Re: [Emc-users] Guidance on usb comp
On Sunday 27 February 2011 14:13:12 Erik Christiansen wrote: If the community decided to go with this, I'd buy a couple, and arc up the gnu toolchain for this target, just for the fun of playing with ARM. These chips have one quadrature encoder input. At 100 MHz, and with hardware counters, I'd figure we could handle a couple. However, I'm much more familiar with AVR (Atmel), have the toolchain up to working temperature, and can fix problems more quickly there. There are AVR-based ehternet card, including ethernut 1.3: I have update the http://wiki.linuxcnc.org/emcinfo.pl?ARM with the links from the last postings. Have also a look to the bottom of the http://wiki.linuxcnc.org/emcinfo.pl?AVR page: it is about AVR32 for motor control (with eclipse based ide for linux). I also prefer udp packets. Maybe we are able to concentrate our forced and avoid 12 or more variants of udp packets. Is there some code available on the emc side? BTW: what do you think about PTP (IEEE 1588)? Joachim -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On Mon, Feb 28, 2011 at 09:12:50AM +0100, Joachim Franek wrote: Have also a look to the bottom of the http://wiki.linuxcnc.org/emcinfo.pl?AVR page: it is about AVR32 for motor control (with eclipse based ide for linux). Seeing that they are also supported by the gnu toolchain, and the UC3 devices have useful on-chip peripherals, I'd consider these 32 bit chips a viable alternative to ARM, for our uses. I also prefer udp packets. Maybe we are able to concentrate our forced and avoid 12 or more variants of udp packets. Is there some code available on the emc side? That would be inbuilt in the kernel, I'd expect. And in /boot/config-2.6.24-16-rtai, I see: CONFIG_NET_ETHERNET=y So it looks like a compile-time option, and the EMC2 application just uses that. And our RTAI environment probably wouldn't help much toward RT sending of UDP packest, because we'd still be using the linux kernel (non-RT) to push them out. But there's likely to be various open-source UDP implementations around. BTW: what do you think about PTP (IEEE 1588)? I'd not previously seen that. However, high timing precision (or even resolution) doesn't necessarily indicate high throughput, or even frequent transmissions, I think. Hold on a moment ... Not using parport myself, it's only belatedly occurred to me: why not just drop in a PCI parport card? (And skip all this fun serial stuff? :-) Erik -- In all affairs it's a healthy thing now and then to hang a question mark on the things you have long taken for granted. -- Bertrand Russell -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 02/27/2011 11:42 AM, Jean-Paul Moniz wrote: Ethernet medium can be in realtime take a look at sercos3 for example. It uses standard ethernet hardware. Sent on the TELUS Mobility network with BlackBerry According to the wiki, you can't use standard ethernet hardware: SERCOS III is designed in such a way that no additional network infrastructure (standard Ethernet switches, Hubs, etc.) is required to operate. In fact, no additional standard Ethernet (non-SERCOS III capable) components may be placed within a SERCOS III network, as their presence will adversely affect the timing and synchronization of the network. http://en.wikipedia.org/wiki/SERCOS_III Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 02/27/2011 04:22 PM, Peter Blodow wrote: Peter, I work on a daily basis with Ethernet as a system and network administrator. We're talking about a Real Time Ethernet, with a dedicated master and slave network interface, which should preclude any worry of collisions on a duplex connection between the two. Modern systems with full duplex connections no longer really rely on CSMA/CD. As defined in 802.3, that's just one of the protocols that have been used, not the only way. The reason we're talking about Ethernet as one of the possibilities for communications between the computer and the controller is the possibility of the ultimate demise of the parallel port, and the difficulties of real time communications on the USB bus. So, Real Time Ethernet ain't quite your Granddaddy's TCP/IP or USB protocol. Well, Mark, my granddaddy died in 1960. Concerning Ethernet, I was thinking of the times when I worked close to the IEEE 802 group way back when Siemens participated in technical innovation... It's just that I was so happy that PC's, being cheaply available and relatively easy to write programs for, have such a simple port as the parallel is. Well, I either will have to get used to the idea of writing special programs for dedicated processors again or else?? Whenever I had the impression to have cought up on technological development, it was running away again... Peter Blodow Peter, The Granddaddy's TCP/IP was just a figure of speech. We use it a bunch over here when we talk about new and improved things. It was first coined in the US auto industry. I hope I haven't given you the impression I dislike the use of the parallel port for what we are doing with it, far from it. This is just a look ahead, to get some ideas for when the parallel port is no longer supported in the industry. It would be nice if there were some other port available that would work in real time, and deliver the throughput and versatility we need as good or better than the parallel port. Problem with PC's you buy off the shelf, they really aren't designed with industrial use or motion control in mind. They're for Sally to read her email, and play around on Facebook, or the gamers, or the Word/Excel/Powerpoint crowd. We make the best with what we have, but we also need to look to the future when certain features we've become used to having around are no longer there. I've got a bunch of old parallel port cables laying around with Centronix connectors at one end. Who knew they'd stop supplying interfaces that matched up with those cable ends? Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On Monday 28 February 2011 10:31:32 Erik Christiansen wrote: And our RTAI environment probably wouldn't help much toward RT sending of UDP packest, because we'd still be using the linux kernel (non-RT) to push them out. Thats exactly my question: how to send/receive udp packets on a (additional) rt-eth interface? Has someone done something allready? I have here on my destop 3 variants: - FPGA (http://wiki.linuxcnc.org/emcinfo.pl?FPGA) - RDK-BLDC from LM - EVK1100 which receive and answer udp packets. So I can test this. (And I expext Intel is not telling us ethernet is deprecated.) Do not understand my hardware as a decision. My preference is to select one microprocessor to get something working in short time. And if the comunity decides to go with hardware xy, I buy this. I think there are more people familiar with programming micros compared to the number familiar with VHDL. I think we need some structure for the udp packet! And it is also possible to use this protocoll over the parallel port with micros or FPGA's. Joachim -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 28 February 2011 10:19, Mark Wendt mark.we...