logo elektroda
logo elektroda
X
logo elektroda
Advanced Search
logo elektroda

DIY // CNC welding robot

maciek95k 35283 37
ADVERTISEMENT
Treść została przetłumaczona polish » english Zobacz oryginalną wersję tematu
Reply Cool? Ranking DIY | New topic
Notify about new articles
  • DIY // CNC welding robot
    Hello
    After a long time, I finally gathered myself to collect all the photos and videos of my project in one place and publish it on the forum. The construction is about two years old. I did the job when I was in the 3rd class of technical secondary school in mechatronics profile. Let me mention at the beginning that this is my second project of this type. The previous robot was made of plexiglass, controlled by arduino, and the actuating mechanisms were servos. Below is a photo of the previous project:
    DIY // CNC welding robot

    But getting to the point. The robot that I have built, and in fact I am still building (although I have not done anything with it for a year, I do not consider this project finished) was not created for any specific purpose. As I mentioned earlier, I did something similar for the Olympics in the 2nd class of technical school and it caught me so much that another construction was created and I have another one in my head. Basically, not knowing why, I did not follow any robot.

    All the mechanisms that are there are recovered, and more specifically in the junkyard, I found used medical equipment from which I twisted out the more interesting elements and replaced it with a few old burnt power tools. Due to the lack of any large resources, everything was rather done as it was.
    The whole structure was created so that as many elements as possible could be ordered to be burned on the laser.
    All moving parts are mounted on ball bearings, thrust bearings or bronze bushings and it is possible to replace them.
    The robot has 4 degrees of freedom, and the possibility of expanding to 7 if I add another controller and currently two relay outputs. As I did not have elementary knowledge of inverse kinematics at that time, the whole thing is controlled on the Teach In principle. The robot is controlled by the Mach 3 program. After starting the whole operation, drive each axis to the limit switch and reset the readings in Mach 3. Then we drive to the desired points one by one and enter their coordinates with GCODU.
    I know that this is not a professional control, but the only one I was able to handle and adjust at that time.

    As for the control system, I use a PC for this purpose. It is connected to an electrical box with a transformer, a 4-axis CNC controller, and a start-stop contactor relay system. Then, by means of high-class shielded cables, the whole is connected to the manipulator.
    The structure is quite complex, so it is impossible to describe the whole in just one elaborate. I am happy to answer any questions. Currently, I am a student of Automation and Robotics, so I am planning another construction in which the programming interface itself will be much more user-friendly. In addition, I have an idea for several improvements, such as automatic maintenance of the welding tip during operation as well as the systems for detecting "sticking" of the welding wire in the tip. The welding tip itself is a manual holder that has been reworked (shortened). The robot works with any MIG welding machine equipped with a Euro connector. So far, I have not done any serious welding tests with it because, as I wrote, it is more of a hobby work.
    Here is a photo of the first welding trials:
    DIY // CNC welding robot
    And a video from one of the robot's exhibitions. He then welded several dozen samples without any failure. Since I did it once, I had to fix one February and that's it.
    [movie: 4204f6cd01] https://filmy.elektroda.pl/33_1456778198.mp4 [/ movie: 4204f6cd01]

    This work was presented to many different interesting competitions and it took quite good places, but I know that it is not yet the top of the technology, so I would like to hear criticism and good advice;

    DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot
    DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot
    DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot DIY // CNC welding robot

    Cool? Ranking DIY
    Do you have a problem with Arduino? Ask question. Visit our forum Arduino.
    About Author
    maciek95k
    Level 10  
    Offline 
    maciek95k wrote 11 posts with rating 91. Been with us since 2016 year.
  • ADVERTISEMENT
  • #2 15485632
    jak.sze
    Level 2  
    Hello,
    I am full of admiration when it comes to the idea and execution, but I have a question, are you not afraid that bare wires may be damaged during welding - e.g. by splinters?
  • #3 15487231
    Soldier_Bk
    Level 10  
    The idea is cool, especially that when welding with a robot, the weld accuracy is higher (from what I saw, the presented sample looks quite professional)
    I think an interesting idea for the future would be a robot with a plasma cutter head.

    Even such a: DIY // CNC welding robot
  • #4 15487668
    maciek95k
    Level 10  
    jak.sze wrote:
    Are you not afraid that exposed wires may be damaged during welding - for example by spatter?

