Re: [Emc-users] YAGC
2011/5/19 Tom Easterday tom-...@bgp.nu: On the IRC Chris Radek suggested there was no advantage to using Gantrykins and that it may have some problems and so I changed things to Trivkins. Did he explain and give some arguments? Gantrykins worked just fine for me and other users are using it as well. Gantrykins just links 2 joints to 1 axis, it does not interfere with joint settings. If You had difficulties with motor acceleration and movement, I would suspect that there was something wrong with settings or maybe with the loop in the Geckos. But I have never used servos, so that is just a guess. Yes, You are totally right - if You want each motor to move independently, You have to treat them as separate joints. My suggestion would be - set up gantrykins without THC to work properly and then add THC capability. If there realy is some kind of problem, You can easily customize trivkins module to have 2 joints hardcoded to Y axis. Gantrykins allows flexibility to cover all kinds of situations, but You can adjust a module for Your particular machine. That is what I ended with, because I needed a module for 5 axis gantry machine. Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] (no subject)
Chris, was this a genuine posting or your personal spambot?? jan. On Wed, May 18, 2011 at 8:39 PM, Chris Reynolds c_reynolds2...@yahoo.comwrote: It’s awesome!!! In no time you’ll rid of your problems!!!.. http://news-trial.com/friends_links.php?coPage=14pj1 -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] 5-axis kinematics XYZAC
Not finding any other way to upload the XYZACkins that I recently developed, I added a small description to the Wiki under HAL user contributed components: Kinematics. A link to a tar.gz file is given which contains the kinematics code as well as the ini and hal files, a vismach simulation model and a demo ngc file. Comments will be appreciated once you have tried it out. The remaining problem I have is the homing and setup to start a job. I am still working on a solution. Rudy -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] 5-axis kinematics XYZAC
2011/5/19 Rudy du Preez r...@asmsa.co.za: Comments will be appreciated once you have tried it out. The remaining problem I have is the homing and setup to start a job. I am still working on a solution. What is wrong with homing? Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] YAGC
On May 19, 2011, at 3:19 AM, Viesturs Lācis wrote: 2011/5/19 Tom Easterday tom-...@bgp.nu: On the IRC Chris Radek suggested there was no advantage to using Gantrykins and that it may have some problems and so I changed things to Trivkins. Did he explain and give some arguments? No, but he was busy and had to go just after telling me this... Gantrykins worked just fine for me and other users are using it as well. Gantrykins just links 2 joints to 1 axis, it does not interfere with joint settings. If You had difficulties with motor acceleration and movement, I would suspect that there was something wrong with settings or maybe with the loop in the Geckos. But I have never used servos, so that is just a guess. Yes, You are totally right - if You want each motor to move independently, You have to treat them as separate joints. My suggestion would be - set up gantrykins without THC to work properly and then add THC capability. If there realy is some kind of problem, You can easily customize trivkins module to have 2 joints hardcoded to Y axis. Gantrykins allows flexibility to cover all kinds of situations, but You can adjust a module for Your particular machine. That is what I ended with, because I needed a module for 5 axis gantry machine. Ok, I redid the config using gantrykins and the parameters I used to get it working in trivkins and now it seems to work. The only problem now is that I see three joints 1,2,3 and when I try to switch to world mode it says I have to home all joints. I did home all (3) joints, how do I home the 4th - or am I missing something more basic? -Tom -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Joints_axes branch
See my responses embedded below: On 5/17/2011 11:09 AM, Viesturs Lācis wrote: 2011/5/17 Davee...@dc9.tzo.com: If you can create the stepgen and a few other components you can go beyond the interpreter limit. Thanks for the sample! It really seems doable. But I do have few questions / comments. I hope that You can help me clarifying them. This is not really an X axis although I used the X axis values out of the INI file (Axis_0 - not necessary).This axis is not connected to the interpreter. It is driven by CL. Everything accels and decels like it should. Very simple once you try it. I'm sure this could be cleaned up but I just snipped it out of a existing hal file. The is actually running on a machine so I know it works. Dave :-) # # X [0] Axis # Drive 1 - X axis #Divide CL floatout by 100 to get proper scaling - actually multiply by .01 setp mult2.0.in0 0.01 net mult201 mult2.0.in1= classicladder.0.floatout-00 What data is feeded to classicladder.0.floatout-00? Commanded position of motor? Then I am curious, how is that commanded position calculated in