DIY Physical Automation - Robot Arm Edition

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • DIY Physical Automation - Robot Arm Edition

    Hello!

    For the last several weeks, I have been running the same style of part; this is the fifth 3,000 piece order I have manually sorted on the CMM (by "manually" I mean I manually chuck one part into a chuck [rotary table] at a time. several thousand times over)

    As you can imagine, this is taking forever.

    So, I got a brilliant idea; what if we get a robot arm or something to just automatically stick stuff in the chuck?

    Now, my first idea would have been a fixture that I can stick, say, 10 or 30 parts onto, and measure them in a loop.
    However, what if the next set of parts is a little bigger?

    Now I need to commission another fixture, which could take awhile; combine this with the other couple thousand different pieces we make, and....


    So, back to the arm...
    As it turns out, having a company come in and retrofit your machine to automatically measure a buttload of parts is really expensive.

    I thought I'd take the less expensive, DIY approach.
    I'm a programmer at heart, and I love robots so I mean, why not.
    ----------------------------------------------------------------------------------------------------------------------------------------------------------------------
    I was curious if anyone has taken on such an endeavor? What sort of options are out there?

    For those that have full-fledged automation cells, what sort of technology are you using?

    I haven't even been to college yet, so if something like PLC is the "only" way to go, I can add that to my list of classes.

    I'm looking to build a robot that can load a theoretically infinite number of parts into a rotary table, or even just hold it straight in CMM space.

    Any help would be appreciated!

  • #2
    Sooo many ways to skin this cat...

    Hexagon offers some solutions. We use their CMM Supervisor.
    PcDmis 2015.1 SP10 CAD++
    Global 7-10-7 DC800S

    Comment


    • InspectorJester
      InspectorJester commented
      Editing a comment
      I agree, but my current company isn't willing to shell out a dime for this project; I'm working alone here.
      I'll get to take it with me if we don't work out but I'd much rather get a solution straight from Hexagon (we just won't afford it!)

  • #3
    What's your budget range?

    For 2yrs, now, I've been part of the 'automation' department here at my work. I program the CMM (not that much different from any CMM program but with a really dumb operator: a robot) and on my down time I learn about the cell. I know enough about Fanuc Robotics to create programs online and offline (much easier coming from a CMM background IMO) took me about a solid year. Along with robotics I picked up on SQL databases, they're so much fun, MYSQL to be exact and it's freeeee! At the moment I'm going thru the software that handles the HMI along with the PLC programming (structured text style but still want to pick up on ladder logic)...it'll be a while before I put these 2 in my back pocket. All on my down time, I cut everything else out except for reading physical books here and there. You can do it InspectorJester

    Sooooo much to learn. I'm encoureaged to attend college, now, just to get that paper that says "hey this guy is OK to work on your expensive stuff."! Otherwise I'll stay as the "CMM guy that self taught himself workarounds to offer simple solutions for us" but never actually got the title and the mula.
    Last edited by Kp61dude!; 03-14-2019, 02:28 PM.
    PcDmis 2015.1 SP10 CAD++
    Global 7-10-7 DC800S

    Comment


    • InspectorJester
      InspectorJester commented
      Editing a comment
      I'm flexible, but I'd rather not spend thousands if I don't need to, especially since I don't plan on marketing the idea.
      I'm looking for a solution I can develop in my garage in my spare time; I only need to pick up and stick up to 40lbs in a hole

  • #4
    Google search Twincat 3, codesys, iec 61131-3. You can convert arduinos to run PLC on the cheap. You can buy industrial PLC controllers by Beckhoff on ebay along with servo motors also on the cheap. My guess would be that the most expensive items are going to be the fixture machining.

    A PLC guy told me that the starting point is the toughest: figuring out the communications (everyone seems to be using ethernet) and layout. The rest almost falls into place.

    This dude's channel is what inspired me the most (I stumbled upon him searching automotive stuff haha):https://www.youtube.com/channel/UC0h...Bpt2r6fyus4GWQ
    Last edited by Kp61dude!; 03-14-2019, 02:35 PM.
    PcDmis 2015.1 SP10 CAD++
    Global 7-10-7 DC800S

    Comment


    • InspectorJester
      InspectorJester commented
      Editing a comment
      I will look into it.
      EDIT:
      With the very little research I've done here at work today, PLC is the way to go for full-fledged enterprise robots.
      I have several Raspberry PI's and Arduino's laying around the house, so I'll figure where I can go with that...
      I even have the CAD file of the rotary table used on my machine, so there's plenty of possibilities.
      In the coming months, I will update with progress.

  • #5
    Nevertheless I think you should still loop this process for the time being. I was on the same boat as you but only had to inspect safety critical characteristics on fairly simple parts. Was a challenge to do it all in PcDmis but worked well and really maximized my output.

    EDIT:
    I used double sided tape and the grid plate the CMM came with, again the parts were simple enough to where I only used a A0B0 tip angle and about 8 tight tolerance features for each part. There were thousands!

    Company went from inspecting 30pcs in a 10hr shift to pumping out 300pcs in the same amount of time. All with no budget, little to no support, and a twice a day nagging Qc manager requesting ECD!!! I would just say hurry up and wait, let me finish this, I don't know how much longer but you'll want to kiss me when I'm done.

