Single Axis Motion Control Project

J

Thread Starter

John Jackson

Good Morning. I have been following the discussions of this project for several weeks. It has my sincere interest. However, I would appreciate some direction on just how to fully understand what is being attempted. I would like to explore the PuffinPLC philosophy on a single axis motion control project that I am working on. Any advice you can provide will be greatly appreciated. Thank You. John Jackson [email protected] _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
Hi, What type of axis control are you working on, and how extensive is it? My personal opinion is that the controller would be best built in small pieces that can be interchanged or left out. For example, maybe one set of functions would control the motion profile / setpoint scheduling. Another module would deal with the PID feedback loop (you will find a PID function in the CVS tree). If somebody wanted something other than a PID controller, then they could swap it for the other routines easily. Hugh
 
J

John Jackson

Hugh Thanks for your response. I have been out of town and am behind as usual. As for the single axis project, it is a linear positioning system. The goal is to experiment with Linux on this project to prove its viability in a rather low tech industry. I have been doing similiar work for 15 years. Fortunatly, Windows is just now starting to become a factor and I am taking on the responsibility to not use it. My initial work was in Unix, year ago, so I am confident with this under taking. In brief, the postioning system will be under PC control, for its data source. Typically the data is sent via network, however I will not rule out floppy disks for this project. Afer receiving data, the Batch is processed on a local PC at the positioning device. A single axis is all that is necessary because we are doing a cut to length application. I want to do as much of the motion control as possible with software, and use minimal control hardware - only where necessary. Hope this gives you a little more insight into the project. Thanks. John Jackson _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
Hi, This sounds like a good concrete application. If you are proving Linux you might just want to go ahead and implement the system in C/C++ for now - but steal code :) Here are some thoughts on this application: 0. (MAJOR HINT): If you are proving this to your management you might want to keep it as simple as possible so that it works well, and then they will let you do it again. 1. The network access to data should be easy if you know the format of what is being sent, and the protocol. You can get some code for network access from my LPC project at http://claymore.engineer.gvsu.edu/lpcd. I will give you help if you need it. 2. If you can use steppers the application will cost a bit less (1-2KUS$), and be easier to control, but probably be slow. 3. If you use a servo controller (2-3KUS$) you can send setpoints to it, and it takes care of the rest. You controller then becomes more of a translator between the network and servo drive. 4. Depending upon the approach taken a simple DAQ board (500-1000US$) might provide all the IO you need. 5. If you need high speed response you might also want to use a realtime kernel - RTLinux or RTAI. Hugh _______________________________________________ LinuxPLC mailing list [email protected] http://linuxplc.org/mailman/listinfo/linuxplc
 
M

Mario de Sousa

Top