Advanced Search

Thanks Tommy.

FWIW I've been digging into isolcpus and becoming convinced it's not a good idea on this 2 core CPU at least. Maybe it helps on CPUs with more cores but I am finding very bad results with it here.

I discovered even tho I have isolcpus=1 and I locked the NIC IRQ to CPU 1 rtapi_app was running on CPU 0 during the latency measurements above. When I ran latency-histogram using taskset to lock it on CPU 1 I got horrible latency spikes.

I also found that *all* the IRQs are on CPU 1 which defeats the whole purpose of isolating it.

So I'm going to remove all the CPU isolation optimizations and just run this 2 core setup with the CPU scheduler left to do whatever it decides to do and see what happens. Of course I'll keep all the other optimizations. i.e. power management disabled, governor on performance, etc.

Ich habe inzwischen meinen A6-Servo 750 W EtherCAT von Stepperonline in einer kleinen Sondermaschine zum Laufen gebracht, in der sich ausschließlich dieser Motor befindet.

Dazu lese ich das Drehmoment aus und lasse mir das höchste Drehmoment anzeigen und mitloggen. Außerdem habe ich den Signalstatus der wenigen Pins, die der Motor hat, auf den Monitor gelegt. Sollte der Motor einen Fehler werfen, wird dieser ebenfalls angezeigt.


Was ich nicht verstehe, ist: Ich kann die Maschine freigeben, entriegeln etc. Wenn ich den Motor dann drehen lassen möchte, passiert jedoch nichts. Erst beim zweiten Mal (manuell auf Dauerbetrieb) beginnt er, sich zu drehen. Das finde ich etwas merkwürdig, aber so ist es eben.

Als Nächstes werde ich diesen einzelnen Motor durch einen Schrittmotor ersetzen. Das wird spannend, da alle anderen Schalter und Motoren (Schrittmotoren) über eine Mesacarte kommunizieren.

In der Anlage habe ich euch meine komplette Konfiguration beigelegt – für den Fall, dass jemand Interesse hat, darin zu stöbern.

Diese Nachricht ist als kurzes Feedback gedacht und nicht als Fragestellung der nächsten Schritte.


Was für mich bei der Installation von EtherCAT Master etwas verwirrend war, ist, dass nicht LinuxCNC der Master ist, sondern dass man auf dem Rechner noch weitere Dienste und Programme installieren muss, die man natürlich dann auch laden muss. Es ist also keine Erweiterung von LinuxCNC, sondern ein weiteres „Tool” auf dem Rechner, auf das LinuxCNC zugreift. Zumindest habe ich mir das so zurechtgelegt.

Gruß Mücke 

Hi Andy,The tool changer uses two motors: one to rotate the carousel into position and the other to actuate the tool change mechanism.The carousel rotation has a proximity sensor fitted, which could be used to establish a reference or home position. Once the correct tool position has been indexed, the second motor drives the tool in and out to allow the machine to perform the tool change.I’ve added a video link below showing an example of the system running on Mach4. It should give a better idea of how the sequence operates.


thanks for the help !

man pages built-in.

Now working on real hal files import/export.

Es riecht stark nach zu hoher Latenz. Wie sind die Werte vom Base und Servo-thread?

I don't think moving back to the original position after an M6 is standard practice, all post should have safe move inserted after a tool change, this is not something dealt with on the machine side IMO
If your post expects the tool change to happen and the machine to not have moved and does not give you a way to implement safety moves to re-approach the work-piece after a tool change, this is just bad design of the CAM/Post the fix should be there IMO

Not that i know of, except the included with LinuxCNC, but i can prove some for plasma and for engravers/scratchers i build:
forum.linuxcnc.org/show-your-stuff/46752...ver-etcher-scratcher

current CPU frequency: 2.10 GHz (asserted by call to kernel)

All is good, move on to testing on a machine.

Uthayne wrote:

This also occurs with Qtplasmac. I can replicate it if I use the keyboard to jog and click to another tab while jogging. It loses focus/control even if I release the keyboard jog arrow and I have to go back to the main tab and tap a jog button again to get it to stop jogging.

I never realized this. I can replicate.

I will definitely be working on a fix for this. "Stay tuned."

I tried to compile this old working 4Axis kinematics (originally 5Axiskins) but getting compiling errors

forum.linuxcnc.org/10-advanced-configura...p-with-rtcp?start=80

Not able to understand current 5Axiskins

Can you throw some example? Zero in what?

Hello,

is there any gentelman which has running Xenomai or Cobalt kernel for his iron?

I need some inspiration or talk what, how, prons and cons ... experiences !

thanks 

Tool changers are often more difficult than the test of the machine, if only because there are more variations.
What do the two motors do? What manner of toolchanger is it?

This also occurs with Qtplasmac. I can replicate it if I use the keyboard to jog and click to another tab while jogging. It loses focus/control even if I release the keyboard jog arrow and I have to go back to the main tab and tap a jog button again to get it to stop jogging.

I would leave the kinematic as is and simply set the unused joint position to zero.