@nrl.navy.mil wrote: It would be nice if there were some other port available that would work in real time, and deliver the throughput and versatility we need as good or better than the parallel port. There is a realtime USB extension for RTAI: https://www.rtai.org/RTAICONTRIB/ However, it does not compile with the current kernel version due to a number of kernel macro changes, and I have a feeling that the version at that link has _never_ compiled as there seems to be some mismatch between the structs in the headers and some of the functions that use them, -- atp Torque wrenches are for the obedience of fools and the guidance of wise men -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Siemens DriveCliq works the same way. Everything is point to point. No external hubs or switches. I think the controller I was using recently had 6 or 8 Ethernet jacks. The drive module had 3, one for the connection to the controller and two for encoder connections. I think the line module (Ac to DC smart power supply) had 3 Ethernet jacks. Hooking everything up was very quick. Dave On 2/28/2011 5:07 AM, Mark Wendt wrote: On 02/27/2011 11:42 AM, Jean-Paul Moniz wrote: Ethernet medium can be in realtime take a look at sercos3 for example. It uses standard ethernet hardware. Sent on the TELUS Mobility network with BlackBerry According to the wiki, you can't use standard ethernet hardware: SERCOS III is designed in such a way that no additional network infrastructure (standard Ethernet switches, Hubs, etc.) is required to operate. In fact, no additional standard Ethernet (non-SERCOS III capable) components may be placed within a SERCOS III network, as their presence will adversely affect the timing and synchronization of the network. http://en.wikipedia.org/wiki/SERCOS_III Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 02/28/2011 08:19 AM, Dave wrote: Siemens DriveCliq works the same way. Everything is point to point. No external hubs or switches. I think the controller I was using recently had 6 or 8 Ethernet jacks. The drive module had 3, one for the connection to the controller and two for encoder connections. I think the line module (Ac to DC smart power supply) had 3 Ethernet jacks. Hooking everything up was very quick. Dave Looking again at the Sercos3 implementation, each host requires at least a two-port network interace. That's how they implement either the Line, Ring or other network topology. Each host has to process the data and then pass it on to the next host in a host 2+n topology. The Ring topology looks to be the most efficient, since two channels are being used to move the data around the circuit, and the second channel offers a bit of redundancy of the primary channel goes down. For environments that have multiple machines, this could be an issue with multiple, long network cables going from the master controller to the slave boxes. With that in mind though, it's kind of hard to imagine multiple instances of EMC2 running on one controller PC controlling different machines. For a single control PC sending data to a single control box, this shouldn't really be an issue. Interesting also how the protocol utilizes multiple length packets (or as they call them, Telegrams), anywhere from 84 bytes to 1538 bytes in length. Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 02/28/2011 07:37 AM, Mark Wendt wrote: On 02/28/2011 08:19 AM, Dave wrote: Siemens DriveCliq works the same way. Everything is point to point. No external hubs or switches. I think the controller I was using recently had 6 or 8 Ethernet jacks. The drive module had 3, one for the connection to the controller and two for encoder connections. I think the line module (Ac to DC smart power supply) had 3 Ethernet jacks. Hooking everything up was very quick. Dave Looking again at the Sercos3 implementation, each host requires at least a two-port network interace. That's how they implement either the Line, Ring or other network topology. Each host has to process the data and then pass it on to the next host in a host 2+n topology. The Ring topology looks to be the most efficient, since two channels are being used to move the data around the circuit, and the second channel offers a bit of redundancy of the primary channel goes down. For environments that have multiple machines, this could be an issue with multiple, long network cables going from the master controller to the slave boxes. With that in mind though, it's kind of hard to imagine multiple instances of EMC2 running on one controller PC controlling different machines. For a single control PC sending data to a single control box, this shouldn't really be an issue. Interesting also how the protocol utilizes multiple length packets (or as they call them, Telegrams), anywhere from 84 bytes to 1538 bytes in length. Mark Sounds like standard ethernet frames. I would imagine they would use standard frames because ethernet hardware is expecting that. http://en.wikipedia.org/wiki/Ethernet_frame Moses -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Sercos 3 was one example of the use of ethernet for real time control. Another example is ethernet i/p. Again it uses standard ethernet technology and allows the use of switches between the host and slave. It all comes down to timing. AB claims they do +/- 100 uS without modified hardware. With IEEE 1588 compliant hardware they can get it down much lower. It all comes down to the application and what is required. Refer to ODVA's CIP Motion and CIP Sync for more information. I use AB PLC's and Bosch Rexroth drives using ethernet I/P communication on a daily basis without 1588 hardware and have found it to offer exceptional performance. My comments were relating to someone indicating that ethernet could not be used in a real-time application. Thanks Jp Sent on the TELUS Mobility network with BlackBerry -Original Message- From: Mark Wendt mark.we...@nrl.navy.mil Date: Mon, 28 Feb 2011 08:37:19 To: Enhanced Machine Controller (EMC)emc-users@lists.sourceforge.net Reply-To: mark.we...@nrl.navy.mil, Enhanced Machine Controller \(EMC\) emc-users@lists.sourceforge.net Subject: Re: [Emc-users] Guidance on usb comp On 02/28/2011 08:19 AM, Dave wrote: Siemens DriveCliq works the same way. Everything is point to point. No external hubs or switches. I think the controller I was using recently had 6 or 8 Ethernet jacks. The drive module had 3, one for the connection to the controller and two for encoder connections. I think the line module (Ac to DC smart power supply) had 3 Ethernet jacks. Hooking everything up was very quick. Dave Looking again at the Sercos3 implementation, each host requires at least a two-port network interace. That's how they implement either the Line, Ring or other network topology. Each host has to process the data and then pass it on to the next host in a host 2+n topology. The Ring topology looks to be the most efficient, since two channels are being used to move the data around the circuit, and the second channel offers a bit of redundancy of the primary channel goes down. For environments that have multiple machines, this could be an issue with multiple, long network cables going from the master controller to the slave boxes. With that in mind though, it's kind of hard to imagine multiple instances of EMC2 running on one controller PC controlling different machines. For a single control PC sending data to a single control box, this shouldn't really be an issue. Interesting also how the protocol utilizes multiple length packets (or as they call them, Telegrams), anywhere from 84 bytes to 1538 bytes in length. Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 02/28/2011 08:56 AM, Moses McKnight wrote: Looking again at the Sercos3 implementation, each host requires at least a two-port network interace. That's how they implement either the Line, Ring or other network topology. Each host has to process the data and then pass it on to the next host in a host 2+n topology. The Ring topology looks to be the most efficient, since two channels are being used to move the data around the circuit, and the second channel offers a bit of redundancy of the primary channel goes down. For environments that have multiple machines, this could be an issue with multiple, long network cables going from the master controller to the slave boxes. With that in mind though, it's kind of hard to imagine multiple instances of EMC2 running on one controller PC controlling different machines. For a single control PC sending data to a single control box, this shouldn't really be an issue. Interesting also how the protocol utilizes multiple length packets (or as they call them, Telegrams), anywhere from 84 bytes to 1538 bytes in length. Mark Sounds like standard ethernet frames. I would imagine they would use standard frames because ethernet hardware is expecting that. http://en.wikipedia.org/wiki/Ethernet_frame Moses Almost, except that standard Ethernet frames are typically of non-variable length, adjustable only by changing the MTU, and are padded to maintain the frame length being used. The Sercos3 seemingly allows variable length frames. Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Mark Wendt wrote: On 02/28/2011 08:56 AM, Moses McKnight wrote: [snip] Sounds like standard ethernet frames. I would imagine they would use standard frames because ethernet hardware is expecting that. http://en.wikipedia.org/wiki/Ethernet_frame Moses Almost, except that standard Ethernet frames are typically of non-variable length, adjustable only by changing the MTU, and are padded to maintain the frame length being used. The Sercos3 seemingly allows variable length frames. This isn't how it works. There is a minimum frame size (512 bits), and a maximum transfer unit (usually 1500 bytes or so, sometimes up to 9000 bytes on a gigabit line with jumbo frame support). As long as they're between those limits, the packets can be any size. -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Sercos 3 was one example of the use of ethernet for real time control. Another example is ethernet i/p. Again it uses standard ethernet technology and allows the use of switches between the host and slave. It all comes down to timing. AB claims they do +/- 100 uS without modified hardware. With IEEE 1588 compliant hardware they can get it down much lower. It all comes down to the application and what is required. Refer to ODVA's CIP Motion and CIP Sync for more information. I use AB PLC's and Bosch Rexroth drives using ethernet I/P communication on a daily basis without 1588 hardware and have found it to offer exceptional performance. My comments were relating to someone indicating that ethernet could not be used in a real-time application. RT-Net is a good candidate for an open source implementation. (http://www.rtnet.org/) RT-Net provides drivers for a collection of network cards that work under a real time kernel (RTAI). In its raw form, it can just be used to schedule delivery of Ethernet packets. You implement the protocol. To support real time communications, there are 3 specifications defined - rtmac, which defines the frame format, rtcfg, which is used by the master to find, and configure the slaves, and tdma (time division multiple access) which slices the bandwidth of the line into packets, then these packets sliced into slots. It also does clock synchronisation. The rtnet protocols can coexist with other Ethernet protocols on the line at the same time, but that seems a bit risky in a real time environment. There are some issues with it - the network card drivers seem to require a dedicated IRQ, so it can be hard to set up. The rtnet rtai software wants to control the rtai internal clock, so it needs to be made to play nicely with emc. It is open source, so theoretically a slave could be built on any microcontroller powerful enough to run an embedded linux, but it would still a realtime kernel and network drivers - I was looking and writing a dedicated slave firmware for the microcontroller. -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 2/26/2011 3:35 PM, cogoman wrote: On 02/24/2011 01:35 PM, emc-users-requ...@lists.sourceforge.net wrote: It would be nice if there was a way to interface CNC equipment via USB such that the timing of actions was being scheduled in the controller rather than by the host O/S. That way it would not be necessary to run a specific version of the O/S hacked up to provide real time scheduling. All that would be necessary would be to have the host able to guarantee response within a few seconds. Parallel ports are effectively obsolete these days and finding a motherboard with support is starting to confine selection. So this is effectively meaning I have to use an obsolete machine that I am not sure I could replace in a hurry. Sounds like the NGX BlueBoard-LPC1768-H might be all the hardware you'd need. http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id=65 http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id=65 Hmmm, wonder if that board could be programmed to be a real time ethernet slave? Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Hello Mark, Ethernet is a sloppy name for IEEE 802.3, the real protocol name is Carrier Sense Multiple Access/Collision Detection (CSMA/CD). In local wireless networks, /CA is used (Collosion Avoidance) instead of /CD. Other Bus protocols are based on the same idea such as CAN bus (CSMA/CR, Collision Resolution) and others. All these protocols have in common that they are asynchronous and data transmission is randomly timed, depending on network load. Just the same especially with serial busses like USB. Since EMC is based on exactly timed pulse generation, there is PRINCIPALLY no use trying to adapt it to the protocol systems mentioned. Why not stick to the parallel port? Best regards Peter Blodow Mark Wendt (Contractor) schrieb: On 2/26/2011 3:35 PM, cogoman wrote: On 02/24/2011 01:35 PM, emc-users-requ...@lists.sourceforge.net wrote: It would be nice if there was a way to interface CNC equipment via USB such that the timing of actions was being scheduled in the controller rather than by the host O/S. That way it would not be necessary to run a specific version of the O/S hacked up to provide real time scheduling. All that would be necessary would be to have the host able to guarantee response within a few seconds. Parallel ports are effectively obsolete these days and finding a motherboard with support is starting to confine selection. So this is effectively meaning I have to use an obsolete machine that I am not sure I could replace in a hurry. Sounds like the NGX BlueBoard-LPC1768-H might be all the hardware you'd need. http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id=65 http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id=65 Hmmm, wonder if that board could be programmed to be a real time ethernet slave? Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On Sun, Feb 27, 2011 at 06:56:18AM -0500, Mark Wendt (Contractor) wrote: On 2/26/2011 3:35 PM, cogoman wrote: Sounds like the NGX BlueBoard-LPC1768-H might be all the hardware you'd need. http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id=65 http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id=65 Hmmm, wonder if that board could be programmed to be a real time ethernet slave? Yup, especially if an open-source copy of the protocol stack is to hand. There's no copyright message in EMAC.c, the ethernet packet-driver downloadable from the site, so I don't know its status. If OK, it could form a basis for UDP communication with an EMC2 host. Given no other traffic on the cable, it wouldn't be hard to keep that RT from the slave end. (There's tcpip software there, but we'd want UDP.) The board is only a breakout board for the CPU, but that's worth something, because 0.25mm wide pins at 0.5mm spacing are a bit tricky to hand solder. (Too easy to misalign or bend, for one thing.) I notice that 50% of the reviews, and the marketing page warn that the soldering on the board is crap. Some rework with a soldering iron needs to be added to the cost. Yes, it's cheap, and very convenient for experimenting, but hard limit switches and charge pump would be doubly mandatory with something so inherently unreliable, I expect. If the community decided to go with this, I'd buy a couple, and arc up the gnu toolchain for this target, just for the fun of playing with ARM. These chips have one quadrature encoder input. At 100 MHz, and with hardware counters, I'd figure we could handle a couple. However, I'm much more familiar with AVR (Atmel), have the toolchain up to working temperature, and can fix problems more quickly there. There are AVR-based ehternet card, including ethernut 1.3: http://www.egnite.de/egnite-shop/ethernut-13/ethernut-bulk-13h.html Click on Technische Daten. (Sorry, I don't see an English link. It's surely there somewhere.) There's an open-source project at: http://www.ethernut.de/ At least that 4-layer board won't need rework before it can be used. Erik -- Meskimen's Law: There's never time to do it right, but always time to do it over. -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
2011/2/27 Erik Christiansen dva...@internode.on.net: http://www.egnite.de/egnite-shop/ethernut-13/ethernut-bulk-13h.html Click on Technische Daten. (Sorry, I don't see an English link. It's surely there somewhere.) Take a look at top right corner :) Viesturs -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 2/27/2011 8:03 AM, Peter Blodow wrote: Hello Mark, Ethernet is a sloppy name for IEEE 802.3, the real protocol name is Carrier Sense Multiple Access/Collision Detection (CSMA/CD). In local wireless networks, /CA is used (Collosion Avoidance) instead of /CD. Other Bus protocols are based on the same idea such as CAN bus (CSMA/CR, Collision Resolution) and others. All these protocols have in common that they are asynchronous and data transmission is randomly timed, depending on network load. Just the same especially with serial busses like USB. Since EMC is based on exactly timed pulse generation, there is PRINCIPALLY no use trying to adapt it to the protocol systems mentioned. Why not stick to the parallel port? Best regards Peter Blodow Peter, I work on a daily basis with Ethernet as a system and network administrator. We're talking about a Real Time Ethernet, with a dedicated master and slave network interface, which should preclude any worry of collisions on a duplex connection between the two. Modern systems with full duplex connections no longer really rely on CSMA/CD. As defined in 802.3, that's just one of the protocols that have been used, not the only way. The reason we're talking about Ethernet as one of the possibilities for communications between the computer and the controller is the possibility of the ultimate demise of the parallel port, and the difficulties of real time communications on the USB bus. So, Real Time Ethernet ain't quite your Granddaddy's TCP/IP or USB protocol. If it can be worked into a real time communication scheme, unlike USB, it may be a possible alternative to the parallel port when the parallel port is no longer supported. And, if it's a dedicated port, like the parallel port is, and can be controlled by the Real Time kernel, it can handle the timed pulses in 100 Mb, Gig E, or 10 Gig E depending on the network interface that's supported by the software. Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Ethernet medium can be in realtime take a look at sercos3 for example. It uses standard ethernet hardware. Sent on the TELUS Mobility network with BlackBerry -Original Message- From: Mark Wendt (Contractor) mark.we...@nrl.navy.mil Date: Sun, 27 Feb 2011 11:10:15 To: Enhanced Machine Controller (EMC)emc-users@lists.sourceforge.net Reply-To: Enhanced Machine Controller \(EMC\) emc-users@lists.sourceforge.net Subject: Re: [Emc-users] Guidance on usb comp On 2/27/2011 8:03 AM, Peter Blodow wrote: Hello Mark, Ethernet is a sloppy name for IEEE 802.3, the real protocol name is Carrier Sense Multiple Access/Collision Detection (CSMA/CD). In local wireless networks, /CA is used (Collosion Avoidance) instead of /CD. Other Bus protocols are based on the same idea such as CAN bus (CSMA/CR, Collision Resolution) and others. All these protocols have in common that they are asynchronous and data transmission is randomly timed, depending on network load. Just the same especially with serial busses like USB. Since EMC is based on exactly timed pulse generation, there is PRINCIPALLY no use trying to adapt it to the protocol systems mentioned. Why not stick to the parallel port? Best regards Peter Blodow Peter, I work on a daily basis with Ethernet as a system and network administrator. We're talking about a Real Time Ethernet, with a dedicated master and slave network interface, which should preclude any worry of collisions on a duplex connection between the two. Modern systems with full duplex connections no longer really rely on CSMA/CD. As defined in 802.3, that's just one of the protocols that have been used, not the only way. The reason we're talking about Ethernet as one of the possibilities for communications between the computer and the controller is the possibility of the ultimate demise of the parallel port, and the difficulties of real time communications on the USB bus. So, Real Time Ethernet ain't quite your Granddaddy's TCP/IP or USB protocol. If it can be worked into a real time communication scheme, unlike USB, it may be a possible alternative to the parallel port when the parallel port is no longer supported. And, if it's a dedicated port, like the parallel port is, and can be controlled by the Real Time kernel, it can handle the timed pulses in 100 Mb, Gig E, or 10 Gig E depending on the network interface that's supported by the software. Mark -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Parallel ports are effectively obsolete these days and finding a motherboard with support is starting to confine selection. So this is effectively meaning I have to use an obsolete machine that I am not sure I could replace in a hurry. Sounds like the NGX BlueBoard-LPC1768-H might be all the hardware you'd need. http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id = 65 http://shop.ngxtechnologies.com/product_info.php?cPath=21products_i d =65 Hmmm, wonder if that board could be programmed to be a real time ethernet slave? The board doesn't have any