Re: [Emc-users] EMC a digital pantograph milling machine?
Hi, thanks for you comments on my project idea. I played with emcrsh a litle today and would like to show you the result: A copy of a tee spoon milled: http://128.130.120.105/_DSC5204.lowres.jpg emcrsh is not the ideal solution though. I guess it'd bee better to link the digitiser directly to HAL but I have to find a way to access the MicroScribe from Linux. Flo -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
Neat! Just out of curiosity, what were the problems with using emcrsh? On Nov 30, 2011, at 10:18 , Florian Rist wrote: Hi, thanks for you comments on my project idea. I played with emcrsh a litle today and would like to show you the result: A copy of a tee spoon milled: http://128.130.120.105/_DSC5204.lowres.jpg emcrsh is not the ideal solution though. I guess it'd bee better to link the digitiser directly to HAL but I have to find a way to access the MicroScribe from Linux. Flo -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Sebastian Kuzminsky -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
Il giorno mer, 30/11/2011 alle 18.18 +0100, Florian Rist ha scritto: Hi, thanks for you comments on my project idea. I played with emcrsh a litle today and would like to show you the result: A copy of a tee spoon milled: http://128.130.120.105/_DSC5204.lowres.jpg emcrsh is not the ideal solution though. I guess it'd bee better to link the digitiser directly to HAL but I have to find a way to access the MicroScribe from Linux. Flo I find this very interesting, but i didn't understand how. first: did you mill 'live' while pantographing, or did you record positions? then, how did you modify emcrsh? thanks, davide. -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
On Nov 30, 2011, at 10:18 , Florian Rist wrote: I played with emcrsh a litle today and would like to show you the result: A copy of a tee spoon milled: http://128.130.120.105/_DSC5204.lowres.jpg emcrsh is not the ideal solution though. I guess it'd bee better to link the digitiser directly to HAL but I have to find a way to access the MicroScribe from Linux. Neat! Just out of curiosity, what were the problems with using emcrsh? Does the MicroScribe device connect to the PC via USB? Does it show up as a HID device? If so, it will probably be fairly easy to get raw joint data into HAL, you'd probably still have to do the inverse kinematics to translate that into usable world coordinates. -- Sebastian Kuzminsky -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
Hi davide I find this very interesting, but i didn't understand how. first: did you mill 'live' while pantographing, or did you record positions? I milled 'live' or on-line while moving the digitiser. then, how did you modify emcrsh? Not at all. I just wrote a very simple program (running on a second PC on Windows 7) that take the digitiser readings and sends them as g01 commands to emcrsh. See you Flo -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
Hi Sebastian Just out of curiosity, what were the problems with using emcrsh? See my other posting. Does the MicroScribe device connect to the PC via USB? Yes. Does it show up as a HID device? Hmm, I don't know. At least it's not obviously a HID. I used a DLL provided by the manufacturer as part of the SDK for Windows to communicate with the digitiser. The strange thing is that the documentation says it's also compatible to MacOS X and IRIX but I cant see how to access the MicroScribe from OS X or IRIX. If so, it will probably be fairly easy to get raw joint data into HAL, you'd probably still have to do the inverse kinematics to translate that into usable world coordinates. Kinematics (forward kinematics, actually) seams to be performed by the digitiser hardware. At least some data types defined in the header file for the Windows library suggest that. I found this Python code (part of ROS, Robot Operating System) that seams to communicate with the digitiser directly, but I'm not sure what it relay does: http://www.ros.org/doc/api/microscribe/html/microscribe_8py_source.html Is anyone familiar whit Pythen, USB and ROS? See you Florian -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
