Advanced Search

How to get the AXIS into manual mode at start up?

I have the typical cheap pendant (XYZA select, X1, X10, X100 and 100count MPG) wired into the encoder of a new 7i96s and I've got everything working on the test bench (after some help with an axis scale = 1 issue...)
At first the MPG worked fine until I made any MDI move (home all axis or a G0 x1) and then the MPG would stop responding.
I eventually thought the problem was similiar to one posted here concerning joints vs axis so I went in a swapped a bunch of lines from joint and changed them to be given in axis format and that worked... sort of
The pendant now doesn't respond until I make some form of MDI move (home all or a G0 x1 type thing)
Looking through what I can see, the program is starting in MDI and I really want it to start in Manual
Tried a handful of things relating to force halui.teleop.enabled in the custom.hal and then in the custom_postgui.hal thinking that it was getting overwritten before the program actually got going but with not luck
Tried
setp halui.mode.manual 1
and then
setp halui.mode.teleop 1
in either the custom.hal or post_customgui.hal with no change

I've got forced homing off in the ini and would like to keep it that way
I know the logic behind having the machine not in MPG mode until it's homed but since I can do any MDI command (g0 x1) and it works after than the way is should so it's not a safety thing

As always, any help would be appreciated
M
 

Hello. I am a new user. I've been at this for days and no joy. I just came to the realization that StepConf Wizard does not work on my setup. Here is my logic. 
I have a Raspberry Pi 5 with a :
Byte2Bot Parallel Port Raspberry Pi Hat                  byte2bot.com/products/parallel-port-raspberry-pi-hat
5 Axis CNC Interface Adapter Parallel Breakout Board For Stepper Motor Drivers      byte2bot.com/products/5-axis-cnc-interfa...tepper-motor-drivers.

I am running Linux Bookworm with an installation of LinuxCNC. 

When I run StepConf Wizard and get to the X-Axis page I attempt ot see if I can move that axis - no joy. 

My theory is that it doesn't work. I feel that when I complete the process for a machine which I will call Machine, I end up with a Machine.hal program that has lines containing "hal_gpio.PIN21-out" or "hal_gpio.PIN12-out".

Using the Byte2Bot parallel port, you are NOT allowed to refer to the output on the DB25 connector as PIN1, PIN2, PIN3 etc. Using a Byte2Bot hat with a Raspberry Pi 5, you must refer to the GPIO pin that is connected to PIN2 or PIN3 as it's corresponding GPIO pin on the Raspberry Pi. As an example, PIN2 is actually GPIO9, so the line would have to be "hal_gpio.GPIO9-out" to activate PIN2. 

Seeing how StepConf Wizard is asking me what feature I want to assign to PIN2, I assume it is passing on the command ""hal_gpio.PIN2-out" which doesn't work, rather than "hal_gpio.GPIO9-out" which is required. 

I have noticed that you can pick different interfaces prior to getting to the Axis X page, and the assignments change, BUT there is no corresponding change for the Byte2Bot setup.

I might also point out that I have found examples that PIN2 etc. that reportedly works on the Pi 4  

Having made my case I suspect that I will hear "We can't set it up for every possibility" but I think I am missing something. 

EdWorks


 

I see there wasn't an "actual-position" pdo. There absolutely must be an actual-position PDO when using csp.
I think lcec_configgen didn't do a good job this time and maybe selected wrong PDOs.

There is a long thread about Lichuan LC10E and problems to get that to work.
On page 7 forum.linuxcnc.org/ethercat/52341-proble...rive?start=60#304887
they come to a working setup, a working ethercat-conf.xml
And when I look at it, it's just a standard setup with configurable PDOs, often used.
Combine the ethercat-conf.xml with the hal file from cia402 component github.com/dbraun1981/hal-cia402/blob/main/example/cia402.hal
and I think it should move.

Step/dir at 10000 steps/turn and 6000 RPM (1 MHz step rate) in not an issue for the 7I96S
but it would require a step pulse length < 0.5 usec. Not sure if the drive will work with that.
(the manual seems to lack timing specs) You can change the steps per turn scaling to make this
easier.


