'MANUAL_TOOL_CHANGE = True'  without remap M6, the toolchange popoup is working as expected. When m6 is remaped, the popup is not showing up, but the ui crashes:


  File "/usr/lib/python3/dist-packages/PyQt6/uic/ui_file.py", line 35, in __init__
    document = ElementTree.parse(ui_file)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/xml/etree/ElementTree.py", line 1218, in parse
    tree.parse(source, parser)
  File "/usr/lib/python3.11/xml/etree/ElementTree.py", line 569, in parse
    source = open(source, "rb")
             ^^^^^^^^^^^^^^^^^^


I think its a similar problem i described with touchoff, its looking for the tool_change.ui in my config folder
 

Is there a good .hal tutorial you could point me towards? I'm guessing many of my problems stem from my incompetence there.

I fixed pos-scale so that linuxcnc and the drive match, but the other issues still occur. Odd thing is that it still gives the joint 0 following error at ~60in/min, despite the drive actually moving faster. I suspect, then, that I could make the drive move twice as fast and the jog error would stay at the same logical velocity.

I had loaded the images, must have done it wrong...
 

Okay, with "remove any current tool length offset" you mean to remove the offset saved in the tool table? Shouldn't the tool probing overwrite this offset?

 

I have extended the python code mentioned in forum.linuxcnc.org/24-hal-components/145...with-graphviz#325079 in github.com/JTrantow/LCNC_HAL_to_graphviz

With these changes I can use it to get 99% of the graphviz I use to document my configurations. 

The attached CompleteConfig.svg is an example I whipped up today to help a friend setup a spindle VFD and external estop switch.

Using an online graphvix editor, I can quickly delete the subgraphs and signals that were cluttering up the spindle VFD and estop we were working on. See attached VFDandEStop.svg.

See the readme.md in github for further detail. I have run this on multiple configurations without issue. When your config is divided into multiple files, it's relatively easy to only load the portions you want to document.

When I tried to attach individual .svg files it complains about "Unable to get properties for the image."???
 

Sorry for late reply. I've have made a new rebuild of my wiring, and i would appreciate if you guys would check if everything looks correct. I will make a video of the problem if it still there, if my wiring diagram looks right. 

The unit of pos-scale is increments/distance unit, like steps/mm or steps/inch.
If you motor's encoder has 500 pulses/rev, then it is 2000 increments/rev.
With a axis movement of 5 mm/motor revolution you should set pos-scale = 2000/5 and 2000/180 respectively.
400 and 11.111

It may be possible to set the increments/rev separately from the encoders value, check the drive manual if the above doesn't match.

 

Not sure what you mean by encoder_value.
You have the cia402.x.pos-scale to work with.
Instead of mathematically figuring it out, multiply the current value with 6.5/2 = 3.25 and try again.
If you find that number matches a known relation, use that.

I think the following errors follows from the pos-scale. At least get this right first.

More text here on how to calculate pos-scale
forum.linuxcnc.org/ethercat/22346-etherc...er?start=1360#338150

db1981 wrote:

state machine from my comp is ok . 0x6 -> 0x7 -> 0x7f . Bit 2 (quickstop) is always set.

mine too, do you fix that

Did you start the motor?

Hello,
I'm trying to start the motor but it's not working, I tried via halrun and via sudo ethercat upload -p and motor enters OP, but it doesn't accept OP_STATUS, always 0, and the motor dont move, please can you help me somehow.

I have been trying to solve these problems on my own with very little headway:

When the Servo moves 2 inches, LinuxCNC registers it as ~6.5inches regardless of the encoder_scale value

joint following error when I try to jog over ~60in/min (linuxcnc units)

joint following error when try to home (probably related)

I haven't even gotten to the real configuration and am already failing :'/
Help is appreciated

1. there is no issue changing the firmware (unless you needed the 7I74)

2. 150V full scale should be OK though it may give less margin for sensing ARC OK if you don't have an
ARC OK signal from the plasma source. You can use any divider but /32 is suggested for best noise immunity
If you use /1 you must reduce the encoder filer time constant (so not suggested)

3. Yes, for 7i96s_7i74 or 7i96s_d or 7i96s_dpl firmware, the "A0" encoder input
is always on TB2 pins 7,8 (you can check by looking at the .pin files in the firmware directory)

Backlash not working anymore in LinuxCNC v2.9.x

Build my own laser CNC machine about 4 years ago, which worked fine with LinuxCNC 2.8.2.
Using an RPi 3B+, laser and 2 x L293D controllers and 28BYJ-48 steppenmotors.

After trying to update Debian to trixie the SD-card would not boot anymore, so build new one (back to Bookworm).
(Although i can get the old data from LinuxCNC of the card.)
Debian Trixie is not that RT compatible as written all over, still need an RT version to run LinuxCNC on RPI 3B+.
Pitty there is no complete RPi 3 (only 4 and 5) version of LinuxCNC available anymore?

After installing LinuxCNC v2.9.6 (on Debian Bookworm) I could not get the machine working as it did with v2.8.2.
I also changed X-axis steppenmotor with a spindle drive to make it more accurate, this works fine (backlash between 0.1 - 0.3),
but after this I still have trouble with the Y-axis (on which i did not changed anything).
Before (v2.8.2) the Y-axis had a backlash of 1.315, but now I can not get it right anymore.
It looks like the backlash is now only working in 1 direction and I do not see the backlash correction back on the motor movement.
Even if I make the backlash much bigger, just to see if is really done by the motor.
It only draws (i.e. circles) bigger, but the error stays (see IMG_4618a.JPG with BL of 1, 2 and 3)

My question is what has been changed in v2.9.x according the backlash, as I only find 'improved'.
In my case it is not improved but getting worse.
Maybe the new changes do not work with my controllers (L293D)?
And how to solve this weird issue.

Reagrds, DG