of the Ethernet hardware - RJ45 connector or PHY chip. The MAC they speak of is built into the chip. Here's a better source of Ethernet+microcontroller breakout boards... http://www.propox.com/products/t_245.html?lang=en That one is basically the same with the Ethernet hardware. They have other models with faster microcontrollers. I've been looking at this for a rt-net implementation. I built a demo slave using an Atmel NGW100. My software was sloppy and could only get about 200Hz update rate. The real problem was doing any interfacing. I want to basic PWM out, and encoder count in, plus homing and some IOs. My plan is to use a microcontroller to do the Ethernet bits, then use an FPGA with a memory mapped interface to do the encoder reading and PWM generation. Just need some time to learn FPGA design, then implement it. -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Peter, I work on a daily basis with Ethernet as a system and network administrator. We're talking about a Real Time Ethernet, with a dedicated master and slave network interface, which should preclude any worry of collisions on a duplex connection between the two. Modern systems with full duplex connections no longer really rely on CSMA/CD. As defined in 802.3, that's just one of the protocols that have been used, not the only way. The reason we're talking about Ethernet as one of the possibilities for communications between the computer and the controller is the possibility of the ultimate demise of the parallel port, and the difficulties of real time communications on the USB bus. So, Real Time Ethernet ain't quite your Granddaddy's TCP/IP or USB protocol. Well, Mark, my granddaddy died in 1960. Concerning Ethernet, I was thinking of the times when I worked close to the IEEE 802 group way back when Siemens participated in technical innovation... It's just that I was so happy that PC's, being cheaply available and relatively easy to write programs for, have such a simple port as the parallel is. Well, I either will have to get used to the idea of writing special programs for dedicated processors again or else?? Whenever I had the impression to have cought up on technological development, it was running away again... Peter Blodow -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 02/24/2011 01:35 PM, emc-users-requ...@lists.sourceforge.net wrote: It would be nice if there was a way to interface CNC equipment via USB such that the timing of actions was being scheduled in the controller rather than by the host O/S. That way it would not be necessary to run a specific version of the O/S hacked up to provide real time scheduling. All that would be necessary would be to have the host able to guarantee response within a few seconds. Parallel ports are effectively obsolete these days and finding a motherboard with support is starting to confine selection. So this is effectively meaning I have to use an obsolete machine that I am not sure I could replace in a hurry. Sounds like the NGX BlueBoard-LPC1768-H might be all the hardware you'd need. http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id=65 http://shop.ngxtechnologies.com/product_info.php?cPath=21products_id=65 -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Phil, what makes me appreciate EMC2 so much is the fact that I can use a regular off-the-shelf-PC and it will perform all there is needed to run a machine. Basically, no controller box, no C programming, no stepper indexer etc. re necessary. The price is that you can't upgrade your Linux. So what? I don't see the use of upgrading a perfectly running operating system if it does all it is supposed to (never change a winning team!). In case you feel the need for an upgrade it will be for another purpose aside of EMC - well, take another PC for that other purpose as you can find them thrown out at every company's junkyard. If you are so good in electronics and programming that you don't need EMC's real time features, why don't you build your own specific controller, write your code by yourself and forget EMC? EMC comprises great features for running CNC machines, almost for free, and it is not supposed to substitute expensive dedicated machine control computers (although it seems to be able for this, too, as we have seen in examples). Peter My concern is the following: Note: Do Not upgrade Ubuntu from the installed version as it will prevent EMC from working. If EMC was running on pure stock Linux rather than Linux+Realtime extensions it would be able to tollerate upgrades through the regular process. If the responsibility for timing was shifted from the kernel to the controller it would not be necessary to run with the 'realtime' package installed in the kernel. However if it was a windows os there would be lots of hacking going on to get something that appears to be real time Windows is not built to support real time. Although rather ironically the original kernel was. It is pretty difficult to timeshare a machine between real time tasks and supporting a GUI with real-time response. One or the other is going to end up taking precedence. The problem with USB is that it is a contended bus. So even if your host is running in real time there is no guarantee that signals get out to the end points in real time. That is why there is still a demand for MIDI interfaces despite them being much slower than USB. I would really prefer to have separate CPUs dedicated to scheduling and GUI. That way I don't have to worry about keeping the timing clean over the data wires, the only timing sensitive wires are carrying the signal. If I could eliminate the parallel port requirement, I can eliminate the need for a separate controller box entirely and just bring down one of the laptops when necessary. It also makes it possible to use the mill and the lathe simultaneously from a single host. -Original Message- From: Phillip Hallam-Baker hal...@gmail.com Date: Thu, 24 Feb 2011 12:03:05 To: Enhanced Machine Controller (EMC)emc-users@lists.sourceforge.net Reply-To: Enhanced Machine Controller \(EMC\) emc-users@lists.sourceforge.net Subject: Re: [Emc-users] Guidance on usb comp On Thu, Feb 24, 2011 at 6:44 AM, John Thornton bjt...@gmail.com wrote: Frank Tkalcevic wrote: The PoKeys55 docs claim that the device implements a standard USB HID keyboard and joystick. Based on that, I'd expect it to just work with hal_input. Is that not the case? You are correct. It doesn't use the HID Descriptor reports, so generic software can't query the device for its capabilities or message formats. Their messaging is specified in their document, but needs to be code specifically. The pokeys makers approached me to add support for it to hidcomp, but given how little spare time I have, it was too much work. What would be involved in getting it to work? When I tested it I only tested the analog inputs then lost track of it for a while. It would be nice if there was a way to interface CNC equipment via USB such that the timing of actions was being scheduled in the controller rather than by the host O/S. That way it would not be necessary to run a specific version of the O/S hacked up to provide real time scheduling. All that would be necessary would be to have the host able to guarantee response within a few seconds. Parallel ports are effectively obsolete these days and finding a motherboard with support is starting to confine selection. So this is effectively meaning I have to use an obsolete machine that I am not sure I could replace in a hurry. Making such an implementation safe would require a little more. There really should be some form of safety switch in the loop and that really should be capable of shutting the machine down dead, cutting power to the stepper motors and the spindle. A safety stop should be hardwired and failsafe. I am just making a box for my xylotec controller and I am seriously thinking about adding a safety stop and routing power to the machine being driven through the box. But in the longer term
