gmane.linux.distributions.emc.user

Re: Spindle Motor (was "Joint 0 Following Error")

Gene Heskett — 2008-12-01T23:49:02

You may want to have the guys at Lowes plug in one of Hitachi's std size units with the variable speed and soft start. It, even cutting, is probably 10, maybe 20 db quieter than my similarly sized PC, and 30 db quieter than any of my old B&D;'s or Skil's. I've been using a 9/32" dovetail bit in mine to run the grooves for some sliding dovetail joints and it is hard to tell when its cutting air, or cherry wood at times. Sweetest router I ever laid a pair of hands on. Absolutely zero run out on my copy. Turn it down to the 10k range and bit burns seem to be a thing of the past if the bit is sharp. Just keep it under control when running that slow and doing climb cuts. :)

Re: Spindle Motor (was "Joint 0 Following Error")

Andrew Ayre — 2008-12-01T23:25:05

Gene - I'm now going off the idea of a router. I don't have a separate workshop, so the best I can do is a foam lined dust box to keep the noise down. Don't think that will work well for a router. Andy

Re: Spindle Motor (was "Joint 0 Following Error")

Andrew Ayre — 2008-12-01T23:23:53

Doug - looks very interesting. One thing that attracted me to a router is being able to use router bits for edging, etc. How many hours do you have on your Proxxon and is it showing any signs of wear? Andy

Re: Multi-core processors with EMC

Leslie Newell — 2008-12-01T23:01:35

This is very interesting. I only started thinking about it because I set up my house machine using the EMC live cd and noticed only one core was being used. As I don't actually need RTAI on this machine I simply switched to the generic kernel which has SMP. Les Alex Joni wrote: ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Control of AC motor in antenna rotor (Kirk Wallace)

Bob Freeman — 2008-12-01T22:40:46

Hello, EMCers, Am back from Mom's and Thanksgiving--sorry to see I missed out on the conversation regarding antenna rotor control! Thanks, Kirk, Gene, Jim, and others for great comments on these rotor/controllers. There was some question about how the controller would be used. My hope is to use a pair of rotators for AZ/EL antenna positioning in the EMCAR project for antenna measurements. And, I would like to pipe position commands from a Predict daemon to EMC and use this for LEO satellite tracking for Amateur Radio communications (of course, real-time control would help here a lot). Commercial controllers for these applications are relatively expensive; so, I think other hams and antenna folks would be interested in this approach, as well. As for control, there is one article in QST, 1998, p42, that describes a homebrew controller for the U-100 rotor (that I have). The circuit is microcontroller based and controls a Crydom CTX240D30 solid-state relay module that serves to power the rotor motor. Direction is controlled by which of two motor windings are used. My hope would be to keep the smarts inside EMC and build a motor interface based on either a triac or an off-the-shelf solid-state-relay, similar as the one above. I need to make a few measurements to see how/if the motor responds to pulsed control signals. I think that is my next step before proceeding with any circuit design. Thanks again for all the comments. Regards, Bob ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Multi-core processors with EMC

Stephen Wille Padnos — 2008-12-01T19:23:25

Though that is possible, I got the same results using the kernel "no-hlt" parameter (or however it's spelled), which is supposed to prevent that. - Steve ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Multi-core processors with EMC

John Kasunich — 2008-12-01T19:20:58

It didn't occur to me at the time, but I bet I know why the "do a lot of nothing" task helped. I bet the Linux idle task puts the CPU into a HALT or some other low power state. Even though the RT stuff is running on the other core, there might be some side effects of that state that hurt latency. I could certainly see how it would help on a single core system - if the CPU halts, then the latency will include whatever time is needed to get it out of that halt state, which might be non-trivial. Regards, John Kasunich ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Multi-core processors with EMC

