Advanced Search

I would use all cores but set isolcpus=3  in the kernel command line
otherwise you run all realtime and non-realtime code on the one core.
(plus reducing overall performance)

In addition, pinning the Ethernet IRQ to the last CPU (LinuxCNCs CPU)
will help with network latency.
 

Right now I've got it turned down to only a single. One of the many documents on tuning I came across seemed to want the 7010 to run with it set to single. I'll reset it to all in the bios and re-post


--- 192.168.1.121 ping statistics ---
60000 packets transmitted, 60000 received, 0% packet loss, time 59999ms
rtt min/avg/max/mdev = 0.065/0.102/0.281/0.018 ms

M

 

There's always the possibility of installing MATE, but unfortunately the Debian setup of the Menus and panels isn't as nice as Mint's setup.

The i5 3470 is a four core chip.

Might be worth checking your BIOS to see if there is a setting that either enables or disables the 4 cores.

Thumbs up to the old timey British sayings.

Another I like is "proper poorly" but I've been told is a more Northern thing.

I ran it a couple of times
cnc@router:~$ sudo chrt 99 ping -i .001 -q -c 60000 192.168.1.121
PING 192.168.1.121 (192.168.1.121) 56(84) bytes of data.
--- 192.168.1.121 ping statistics ---
60000 packets transmitted, 60000 received, 0% packet loss, time 59999ms
rtt min/avg/max/mdev = 0.074/0.086/0.256/0.008 ms
cnc@router:~$ sudo chrt 99 ping -i .001 -q -c 60000 192.168.1.121

PING 192.168.1.121 (192.168.1.121) 56(84) bytes of data.
--- 192.168.1.121 ping statistics ---
60000 packets transmitted, 60000 received, 0% packet loss, time 59999ms
rtt min/avg/max/mdev = 0.074/0.086/0.377/0.008 ms

The histogram l ran on it's own and with 5 gears running

Nope, the network driver needs to be changed, installing the DKMS version of the Realtek drives should fix it.
Give this a try:

I have after many, many tries still cannot get the versaProbe actions to work properly. So I added some logging.

Please refer to this image inline.

 

1 issue and another oddity.

1st issue, note the lines that I have added to the qt_auto_probe_tool.ngc file in order to log and hopefully put an end to the first probe try failing.
The question, why is there a space that's being added to the filename when it's written to file?

The second is similar to a thread that I started many moons ago and still exists. In the image I point out what is actually logged in the file. It begins with this code in the aforementioned file:

touch_result= 161.824940 
calculated_offset= [161.824940 - -132.110000 + -106.700000] 
Probe Height: -132.1100
Block Height: -106.7000
Probe Result: 161.8249
Calculated Offset: 187.2349

I don't understand the calculated_offset line which includes a double negative.

Now then, in this same file I see the following reference prior to code being executed and for some reason it's almost as if these instructions are either out of order or are incorrect. Preceding numbers are actual line numbers.

16 ; The following settings should be enabled in the QtDragon interface:
  17 ;    A -  Under Probe Tool screens: insure that "Tool Measure" is enabled
  18 ;    B -  Under Settings: insure that "Use Tool Sensor" is enabled
  19 ;  The following workflow assumes using both a XYZ probe and a Z Toolsetter:
  20 ;    1 - Initial Setup: Before beginning the program, setup the Probe Tool as
  21 ;        having a zero z-offset in the tool table. (Non-zero tool lengths
  22 ;        for the probe tool can be made to work, but have multiple extra
  23 ;        steps required, and it is easy to have the tool offset interfere with
  24 ;        the remap coding and cause incorrect adjustments to your offsets.
  25 ;        This procedure assumes a zero length.)
  26 ;    2 - Initial Tool: Load the probe tool with M61 Qx, where x is the Probe tool's number
  27 ;        in the tool table. (Do not use TxM6)
  28 ;    3 - Toolsetter measure: Use button under the Probe Screens for "Probe Tool Setter
  29 ;        Z Height:" this will set and display on the Probe Settings screen the
  30 ;        "Probe HT" = #<_hal[qtversaprobe.probeheight> value in ABS coordinates.
  31 ;    4 - Workpiece Measure: Use the button under the Probe Screens for "Probe Z height
  32 ;        of material:" this will set and display on the Probe Settings screen the
  33 ;        "Block Ht" = #<_hal[qtversaprobe.blockheight]> value in ABS coordinates.
  34 ;    5 - Set Coord System (G54): Use the Probe Tool and whichever probe screen is
  35 ;        appropriate to set the coordiate system XYZ offsets needed for your job.
  36 ;        NOTE: Return_Option 3 uses the current local coordinate system, not just ABS coord.
  37 ;    6 - Prepare to Run: You may then issue a manual TnM6 command to change the tool
  38 ;        before starting the job, or if the job begins with a TnM6 command before
  39 ;        spinning the spindle, you may leave the Probe Tool installed.

And last I would like to know (if the above instructions are correct) what are "sane" values that belong in the my_machine.ini file?

Mine are as follows:
Max Z travel for Z-Axis is: -190
[VERSA_PROBE_OPTIONS]
 ps_probe_tool = 99
 
 [VERSA_TOOLSETTER]
 X = 14.68
 Y = 43.934
 Z = -55
 # Safe Z-Height, in Absolute coordinates.
 Z_MAX_CLEAR = -5
 # MAXPROBE must be positive!!
 MAXPROBE = 80
 # tool setter diameter for diameter probing
 DIAMETER = 22

 [CHANGE_POSITION]
 X = 200
 Y = 20
 Z = -5

Note that MAXPROBE in the LinuxCNC online documents still indicates a negative value of -20.

Thanks in advance for your feedback.
 

Sorry, but i lost the plot, can you remind me what was i supposed to move?
Thank you.

Try this:and also paste the result of this:

I think the complaint is for line 149. Should be G0 Z20?

OK, so it looks like, because it's in "announcements" only Mods get to vote. Which is jolly silly, if you ask me.

I will try to move it.

Yes, I used the official LinuxCNC ISO. I see what you mean about XFCE, in contrast to the previous Mint GUI.....

Had a few problems installing it, which may have been due to using a 256GB UFD. Changed to a 16GB and it went well, whether that was the problem or not.

Now I have a functioning OS, I will have a go at reinstituting gmoccapy.

I'm setting up a new system and I'm sometimes getting a unexpected realtime delay on task 0 with period 1000000 error (mostly on startup) and there was a few error finishing read ones as well. After much reading and tinkering I've got it to this point and I'm not sure if it's OK or not.
7010, all the extra anything in the bios has been turned off, only using the 1 core
built in intel network controller, onboard graphics
Ping 192.168.1.121 was up around 350ms, made the changes to the networks file and it is about 150ms now, mainly sits just under 100ms so I think the network to card connection is ok.
Histogram looks pretty "fat" though ???
Also had a tri-peak at one point... might have been a glitch though
Is it safe to just bump the period up some or is there something I can actually do to make this better

M
 

At least some honesty.