Re: [Emc-users] Guidance on usb comp
All, I have to agree with Peter. EMC2 has allowed me, as a hobbyist, to build an affordable CAD CAM post processor CNC system that rivals the big name ones I use in my workplace. Tool tables, pre-setter, homing, coolant, cutter comp, etc. I have huge respect for the intellectual curiosity of command line Linux coders, but if you want to make chips using off-the-shelf hardware for a retrofit or new build then EMC2 is perfect. Many thanks to its developers! Jim -Original Message- From: Peter Blodow [mailto:p.blo...@dreki.de] Sent: Friday, February 25, 2011 3:21 AM To: Enhanced Machine Controller (EMC) Subject: Re: [Emc-users] Guidance on usb comp Phil, what makes me appreciate EMC2 so much is the fact that I can use a regular off-the-shelf-PC and it will perform all there is needed to run a machine. Basically, no controller box, no C programming, no stepper indexer etc. re necessary. The price is that you can't upgrade your Linux. So what? I don't see the use of upgrading a perfectly running operating system if it does all it is supposed to (never change a winning team!). In case you feel the need for an upgrade it will be for another purpose aside of EMC - well, take another PC for that other purpose as you can find them thrown out at every company's junkyard. If you are so good in electronics and programming that you don't need EMC's real time features, why don't you build your own specific controller, write your code by yourself and forget EMC? EMC comprises great features for running CNC machines, almost for free, and it is not supposed to substitute expensive dedicated machine control computers (although it seems to be able for this, too, as we have seen in examples). Peter My concern is the following: Note: Do Not upgrade Ubuntu from the installed version as it will prevent EMC from working. If EMC was running on pure stock Linux rather than Linux+Realtime extensions it would be able to tollerate upgrades through the regular process. If the responsibility for timing was shifted from the kernel to the controller it would not be necessary to run with the 'realtime' package installed in the kernel. However if it was a windows os there would be lots of hacking going on to get something that appears to be real time Windows is not built to support real time. Although rather ironically the original kernel was. It is pretty difficult to timeshare a machine between real time tasks and supporting a GUI with real-time response. One or the other is going to end up taking precedence. The problem with USB is that it is a contended bus. So even if your host is running in real time there is no guarantee that signals get out to the end points in real time. That is why there is still a demand for MIDI interfaces despite them being much slower than USB. I would really prefer to have separate CPUs dedicated to scheduling and GUI. That way I don't have to worry about keeping the timing clean over the data wires, the only timing sensitive wires are carrying the signal. If I could eliminate the parallel port requirement, I can eliminate the need for a separate controller box entirely and just bring down one of the laptops when necessary. It also makes it possible to use the mill and the lathe simultaneously from a single host. -Original Message- From: Phillip Hallam-Baker hal...@gmail.com Date: Thu, 24 Feb 2011 12:03:05 To: Enhanced Machine Controller (EMC)emc-users@lists.sourceforge.net Reply-To: Enhanced Machine Controller \(EMC\) emc-users@lists.sourceforge.net Subject: Re: [Emc-users] Guidance on usb comp On Thu, Feb 24, 2011 at 6:44 AM, John Thornton bjt...@gmail.com wrote: Frank Tkalcevic wrote: The PoKeys55 docs claim that the device implements a standard USB HID keyboard and joystick. Based on that, I'd expect it to just work with hal_input. Is that not the case? You are correct. It doesn't use the HID Descriptor reports, so generic software can't query the device for its capabilities or message formats. Their messaging is specified in their document, but needs to be code specifically. The pokeys makers approached me to add support for it to hidcomp, but given how little spare time I have, it was too much work. What would be involved in getting it to work? When I tested it I only tested the analog inputs then lost track of it for a while. It would be nice if there was a way to interface CNC equipment via USB such that the timing of actions was being scheduled in the controller rather than by the host O/S. That way it would not be necessary to run a specific version of the O/S hacked up to provide real time scheduling. All that would be necessary would be to have the host able to guarantee response within a few seconds. Parallel ports are effectively obsolete these days and finding a motherboard with support is starting to confine selection. So this is effectively meaning I
Re: [Emc-users] Guidance on usb comp
Frank Tkalcevic wrote: The PoKeys55 docs claim that the device implements a standard USB HID keyboard and joystick. Based on that, I'd expect it to just work with hal_input. Is that not the case? You are correct. It doesn't use the HID Descriptor reports, so generic software can't query the device for its capabilities or message formats. Their messaging is specified in their document, but needs to be code specifically. The pokeys makers approached me to add support for it to hidcomp, but given how little spare time I have, it was too much work. What would be involved in getting it to work? When I tested it I only tested the analog inputs then lost track of it for a while. John -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On Thu, Feb 24, 2011 at 6:44 AM, John Thornton bjt...@gmail.com wrote: Frank Tkalcevic wrote: The PoKeys55 docs claim that the device implements a standard USB HID keyboard and joystick. Based on that, I'd expect it to just work with hal_input. Is that not the case? You are correct. It doesn't use the HID Descriptor reports, so generic software can't query the device for its capabilities or message formats. Their messaging is specified in their document, but needs to be code specifically. The pokeys makers approached me to add support for it to hidcomp, but given how little spare time I have, it was too much work. What would be involved in getting it to work? When I tested it I only tested the analog inputs then lost track of it for a while. It would be nice if there was a way to interface CNC equipment via USB such that the timing of actions was being scheduled in the controller rather than by the host O/S. That way it would not be necessary to run a specific version of the O/S hacked up to provide real time scheduling. All that would be necessary would be to have the host able to guarantee response within a few seconds. Parallel ports are effectively obsolete these days and finding a motherboard with support is starting to confine selection. So this is effectively meaning I have to use an obsolete machine that I am not sure I could replace in a hurry. Making such an implementation safe would require a little more. There really should be some form of safety switch in the loop and that really should be capable of shutting the machine down dead, cutting power to the stepper motors and the spindle. A safety stop should be hardwired and failsafe. I am just making a box for my xylotec controller and I am seriously thinking about adding a safety stop and routing power to the machine being driven through the box. But in the longer term, I am thinking of maybe programing an ardulino or the like to replace the parallel port i/f. -- Website: http://hallambaker.com/ -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