Stephen Wille Padnos — 2008-12-01T19:12:17

RTAPI already does this - it uses the highest numbered active CPU for RT tasks when compiled for SMP. You need to use a kernel command-line parameter, isolcpus. This is a list of CPUs that the normal scheduler should not use. If you specify isolcpus=1 (for a two-processor system, use 3 for a quad), Linux will not use the second core. I have done reasonably extensive testing on an Intel Core 2 Duo CPU, and I found that the latencies weren't improved much simply by using an isolated CPU. What did help (a lot!) was making the Linux-managed core do a lot of nothing. Running the bash script `while true ; do echo "nothing" > /dev/null ; done` improved matters a lot. Some additional trimming of loaded modules, combined with using ext2 instead of ext3 (kjournald made a blip every 5 seconds) made that machine get 200-400ns average latencies, with the highest spikes still in the 2 us range. I suspect that a major improvement could be made on single or multiple core CPUs by locking some HAL data in CPU cache. Ideally the code would also be locked in cache, but that's a much harder problem. In some old testing, we found that the average latency dropped sharply when the BASE_PERIOD was reduced below some threshold (which was different depending on the CPU and system). Our assumption was that at very high interrupt rates, the code and/or data was never evicted from the CPU cache. - Steve ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Multi-core processors with EMC

Alex Joni — 2008-12-01T19:08:04

Here's an example using cpusets: http://kerneltrap.org/mailarchive/linux-kernel/2008/6/4/2031744 Regards, Alex ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Multi-core processors with EMC

Alex Joni — 2008-12-01T19:03:38

when booting the linux kernel you use isolcpus (http://www.linuxtopia.org/online_books/linux_kernel/kernel_configuration/re46.html) Note: isolcpus is supposed to get deprecated in favour of another mechanism. Once you've isolated linux from one core, you can run rtai specifying that core at startup (something like: "insmod rtai_hal.ko IsolCpusMask=2") Regards, Alex ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Multi-core processors with EMC

Michael Buesch — 2008-12-01T18:47:17

Nice. What are the basic config settings you have to do? How to bind rtai to a CPU?

Re: Multi-core processors with EMC

Alex Joni — 2008-12-01T18:30:47

I think the main problem is getting a kernel+RTAI combo that works on as many SMP machines as possible. We have packages (not that used as the x86 packages) for x86_64 that are compiled as SMP capable. Maybe we should have a SMP kernel for the x86 hardy users aswell. Regards, Alex ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Multi-core processors with EMC

Chris Radek — 2008-12-01T17:57:50

I did this a while back on a dual P3 system. It gave excellent latency. Like you say, you can isolate a CPU so nothing but realtime tasks run on it. No special support is needed in EMC2; it is just a kernel and RTAI configuration issue. It was quite a job to get everything compiled right. I hope it's easier now but I wouldn't bet on it. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Multi-core processors with EMC

Leslie Newell — 2008-12-01T17:46:48

Is it possible to use EMC with multi-core processors? I see that by default the kernel is compiled to only support one processor but what would happen if SMP support was compiled in? Would RTAI choke? As multi-core processors become more prevalent, would it be practical to dedicate one core to the realtime stuff and let the other(s) do the rest? It seems wasteful but I would have thought it would give very low latency. Les ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Tune up- HAL Oscilloscope.

Jon Elson — 2008-12-01T17:26:58

I still don't know what interface you are using, but that MAY be correct. It would certainly limit your speed if the drive was assuming 10 V = 100%. Yes. Jon ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: general control thoughts

Jim Combs — 2008-12-01T15:17:07

Actually why not create a network loop test much like the interrupt latency test. Measure the Ping time and get a known good "fast" and "Reliable" network. ie Low utilization of the bandwidth. Then run UDP vs TCP/IP and you should be able to run a control system that would be usable. Granted you would not have the high speed parallel port response, but you might get something useful. I think most drivers would provide some usable bandwidth. Maybe a minimum requirement would be a 100Mb network (Wired) for starters. Figure out what that takes to make that work then evaluate and see if something slower would work. That will take you into the wireless 802.11 envelope. Jim Combs Lexmark - Lexington, Ky ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: general control thoughts