    As for that, I had it in mind, although the robot is not used for any works as I wrote, only exhibitions and presentations. Of course, if he was to work at a welding station, it would be necessary because dust has a negative effect on moving parts and it would be necessary to cover the wires. However, as I traveled with it to exhibitions and competitions, I left it in this form because you can look at its construction nicely. As for the covers themselves, there were two options. The first is to get some 3D printers and make them quite professionally so that they are not only functional but also look good. The second, however, is to sew them out of some fireproof material and pull a flexible sleeve over the whole thing. It is possible that the next incarnation of this design will already have something like this.

    As for the plasma torch, I did not think about it to be honest and I even have exactly the plasma as shown in the picture. Certainly an attention worth considering because together with the rotary table you could get a very interesting effect.
  • #5 15488733
    _JAG_
    Level 15  
    Very nice designed, but if not it was easier and cheaper to try to drive wiper motors connected directly to the axis of rotation.
    It is very easy to convert such a motor into a servo drive.
    The system still requires setting the starting point before starting work.

    At the burner itself, the axis transverse to the last axis is missing.
  • #6 15488801
    rcnitros
    Level 16  
    Nobody would despise such a construction, certainly not me. In my opinion, the design is flawless. The welds are of average quality, but it's probably too fast arm movements.
  • #7 15488889
    maciek95k
    Level 10  
    The welds you see there are just a sample of his first welding test.
    I was able to sort it out later. However, the problem with the structure is the lack of proper control, which means that the movement between the points is not perfect in a straight line. When I was trying to weld two flat bars, the robot easily connected them at the beginning point-wise, but when he was already making a weld along the length in the middle, the end, unfortunately, was a bit too close and the material was melted to the other side. This could be corrected by adding more intermediate points, unfortunately there was not enough time for such tests. Currently, I had to leave the structure at home because of my studies.

    As for the wiper motors, I would also need servo drivers. They are not so cheap anymore and the encoders themselves are also costs. The whole thing would be much less stiff. Here I have motors with 200 steps per revolution, plus microstep control. The driver has a maximum step split of 1/16.
    In the next version, I also plan to use stepper motors, but with worm or planetary gears. The robot will then have a much greater range of motion. I also plan to add my own electronics to the control box based on arduino or something similar. Then, after starting, the robot will be able to calibrate itself before it sends a signal to the computer that it is ready for work. In fact, I have already done some tests with arduino and made a small module connected to it, thanks to which it was possible to control the robot without the need for a PC. However, you were only trying so I didn't write any specific software for this purpose.
  • #8 15490403
    marcinsud
    Level 11  
    Instead of Arduino, I would use Texas Instruments launchpad in the next version. The price does not deter much larger options. You can program in an environment identical to Arduino. There are models that are immediately dedicated to controlling drives.

    I admire the design and enthusiasm.
  • ADVERTISEMENT
  • #10 15490826
    maciek95k
    Level 10  
    I have seen this project for a long time. However, there is no question of any accuracy here. It's once. Two are the same DC motors and encoders okay but not from the wipers in my opinion. The accuracy of the gears is negligible. I also do not know how much strength this robot has. I have a typical engine mount, so I can successfully replace them and the arm lifts about 3 kg when fully extended. In addition, control of the mac, so I have a program-configurable slow start and a slow stop.
    Such self-made servos cannot be accurate because at different speeds the rotor remains inertia and does not stop in place when power is removed.
    In addition, there is an option to remove the backlash.
    Therefore, I will probably continue to use this program as well as stepper motors, but I will write a second program for it that will convert the code written in Cartesian coordinates into the code for the robot's kinematics, which mach does not need to understand anymore. It will only be used to send data to the driver.

    When it comes to launchpad, I didn't know this platform. So far, programming is still in my infancy and I have an arduino, so it will probably sit there, if only for the sake of minimizing costs, but over time I will think about a change. Generally, the controller that I have has inputs to the limit switches that can be easily programmed in Mach, so the module itself will rather perform simple functions and send a low or high state to the limit switches.
  • #11 15492579
    michal.rogaczewski
    Level 28  
    maciek95k wrote:

    Such self-made servos cannot be accurate because at different speeds the rotor remains inertia and does not stop in place when power is removed.
    In addition, there is an option to remove the backlash.