Hi Sebastian Just out of curiosity, what were the problems with using emcrsh? Right now my super simple program tracks the digitiser and once it is moved by 0.5 mm it sends the new coordinates as a g01 command to the mill. So if I move the digitiser faster than the mill can follow a buffer in emcrsh buffers the commands and they will be executed later, following the exact (+/-0.5mm) trace. I'd rather prefer a behaviour like the one of the mouse courser on screen. If you move you mouse to fast the tracing resolution is decreasing, so there are longer linear path segments with more time for the mill to accelerate and speed up. My bee I'll try to use time-inverse feed rate specification mode for the next test. This might improve this behaviour. Another problem with the buffer in emcrsh is that it seams to be rather small, just big enough for a few hundred commands. Once this buffer runs over emcrsh terminates. In addition to these aspects I think that the idear of trying to read the digitiser trace as a polygon and mill this polygon using G01 commands is not the beset way, though playing with G64 might improve it. I have the feeling that I could create a more natural feeling with both high dynamic and accuracy by feeding a continues stream of digitiser coordinate (I could poll the digitiser ever 50 ms) directly to HAL and treat this as an input signal for a PID controller that controls the mill. This approach would also open the way to further extension like force feedback [1]. I guess setting the feedback forse proportional to the p-component of the PID controller would give good results. See you Flo [1] I'm looking for a used Sensable Phantom, buy the way. Mail me if you have an unused device you might want to sell. -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
On Nov 30, 2011, at 12:27 , Florian Rist wrote: Another problem with the buffer in emcrsh is that it seams to be rather small, just big enough for a few hundred commands. Once this buffer runs over emcrsh terminates. Ugh, that doesn't sound right. Could you open a bug in the sourceforge bug tracker for this please? -- Sebastian Kuzminsky -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
Hi Sebastian Another problem with the buffer in emcrsh is that it seams to be rather small, just big enough for a few hundred commands. Once this buffer runs over emcrsh terminates. Ugh, that doesn't sound right. Could you open a bug in the sourceforge bug tracker for this please? I'm not sure that it's an error or just a small buffer and I misuse emcrsh. The buffer holds abut 124 line like this one (about 5.5 kB in total): set mdi g1x0.0123123y1.123123z2.123123f1234 But the buffer overflow is not handled by emcrsh. it crashes and displays this: *** buffer overflow detected ***: emcrsh terminated === Backtrace: = /lib/tls/i686/cmov/libc.so.6(__fortify_fail+0x50)[0x2e8390] /lib/tls/i686/cmov/libc.so.6(+0xe12ca)[0x2e72ca] /lib/tls/i686/cmov/libc.so.6(+0xe05fa)[0x2e65fa] emcrsh[0x804e0ed] /lib/tls/i686/cmov/libpthread.so.0(+0x596e)[0x83b96e] /lib/tls/i686/cmov/libc.so.6(clone+0x5e)[0x2d3a4e] === Memory map: 0011-001f9000 r-xp 08:01 797977 /usr/lib/libstdc++.so.6.0.13 001f9000-001fa000 ---p 000e9000 08:01 797977 /usr/lib/libstdc++.so.6.0.13 001fa000-001fe000 r--p 000e9000 08:01 797977 /usr/lib/libstdc++.so.6.0.13 001fe000-001ff000 rw-p 000ed000 08:01 797977 /usr/lib/libstdc++.so.6.0.13 001ff000-00206000 rw-p 00:00 0 00206000-00359000 r-xp 08:01 786494 /lib/tls/i686/cmov/libc-2.11.1.so 00359000-0035a000 ---p 00153000 08:01 786494 /lib/tls/i686/cmov/libc-2.11.1.so 0035a000-0035c000 r--p 00153000 08:01 786494 /lib/tls/i686/cmov/libc-2.11.1.so 0035c000-0035d000 rw-p 00155000 08:01 786494 /lib/tls/i686/cmov/libc-2.11.1.so 0035d000-0036 rw-p 00:00 0 00645000-0066 r-xp 08:01 790839 /lib/ld-2.11.1.so 0066-00661000 r--p 0001a000 08:01 790839 /lib/ld-2.11.1.so 00661000-00662000 rw-p 0001b000 08:01 790839 /lib/ld-2.11.1.so 00708000-0075 r-xp 08:01 787236 /usr/lib/libnml.so.0 0075-00751000 r--p 00047000 08:01 787236 /usr/lib/libnml.so.0 00751000-00752000 rw-p 00048000 08:01 787236 /usr/lib/libnml.so.0 00752000-00753000 rw-p 00:00 0 00836000-0084b000 r-xp 08:01 786606 /lib/tls/i686/cmov/libpthread-2.11.1.so 0084b000-0084c000 r--p 00014000 08:01 786606 /lib/tls/i686/cmov/libpthread-2.11.1.so 0084c000-0084d000 rw-p 00015000 08:01 786606 /lib/tls/i686/cmov/libpthread-2.11.1.so 0084d000-0084f000 rw-p 00:00 0 009e6000-00a0a000 r-xp 08:01 786547 /lib/tls/i686/cmov/libm-2.11.1.so 00a0a000-00a0b000 r--p 00023000 08:01 786547 /lib/tls/i686/cmov/libm-2.11.1.so 