distance using sensor, why you need to perform a search?



    I guess it depends on what information you get from the sensor....
    Most inexpensive laser sensors have a fixed trigger point, so they don't measure distance per se.
    One step up is the ones where you can set the trigger distance through a little adjuster (semi sketchy for production imho).
    Next step up (still referring to "dumb" 1D lasers) is the ones with a true distance measurement, e.g. "analog" output.


    All of them can be used with SearchL, even though the 3rd one is a bit overkill for SearchL, but then again, most of those
    you can still set a trigger and then you can use the "analog" for verification / other stuff, so I wouldn't disqualify the sensors or SearchL.


    SearchL is used to find positions... ( duh ! :winking_face: )
    You tell it where to start and where to go and if it detects an input along the way it will register that position, what you do with it is up to you...


    Most widely used is Smartac where you have a reference point that is compared with the found point and the difference is your offset.
    But you can also use it to define object frames if you for example is searching a whole fixture frame.


    The solution really depends on your application and setup...

  • Thanks for your suggestions guys,


    My plan is simple


    1. I would like to connect a sensor eg:3D laser triangulation Sick WT18 or Omron E3ZL66 or some other to the tool holder pointing TCP XYZ


    2. I will create a program and run the program 1st time.


    3. If part having variation (2nd time) robot need to adapt to the variations of the part using sensor values only linear no orientational changes If orientation is possible its an added bonus.


    4. So, How to perform this test and which sensor you would recommend ?


    5. how to establish communication between controller and sensor to adapt path according to sensor values ?


    6. I have following options in ABB S4C+ robot : Arcware (I know its arc welding package)
    Arcware Plus (Advanced seam tracking I guess)
    Sensor Interface (serial communication RS 232)


    So, Let me know your views about it :help:

  • 1. These type of inexpensive (by comparison to Laser trackers, e.g. Meta, Servorobot, Binzel, etc.) are really sensitive to reflections (especially on machined beveled joints, and/or sheared/cut joints) so I would recommend testing them first before you commit to using one or the other (if you have that luxury).


    2/3 Depending on your fixture, part and what the joint / path looks like there's a number of approaches you can take.
    Does the seam / part vary so much that you need to offset every position, or is it OK to move the whole path / weld?


    Because rapid reads the program ahead of the execution it's a bit tricky to apply offsets "on the fly" (for example MoveL p10, Pdisp X, MoveL p20, Pdisp Y...)
    ...

  • Hi SAABoholic,


    I already discuss with Servo-robots they said there are not supporting s4c+, then Binzel is expensive to our budget and Meta? Do u mean meta-vision???


    I hope I'm luxury but unfortunately not so I'm using these laser sensor cost 500$ only. Now what I'm trying to do is buy a smarTac PIB and connect a laser sensor to the robot and take the offset values and adapt the program to it.


    Traditionally, smarTac PIB uses gas cup search or contact tip search which will take a lot of time. Rather I wanna reduce the time using this laser(Cheap) sensors to get offset values and get done.


    So, tell me your view on it. I'm only bypassing the values of sensor 1 and sensor 2 in smarTac PIB :help:

    sandeep

  • It's unfortunate that they stopped supporting the S4 platform, they did support them as early as S4 M94A (two generations prior to S4C+).
    Yes, it's Meta (Vision), they have been around for quite some time and is a cost efficient alternative to Servo Robot.


    Granted, there's a BIG difference in price between a laser tracker (SR / META) and the sensors you're looking at, but there's also a BIG difference in functionality and at the end of the day it's the application that dictates the solution.



    If you plan on using an external sensor, then I'd NOT recommend SmarTac PIB as you're paying a lot for HW you'll never use.



    What I meant with the luxury of testing is that hopefully you have suppliers kind enough to let you try one before you buy it
    so that you're not spending money on a sensor that eventually turns out to be unsuitable for your application.


    These sensors would be wired directly to a digital input (recommend staying with ABB I/O's for performance/response time purposes).




    What about my other question ?

    Quote

    Does the seam / part vary so much that you need to offset every position, or is it OK to move the whole path / weld?


  • It's unfortunate that they stopped supporting the S4 platform, they did support them as early as S4 M94A (two generations prior to S4C+).
    Yes, it's Meta (Vision), they have been around for quite some time and is a cost efficient alternative to Servo-Robot.


    Okay, I will check with Meta Vision but I'm thinking it is expensive. I already contact with Axson seamfinder but didn't work out well.


    Quote


    If you plan on using an external sensor, then I'd NOT recommend SmarTac PIB as you're paying a lot for HW you'll never use.


    Okay, So without smarTac how the external sensor communicates with the controller. I need more idea about it, Please can you explain ???


    Quote


    What I meant with the luxury of testing is that hopefully, you have suppliers kind enough to let you try one before you buy it
    so that you're not spending money on a sensor that eventually turns out to be unsuitable for your application.


    Yes, Of course, you're right. I don't need equipment which doesn't suitable for my application. So, without smarTac and only using sensor how the sensor will communicate with the controller about part variation?


    Quote


    These sensors would be wired directly to a digital input (recommend staying with ABB I/O's for performance/response time purposes).


    Do you mean DSQC 328? Okay, I need more explanation about it.


    Quote


    Does the seam / part vary so much that you need to offset every position, or is it OK to move the whole path / weld?


    Yes, my parts manually fabricated so the part variations is a big hectic. Everytime modification required generally 4-10mm variation. I want to do automatic everything. What would recommend for this? SmarTac or Sensor or any other. I Need a solution to put stopper for this problem. So, Tell if you're in my position what your solution about it. I can PM if you like to discuss it, Thanks for reply :backtotopic:

    sandeep

  • SmarTac is SearchL on steroids, plus (sometimes) the additional HW to generate a digital input through the Gas Cup/Nozzle/Shroud.
    In your case you already have something that can generate an input / 24VDC to the robot (e.g. your sensor) so you don't need the SmarTac hardware.


    Your sensor needs to be wired directly to an Digital Input (328 is fine) and then you use that input as the trigger for SearchL.
    SmarTac does not have any functionality that you can't re-create with SearchL, in fact, if you're searching positions rather than an offset
    then I prefer SearchL as you don't have to deal with the "intelligence" of SmarTac.



    What type of joints, material, material thickness etc. are you dealing with ?
    Sadly your solution might not be making the robot smarter, it's 9/10 times ensuring the incoming parts are "perfect"....


    Any chance you can PM a picture of your part ?

Advertising from our partners