    You are probably a bit confusing the concepts, a servo based even on wiper gears can be quite accurate. The precondition, however, would be a more precise encoder. The rotor obviously doesn't stop in place, but it's a matter of using the right governor for each degree of freedom (which is what any commercial servo driver does, anyway). For even greater accuracy, you could play with control with two encoders, you could then fight with backlash on the gear. This, of course, requires much more work and knowledge.
    Additionally, the worm gear has several useful properties in such an application.

    For you, backlash is rarely a problem because forces are applied in most joints in one direction anyway, and this design does not provide for a working range where this direction could change.

    As for the robot itself, the design looks professional, although it probably lacks 1 degree of freedom to be really useful for welding.
  • #12 15496080
    maciek95k
    Level 10  
    All in all, you are probably right. However, as I wrote, the main goal was to use what I have to minimize costs. I had the controller at home, and so did the motors. Another thing is that when I was building it, my knowledge in this topic was just getting started and I didn't really know what it was. I have never worked with servomotors, so I do not feel confident enough in this topic to use such a drive for now.
  • #13 15496476
    Freddy
    Level 43  
    maciek95k wrote:
    Such self-made servos cannot be accurate because at different speeds the rotor remains inertia and does not stop in place when power is removed.
    In addition, there is an option to remove the backlash.
    You're wrong, buddy. The accuracy depends on the encoder. The Mach backlash resetting option is used to "resetting the gear backlash".
  • #14 15496816
    maciek95k
    Level 10  
    Like I said, I have never had the opportunity to work with servo drives. How does something like this come out? Is it possible to build only the actuators by yourself, as with a good drive based on stepper motors? Since the next robot would have a greater mass, the drives would have to be much stronger.
  • #15 15501795
    Falco76
    Level 25  
    Let go of the welded function. This is such a complex topic that you won't get it.
  • ADVERTISEMENT
  • #16 15504345
    Tomekkoko
    Level 15  
    Congratulations on your project. Have you ever had contact with professional 6-axis robots such as fanuc or abb? Maybe you can see something in the way of control and construction.
    I recommend the maintenance manual, for example of IRB robots, you can learn a lot of interesting things.
    Are you planning an external manipulator? I understand that you steer the robot "linearly" with coordinates, and what if you want to move one of the axes?
  • #17 15504565
    wik12
    Level 14  
    Nice design, but when welding you need the ability to rotate the welding torch so that it can reach the welded material at different angles. Therefore, the last member in the robots is rotatable (about its axis) and the torch is attached at an angle to the axis of the motor.
  • #18 15504791
    TMK_M
    Level 22  
    Congratulations on the project. I have a few questions.
    What does your coordinate system look like? Do you have any idea (program) to generate gcodes or do you write them by hand?
  • ADVERTISEMENT
  • #20 15512891
    maciek95k
    Level 10  
    Falco76 wrote:
    Let go of the welded function. This is such a complex topic that you won't get it.

    Something more? Is it just not because no? I have a fairly large understanding of welding and I am well aware that it is a very complex technological process, but the robot is built more for fun and the joy of creating. However, if you have any knowledge about automatic welding that could be useful to me, I would like to hear something because I have not dealt with welding robots.
    Tomekkoko wrote:
    Have you ever had contact with professional 6-axis robots such as fanuc or abb?

    Unfortunately. The first and only robot I had contact with was the one I built for myself.
    Tomekkoko wrote:
    Are you planning an external manipulator?

    External? Ie? Because I don't really understand the question.
    Tomekkoko wrote:
    I understand that you steer the robot "linearly" with coordinates, and what if you want to move one of the axes?

    Control is a problem with this system. The drives are not entirely linear (planetary / worm gear) would be better so unfortunately the movement between points is almost always a slight arc. More points can be added to increase accuracy. However, each axis is programmed and controlled separately. I would like to make a system that will connect these axes. However, I do not have enough knowledge myself, and even here at the university, not everyone is interested in helping "charity", so I keep trying :)
    wik12 wrote:
    Nice design, but when welding you need the ability to rotate the welding torch so that it can reach the welding material at different angles.

