I have an application where I am required to move a component weighing 2kg in all the 3 directions along a fixed path (available in a IGES file or 3D DXF File), which involves motion of component in 3 directions + rotating the component.
First Idea was to build a suitable multiple axes CNC system but after looking at Lynx robot videos, realised that the robot arm which already has all these motions would be the ideal solution.
I have a budget of around 5000US$; but when I checked a few of the industrial robots, they are certainly a lot more than my budget.
So I was wondering if I could build a robot using the controller used in AL5D robots. I am new to building robots, but I have a good experience of programming for PIC microprocessors, I have built my own 3 axis CNC controller and already have over 300+ dot pin marking machines being used in the industry which use stepper motors and my designed controller and software.
Tried first to go through all the forum topics… went through a few, found out that the components used in construction of the AL5D are not suitable for 24x7 working. Can I make a robot arm using industrial duty DC servo motors and use the Lynxmotion controller and software to control it?
WOuld like to have your guidance on this.
Moving 2kg 24/7 will require a heavy duty good quality setup. I’d say a hobby servo based arm probably won’t get the job done. You may want to look at a linear actuator based arm that would duplicate the hydraulic arms on large excivators and such. Pots on the arm joints, an ssc-32 for the I/O work, and a computer for the brain might get the job done. just some quick thoughts.
Still sounds like a CNC project to me. Way too much weight for a hobby arm. Check out the CAD_CAM_EDM_DRO list on yahoo. You can put together a pretty good 4-axis CNC for that.
That is precisely the thought behind this post. Want to know whether it is possible to do this way.
I can make the mechanical arm links, use industrial duty stepper motors or servo motors and use the SSC-32 controller for controlling the robot with the PC acting as the brain and also for storing the standard paths that I intend to use.
Is this really possible and has any one done it already?
Well, I don’t know anybody who has done this as this is generally a hobby forum and most stick to gizmos that use regular hobby servos. From what I’ve seen, industral servos operate differently than hobby ones and are in a different world. Industral robotic arms operate reliably 24/7/365 which results in the high cost to make them. Fairly often persons come to the forum wanting to duplicate a $50k arm for $5k or even $500. That probably ain’t gonna happen. The old land speed car builder Mickey Thompson use to have people come to his shop wanting him to build them a car so they could set a new speed record. The first thing he would tell them was “Speed cost money. Now, with that understanding, just how fast do you want to go?”. You should be able to build an arm that does what you want it to do, but the expected life expectancy will probably be much shorter than a very $$$ industral arm. With that understanding, below are some links to look at for using linear actuators. I’d start by making a full size mockup of your desired arm using cardboard or something similar. Using that decide what and how the motor/actuator will fit to the arm. If you decide on linear actuators, then purchase the needed components for one and do some testing to verify you can operate it as desired. Servo city makes 6v linear actuators for $299, but you should be able to something similar for much less $$$. I think using a combination of the HS-422 servo, the multiwatt15 L298 H-bridge, and a 12V linear actuator for each arm joint, you should be able to make your arm.
Hmm, well I was thinking more like the Tonegawa industrial servos or linear actualtors such as the Firgelli ones. ServoCity does sell some now but… I’m not sure how they justify being 2x-3x the price. You may want to make certain the Firgelli and/or ServoCity ones are actually rated continuous duty as they are both catering towards hobby customers/applications more so than industrial customers. $5KUSD is not a lot to work with for a 24/7 duty arm… the real danger being you could eat that much or more per year in keeping it going if you are not careful about what you design. Remember that if you put this online and it breaks down there is usually a cost of NOT doing the task it was assigned while being repaired which needs to be considered in the total budget.
However commercial/industrial arms usually use DC servo motors that can be controlled close loop (I don’t know about the Tonegawa) And with gearing, timing belts, cables or similar means of transmitting the torque up the arm (keeps supported weight to the minimum). The arms that I have designed and built have several motors near the base, and transmit their torque via timing belts.
Did you say what you need to position/move? a 3-axis mill setup with a rotary table can probably handle more weight then an arm of comparable size, and would be easier to build using linear rails, leadscrews, and stepper/servo motors.
This forum normally talks about hobby Robots and not industrial Robots : “Can you guide me to a forum which talks about Industrial Robots?”
One can use the SSC-32 and PC provided the Servos accept PWM Control Signal… and the standard DC Servos do not use this PWM Control Signal: “Can anyone Guide on any alternate similar controller with software to control an robot using closed loop DC Servos?”
Alan, I am also thinking of building a simillar robot and I saw a good video of its construction at youtube.com/watch?v=tkDbmWAyHYw . What sort of controller do you use for this robot. I have started doing the CAD work for the mechanical structure, will build the complete structure and then choose the servos depending on the actual weights of the arms. Eager to have tips from you about the points to take care in building this robot.
Well, regarding CNC, I can anytime go back to building the 5 axis CNC, which I am most comfortable with, but programming the path of component movement is going to be tricky as it involves a lot of twisting and turning of the component, which would be very easy in the robot (teach mode).
That’s a Lab-Volt arm, there were many similar arms built for teaching. The motors can be either DC servo or stepper motor. Appropriate motor controllers are needed for each. In the commercial, CNC and motion control arena, these motor controllers or motor drivers would often be driven by step/direction pulses. these pulses would be created by a controller program, which could run on a PC.
You may want to make certain the Firgelli and/or ServoCity ones are actually rated continuous duty as they are both catering towards hobby customers/applications more so than industrial customers. $5KUSD is not a lot to work with for a 24/7 duty arm… the real danger being you could eat that much or more per year in keeping it going if you are not careful about what you design. Remember that if you put this online and it breaks down there is usually a cost of NOT doing the task it was assigned while being repaired which needs to be considered in the total budget. 