    I received faith and support after that, even the raised with a silver spoon owner came down to check it out. The program is still used today tough no one there understands my spaghetti code and neither would I...I was so new at loops hehe so there aren't many parts that are looped other than this one and one other one that I made before I departed. Management can't see passed the horizon and you can't convince them that there are sunny skies and green grass either, sometimes you just have to do it behind their backs (being super careful of course) and then show them. My 2 cents.
    Last edited by Kp61dude!; 03-15-2019, 08:49 AM.
    PcDmis 2015.1 SP10 CAD++
    Global 7-10-7 DC800S

    Comment


    • InspectorJester
      InspectorJester commented
      Editing a comment
      I'd love to, but I'm no engineer, and they won't spend money for a fixture
      EDIT:
      This is not a typical thing for us; typically I'm just a programmer, but for the last few weeks I've been an operator!

  • #6
    We have made many (both simple and advanced) automation cell installations during the years, and my first advice would be: Don't try to do it by yourself - there are too many things that can (=will) go wrong, there are many things to learn, there are safety aspects involved, etc.

    With that said, I'll also have to agree with everything Kp61dude! has said - you can do it, but it will be LOTS of work!

    My second (and third, and 4:th, ...) advice is: Start small! Go for small steps. Make small things that speeds up the process a little - a flexible fixture, a support stand so you don't have to do so much lifting, etc. Wait with the 'complete solution' - your first attempt will almost probably be wrong, so don't spend too much time on it.
    AndersI
    SW support - Hexagon Metrology Nordic AB

    Comment


    • #7
      These might be able to do what you're looking for.... Don't have any experience with them, just thought they sounded cool....

      https://www.rethinkrobotics.com/

      Comment


      • #8
        DIY? Buckle up, Buttercup.

        Robots that can place a part with the precision we need are super expensive. Right there your idea is shot down by mgmt.
        On the off-chance that your idea is NOT shot down by mgmt., there are two big hurdles that have not been mentioned in this thread:
        A. Technical issue of Safety.
        B. Non-technical issue of Safety.

        The Technical Safety is how the hardware is fabricated and the communication programming is done to make it near-impossible to injure a human, what with prox sensors on the mandatory cage enclosure doors and light curtains, dwell times, double-checks, all input being fed into the PLC control before automatic motion is allowed to occur.

        The non-technical is all the paperwork and blessings from HR, plant safety, lower mgmt., middle-mgmt., upper mgmt., and all contractors involved - and a significant amount of this is explaining chunks of the technical stuff to non-technical people over and over again. Do not underestimate the time resources this will drain, or the medium-high probability of the whole project being scrapped due to some little detail after mgmt spends a buttload of money.

        As CMM guys and gals, we are used to running automated equipment that can barely injure us and is pretty much guaranteed to never kill us. Robotic arms are a much more dangerous game and require a different approach.

        For what it's worth: Hexagon sells versions of their controllers built for automated cells, the Safety Controller, which comes ready for integration into pro-level industrial PLC systems.
        This isn't required but it hugely simplifies the process.

        Comment


        • Kp61dude!
          Kp61dude! commented
          Editing a comment
          Oh yeah that ^^^^. It's easier to "home-brew" stuff in mom and pop shops and get away with as chances are you're not only the CMM guy but also HR. That being said take safety very serious as what I think Ego Murphy is trying to convey here along with other points.

      • #9
        well i have a lot of the same ambitions as you it sound like and after about 2 mounts at my new job i was tasked with coming up with a way to check a 8 dimensions on 100 parts in a row wile a customer was in the building the issue i was having was part orientation so after figuring out at what angle i needed the parts to be and what kinda goofy star probe i would need to get at all the holes with only on probe rotation i opened up solid works and got designing a fixture i could 3d print at home on my cheap ebay 3d printer that i had been playing around with with my son and his nerf guns. any ways the project went off with out a hitch and my boss looked like a supper star for hiring me lol and the owner of the company a week later come up to me and said he was real impressed and to pick a 3d printer out that i thought might work out well for CMM fixtures a week late it was on order and since i have been giving some freedom to at least per sure new ideas such as automation of some of our production parts. agreed this is one deep hole but at the end of the day i think this is where our path as programmers is heading . as for a robot arm there are alot of one out there that can be 3d printed on a 200 dollar 3d printer and with a close loop stepper system it will be plenty of accurate for what you want to do at least thats what im planning on trying . good luck and hope this thread grows

        Comment


        • #10
          Very interesting thread. Bless y'all for reaching to the stars! At the last company was at, we had a vendor (one of many) who supplied medical devices. These guys ended getting into robot/automation pretty seriously. IMHO, these guys were always a shade brighter than our other suppliers. Seemed to me that they had an excellent automation project approach:

          https://lowellinc.com/wp-content/upl...hite-Paper.pdf

          (hope Jim doesn't mind the link)

          Comment


          • Kp61dude!
            Kp61dude! commented
            Editing a comment
            Oooh yeah, reading material, thanks for sharing!!

          • sealevel
            sealevel commented
            Editing a comment
            another sicko soooo bored..... picks up a white paper to gander!!

        Related Topics

        Collapse

        Working...
        X