Advanced Search

-1mm might not be enough to move off of the switch so try -2 or -3.
If it still does not work, try 2 or 3.
Also, hoome sequence should be 0 for Z axis so it moves up before other joins/axis start homing.

 You can check the irqmask ( and grub setup) with this script



usage

first

chmod + checkmask,txt

then

./checkmask.txt  [ethernet_device_name]


You can pin  the Ethernet IRQ to the last processor with this script


 

usage

first

chmod + pinirq,txt

then

sudo ./pinirq.txt  [ethernet_device_name]

To disable irq coalescing temporarily for testing:

sudo ethtool -C [ethernet_device_name]  rx-usecs 0

To permanently disable IRQ coalescing, you setup the RT ethernet interface in the interfaces file, something like:

auto eth0
iface eth0 inet static
address 10.10.10.1
hardware−irq−coalesce−rx−usecs 0





 

Any Idea why my Z Axis won´t home anymore? I changed the negative travel distance. But before i noticed that i can´t change some values anymore -> see attached file

Thatś it. Thanks

Here you are...

Yes, but only for the network Mesa is attached to.

It is the included one yes. So you say I must select the option of "Use this connection ..."

Has anyone used the io pins on the drives as io pins in linuxcnc?

mariusl wrote:

Tommy I usually make use of nmtui or the network manager

Is that the included Network Manager ?
If yes, then make sure the below is selected:

Final home position from 0 to -3

please share your hal and ini files

Hi,
i get the "Joint 2 on limit switch error". I have a Mesa 7i76e and you can find my configuration for Joint 2 attached. 
I tried several things but still with no success.
Please help

Sounds good. Let's hope it keeps that way.

I ran the RPi4 for some time and it was ok, so the Rpi5 should be ok too?
Latency isn't so much of an issue now with the "-" sign.
That "should" keep it on track (sync).
You can check if it isn't doing so great with the value of the pin lcec.0.pll-reset-count
If that is way above 1 there may be a problem.
First thing to do then is to increase lcec.0.pll-step to 10000 which is max, or something in between say 5000,
setp lcec.0.pll-step 10000
If pll-reset-count then keeps increasing you need to work on the latency.

The main reason I switched to Intel-based is that the gui became more and more
sophisticated and needed more and more processing power to function smoothly.
I am hesitant to recommend a setup, there are many that works fine.
For ethercat I would definitely look for an Intel network adapter,
the rest is more form factor, power draw, cpu performance.
I have had good luck with the N100/N150 mini pcs. Intel network adapter though.

Having tested the machine this morning , the Juddering on the servo drives appears too of stopped for the time being?

Could it be due to the pi not being powerful enough and having bad latency? Yesterday I had a few windows open running whilst running linux .

If this is the issue is it worth upgrading from the pi5 to something with more powerful ? .
Any suggestions on reasonable priced upgrade options

Looks fine so far.
Can you start linuxcnc to see if any useful message comes out to explain jerkiness.
You can also try to temporarily remove/comment out the dcconf line for the spindle driver only in the ethercat.conf.