Advanced Search

Couple more questions/observations
I see this line in both the 5i25 .hal file and the 7i76e .hal file
sets spindle-at-speed true

Is sets correct or should it be setp?

In the 5i25 configs it starts the x-axis with stepgen.05
but in my 7i76 config it starts the x-axis with stepgen.00
What should I use with the 7c81 configs?

To answer your last question, yes
it reports boardname 7c81
9k gates
FPGA pins 144 and the rest of the info looks good to me (10 stepgens)

This 8 spindle gang router with 8 independent Z axis (using closed loop stepper motors), tandem servos for the Y axis gantry, and a servo on they X axis.
 
This machine has evolved a bit over time.  It was my very first Linuxcnc(EMC2) retrofit.  I have shared some details of it through the years.  Here is a link to what it looked like when I was first setting it up on Linuxcnc.
forum.linuxcnc.org/38-general-linuxcnc-q...-help?start=50#11052
The machine originally only had 3 joints, with a single servo moving a large (2ft x 8ft) aluminum plate up and down via a centrally mounted lead screw.  All 8 spindles were bolted to that plate with a hand actuated screw to fine tune each spindle's depth.  The Y axis was moved by a set of helical rack and pinion gears on each end of the gantry, connected together by a 12ft long torque tube drive shaft, then a 12:1 belt reduction to the servo motor.  The X axis was driven by a servo/lead screw combo.  After 20years of use all of the linear components needed refreshed, so we decided to redesign most of it.

X axis received the least change, changing from a high lead lead screw connected to the servo via a belt reduction, to a new 25mm x 5mm lead ball screw directly coupled to the servo, and new linear ways.

Y axis, got new larger linear ways, and replaced the rack and pinions with two 25mm x 10mm lead ball screws and 2:1 belt reduction to the old Y and former Z servos.

For the Zs, I removed the large 3/4inch thick by 2ft x 8ft plate, cut off 8 pieces 5inch x 24inch and milled them to mount a pair of linear ways and a ball screw for each spindle and a NEMA 23 closed loop stepper motor for each.
Here is a link to a Google photo album containing some photos of the pieces I milled for each of the 8 Z-axis joints.
photos.app.goo.gl/Tx2oYkS272xBUMvZ7

Then I'm attaching a copy of my config directory so others can see how I implemented the individual Z axis joints.  The Z axis is set up as a dummy master joint that automatically homes.  Each spindle joint is set up as an "extra" joint.  There is a GladeVCP panel that lets the operator enable and disable each of the spindles.  When a spindle is enabled, that spindle's Z-joint is connected to the dummy master's position plus a tool offset.  When it is disabled the spindle's joint is moved back up to it's home position.  This way only the spindles actually being used are ever moved with the Z axis. 

There are also buttons on the Glade panel for tool probing routines for each of the spindle Z-axis.  These disable all of the other spindle Z-axis and enable the one being probed, then sets the tool offset for that spindle according to the probe results.

Then I created a GladeVCP tab window for displaying and manipulating the tool offsets for each of the spindles.
 

 
 

ich habe in der letzten Anwort  ein screen von probe-tab eingefügt,
schau mal da drauf, da ist Halui = halui zu sehen.

ich glaube wir kommen so nicht weiter..

viele grüße und danke

karlhe

This feels like a gateway to spam.  But I’m going to allow it … for now.  

EDIT:  After more than one account was made from the same location, I decided this was spam, and deleted it.

What about the other question that I had? Not sure if you noticed that?

"Do I need both lines 41 and 46?"

The original error is a watchdog bite.
This means you have a latency issue
(or the watchdog timeout is set way too fast)

I would do a ping test:

ping -c 4 [card_ip_address]
sudo ping -c 30000 -i .001 -q [card_ip_address]

Where [card_ip_address] is the IP address of the 7I92 card
(probably 10.10.10.10 or 192.168.1.121)

The last command will take 30 seconds to complete

how did you get the ECR moving in the first place? did you add the lcec_rtec.c to the lcec_main.c ? I'm a bit stuck and out of ideas

1. I can't really tell you as I generally use the newer .comp method for kinematics:
linuxcnc.org/docs/devel/html/motion/kine...erkins_comp_template

2. If I need to use the classic C file method I modify one of the built in kinematic files.

3. If I really wanted to figure it all out then I would look for an installed kinematic file (eg 'scarakins.c') and look where that pops up in the source code and copy the lines in the relevant files and replace 'scarakins.c' with my .c file (likely '/src/emc/kinematic/meson.build' and '/src/Makefile'):

 
 

I think in .hal there should be:
setp cia402.0.csp-mode 1
setp cia402.0.pos-scale 800

 


setp cia402.1.csp-mode 1
setp cia402.1.pos-scale 800

 

 

Hello community, I have tried github.com/jzolee/HAL2UDP, an external step generator: ESP32 for 3-axis in a single ESP32. But, I want to use individual ESP32 for individual motors to have extra I/O pins that can be used for other purposes.
See the image attached below for your understanding.
  
Please let me know if anyone has any idea.

Thankyou

The problem with `ethercat slaves` is presumably because your kernel is running a different version of the EtherCAT code than the one that the `ethercat` command-line tool expects. If you'd recently installed an update without rebooting, then I'd expect to see the kernel version lower than the command-line version, but that's not what I see here. So that's kind of odd. The best fix would be to update and reboot. But...

IIRC they recently changed the keys for the OpenSUSE build server. I *think* the instructions in forum.linuxcnc.org/ethercat/45336-etherc...ep-by-step?start=290 should fix this. Try that, and if it runs successfully then reboot and see if everything works?

2.9.3 is a branch inside the git tree that you cloned to your local machine. Default branch is 'master' ie currently 2.10.xx.
To compile 2.9.3 you need to switch to the '2.9' branchand then recompile

I had to ask because I don't understand the core of electronics at all, my level is that I know how to connect the wires, set the pid and so on.. I thought that maybe by chance there is some limitation related to the 7i96s in terms of the digital signals it would give to this driver.

Why do you think it won't be compatible? I don't see anything to worry about. You can just buy an extra board for the encoders and you will be able to close the loop in LinuxCNC

sry my question was ambigous.

I have already built linuxcnc to RIP

the question was how to put this extra kinematic into the build process so that it is recompiled linked and put in the right place with make