classicladder? There will be some starting position and on some input trigger the movement should start. How does classicladder it? Calculates commanded position each servo cycle? Could You, please, expand on this topic a little more? It is even easier than that. Classic Ladder (CL) only need to send out the end position of the move and the limit3 will accelerate the axis and decel the axis to the new end position. If you want it to go to another position, simply change the end position value. #Tie classic ladder drive control out (/100) to Limit3 input net limit3_0in limit3.0.in= mult2.0.out #setp limit3.0.maxv 5 # to be set via Modbus net limit30maxv classicladder.0.floatout-01 = limit3.0.maxv #setp limit3.0.maxa 25 # to be set via Modbus net limit30maxa classicladder.0.floatout-02 = limit3.0.maxa Is this something specific for that particular machine that I should skip? The Modbus notes were just notes for me as I put the controls together. The velocity and accel was set via a Modbus connection to a PLC which had a screen attached. The PC runs EMC2 as a headless slave to a PLC. There is no operator interaction directly with the EMC2 PC. net emcmot.00.pos-cmd= limit3.0.out net emcmot.00.pos-cmd = hm2_[HOSTMOT2](BOARD).0.stepgen.00.position-cmd Is emcmot.00.pos-cmd a signal name or a pin? I googled this phrase and got several links to HAL files for some Mesa card configs. My sense tells me that this should be just a signal name, but I am confused. I think that is just a signal name. I tend to adapt existing configs to what I need rather than renaming everything. the point of those two lines is that the limit3 #0 output is connected to the input of the stepgen.00 position command. The limit3 takes in a position and accels, limits the velocity, and decels so that the output arrives at the input command position via the set parameters (accel, max vel). net motor.00.pos-fb= hm2_[HOSTMOT2](BOARD).0.stepgen.00.position-fb net motor.00.pos-fb = axis.0.motor-pos-fb =mult2.1.in0 Why do You feed motor position feedback to axis.0.motor-pos-fb, if You do not use axis.0.motor-pos-cmd? Doesn't that cause following error, because pos-cmd and pos-fb would differ a lot? Or am I missing something here? I believe that I did that so I could see the actual position on the Axis screenset screen when I was debugging the machine. Following errors are not an issue. setp mult2.1.in1 100 net motor0pos mult2.1.out = classicladder.0.floatin-02 Ok, You get motor position feedback back into classicladder. What happens then? I mean - does it somehow compare cmd and fb values? Can You, please, explain more, what happens with this data? I believe I did that so I could get the axis position data back into CL so I could tell when the position was reached . I believe that the actual position for a couple of these axes are also fedback to an actual PLC for display and monitoring purposes. It works very well. Try mocking something up and if you get stuck I'll try and help you out. I think you working on some custom graphical interfaces for EMC2 a while back... where did that all end up? I think there were some additions made to the EMC2 dev version?? And perhaps the Wiki? but I don't remember what they were... I was busy doing some other non-EMC2 related work at the time and lost track of things. Dave Viesturs -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters.
Re: [Emc-users] Joints_axes branch
2011/5/19 Dave e...@dc9.tzo.com: See my responses embedded below: It works very well. Try mocking something up and if you get stuck I'll try and help you out. Wow, thanks a lot!!! Honestly, I am pretty far from actually building something. because that machine is pretty complicated and although it seems that control side can be handled, the big issue is to get the data in the EMC. I have not solved that with my client. Saying that they are reluctant to use cnc control software, which is considered to be a hobby level solution, is saying just a little. But I hope that robot will make them reconsider. I think you working on some custom graphical interfaces for EMC2 a while back... where did that all end up? What I tried to dowas: move pyvcp buttons from the right side of the screen to the left side. I did that and managed to tie them with Manual control tab, so that they are not visible in MDI tab. It was successful!!! I think there were some additions made to the EMC2 dev version?? And perhaps the Wiki? Sorry, I did not add anything about it to documentation of any kind, because I did not feel any kind of interest at all from other users. Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] YAGC
2011/5/19 Tom Easterday tom-...@bgp.nu: I did home all (3) joints, how do I home the 4th - or am I missing something more basic? Just the same way as You home all the remaining joints. I do not understand, where is the problem. Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] YAGC
