Advanced Search

Try 2.4 GHz. 5 GHz is not as reliable in an industrial setting and has much shorter range. I turned off all my 5G wifi

Failing that use a wifi USB dongle that has an external antenna and get a SMA bulkhead mount cable to extend it outside of the cabinet

Enjoy your trip, and no, not network, the Axis GUI.

Raspberry Pi 4B 4GB has operated a very conventional x,y,z mill with a Mesa 7i96 for a couple of years.
WiFi never worked and has been a back-burner issue for a year or more.

Two cubicles:
 - Power: VFD, stepper supply, general supplies, AC/DC relays,
 - Electronics: RPi, stepper drivers, DC relays, power supplies,
     - minimal 240VAC cabling for power supplies, monitor.
 Monitor, keyboard and mouse cabled.

WiFi-Extender for workshop 4m from RPi.
Test set-up on workshop bench 6m from Extender.
Bench desktop ethernet cabled to extender, has always had reliable WiFi.

I have two RPi; 4GB and 8GB.
The 4GB is currently LC2.9.2, and I downloaded LC2.9.4 recently to a separate SD card.

Investigated:
 - both 4GB and 8GB operate WiFi with either L2.9.2 or LC2.9.4 on the bench no problem,
 - confirmed linked to the Extender by switching it Off (house WiFi 5G is at -70dBm),
 - both operated with Mesa card Off,
 - both fail to launch WiFi in electronic cubicle when mounted or hung in free space.

Also:
 - both have identical (as close as I can tell) WiFi set-up,
 - both detect the Extender and show signal level at -40dBm,
 - both fail to scan for ssid using menu-config in LC2.9.4,
 - both respond to "iw dev wlan0 scan" with response for Extender at -40dBm.

Note: test in cubicle has all non-essential supplies Off: VFD, steppers, Mesa, general control,
nothing else in workshop energised, other than air conditioner and desktop.

Need suggestions on which stone to turn over next.
John.

Hat zufällig jemand erfahrung mit Modbus beim T3D Servotreiber?

Ich habe ein kleines testprogramm, aber ich bekomme keine unterschiedlichen geschwindigkeiten hin. es wird immer die Drehzahl aus Parameter P-076 ( Jog Geschwindigkeit) genommen.

Anbei die Anleitungen die ich vom Verkäufer bekommen habe und mein Demo Programm.

 
 
 
 

Hi Guys,

greetings from New Zealand. First post please don't judge too harsh.

Wanted to share some results of my experiments with a cheap fanless pc i got from AliExpress for 89usd including shipping.

Don't know the seller just chose the cheapest. 

Specs: Pentium J2900 4 cores, 2 Lan ports, 4Gb Ram, 128GB SSD, Wifi. 
Mesa board: 7i96s

Used standart 2.9.4 Debian 12 Preempt-rt ISO with 6.1.0-30-rt-amd64 #1 SMP PREEMPT_RT kernel.

List of changes i made:

1. Physically removed internal Wifi board. (plan to use usb dongle)

2. Disabled everything in BIOS.

3. Network is Realtek so had to replace the driver. 

Included r8169 driver is a big no no, so here is what i did:

•     deleted all r8169 related ko files from /lib/modules/6.1.0-30-rt-amd64/kernel/drivers/net/ethernet/realtek (there were 3)
•     Since this machine is not connected to the network i compiled r8168.ko using dkms in a VM with the same kernel and copied this one file to the above folder.
•     ran sudo depmod to refresh the tree.

4. Kernel adjustments are all in /etc/default/grub
    
    GRUB_CMDLINE_LINUX="irqaffinity=0-1 kthread_cpus=0-1 isolcpus=domain,managed_irq,2-3 rcu_nocb_pol pcie_aspm=off r8168.aspm=0 r8168.eee_enable=0 modprobe.blacklist=r8169,mei_txe"

5. changed /etc/lightdm/lightdm.conf for autologin - autologin-user=dm


the results with latency, pings and temperature are attached.


PS. I did see some occasional single pings over 1ms, very rare, may be one in an hour or even less. interestingly there are no requests in 0.5-1 range, it stays well under 0.5ms and then has 1.2-1.5 ms or so just for a single request. Could be just noise. I suspect the no name SSD does not help, but at this point don't see the need for upgrade.  Will update if this will get worse or cause any troubles in real use. 

Cheers,
DM

tommylight post=330048 identifiant utilisateur=17274Capture d'écran ?


I'm traveling, do you want a screenshot of the network configuration?