    I know. In general, the title is a bit confusing because the project was rather focused on building a robot. It was supposed to be a 4-axis robot with general purpose. Until one day, my father burst out with the text "fasten a welder to him" and that's how it stayed. However, the design is still evolving. I plan to add one more 3-axis driver to the control box, so I will be able to control 7 axes or 6 in the case of waving.
    TMK_M wrote:
    What does your coordinate system look like? Do you have any idea (program) to generate gcodes or do you write them by hand?

    As I wrote above. I start the robot. I drive up to base points. Then I reset the Machu readings. Later, I drive to a given point and take its coordinates from the readings in Mach and enter it into Gcode. I have only been learning programming for a few months, so a simple conversion program is planned.
    nelik1987 wrote:
    Cool design, by the way, very similar to mine from years ago

    Really! I didn't even know there was something like that on the electrode. And the constructions are really very similar, although yours has a greater range of movements.
  • #21 15515598
    Falco76
    Level 25  
    Quote:
    Something more? Is it just not because no? I have a fairly large understanding of welding and I am well aware that it is a very complex technological process, but the robot is built more for fun and the joy of creating. However, if you have any knowledge about automatic welding that could be useful to me, I would like to hear something because I have not dealt with welding robots.


    I wrote this because the title of this post is "welding robot". Since you know a lot about welding, you should come to these conclusions yourself. The photos you have attached are not welding tests, but attempts to make a weld on a flat surface. Watching the project (where I am full of praise), the degree of advancement of the robot, control, no software for generating a gkod (welding) etc ... I would not call welding. Have you already tried to use a robot to weld (I will highlight WELDING!) Two pieces of fabric? As you have noticed, you must have encountered problems when trying to put the weld seam on (as can be seen in the photo). Maybe I'm too stupid, but explain to me how you want to automate it. How do you get it: generating paths (groin, butt ... that's enough for now) along the curves, how you can grasp the feeds, as in the case of sticking the wire to the material, ... and a lot of other problems related to welding. You can only use it for punching or laying short sections of welds in a straight line.
    If you know how to handle it, you are great.
  • #22 15515622
    wik12
    Level 14  
    Each robot in production plants is incomparable to the presented project, which is only a hobby curiosity.
    80% of the robots are powered by stepper motors and planetary gears.
    They are equipped with a mass of sensors detecting overloads, collisions, etc.
    Welding robots are built in such a way that they are not damaged by welding spatter.
    I have the impression that the author is asking for ideas on how to develop it as an amateur, on an amateur level, and not to compare it to professional machines.
  • #23 15516856
    Tomekkoko
    Level 15  
    maciek95k wrote:
    Tomekkoko wrote:
    Are you planning an external manipulator?

    External? Ie? Because I don't really understand the question.

    It is a penden like this one, for example: DIY // CNC welding robot
    It would certainly improve the manipulation of the robot "live"
    Probably every robot has the possibility of linear motion and "axis joging" is also a big improvement.
    I would go to the service from some pandent, it may be, for example, a tablet and a joystick.
    I connect to pc as needed to rewrite the program. It's faster and more convenient to modify from a pandent
    An even more interesting thing is to move the robot "by hand", like here:


  • #24 15521487
    maciek95k
    Level 10  
    Falco76 wrote:
    Have you already tried to use a robot to weld (I will highlight WELDING!) Two pieces of fabric?
    maciek95k wrote:
    When I tried to weld two flat bars, the robot easily connected them at the beginning point-wise, but when he was already making a weld along the length in the middle, the end, unfortunately, was a bit too close and the material melted to the other side.

    Falco76 wrote:
    how do you handle the feeds, as in the case of sticking the wire to the material

    Personally, it has never happened to me that a wire stuck to the material. Unless a really strange move is made, but it is always due to human error, so UNLESS (I am not sure) in the case of a robot, this problem disappears. However, the sticking of the wire in the welding tip is different. I would like to make a simple system that will detect this. This can be done on the principle of a clamp meter. If the wire gets stuck, it immediately results in a zero welding current (wire break). The system gives a signal to the limit switch that can be programmed in the machine and the robot will stop signaling the problem. Another thing is that I would like to add a maintenance system. Automatic lubrication every few cycles plus spraying the welding nozzle with an anti-sticking agent. Of course, this will also include the compressed air system.
    I also have an idea to add gas cylinders. As a result, I will be able to greatly relieve the engines and reduce gear ratios, so I will get higher speeds.
    In the next work I would like to connect the elements on ball bearings with large diameters (about 100mm). Backlash-free connection and in addition, using planetary gears, I can place the drives in the axis and route the cables.
    Falco76 wrote:
    How do you get it: generating paths (groin, butt ... that's enough for now) along the curves as you embrace the strokes,
    wik12 wrote:
    I have the impression that the author is asking for ideas on how to develop it as an amateur, on an amateur level, and not to compare it to professional machines.