00a0b000-00a0c000 rw-p 00024000 08:01 786547 /lib/tls/i686/cmov/libm-2.11.1.so 00a76000-00a79000 r-xp 08:01 787230 /usr/lib/libemcini.so.0 00a79000-00a7a000 r--p 2000 08:01 787230 /usr/lib/libemcini.so.0 00a7a000-00a7b000 rw-p 3000 08:01 787230 /usr/lib/libemcini.so.0 00ab2000-00ab3000 r-xp 00:00 0 [vdso] 00d5e000-00d7b000 r-xp 08:01 786529 /lib/libgcc_s.so.1 00d7b000-00d7c000 r--p 0001c000 08:01 786529 /lib/libgcc_s.so.1 00d7c000-00d7d000 rw-p 0001d000 08:01 786529 /lib/libgcc_s.so.1 08048000-0805a000 r-xp 08:01 787287 /usr/bin/emcrsh 0805a000-0805b000 r--p 00012000 08:01 787287 /usr/bin/emcrsh 0805b000-0805c000 rw-p 00013000 08:01 787287 /usr/bin/emcrsh 09f88000-09fa9000 rw-p 00:00 0 [heap] b70c5000-b70c6000 ---p 00:00 0 b70c6000-b78c9000 rw-p 00:00 0 b78d-b78d1000 rw-p 00:00 0 b78d1000-b78d3000 rw-s 00:04 557070 /SYSV03eb (deleted) b78d3000-b78d7000 rw-s 00:04 524301 /SYSV03ea (deleted) b78d7000-b78d9000 rw-s 00:04 491532 /SYSV03e9 (deleted) b78d9000-b78db000 rw-p 00:00 0 bfaa-bfab5000 rw-p 00:00 0 [stack] Aborted Bug or not? I guess emcrsh should handle the buffer overflow and prevent the reported buffer overflow in libc, so it's bug, right? See you Flo -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
- is the stream of xyz values in the order you would use it for cutting? - what about relative speeds? - is it ok to generate a ngc file from the digitiser stream and executed that eventually, or do you need to do this in paralllel? -m Am 25.11.2011 um 22:01 schrieb Florian Rist: Hi, I wonder if it would be possible to extend EMC in a way that I could ad a 3D digitiser as an input device and control the position of a 3aces CNC mill live via that digitiser so that I could use the CNC mill just like a digital pantograph milling machine. The digitiser, a Immersion Microscribe II, gives me a stream of xyz coordinates, either via USB or a serial RS-232 interface. I guess the serial connection is much easier to use and probably fast enough, too. My problem is that I have no clue how I might be able to feed these coordinate into EMC, to make EMC move according to these coordinates. I haven't done any development vor EMC yet, except some minor extensions to trivkins, to compensate some errors on my machine. How could I link the digitise and EMC? I hope you can give me some ideas, see you Flo -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
Hi Michael - is the stream of xyz values in the order you would use it for cutting? Yes. - what about relative speeds? It would be ok if I could set a maximum speed and until that speed is reached the speed of the digitizer movements is used. - is it ok to generate a ngc file from the digitiser stream and executed that eventually, or do you need to do this in paralllel? No, recording is not an option. I want the mill to follow the movement of the digitizer instantly, just as the were coupled mechanically. See you Florian -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] EMC a digital pantograph milling machine?
Hi Another option would be to run emc2 with emcrsh, and send the xyz coordinates from the digitizer as mdi g1 commands to emcrsh. Hmm, that sounds simple. So I could even connect the digitiser to a Windows machine, use the Immersion Windows SKD to communicate with the Microscribe via USB, and just send a steam of g1 commands via telent to EMC. I think I'll try this. See you Florian -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
[Emc-users] EMC a digital pantograph milling machine?
Hi, I wonder if it would be possible to extend EMC in a way that I could ad a 3D digitiser as an input device and control the position of a 3aces CNC mill live via that digitiser so that I could use the CNC mill just like a digital pantograph milling machine. The digitiser, a Immersion Microscribe II, gives me a stream of xyz coordinates, either via USB or a serial RS-232 interface. I guess the serial connection is much easier to use and probably fast enough, too. My problem is that I have no clue how I might be able to feed these coordinate into EMC, to make EMC move according to these coordinates. I haven't done any development vor EMC yet, except some minor extensions to trivkins, to compensate some errors on my machine. How could I link the digitise and EMC? I hope you can give me some ideas, see you Flo -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users