You could also use analog in unipolar mode (0 to +10V  plus a direction signal)
This would require an external 10V power supply (12V would be OK with the proper hal setup)

Hello everyone,I am currently trying to commission an EtherCAT servo drive (LC10E series) using LinuxCNC + LCEC + CiA402.
The slave enters OP-state correctly, torque is enabled, the shaft is held, but the servo never moves.
Any motion command (homing or G0/G1) results instantly in a following error.Below is a complete summary of everything that was tried so far.System Setup

• LinuxCNC 2.10~pre0 (uspace), clean installation via

  .deb

  packages
• EtherCAT master working
  • Slave recognized

  • ethercat slaves → PREOP / SAFEOP / OP

    all OK
• Servo drive: LC10E series (Chinese EtherCAT servo)
• EtherCAT mapping generated via

  lcec_configgen

• CiA402 component compiled and installed from GitHub
• Motor shaft is energized and held when LinuxCNC is activated
• Mechanical disturbance triggers an error from the drive → feedback works

What has already been done1. XML generated correctly

• Using:

  lcec_configgen > ethercat-conf.xml

• lcec_conf ethercat-conf.xml

  loads without errors

• halrun → show pin lcec

  lists all expected pins

2. CiA402 integration

• CiA402 compiled successfully
• Configuration created using:

  • cia402.ini

  • cia402.hal

  • ethercat-conf.xml

• CSP mode enabled:

  • setp cia402.0.csp-mode 1

• Controlword, Statusword, Velocity pins linked correctly

3. HAL issues fixed

• Fixed type mismatch (drv-fault bit vs u32) by removing problematic nets
• Removed unused / nonexistent pins
  • Example: the LC10E has no

    srv-actual-position

    PDO

4. halrun diagnostics

• EtherCAT pins work
• OP-state reached

• srv-actual-velocity

  ,

  srv-cia-controlword

  etc. update correctly
• BUT actual velocity always stays 0

5. LinuxCNC behavior

• Servo enables and holds torque
• But no movement at all
• Homing → immediate following error
• MDI

  G0 X50

  → following error
• Trying as spindle (

  M3 S1000

  ) → same issue

6. OP-state OK but drive ignores setpointsThe drive enters OP without issues, but does not react to:

• target position
• target velocity
• target torque (CST also tested)

Setpoints are written by LCEC, but the servo does not act on them.Current suspicionThe LC10E likely does not accept the setpoint PDOs generated by default, or requires:

• additional manufacturer-specific objects
• a different operation mode (CSV instead of CSP?)
• special enable/transition sequence
• custom PDO mapping not included in the ESI

The servo provides status feedback, but ignores motion commands.QuestionsHas anyone successfully used:

• LC10E, LC-EH or similar Chinese EtherCAT servos with LinuxCNC?
• CiA402 in CSP mode on these drives?
• CSV or PP mode with these drives?
• Custom PDO configuration?

Also helpful would be a minimal example of:

• which PDOs are absolutely required for motion
• how to correctly modify

  ethercat-conf.xml

  for this family of drives

Any input is greatly appreciated – I am stuck at this very last step.Thanks,
MatthiasDeutsche VersionHallo zusammen,ich versuche derzeit, einen EtherCAT-Servotreiber (LC10E-Serie) mit LinuxCNC + LCEC + CiA402 in Betrieb zu nehmen.
Der Servo geht korrekt in den OP-Zustand, das Haltemoment ist da, aber er bewegt sich nicht.
Jeder Bewegungsbefehl (Homen oder G0/G1) führt sofort zu einem Schleppfehler.Hier eine vollständige Übersicht über alle bisher getesteten Schritte.System

• LinuxCNC 2.10~pre0, frisch installiert
• EtherCAT läuft
• Slave wird korrekt erkannt
• Servotreiber: LC10E-Serie
• XML erzeugt mit

  lcec_configgen

• CiA402 kompiliert und installiert
• Motorwelle wird gehalten
• Manuelles Verdrehen erzeugt Fehler → Feedback vorhanden

Bereits getestete Schritte1. XML korrekt erzeugt

• lcec_configgen > ethercat-conf.xml

• lcec_conf

  lädt Datei ohne Fehler

• halrun → show pin lcec

  listet alle Pins

2. CiA402 korrekt eingebunden

• CSP-Modus aktiviert:

  setp cia402.0.csp-mode 1

• Statusword, Controlword, Velocity verknüpft
• Nicht vorhandene Pins (z. B. actual-position) entfernt

3. HAL-Fehler behoben

• Typfehler bei

  drv-fault

  (bit vs u32) korrigiert
• Fehlerhafte nets entfernt

4. Diagnose mit halrun

• OP-State erreicht
• EtherCAT-Pins aktiv
• Tatsächliche Geschwindigkeit bleibt immer 0

