> However, I started to wonder last year what's the future of LinuxCNC? > What concerns me most is the computing part mainly based on silly old PC > architecture from the 80's. Using PCs for that purpose made sense in the > early 1990s to about 2010 but now it's time to move on IMO. I don't see > much usable discussions on this mailing list, sorry guys.
I'm still in the middle of my CNC conversion. I'm far enough along that the motors are mounted and now I'm working on the power supply cabinet. Control has aleready been done with MachineKit on a BBB. Here's a video of two axis control published last year in May. I'm not much further along other than the third axis(X) now has a motor mount and I've created motion on all three axis using a surplus PC on which I've booted both MACH3 and LinuxCNC. I have a new dual parallel port BoB for it. The interface board I was using for the Beagle expected NO limit switches and I'm uncomfortable with that concept for machine control. Anyway here's a link to the BBB with MachineKit short video. https://youtu.be/9GF709ZfLRQ This weekend I'm working on the power supply cabinet and a place to mount the PC screen, keyboard and mouse. There's a pull out drawer for a surplus PC or I suppose a BeagleBone. Haven't decided ultimately what I will use. During the last year I also acquired a harmonic drive and STMBL servo controller but I still need to cast the mount which means I still need to finish the 15 lb crucible all for a fourth axis. There's no end to projects to make the project. I won't get into the Linux verses Windows religious discussion. I use windows because my job requires I use development tools that aren't available on Linux. So there's no discussion really. Dedicated hardware is an interesting question. I designed and built an controller for old lathes that were missing gears. Called an Electronic Lead Screw (E-Leadscrew group on Yahoo) it uses a small PIC18F series processor, has 2 line LCD display, MPG knob and 35 key buttons. And before the stepper driver devices became really expensive also a 3A 48V micro-stepper driver for the Z axis (carriage). Single pulse per revolution sensing means that it needs a very stable spindle speed. The user interface is less CNC like and more manual lathe like. I use it on my 1942 South Bend with only single axis control and can cut both metric and imperial threads and safely bore a hole to depth without hovering over the half nut. I mention the ELS only because it's a dedicated machine controller and there have been requests to improve it. And there's the problem. I could use a BeagleBone with Machine Kit and build a new interface. I could use a PIC32 mounted on a small carrier board to plug in and replace the PIC18F 40 pin DIP processor. I even started down that road with modifications to my Gingery home built lathe. I've 3D printed mounts for 3 sensors and a slotted tooth sensor for the Beagle to better capture spindle speed and position. But like the CNC conversion of my mill these projects have all been stalled due to lack of time. The problem, and sorry to take so long to describing this, is that the market has pretty well dried up. Those who have CNC on their mills are using it. Those who don't may dream of it and some do upgrade but very few now. If someone has a working conversion in their shop running MAHC3 on a WIN-XP single core processor why change just for the sake of change. Especially if their desire it to make parts, not play with PCs. There are probably thousands if not 10s of thousands using MACH3 (not the newer MACH4) for their CNC routers in their home shop. Similarly if someone is running LInuxCNC on working hardware, why change. Why upgrade? To what end? Then there is the Far East solutions like the one in this link. https://www.alibaba.com/product-detail/Low-cost-New-Product-3Axis-4Axis_6034 3603384.html?spm=a2700.7724857.normalList.37.3701292eTmjk5p As I understand it there are lots of systems like this out there. I suspect, that like my ELS, the _time_investment_ to make a LinuxCNC something on dedicated hardware only slightly better just isn't available anymore. If I spent another 1000 hours redeveloping my ELS with a 32 bit custom processor or ported BBB MachineKit to make a better ELS the question is how many more would I sell to recover the say 1000hr * $100 per hour consulting R&D costs? I think the support on this group is awesome. I suspect though that none of the members have the time nor interest in re-inventing a wheel that is already turning. What does one do with that 4 hours of free time. Take their CNC machine apart to try something new? Or just go out and make something with what they have and what is familiar? John Dammeyer _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
