Advanced Search

I do not use Debian, except on some industrial machines.
I use Mint Mate,
-and no i do not use it because it is similar to windows (someone here wrote that a while back),
-and no i do not use it because it looks better (been told several times on youtube) as i rarely see it - i have everything i use start at boot, so i never see the desktop,
-and no i will not use Debian as long as they do not fix the "repository self destruction",
-
I use it for a very simple and very important reason: It just works, and never fails.

With a new dual NIC PCIe intel 350 CPÜ intel i5 6 cores
I test this file to isolate some cpu tasks . Please have a look on it, make you comments please

this script acts as a fail-safe backup. It sweeps up the parent application threads, locks them to Core 3, and ensures that the human-machine interface (HMI/GUI) stays back on Cores 0–2 so it cannot starve the machine loop of processor time.
Automation Script
Open or make a script file (sudo nano /usr/local/bin/linuxcnc-rt-tune.sh)
#!/bin/bash
### ========================================================================
### LinuxCNC Real-Time Optimization Script (6 Threads, 3 Isolated Cores)
### Core Mapping: Core 3 (LinuxCNC), Core 4 (EtherCAT IRQs), Core 5 (Mesa IRQs)
### ========================================================================

# 1. Wait for network interfaces to fully come online
sleep 5

# 2. Bind EtherCAT Port Interrupts (Port 1: enp1s0f0) to Core 4 (Hex Mask 10)
for irq in $(cat /proc/interrupts | grep "enp1s0f0" | awk '{print $1}' | tr -d ':'); do
echo "Routing EtherCAT IRQ $irq to Core 4"
echo 10 > /proc/irq/$irq/smp_affinity 2>/dev/null
done

# 3. Bind Mesa Port Interrupts (Port 2: enp1s0f1) to Core 5 (Hex Mask 20)
for irq in $(cat /proc/interrupts | grep "enp1s0f1" | awk '{print $1}' | tr -d ':'); do
echo "Routing Mesa IRQ $irq to Core 5"
echo 20 > /proc/irq/$irq/smp_affinity 2>/dev/null
done

# 4. Wait for EtherCAT Master Daemon to launch, then pin it to Core 5 (Hex Mask 20)
for i in {1..30}; do
ETH_PID=$(pgrep ethercat)
if [ ! -z "$ETH_PID" ]; then
echo "Found EtherCAT Master (PID: $ETH_PID). Tasksetting to Core 5."
taskset -p 20 $ETH_PID
break
fi
sleep 1
done

# 5. Continuous Loop: Safely lock any running LinuxCNC instances to Core 3 (Hex Mask 08)
# This loops indefinitely in the background to catch LinuxCNC whenever it is started by the user.
echo "Starting continuous background monitor for LinuxCNC instances..."
while true; do
# Find the process ID for the real-time interaction layer (linuxcnc or rtapi)
LCNC_PID=$(pgrep -f "rtapi_app" || pgrep -f "linuxcnc")

if [ ! -z "$LCNC_PID" ]; then
# Check if it is already pinned to Core 3 to minimize CPU usage of this script
CURRENT_AFFINITY=$(taskset -p $LCNC_PID | awk '{print $NF}')
if [ "$CURRENT_AFFINITY" != "08" ] && [ "$CURRENT_AFFINITY" != "8" ]; then
echo "LinuxCNC instance detected (PID: $LCNC_PID). Locking safely to Core 3."
taskset -p 08 $LCNC_PID >/dev/null 2>&1
fi
fi
# Sleep for 5 seconds between checks to keep system overhead near zero
sleep 5
done

exit 0

Because we added an infinite loop to the end of the script to monitor for LinuxCNC starting up at any time, we need to change the systemd service type from oneshot to simple. This ensures systemd knows the script is intentionally running in the background.
• Open the service file:
sudo nano /etc/systemd/system/linuxcnc-rt-tune.service
• Update the contents to match this configuration:
[Unit]
Description=LinuxCNC EtherCAT and Mesa CPU Thread Affinity Tuner
After=network.target ethercat.service
Wants=ethercat.service

[Service]
Type=simple
ExecStart=/usr/local/bin/linuxcnc-rt-tune.sh
Restart=always
RestartSec=5

[Install]
WantedBy=multi-user.target
• Save and exit the file, then apply the changes and restart the background tuner:
• sudo systemctl daemon-reload
• sudo systemctl restart linuxcnc-rt-tune.service