    That's it. Relax. I don't swing my head in the clouds and I know that even in 20% I will not come close to professional constructions. However, it happened to me that I had orders to perform where hundreds of the same elements had to be welded. All these were elements that were not part of any serious structure and did not carry heavy loads. So by devoting time and adapting it to one less responsible task, we can make our work a little easier and gain more aesthetics. But these are just plans for a very distant future.

    Tomekkoko wrote:
    It would certainly improve the manipulation of the robot "live"
    Probably every robot has the possibility of linear motion and "axis joging" is also a big improvement.
    I would go to the service from some pandent, it may be, for example, a tablet and a joystick.

    Certainly the idea is very good and would be useful. Yesas I wrote, I am just at the beginning of learning to write programs and in this case, the biggest challenge would be to embrace it programmatically, everything is still ahead of me.
  • #25 15521791
    Tomekkoko
    Level 15  
    maciek95k wrote:
    I also have an idea to add gas cylinders. As a result, I will be able to greatly relieve the engines and reduce gear ratios, so I will get higher speeds.

    ABB uses these in "large" over 30 kilos "payload".
    However, I would go in the direction of more efficient drives. The actuators will compile the structures significantly.
    For example, how would you like to synchronize the piston's movements with the engine?
  • #26 15522944
    maciek95k
    Level 10  
    I meant gas actuators such as are used, for example, in cars to support the clown. I would add them only on the first segment to balance the mass of the entire tonearm which rests unnecessarily and only loads the gear.
  • #27 15523116
    TMK_M
    Level 22  
    From my experience with cnc plotters, I advise against these gas carriers. They break down quickly in this application and it is practically impossible to regulate the force. It is better to use a pneumatic cylinder with a reducer, then you can accurately set the force, but the downside is the need to supply compressed air.

    I suggest to use a servo driven with a brake of higher power together with a large gear. The actuator in professional robots, such as KUKA, is only on one axis and is used to relieve the load as a robot carries fifty kilograms.

    DIY // CNC welding robot
  • #28 15523235
    maciek95k
    Level 10  
    I see. I am also interested in how the rotation mechanism is built in such robots. I was thinking about making a large base and giving a large thrust bearing around the entire circumference to give me a lot of stability. Unless, of course, I find such a bearing. Anyone here has contact with such robots and could someone tell me something?
  • #29 15524544
    TMK_M
    Level 22  
    DIY // CNC welding robot

    You can't see much and I don't have time to take off the rest of the covers to see how it is built, but you can see that there is one large bearing in the center.

    DIY // CNC welding robot

    The servo drive is mounted vertically and it looks like the planetary gear runs on a gear wheel mounted on the side of the fixed base.
  • #30 15536240
    cooltygrysek
    Conditionally unlocked  
    Hello, the weld is made of poor quality and caused by a rather bad angle of the wire advance in relation to the material being welded. Please read about welding, in particular about the types of seams and the ways of conducting them, as well as the speed of wire and pipe feed. The selection of the appropriate voltage and current is also of great importance for smelting. As for me, the robot is great for DYI. How many kg can the arm lift? It would also be worth using a small counterweight on the rear axle of the arm, which would improve its movement stability.
Reply Cool? Ranking DIY | New topic
Notify about new articles
Report a violation of the law

Topic summary

The discussion revolves around a DIY CNC welding robot project, initially built by a technical school student. The robot, which is still under development, was designed for exhibitions rather than practical welding applications. Participants express admiration for the design while raising concerns about the robot's wiring safety during welding and the quality of welds produced. Suggestions include using plasma cutting capabilities, improving control systems, and considering alternative motor types like wiper motors or stepper motors for better accuracy. The conversation also touches on the importance of proper welding techniques, the need for a more robust control system, and the potential integration of sensors and encoders for enhanced functionality. Participants share insights on professional robots, including brands like ABB and KUKA, and discuss the complexities of welding automation.
Summary generated by the language model.
Popular Topics
ADVERTISEMENT