Jim Combs — 2008-12-01T15:35:53

Ken, I needed to do something like this for a printer network interface a long time ago. By using UDP vs TCP/IP you remove all the overhead of the TCP/IP stack. At the HAL parallel port interface do a Network UDP send using a very small packet with the 8 bit parallel port output data (or maybe several sets of port data). Always expect an immediate response from the addressed device with the port status. You can tag the UDP packet with a running count to verify the response is from the sent packet. It also lets you know when the network went south for some reason. Now when HAL reads the parallel port status, its just the last received data from the network device. The throughput is quite good on even a large network. The remote device can be a simple Microcontroller with minimal code. Use UDP broadcast (IP = 255.255.255.255.255) to "Find" a device on the network for initial setup. Their reply will have their IP address. This works and works quite well (at least it did for the printer interface). What we found was the network latency was actually quite small and the response time was constant even with other network protocols running and even with high network usage. Jim Combs Lexmark Kenneth Lerman s e-ltd.com> To "Enhanced Machine Controller 11/28/2008 01:17 \(EMC\)" PM public.gmane.org> cc Please respond to Subject "Enhanced Machine Re: [Emc-users] general control Controller thoughts \(EMC\)" lists. sourceforge.net> I think the easiest way to do this is for a point to point network or even a network with a hub is to use raw ethernet packets. An ethernet packet has a payload of up to 1500 bytes. That should be large enough for most things we would want. On top of the payload we have around 20 bytes of overhead, including a CRC. I think we could come up with a simple polling protocol. The EMC machine would be the master. To me, the thing that needs the most thought is how to build a generic HAL component to handle this. It is the same problem that SWP and I discussed at Fest regarding Modbus. Ideally, you could have a configuration file for the driver that specified the devices (in the case of ethernet if would be by MAC address) and how to map HAL pins to channels on the device. We would need to have a comprehensive set of signal types. There is a decent amount of work involved in spec'ing this out. Offhand, for inputs, I see things line single bit, multi-bit word input, signed word input, counter input, counter that resets on read input. I would make words all the same length; at least thirty-two bits, but possibly sixty four bits. The fact that the transmission time is very small makes that reasonable. Enough, for now. I don't think I have the time to really work on this now. Ken Jon Elson wrote: ... out. challenge prizes world -- Kenneth Lerman Mark Kenny Products Company, LLC 55 Main Street Newtown, CT 06470 888-ISO-SEVO 203-426-7166 ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Emc-users mailing list Emc-users-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f< at >public.gmane.org https://lists.sourceforge.net/lists/listinfo/emc-users ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Tune up- HAL Oscilloscope.

noel — 2008-12-01T14:38:30

Aram, 2 PID loops are ok to use, just a bit more difficult to tune. I use Yaskawa motors/drives with the m5i20 & 7i33. I have to tune the yaskawa drive and motor first with yaskawa software and get them running smoothly. Usually I tune them fairly 'soft', that is not very stiff (in velocity mode). Then I tune them with EMC. It can be a bit tricky, but patience pay off. The motor/drive tuning is very important!! Also, since there are more parameters, it is a good idea to 'write it down' so you don't go around in circles. Good luck. Noel. 'Roguish' -----Original Message----- From: amtb-tcN7jwhIyp60cBpDP109yFaTQe2KTcn/< at >public.gmane.org [mailto:amtb-tcN7jwhIyp60cBpDP109yFaTQe2KTcn/< at >public.gmane.org] Sent: Monday, December 01, 2008 5:26 AM To: Enhanced Machine Controller (EMC) Subject: Re: [Emc-users] Tune up- HAL Oscilloscope. Hi i should be able to go up to +10V to -10V. Am i right? is 1.0 output units means i use only 1.0V out of 10V that system can output? does pic www.conceptmachinery.com/Sh12.jpg looks ok? thanks aram ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Emc-users mailing list Emc-users-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f< at >public.gmane.org https://lists.sourceforge.net/lists/listinfo/emc-users No virus found in this incoming message. Checked by AVG - http://www.avg.com Version: 8.0.176 / Virus Database: 270.9.12/1822 - Release Date: 12/1/2008 8:23 AM ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Tune up- HAL Oscilloscope.

Anders Wallin — 2008-12-01T13:31:33

This looks better, now your Xoutput is not saturating and the error also looks trapezoidal. This is a system you can start to tune - the previous one was hopeless. Check with a multimeter or oscilloscope what you get out from the DAC, and what the servodrive can accept. As someone suggested, there might be a speed/acceleration (or voltage/current) limit in the drive also. AW ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Tune up- HAL Oscilloscope.

2008-12-01T13:26:17

Hi i should be able to go up to +10V to -10V. Am i right? is 1.0 output units means i use only 1.0V out of 10V that system can output? does pic www.conceptmachinery.com/Sh12.jpg looks ok? thanks aram ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Tune up- HAL Oscilloscope./NEW picture

Kirk Wallace — 2008-12-01T05:38:32

... snip I am out of my league here, but I had Jon's new brushless amp in mind. ---------- Kirk http://www.wallacecompany.com/machine_shop/ ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/