The pokeys makers approached me to add support for it to hidcomp, but given how little spare time I have, it was too much work. What would be involved in getting it to work? When I tested it I only tested the analog inputs then lost track of it for a while. In their joystick mode, they report 32 buttons and 8 axis, regardless of how many buttons and axis are connected. If you want to access all the other inputs and outputs, you need to write code to decode and build the packets defined in their specification. It's not a huge project, but it has to be done, then supported. -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 02/24/2011 12:53 PM, Frank Tkalcevic wrote: The pokeys makers approached me to add support for it to hidcomp, but given how little spare time I have, it was too much work. What would be involved in getting it to work? When I tested it I only tested the analog inputs then lost track of it for a while. In their joystick mode, they report 32 buttons and 8 axis, regardless of how many buttons and axis are connected. If you want to access all the other inputs and outputs, you need to write code to decode and build the packets defined in their specification. It's not a huge project, but it has to be done, then supported. I'm currently talking to PoLabs (the makers of PoKeys) about doing this work for them. We'll see what comes of it. -- Sebastian Kuzminsky the moon is full here every night / and i can bathe here in his light the leaves will bury every year / and no one knows i'm gone -- tom waits -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On Thu, 2011-02-24 at 17:29 +, Jean-Paul Moniz wrote: Phil Safety should always be a hardware solution. Safety software applications should be tested and certified to international standards. Signals back to a controller are just a formality imho In the case of emc. The pc is the controller and not hacked up at all. However if it was a windows os there would be lots of hacking going on to get something that appears to be real time Sent on the TELUS Mobility network with BlackBerry In a proper e-stop system, EMC2 (or other) is a peer to other e-stop triggers. EMC2 can trigger an e-stop, as well as BRB's, or other sensors, but the e-stop loop should bring the power supplies, drives, etcetera to a safe condition independently of EMC2, Windows or other software, unless it can not be avoided. An e-stop search found: http://www.hs-compliance.com/uploaded/documents/THE%20EMERGENCY%20STOP.pdf (BTW, It would not bother me at all to see the Sent by my newfangled device go away. Apparently, it is part of the user's signature and is easy to change. Why give them free advertising on your dime?) -- Kirk Wallace http://www.wallacecompany.com/machine_shop/ http://www.wallacecompany.com/E45/index.html California, USA -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Seb, I plan on working on this tonight. After reviewing this all day today. I do not see this a huge bridge to cross. Most of your shuttle comp can be used to test and add some provisions to write hard coded command packets. Basically block read the inputs and then assign to comp pins. That's what I plan to take a stab at tonight. Sent on the TELUS Mobility network with BlackBerry -Original Message- From: Sebastian Kuzminsky s...@highlab.com Date: Thu, 24 Feb 2011 13:32:17 To: Enhanced Machine Controller (EMC)emc-users@lists.sourceforge.net Reply-To: Sebastian Kuzminsky s...@highlab.com, Enhanced Machine Controller \(EMC\) emc-users@lists.sourceforge.net Subject: Re: [Emc-users] Guidance on usb comp On 02/24/2011 12:53 PM, Frank Tkalcevic wrote: The pokeys makers approached me to add support for it to hidcomp, but given how little spare time I have, it was too much work. What would be involved in getting it to work? When I tested it I only tested the analog inputs then lost track of it for a while. In their joystick mode, they report 32 buttons and 8 axis, regardless of how many buttons and axis are connected. If you want to access all the other inputs and outputs, you need to write code to decode and build the packets defined in their specification. It's not a huge project, but it has to be done, then supported. I'm currently talking to PoLabs (the makers of PoKeys) about doing this work for them. We'll see what comes of it. -- Sebastian Kuzminsky the moon is full here every night / and i can bathe here in his light the leaves will bury every year / and no one knows i'm gone -- tom waits -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Kirk, Also depending on possible risk hardware/ software implementations are required to be control reliable. Technically emc or mach or other controllers would fail that requirement unless certified by a testing lab like tuv --Original Message-- From: Kirk Wallace To: Work To: Emc mail list Subject: Re: [Emc-users] Guidance on usb comp Sent: Feb 24, 2011 3:53 PM On Thu, 2011-02-24 at 17:29 +, Jean-Paul Moniz wrote: Phil Safety should always be a hardware solution. Safety software applications should be tested and certified to international standards. Signals back to a controller are just a formality imho In the case of emc. The pc is the controller and not hacked up at all. However if it was a windows os there would be lots of hacking going on to get something that appears to be real time Sent on the TELUS Mobility network with BlackBerry In a proper e-stop system, EMC2 (or other) is a peer to other e-stop triggers. EMC2 can trigger an e-stop, as well as BRB's, or other sensors, but the e-stop loop should bring the power supplies, drives, etcetera to a safe condition independently of EMC2, Windows or other software, unless it can not be avoided. An e-stop search found: http://www.hs-compliance.com/uploaded/documents/THE%20EMERGENCY%20STOP.pdf (BTW, It would not bother me at all to see the Sent by my newfangled device go away. Apparently, it is part of the user's signature and is easy to change. Why give them free advertising on your dime?) -- Kirk Wallace http://www.wallacecompany.com/machine_shop/ http://www.wallacecompany.com/E45/index.html California, USA Sent on the TELUS Mobility network with BlackBerry -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
[Emc-users] Guidance on usb comp