5. Verhalten in LinuxCNC

• Servo geht in Betrieb
• Homen → sofort Schleppfehler
• G0 X50 → sofort Schleppfehler
• Test als Spindel (M3 S1000) → ebenfalls keine Bewegung

6. GrundverhaltenDer Treiber arbeitet grundsätzlich, aber er reagiert nicht auf:

• Soll-Geschwindigkeit
• Soll-Position
• Soll-Drehmoment

VermutungDer LC10E benötigt vermutlich:

• spezielle PDO-Konfiguration
• CSV statt CSP
• zusätzliche Hersteller-Objekte
• besondere Freigabesequenz

Momentan liefert der Servo Statusmeldungen, führt aber keine Sollwerte aus.FragenHat jemand diesen oder ähnliche chinesische EtherCAT-Servotreiber erfolgreich mit LinuxCNC betrieben?Speziell:

• LC10E / LC-EH / vergleichbare Modelle
• CiA402 im CSP-Modus
• CSV / PP als Alternativen
• angepasstes PDO-Mapping

Und:

• Welche minimalen PDOs sind zwingend notwendig, damit Bewegung möglich ist?

Jede Hilfe ist willkommen – ich bin jetzt sehr tief in der Konfiguration drin, aber komme hier nicht weiter.Danke,
Matthias

Or should I switch to unipolar with CW and CCW?
What is the simplest and safest option?

 
 

I think `step/dir` is causing problems.

The incremental encoder has 2500 lines and therefore 10000 pulses. The motor needs 10000 pulses per revolution at 6000 RPM.

That's probably too many.

I need to somehow change the pulses via the parameters. But I'm stuck for ideas.

OK so set the scales to 1000 for now and what do you know.... they all move

Pendant is working on the TB2 encoder with the axis select and feed rate multiplier switches and the MPG doing what it supposed to do.
E-stop triggers an output that isn't connected to anything yet but it works at least
Need to connect the LED to something, probably look for a signal that shows when AXIS is in Manual Control rather than MDI

From what I understand it doesn't look like the Axis indicator "lamps" on the Axis screen will follow the pendant selections, too bad as that would actually be handy.

Next step I think is to get the spindle output working.
I'm hoping to just use a meter to watch the output since the actual spindle is a 220v VFD which I don't have here on the bench

Then I'll have to get the 7i84u hooked up and start looking at limit and home switched

Thanks again guys, more questions will be coming I'm sure
M
 

STEP_SCALE = 1.0 across the board.............. BANG HEAD!!!!
Just setting everything up on the desk so I just hit 1 on every axis in PNCconf without thinking of the effects
There's a couple of hours I'll never get back... 

Thanks guys
 

If it's moving very slowly, (and in the proper direction) that'a a scaling issue
not a wiring/hardware issue.

For step/dir drives with optocoupler inputs (like the DM556)
I would normally suggest single ended wiring:

7I96S   DRIVE
STEP+  STEP+
GND    STEP-
DIR+    DIR+
GND   DIR-
STEP-   NC
DIR-     NC
 

Be happy! you have movement!
you just need to work on your steps per mm/inch and velocities. Note that jogging uses DEFAULT_VELOCICTY so if it is low, motion will be slow. Also try a G0 in the MDI window

Would it be possible for you to give any information regarding the missing module and the installation process ?

Starting a new build for a XYZ router and I went with a 7i96s card because of availability. (Also got a 7i84u but that will create later problems)
Hooked the card up and I thought I had everything set correct for a few simple tests on the bench but obviously missing something.

Hooked the DM556T control to TB1 pins 1,2,3,4  for step -/+ and dir -/+ of the controller
Left the enabled +/- disconnected as per the StepperOnline Manual

in the PNCconf I set the first step/gen for main axis X
net x-enable=> hm2_7i96s.0.stepgen.00.enable type settings show in the machine.HAL

I can jog with the arrow keys and the DRO thinks the axis is moving but the stepper doesn't move. moves incredibly slow it turns out. G0 X100" is turning about 1/4 of a rotation on the motor and takes 15 seconds...... I must have my steps/rev way off somewhere.. back to the drawing board

(The shaft is locked when under power, turns with the power is removed which is what I would expect)

As far as I can figure from the MESA docs this should be the default wiring setup but do I need to configure different for it in the HAL or is it possible the enabled needs to be pulled down on the controller?

Tried various different things most of the afternoon now and I'm just going in circles 

Any suggestions or ways to troubleshoot would be appreciated

Thanks in advance
M

You mean like writing my own hal component?
Where can I find some more info?