.hal fiel
And adding cpu_affinity=3 into your .hal file tells the internal motion controller what to do
loadrt motmod servo_period_nsec=1000000 cpu_affinity=3
addf motion-command-handler servo-thread
addf motion-controller servo-thread

and delate any: loadrt base-thread ( this is not need any more and takes more pc resources )

that is what everyone gas been telling me (the part about home_ignore_limits). that other line i dont understand because i am not a HAL programmer. but there is probly several ways to accomplish this. if your one line per axis accomplishes the same thing, it is certainly more elegant and efficient.

the problem, which is really just an annoyance, is that even though ignor_limits is set to yes, when the axis is homing and touches the switch, it flashes the hard limits tripped error on the status bar for several seconds (At least on my machine it does, apparently im the only one). it doesnt actually stop the machine though, so the logic is properly ignoring the limit switch. I just wanted to cleanup the screen a bit, try and learn some HAL, and yeah even try out this AI thing since ive never been a big fan. its actually kind of cool.

oh, one more thing, i'm using QtDragon

wow.. it works... now i understand my failure with the signs...

interesting part:

you calculate a different arm length for upper arm.. i got a drawing where the 0.75" are not realy present.

but yes its first time the moves are correct. thats so nice... i was fighting with the right hand rule and know i understand my failure... the rotations arround the X axis are not correct.. and thats why i have the TCP upside down.. and the signs for the d3 and d5....

now i am near to get the real ratios... issue are the encapsulated gearboxes.. my solution... add a second encoder to the outputshaft and work with the index signals of the encoder on the motor to get a start and end point for counting but the mesuring with the incinometer was a good starter.

aciara!!! thank you...!!!!!!!! nice lesson! 

rubes wrote:

# =====================================================================
# JOINT 0: X-AXIS CONFIGURATION
# =====================================================================
# Disconnect pncconf's direct mapping to the X limit pins
unlinkp joint.0.neg-lim-sw-in
unlinkp joint.0.pos-lim-sw-in

# Invert X homing status (Output goes FALSE when homing)
net x-is-homing <= joint.0.homing
net x-is-homing => not.0.in

# Mask the X signal during homing
net min-home-x => and2.0.in0
net x-mask-control <= not.0.out
net x-mask-control => and2.0.in1

# Route masked signal back to X limits
net min-home-x-gated <= and2.0.out
net min-home-x-gated => joint.0.neg-lim-sw-in
net min-home-x-gated => joint.0.pos-lim-sw-in
 

I don't understand why and what for ...

In hal file :
net min-home-x <= parport.0.pin-12-in-not => joint.0.home-sw-in => joint.0.neg-lim-sw-in

In ini file :
[JOINT_0]
HOME_IGNORE_LIMITS = YES

i tried
setp hm2_7i96s.0.inm.00.input-00-slow true
but saw no diffference. maybe there are other things that need to go along with that but im no HAL programmer.
as such i made friends with google AI and had him write me some code to block the hard limit status in the status bar. it took several passes, and even i got to call him out when i knew something was wrong. eventually here is what he came up with, so i put it in my custom.hal and it works. figured i would post it in case anyone else runs into this and finds this thread;

# =====================================================================
# OVERRIDE X, Y, Z LIMIT FLASHING DURING HOMING (Safe from pncconf)
# =====================================================================

# 1. Load 3 instances of each logic gate (one per axis/joint)
loadrt and2 count=3
loadrt not count=3

# 2. Add them all to the real-time processing thread
addf and2.0 servo-thread
addf and2.1 servo-thread
addf and2.2 servo-thread
addf not.0 servo-thread
addf not.1 servo-thread
addf not.2 servo-thread

# =====================================================================
# JOINT 0: X-AXIS CONFIGURATION
# =====================================================================
# Disconnect pncconf's direct mapping to the X limit pins
unlinkp joint.0.neg-lim-sw-in
unlinkp joint.0.pos-lim-sw-in

# Invert X homing status (Output goes FALSE when homing)
net x-is-homing <= joint.0.homing
net x-is-homing => not.0.in

# Mask the X signal during homing
net min-home-x => and2.0.in0
net x-mask-control <= not.0.out
net x-mask-control => and2.0.in1

# Route masked signal back to X limits
net min-home-x-gated <= and2.0.out
net min-home-x-gated => joint.0.neg-lim-sw-in
net min-home-x-gated => joint.0.pos-lim-sw-in


# =====================================================================
# JOINT 1: Y-AXIS CONFIGURATION
# =====================================================================
# Disconnect pncconf's direct mapping to the Y limit pins
unlinkp joint.1.neg-lim-sw-in
unlinkp joint.1.pos-lim-sw-in

