[Emc-users] Max speed too fast
Stephen asked: |Is there any particular reason why you chose 2,400,000 for the servo and |traj periods? I'm sure it will work for these speeds (and it does |reduce the CPU load marginally), but it shouldn't be necessary to |lengthen these. Can you share your reason for changing them? Stephen, In answer to your question: I spent a few hours doing the old "only change one thing at a time" routine and I found that as I brought base period below 10,000 even slightly, the display started to slow down and the cursor became "sticky". By 8000 it was freezing up the machine. I decided that even if it "worked" at 1 a little margin wouldn't hurt as long as I could get the number of steps per second I needed. Hence 12,000 I juggled the servo and traj from a high of 5,000,000 down to 1,000,000. I was getting some real time errors at 1,000,000 and they seemed to go away at around 3,000,000 but as I went towards 5,000,000 I started getting following errors when I had the display (TK) in continuous mode but not in any of the incremental modes. When I lowered servo and traj periods down to 3,000,000 the following errors stopped. (note that I didn't say it stopped "because" I lowered it) Again a little margin Hence 2.5 but 2,400,000 is a nice multiple of 12,000 so why not use it since it was arbitrary anyway. In my trials I always left the traj and servo the same even though it's probably not necessary and I might try changing each independently but I don't know what would be the gain since it "seems" to be behaving pretty well now. Cecil - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Max speed too fast
Cecil Thomas wrote: >Thanks Stephen and Manfredi for the help. > >For those with similar setups this might be helpful: >Here is my latest (I hesitate to say final) .ini data >This is for a 1.7 GHz pentium 3 with a 400 MHz bus > > Hopefully it's final :) >2000 steps/rev servo, 5 rev/inch lead screw, 4:1 gear down motor to >screw 40,000 input scale > >it seems to run at 48 ipm (.8 ips) with no problem and there are no >RTErrors or Following Errors (yet) >Base Period 12000 >Servo Period 2,400,000 >Traj Period 2,400,000 >Trajectory Section default vel .2 (12 ipm) maxvel .9 (54 ipm) acc >2.5 maxacc 4.0 >AXIS Section maxvel .8 stepgenmaxvel 1 maxacc 3 stepgenmaxacc 4.4 > > Is there any particular reason why you chose 2,400,000 for the servo and traj periods? I'm sure it will work for these speeds (and it does reduce the CPU load marginally), but it shouldn't be necessary to lengthen these. Can you share your reason for changing them? I ask for two reasons: 1) if you found that something didn't work at 1,000,000 but it did at 2,400,000, I'd like to know. 2) I think it's usually best to keep changes to the INI to a minimum - it makes upgrades easier if you only change things that are specific to your machine and leave the rest at default. - Steve >Thanks, >Cecil > > - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
[Emc-users] Max speed too fast
Thanks Stephen and Manfredi for the help. For those with similar setups this might be helpful: Here is my latest (I hesitate to say final) .ini data This is for a 1.7 GHz pentium 3 with a 400 MHz bus 2000 steps/rev servo, 5 rev/inch lead screw, 4:1 gear down motor to screw 40,000 input scale it seems to run at 48 ipm (.8 ips) with no problem and there are no RTErrors or Following Errors (yet) Base Period 12000 Servo Period 2,400,000 Traj Period 2,400,000 Trajectory Section default vel .2 (12 ipm) maxvel .9 (54 ipm) acc 2.5 maxacc 4.0 AXIS Section maxvel .8 stepgenmaxvel 1 maxacc 3 stepgenmaxacc 4.4 Thanks, Cecil - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Max speed too fast
Cecil Thomas wrote: >Thanks for the help. >I set my base period to 15,000, servo period to 60,000 and traj >period to 600,000. > > Ack! Unless you're running a servo machine at 1000m/min (yes, 1 kilometer per minute), you don't need a 60 microsecond servo rate. For most "normal" machines - those that go less than ~500 IPM - the 1,000,000 (1 ms) servo rate is plenty fast. There is no need to maintain the exact ratio between base, servo, and traj. You can also set traj period back to 10 ms, or set it the same as the servo rate (1 ms). That should clear up the "unexpected realtime delay" errors - those are in the TRAJ thread (I think), and look for a 20% variance in thread period - which is only 12 microseconds with the settings you changed to. - Steve >I lowered my max vel to .75 (45 inches per minute, plenty fast >enough). Everything seems happy except the realtime error which I >have not run to ground yet. >The RT error was there at the larger periods as well. I installed a >video board and attempted to kill the MOBO video but the RT error lives on. > > Hmmm. Maybe going back to the 1ms / 10ms servo/traj periods won't clear up the error then. Take a look at the actual thread times printed in dmesg. If you have numbers <2 ms or so, it's not likely to be a problem. If the error is something like 50 or more ms, then it is likely to be a problem, and we'll have to debug more. >Thanks, >Cecil > > Hope this helps - Steve - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