So, No matter what I try I can't seem to get the pokeys55 board to play nice with emc and hal using hal_input or hid_comp. Looking at seb's shuttle comp I'm thinking that a comp could be written for this device. I have the protocol spec for the pokeys. Could I use hidraw to send request packets to the board and then receive the expected response? I would think this is relatively straight forward for some but not so clear for myself Thanks Jp aka emcrules Sent on the TELUS Mobility network with BlackBerry -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On 02/23/2011 01:31 PM, Jean-Paul Moniz wrote: So, No matter what I try I can't seem to get the pokeys55 board to play nice with emc and hal using hal_input or hid_comp. Looking at seb's shuttle comp I'm thinking that a comp could be written for this device. I have the protocol spec for the pokeys. Could I use hidraw to send request packets to the board and then receive the expected response? I would think this is relatively straight forward for some but not so clear for myself I bet it can be made to work :-) The PoKeys55 docs claim that the device implements a standard USB HID keyboard and joystick. Based on that, I'd expect it to just work with hal_input. Is that not the case? -- Sebastian Kuzminsky the moon is full here every night / and i can bathe here in his light the leaves will bury every year / and no one knows i'm gone -- tom waits -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Sebastian, The pokeys docs list 3 hid instances however I can only ever see two. The pokeys is configured in windows using its own software package where users map various input types to either keyboard or joystick functions. The keyboard hid seems to work fine with hal_input but its the joystick_hid implementation I seem to have an issue with. I can see my hal_input pins within emc but they will change state once and then stop responding. If I unplug and plug the pokeys and restart emc the same cycle repeats. Keyboard_ hid with this device is a waste in my opinion. Why copy a input device that's already connected. The joystick_hid only allows for digital and analog inputs. The goal would be to gain access to the higher functions of the pokeys like matrix keypads, lcd display, encoders and others, Of the 3 hid instances one is used to send commands to the pokeys. Using those commands one could gain direct access to all of those functions bypassing key and button mapping. Thoughts? --Original Message-- From: Sebastian Kuzminsky Sender: high...@q.com To: Work To: Emc mail list ReplyTo: Sebastian Kuzminsky Subject: Re: [Emc-users] Guidance on usb comp Sent: Feb 23, 2011 4:19 PM On 02/23/2011 01:31 PM, Jean-Paul Moniz wrote: So, No matter what I try I can't seem to get the pokeys55 board to play nice with emc and hal using hal_input or hid_comp. Looking at seb's shuttle comp I'm thinking that a comp could be written for this device. I have the protocol spec for the pokeys. Could I use hidraw to send request packets to the board and then receive the expected response? I would think this is relatively straight forward for some but not so clear for myself I bet it can be made to work :-) The PoKeys55 docs claim that the device implements a standard USB HID keyboard and joystick. Based on that, I'd expect it to just work with hal_input. Is that not the case? -- Sebastian Kuzminsky the moon is full here every night / and i can bathe here in his light the leaves will bury every year / and no one knows i'm gone -- tom waits Sent on the TELUS Mobility network with BlackBerry -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
On Wed, 23 Feb 2011 22:26:22 +, you wrote: Keyboard_ hid with this device is a waste in my opinion. Why copy a input device that's already connected. My Xkeys based pendant is basically a programmable USB keypad. The 20 buttons simply emulate various keyboard shortcuts in axis, and double up as a numeric keypad. It uses keyboard_hid. Not a waste - it's small, convenient and totally coolant and swarf proof - more than can be said of a keyboard :) Steve Blackmore -- -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
Steve I agree. My point was that for the pokeys it was a waste. The whole point of using this device is to unload HMI type I/O from more expensive mesa or other hardware. If I was just looking for a device to emulate a keyboard I probably bought exactly what you have. The exercise I am looking at is to bypass the key mapping and just map the physical I/O to hal. From there the user can decide where to send the signals. Sent on the TELUS Mobility network with BlackBerry -Original Message- From: Steve Blackmore st...@pilotltd.net Date: Wed, 23 Feb 2011 23:32:19 To: EMC2-Users-Listemc-users@lists.sourceforge.net Reply-To: st...@pilotltd.net, Enhanced Machine Controller \(EMC\) emc-users@lists.sourceforge.net Subject: Re: [Emc-users] Guidance on usb comp On Wed, 23 Feb 2011 22:26:22 +, you wrote: Keyboard_ hid with this device is a waste in my opinion. Why copy a input device that's already connected. My Xkeys based pendant is basically a programmable USB keypad. The 20 buttons simply emulate various keyboard shortcuts in axis, and double up as a numeric keypad. It uses keyboard_hid. Not a waste - it's small, convenient and totally coolant and swarf proof - more than can be said of a keyboard :) Steve Blackmore -- -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
I thought that Big John T got that HID interface working properly with the Pokeys a while ago. http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/catid,27/id,3980/limit,6/limitstart,6/lang,english/#4014 Here he mentions it. Dave On 2/23/2011 7:01 PM, Jean-Paul Moniz wrote: Steve I agree. My point was that for the pokeys it was a waste. The whole point of using this device is to unload HMI type I/O from more expensive mesa or other hardware. If I was just looking for a device to emulate a keyboard I probably bought exactly what you have. The exercise I am looking at is to bypass the key mapping and just map the physical I/O to hal. From there the user can decide where to send the signals. Sent on the TELUS Mobility network with BlackBerry -Original Message- From: Steve Blackmorest...@pilotltd.net Date: Wed, 23 Feb 2011 23:32:19 To: EMC2-Users-Listemc-users@lists.sourceforge.net Reply-To: st...@pilotltd.net, Enhanced Machine Controller \(EMC\) emc-users@lists.sourceforge.net Subject: Re: [Emc-users] Guidance on usb comp On Wed, 23 Feb 2011 22:26:22 +, you wrote: Keyboard_ hid with this device is a waste in my opinion. Why copy a input device that's already connected. My Xkeys based pendant is basically a programmable USB keypad. The 20 buttons simply emulate various keyboard shortcuts in axis, and double up as a numeric keypad. It uses keyboard_hid. Not a waste - it's small, convenient and totally coolant and swarf proof - more than can be said of a keyboard :) Steve Blackmore -- -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Guidance on usb comp
The PoKeys55 docs claim that the device implements a standard USB HID keyboard and joystick. Based on that, I'd expect it to just work with hal_input. Is that not the case? You are correct. It doesn't use the HID Descriptor reports, so generic software can't query the device for its capabilities or message formats. Their messaging is specified in their document, but needs to be code specifically. The pokeys makers approached me to add support for it to hidcomp, but given how little spare time I have, it was too much work. -- Free Software Download: Index, Search Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