# Invert Y homing status (Output goes FALSE when homing)
net y-is-homing <= joint.1.homing
net y-is-homing => not.1.in

# Mask the Y signal during homing
net min-home-y => and2.1.in0
net y-mask-control <= not.1.out
net y-mask-control => and2.1.in1

# Route masked signal back to Y limits
net min-home-y-gated <= and2.1.out
net min-home-y-gated => joint.1.neg-lim-sw-in
net min-home-y-gated => joint.1.pos-lim-sw-in


# =====================================================================
# JOINT 2: Z-AXIS CONFIGURATION
# =====================================================================
# Disconnect pncconf's direct mapping to the Z limit pins
unlinkp joint.2.neg-lim-sw-in
unlinkp joint.2.pos-lim-sw-in

# Invert Z homing status (Output goes FALSE when homing)
net z-is-homing <= joint.2.homing
net z-is-homing => not.2.in

# Mask the Z signal during homing (Change 'min-home-z' if pncconf used 'max-home-z')
net min-home-z => and2.2.in0
net z-mask-control <= not.2.out
net z-mask-control => and2.2.in1

# Route masked signal back to Z limits
net min-home-z-gated <= and2.2.out
net min-home-z-gated => joint.2.neg-lim-sw-in
net min-home-z-gated => joint.2.pos-lim-sw-in

tommylight wrote:

with RT 6.13. 

Why 6.13 ?
Debian 13 uses 6.12.
 

Yes. You can observe this in the default configuration, e.g., sim.qtdragon.qtdragon_xyz45/qtdragon_xyza.ini.

I've checked it several different ways, both in a virtual machine and in a real one.
You press stop, the machine stops, but to turn off the spindle, you have to press stop again.

auto mode switching = True works fine.
If you want, add it to the GUI.
 

Ryzen 7 7700 integrated, 5850U integrated, 7735HS integrated, Ryzen 9 7900 integrated, RX 6800XT, RX 9060XT, RX 580XT, RX 480XT, none of those will show anything on screen with RT 6.13. Everything works, Linux boots and runs fine, but the screen goes blank as soon as the graphic drives is loaded.
And no, GRUB nomodeset does not help.

Ich bin mir nicht sicher, was du damit meinst, dass das
7I96S vorinstalliert ist

Normalerweise erstellt man eine Konfiguration
für die ausgewählte Karte Mit einem der
Konfigurationsprogramme ( pncconf oder mesact )

Beide unterstützen die 7I96 und 7I96S.

Hallo,
ich habe jetzt nach mehreren Versuchen LinuxCNC auf meinem alten Rechner installiert. Die großen Distros gingen nicht weil es nicht möglich war vom Stick zu booten und für DVD waren sie zu groß. Ich habe dann MiniOs gefunden, das konnte ich problemlos installieren und auch
den Rt-Kernel und LinuxCNC, aber es war nicht möglich im Setup ein deutsches Tastaturlayout zu installieren. Dann habe ich MX Linux 
gefunden, damit konnte ich alles und auch Mesaflash installieren. Jetzt ist im Installationsprogramm von LCNC die 7I96s vorinstalliert,
ich habe aber die 7I96, wie bekomme ich die 7I96 ins Installationsprogramm, oder kann ich einfach die 7I96s nehmen?
MfG
Hans

tommylight wrote:

AMD graphic cards do not work with new RT kernels, 4.19 is the last one that worked.
Does the CPU have an integrated graphic?
Intel works fine with RT kernel, NVIDIA works but causes issues so is unusable, AMD goes to black screen, al this with 6.13 RT kernel.
Had no time to test the 7.nn kernels yet.
 

There is some bug in newer amdgpu driver, but works.
My Ryzens work with all Debian13 RT kernels (computer I write from is Ryzen 5500U, kernel 6.12.95+deb13-rt-amd64).
My AMD A10-7800 work with 6.12.63+deb13-rt-amd64 RT kernel.
With newer than 6.12.63+deb13-rt-amd64 works too, but only with low or medium resolution, enough for Axis or Gmoccapy.
The solution for blank screen is boot with nomodeset kernel parameter (gpu runs but without hardware acceleration) and change GRUB's and system's screen resolution to low, restart and test higher resolutions.

the CPU is a xenon 5650

It’s the same situation for me; I solved it by switching to an AMD "display-only" card.

MOBO do not have integrated graphic since 915 chipset i think, CPU's do have sometimes and more and more lately.