On May 19, 2011, at 11:26 AM, Viesturs Lācis wrote: 2011/5/19 Tom Easterday tom-...@bgp.nu: I did home all (3) joints, how do I home the 4th - or am I missing something more basic? Just the same way as You home all the remaining joints. I do not understand, where is the problem. In Axis I only see joints 1,2,3 (but not 4) as a radio button. I select 1 and click Home. I select 2 and click Home, etc. How do I select joint 4, it is not there? -Tom -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] YAGC
2011/5/19 Tom Easterday tom-...@bgp.nu: In Axis I only see joints 1,2,3 (but not 4) as a radio button. I select 1 and click Home. I select 2 and click Home, etc. How do I select joint 4, it is not there? And what about joint 0? How many joints You have specified in INI file? In [TRAJ] section AXES has to be set to 4, because You have 4 joints and COORDINATES has to contain 4 letters, for example X Y Z A Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Joints_axes branch
Saying that they are reluctant to use cnc control software, which is considered to be a hobby level solution, is saying just a little. I replaced an Allen Bradley Series 9 CNC control with EMC2 on a commercial waterjet machine made by Jet Edge last year. The customer thought that they blew the USB ports on the PC so I visited them yesterday since the machine upgrade is still under warranty. (The computer was fine, some USB connectors external to the cabinet had become corroded - the environment is corrosive) The engineering manager told me that the Jet Edge machine retrofit was the feature of his yearly equipment report to the company board. He said the machine performs much better than it did when it was new. The machine runs 1-2 shifts per day, 5-7 days per week depending on demand. It is automatically fed with a conveyor system. The four waterjet nozzles actually cut over a chain conveyor so the cut pieces can be automatically conveyed out of the machine. The system is very reliable. Properly applied, EMC2 is extremely reliable and capable. Dave On 5/19/2011 11:23 AM, Viesturs Lācis wrote: 2011/5/19 Davee...@dc9.tzo.com: See my responses embedded below: It works very well. Try mocking something up and if you get stuck I'll try and help you out. Wow, thanks a lot!!! Honestly, I am pretty far from actually building something. because that machine is pretty complicated and although it seems that control side can be handled, the big issue is to get the data in the EMC. I have not solved that with my client. Saying that they are reluctant to use cnc control software, which is considered to be a hobby level solution, is saying just a little. But I hope that robot will make them reconsider. I think you working on some custom graphical interfaces for EMC2 a while back... where did that all end up? What I tried to dowas: move pyvcp buttons from the right side of the screen to the left side. I did that and managed to tie them with Manual control tab, so that they are not visible in MDI tab. It was successful!!! I think there were some additions made to the EMC2 dev version?? And perhaps the Wiki? Sorry, I did not add anything about it to documentation of any kind, because I did not feel any kind of interest at all from other users. Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] YAGC
On May 19, 2011, at 12:26 PM, Viesturs Lācis wrote: 2011/5/19 Tom Easterday tom-...@bgp.nu: In Axis I only see joints 1,2,3 (but not 4) as a radio button. I select 1 and click Home. I select 2 and click Home, etc. How do I select joint 4, it is not there? And what about joint 0? How many joints You have specified in INI file? In [TRAJ] section AXES has to be set to 4, because You have 4 joints and COORDINATES has to contain 4 letters, for example X Y Z A Here are my hal and ini file: a2msg.hal Description: Binary data a2msg.ini Description: Binary data -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] YAGC
Try changing the second Y to A, it does not like the same letter twice. COORDINATES = X Y Z A You will have x, y, z and a axes in Axis gui, but do not worry - if You have done kinematics.hal file correctly, then A commands in g-code will do nothing. Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Joints_axes branch
2011/5/19 Dave e...@dc9.tzo.com: The system is very reliable. Properly applied, EMC2 is extremely reliable and capable. That is what I believe in and what I have seen so far on my machine. But I have yet to prove it to convince others. Especially, if I am facing resistance from their current supplier of controls, from whom I expect a lot of things just to show that my solution is not good enough. Actually it already has started by arguing that PC-based solution is not reliable enough, because the hardware is not as reliable as some PLCs. Since I do not have experience with PLC, unfortunately I can't argue with that. Anyway, I think that best argument will a working machine :)) Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] YAGC