Controlador: PLC 1200
Sistema operativo: Ubuntu con Linux 2.6.32-122-rtai
Versión LinuxCNC: 2.5.0

 
 

Dear Community,
i hope this is the right place to describe a bug on the XHC-WHB04B-6 Driver.
My environments are LinuxCNC 2.8.4 with Debian 10.10 and 2.9.1 with Debian 12 ( in both the same issue )
If I use the XHC-WHB04B-6 part , are all functions working very well ... so far so good.
If I change from man Mode to MDI and use the jogwheel on the XHC-WHB04B-6, so change the mode from MDI back to man. Mode.
In Automatic exactly the same behavior ...... so far so good, but now the bug ....... if I do it the same on a running program, so will freeze in the XHC-WHB04B-6, because the Automatic Mode can not switch back to the man Mode.
To get open again,  ... switch off the Machine, switch off the XHC-WHB04B-6, start again the LinuxCNC, switch on the
XHC-WHB04B-6 and all will works again very well.
I looking about the HAL to find solutions about the problem, but in the HAL its impossible to fix this issue, its seems to be hard coded.
I looking to the source code, to find anything ( my knowledge in C+ or C++ have tight borders ).
I found something to fix the issue, but its not really working.
Result:
-The switch back to the man Mode is only working in Step Mode or in Con Mode, not in MPG Mode.
-It is possible to switch to Con Mode or to Step Mode on the XHC-WHB04B-6, its impossible to switch to the MPG Mode.
To switch to the MPG Mode has the driver no solution, just in the start is the XHC-WHB04-6 in the MPG Mode.
My idea, to fix this issue was in the hal.cc of the source code to switch in the MPG Mode if the CNC is in MDI or automatic Mode
and nothing will freeze in and working well.
I tried to programing a switch back to MPG Mode, without any faults in compiling the code, but isnt works.
Is anyone there in the community, with experience in C+ ; C++ for help to fix this issue ?
Thank you so much in advance.
Best regards
Reinhard

I would say the output of lcec_configgen is syntactically correct for an ethercat-conf.xml file.
And if you only use recognized devices you can use the generated output directly.
You may still want to change the name"Dx" parameter to something sensible and possibly more.
So it is a really good starting point.

The challenge comes with devices/slaves that are not recognized by lcec and the generic driver is used.
The output from lcec_configgen is syntactically correct as mentioned but it lacks context. It doesn't know
which PDO should be connected to which linuxcnc module and pin. The work to look in the manual
and see what the PDOs actually do remains. With cia402 drives there are boiler-plate examples usually
with great success. Again, a starting point that fixes the syntax of the config file, but the work to understand
how to use the unit still remains.

Doesn't immediately sound related. Would expect a digital IO from the unit to be involved
and not only movement of an axis.

There are some ways to troubleshoot
- any error or info message, if any, in a popup window from the hmi
- any error or info output from linuxcnc in the terminal window where it is started
- the last say 20 line from the command "sudo dmesg" especially when ethercat related 
and one then need to match that with the configuration
- hal file
- ini file
- ethercat-conf.xml
Sometimes the error is obvious, more often the info doesn't directly identify the problem.
You can have a go and look at the info and see if you can spot where it goes wrong, and why.
I can have a look too but then I need all listed info.
 

Not easily at the moment.
That particular button uses multiple images for each LED state, Qtvcp draws LEDs on top of the standard button drawing, adding another could easily be added to the code.
Qtvcp could use either technique.

What would control the LEDs? HAL pins ? Toggle state?

You add a new pdoentry for 606C to the 1A00 PDO in your ethercat-conf.xml file. In this entry idx, subidx, bitlen, is from the A6 manual, you chose a name for "halPin".
Inside you hal file you can then use that name. E.g. like lcec.0.X.yourname
In this example 0 is your first ethercat interface, X is the name of your slave. See "halshow" -> Pins ...

Hi all,

After upgrading from LinuxCNC 2.9.3 to 2.9.4, I encountered an issue in QtDragon where the tool probing routine (qt_auto_probe_tool) starts before I press the “OK” button in the manual tool change dialog (M6).

This behavior was not present in 2.9.3. Previously, the probing would only begin after confirming the tool change.

I found that explicitly disabling the automatic probing behavior in the INI file resolves the issue:

ini
[DISPLAY]
AUTO_TOOL_PROBE = False

This stops QtVCP from launching the tool probe automatically right after M6, allowing me to call the probing subroutine manually after user confirmation.

LinuxCNC 2.9.4 (Debian Bookworm)

QtDragon UI

Manual tool change with M6 remap (calling qt_auto_probe_tool.ngc manually after M6)

I hope this helps others running into the same problem.
If this change in behavior is intended in 2.9.4, it might be worth documenting.

Thanks!

Hi taloot, I am very new to CNC and am trying to setup my home made machine with QTPlasmaC. This post processor is a .doc file, Fusion is prompting me for a .cps post configuration file. Am I missing another step or something really obvious for how to load this post into my Fusion library? Appreciate any advice!

Bump to remove it from "No replies" section.
This was also posted somewhere else on the forum and replied, so this remains a test post.
Thank you.