[Emc-users] Max speed too fast
Thanks for the help. I set my base period to 15,000, servo period to 60,000 and traj period to 600,000. I lowered my max vel to .75 (45 inches per minute, plenty fast enough). Everything seems happy except the realtime error which I have not run to ground yet. The RT error was there at the larger periods as well. I installed a video board and attempted to kill the MOBO video but the RT error lives on. Thanks, Cecil - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Max speed too fast
Well, it is hard to say, I think there is a limit to the base period (at least I saw that in my machines) and it is about 1 to 15000 ns. If I decrease it, everything froze. Said that, an approximated formula that gives you an idea of the needed base period is: BASE_PERIOD = 1 / ( INPUT_SCALE * MAX_VELOCITY * 2 ) * 10^9 The value of INPUT_SCALE * MAX_VELOCITY is the max number of step per second you will need, you double this value to be sure you will be able to obtain the step per second you need. The inverse is the time in second between a step and the successive at the max speed doubled. You multiply for 10^9 to obtain ns. With your data, you should be ok with a Base-Period of 12500, and that should be ok for your machine. Dont' forget to update the STEPGEN_MAXVEL value accordingly to the MAX_VEL value. It should be a little more than the MAX_VEL. Also remember that in the formula above, you should put the STEPGEN_MAXVEL value and not the MAX_VEL... I hope this helps. Regards, Manfredi My websites: www.m24-pro.com www.emc2cnc.altervista.org >From: Cecil Thomas <[EMAIL PROTECTED]> >Reply-To: "Enhanced Machine Controller (EMC)" > >To: "Enhanced Machine Controller (EMC)" >Subject: [Emc-users] Max speed too fast >Date: Thu, 23 Aug 2007 01:11:15 -0500 > >I am using EMC2 and a gecko servo drive to drive a servo I have >mounted on my Millrite. >When I set the max speed to 1.0 (I think that is 1 ips or 60 ipm) I >get an error saying that the velocity is too high. >I the motor/encoder/gecko drive require 2000 steps per rev and the >lead screw is 5 tpi and I have a 4 to one timing belt pulley >reduction between the motor and the axis so the input scale is 2000 x >4 x 5 or 40,000 steps per inch. >This requires 40,000 steps per second for 1 inch per second. > >I am using a 2 GHZ machine and I think I can get a higher step per >second output by changing the base period and the associated periods >but I am a little fuzzy on the relationship. I saw this explained >here sometime in the last few months but I can't find it in my >inbox. How do I relate max steps per second to base period? What >is the minimum base period for a given cpu clock speed? > >Can someone refresh me on maximum speeds and minimum base periods etc. ? > >BTW I really don't need 60 ipm but it looked like a good max speed >and the servo will do it with a squarewave generator input to the >gecko drive. I could live with half that or so as the max traverse speed. > >Thanks, >Cecil > > >- >This SF.net email is sponsored by: Splunk Inc. >Still grepping through log files to find problems? Stop. >Now Search log events and configuration files using AJAX and a browser. >Download your FREE copy of Splunk now >> http://get.splunk.com/ >___ >Emc-users mailing list >Emc-users@lists.sourceforge.net >https://lists.sourceforge.net/lists/listinfo/emc-users _ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
[Emc-users] Max speed too fast
I am using EMC2 and a gecko servo drive to drive a servo I have mounted on my Millrite. When I set the max speed to 1.0 (I think that is 1 ips or 60 ipm) I get an error saying that the velocity is too high. I the motor/encoder/gecko drive require 2000 steps per rev and the lead screw is 5 tpi and I have a 4 to one timing belt pulley reduction between the motor and the axis so the input scale is 2000 x 4 x 5 or 40,000 steps per inch. This requires 40,000 steps per second for 1 inch per second. I am using a 2 GHZ machine and I think I can get a higher step per second output by changing the base period and the associated periods but I am a little fuzzy on the relationship. I saw this explained here sometime in the last few months but I can't find it in my inbox. How do I relate max steps per second to base period? What is the minimum base period for a given cpu clock speed? Can someone refresh me on maximum speeds and minimum base periods etc. ? BTW I really don't need 60 ipm but it looked like a good max speed and the servo will do it with a squarewave generator input to the gecko drive. I could live with half that or so as the max traverse speed. Thanks, Cecil - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