I think that V would offer a better choice, often when creating G-code for a rotary device, your CAM package will default to A for the rotary axis. Greg - Original Message - From: Viesturs Lacis viesturs.la...@gmail.com To: Enhanced Machine Controller (EMC) emc-users@lists.sourceforge.net Sent: Thursday, May 19, 2011 8:21 PM Subject: Re: [Emc-users] YAGC Try changing the second Y to A, it does not like the same letter twice. COORDINATES = X Y Z A You will have x, y, z and a axes in Axis gui, but do not worry - if You have done kinematics.hal file correctly, then A commands in g-code will do nothing. Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users No virus found in this incoming message. Checked by AVG - www.avg.com Version: 8.5.449 / Virus Database: 271.1.1/3640 - Release Date: 05/15/11 18:34:00 -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Joints_axes branch
I work with PLCs daily. They are a different animal entirely. Most of the EMC2 implementations I have done on custom machines have included a PLC simply because they have been I/O intensive machines. There is no reason why you cannot use them together. The PLC vs PC reliability argument still pops up once in a while in the US, but not nearly as much as it did 10 years ago. Now the the PLC is better argument usually dies quickly when it is determined that PCs are required as operator consoles and data servers in conjunction with the PLC system. Since PCs are already required - adding another one for motion control is oftentimes no longer an issue. Another thing; if the PC does not look like a desktop unit, it will not be perceived in the same way. Find a PC case that is industrial looking and use good quality PC hardware and they will forget that it is a PC. It is really all about perception. Dave On 5/19/2011 3:44 PM, Viesturs Lācis wrote: 2011/5/19 Davee...@dc9.tzo.com: The system is very reliable. Properly applied, EMC2 is extremely reliable and capable. That is what I believe in and what I have seen so far on my machine. But I have yet to prove it to convince others. Especially, if I am facing resistance from their current supplier of controls, from whom I expect a lot of things just to show that my solution is not good enough. Actually it already has started by arguing that PC-based solution is not reliable enough, because the hardware is not as reliable as some PLCs. Since I do not have experience with PLC, unfortunately I can't argue with that. Anyway, I think that best argument will a working machine :)) Viesturs -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
[Emc-users] Required POSIX exensions?
Hello all, I have been searching the documentation and HAL User Manual but have not found a simple overview of what POSIX/realtime API's are required if porting EMC2 to another RT kernel. Could someone please summarize the the requirements? Thanks, Michael -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] YAGC
On May 19, 2011, at 2:21 PM, Viesturs Lācis wrote: Try changing the second Y to A, it does not like the same letter twice. COORDINATES = X Y Z A You will have x, y, z and a axes in Axis gui, but do not worry - if You have done kinematics.hal file correctly, then A commands in g-code will do nothing. Thanks I will give it a try. -Tom -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Joints_axes branch
Dave wrote: The system is very reliable. Properly applied, EMC2 is extremely reliable and capable. Yup, I have a number of CNC control interfaces in commercial shops all over the world, and they are well-liked by their users (nearly all EMC2-based). Jon -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Required POSIX exensions?
RTAPI is the name of the real-time portability APIs in emc. Currently, we provide two implementations of RTAPI that are known to work: RTAI's kernel-moade interface (the one everyone uses when actually controlling hardware) and sim (based on gnu pth for threading primitives, no realtime guarantees, not suitable for controlling hardware--great for developing GUIs and debugging non-driver realtime code). There is no working RTAPI implementation in terms of POSIX threading primitives in the linuxcnc source tree, though there are at least two earlier efforts that each achieved some level of functionality. See e.g., http://mid.gmane.org/02c101c9908d$4657afb0$6400a8c0@danalappy http://axis.unpythonic.net/01190912545 http://mid.gmane.org/200910110020.00011...@bu3sch.de http://bu3sch.de/patches/misc/emc-on-linux-rt.patch RTAPI is a fairly small set of APIs (about 50 or so, with maybe 40 actually required to run emc). Most are documented in the unix manpage system, section 3rtapi, but a few may only be in the header rtapi.h. In addition to implementing the APIs, you'd have housekeeping things to do like the build system, the packaging system, and the realtime start/stop script. Finally (but the biggest item if you want to control hardware!), since the only hardware-controlling RTAPI implementation runs in the kernel, hardware drivers freely use Linux kernel APIs for e.g., PCI device detection and I/O port allocation during setup, and inb/outb for I/O in realtime code. For whatever hardware drivers you wanted, you'd have to invent new RTAPI APIs that provide the same functionality, and then change them from calling kernel APIs to RTAPI APIs. (I imagine that these RTAPI APIs would be thin wrappers for the underlying kernel codes in the RTAI RTAPI implementation, so that part wouldn't be trouble) Jeff -- What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
