Advanced Search

Testing this right now on a Lenovo ThinkPad Yoga 370, just added isolcpus=2-3 to boot options, still excursions after a while on batteries.
Stay tuned, testing isolcpu=3-4 now

Is that the default mode?
I did test with stock Mint22, but the latency was not ok, and i waguely recall mentioning Dynamic somewhere, but i had no isolcpu in use.
Will have another look later, thank you.
BTW i did install 6.10RT and that works, did not boot with 6.11 or 6.12 at all.

7i76 oder 7i76E?
Die 7i96S hab venig IO und es gibt beim eusurplus.com immer, nem ein 7i84 mit 32 eingange und 16 ausgange und einfach zu verkablen.

  cnc@cnc:~$ uname -a
Linux cnc 6.8.0-48-generic #48-Ubuntu SMP PREEMPT_DYNAMIC Fri Sep 27 14:04:52 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux
cnc@cnc:~$ linuxcnc
LINUXCNC - 2.10.0~pre0
cnc@cnc:~$ sudo chrt 99 ping -i .001 -q -c 20000 10.10.10.10
[sudo] password for cnc:    
PING 10.10.10.10 (10.10.10.10) 56(84) bytes of data.

--- 10.10.10.10 ping statistics ---
20000 packets transmitted, 20000 received, 0% packet loss, time 20034ms
rtt min/avg/max/mdev = 0.047/0.050/1.014/0.007 ms

cnc@cnc:~$ latency-histogram --nobase --sbinsize 1000







 

what about the availability? i haven't been active here for a long time. 2 years ago the 7i76e was not available, currently it is not available either. did it ever come back? or should i better use another card? i already have some 7i76s in use and would like to use the same hardware on all machines...

wie ist das mit den 7i76? kommen die irgendwann wieder bei Mesa, oder sollte man auf etwas anderes wechseln?

PCW wrote:

It's not a issue with Intel MACs
If you disable IRQ coalescing
and pin IRQs, You get pretty decent
performance.

For example on this HP8300 I have been running a test on a 7I76EU
with a 3 KHz servo thread for about 3 days while using the PC for general
web browsing, email etc and have not had any issues.
(CPU frequency is 2.9 GHz and times are in CPU clocks)


 

 

I am getting latency with an intel adapter on just about every image release after 6.1.0-13 for some reason. so it seems something got worse after that point in time.  I can repeatably return to the above image and no more delays.  i do check whenever a new release comes out to see if that changes but so far it does not.  im not certain what the difference might be but it seems maybe it could be identified with a narrowed down image update timeline.  I tried searching

not sure if that is helpful, 6.1.0-13 is no where near as good as the 4.9 stretch version was latency wise but it at least did not throw warnings.  

I tried searching for a repository of the kernel to see if i could dig through and see what changes were made but its very difficult, it looks like kernel.org was the most informative but i still could not find an actual repository with a history tree to hunt through.  anyone that could direct me to the spot to start digging?

Chris

Try running this script to see if it helps:

 

first chmod +x pinirq.txt

then

 sudo ./pinirq.txt enp0s25

(from wherever you browser downloads stuff, probably ~/Downloads)

(this is temporary and will get reset by rebooting)


The best current PC I have for network latency is a HP 8300 USDT
(AMT needs to be turned off in the BIOS) It will run for weeks at a 3 KHz
servo thread rate while doing normal PC web/compiling/etc stuff

Might be a stab in the dark, but i would change the 5V and 24V power supplies first.

My problem just got worse. Now the error has returned even while using "sudo ethtool -C enp0s25 rx-usecs 0". Its just like before popping up every 5-10 minutes and stopping my cut in the middle of the process forcing me to recover the cuts and it is very difficult. Any ideas of what can be causing this error? and also can you guys maybe recommend working pc models so I could maybe swap the computer for one that works how it is supposed to.

Start here and continue on
forum.linuxcnc.org/38-general-linuxcnc-q...ith-mesa7i76e#124688
Edit card name accordingly.

It's not a issue with Intel MACs
If you disable IRQ coalescing
and pin IRQs, You get pretty decent
performance.

For example on this HP8300 I have been running a test on a 7I76EU
with a 3 KHz servo thread for about 3 days while using the PC for general
web browsing, email etc and have not had any issues.
(CPU frequency is 2.9 GHz and times are in CPU clocks)


 

 

First up i am apologizing for the silly thing that i have missed or dont yet understand but i hope i have tried.  

I have my Probe basic (now will be previous stable on 2.9.3) setup mostly working for what i wanted with Pendant etc.. but wanted to customize the physical buttons that i have for start / resume, Step / feed hold and well stop works fine.

(Currently 7i76 with physical buttons mapped through to the following
I started with forum.linuxcnc.org/47-hal-examples/13201...ume-buttons?start=40 and thought that i would like to have it when i press the start button and the system is in a feed hold state to resume.  So first part was to change from the single mapping to using an And to only allow start to happen when idle. (I know that is how the above cycle-start works but i was trying to do it to just prove i could get it working)
#Mapped physical button with no issues to one leg of And
net my-start-resume-button-in1 and2.pause-resume-0.in0 <= hm2_7i76e.0.7i76.0.0.input-07
#Then when trying to map one of the existing pins / signals is where  i get all screwed up.  
net my-start-resume-button-in2 and2.pause-resume-0.in1 <=  halui.program.is-idle

I thought that read as the following Netname = my-start-resume-button-in2, name of signal wire to set and2.pause-resume-0.in1 (the 2nd leg of the And), The trigger signal = halui.program.is-idle 
Instead of that working (Knowing i have no output yet) i get the following error.
custom_postgui.hal:67: Pin 'halui.program.is-idle' was already linked to signal 'prog-running'

So i was well lets change the trigger to be prog.running and see what happens then.
#net my-start-resume-button-in2 and2.pause-resume-0.in1 <=  prog.running
#.../custom_postgui.hal:68: Pin 'prog.running' does not exist (I can see that prog.running is a signal not pin but i am then back to confused about how to do the 2nd part of the and)

Then i was trying to even take the output of the and to trigger 'cycle-start'
net my-start cycle-start <= and2.pause-resume-0.out
#.../custom_postgui.hal:75: Pin 'cycle-start' does not exist

i hope that i am so close but yet so far.  (The reason that i posted this in the Probe basic is my aim is to use qtpvcp.feed-override.reset to remove from feed hold (I hope that is the correct way))
Thank you for reading this far and hoping someone can fix my broken understanding.  Full code block of what i was trying to do below with no extra typing.

Hard to see on the PDF and my glasses, but those look like analog drives, not step/dir so 7i76E would not be the best option.
Can you take a picture of the drives where the pin naming is?

Hello together.

i am Martin form Germany. I have bought a Knuth Mark Super CNC with a fault Controller (GPlus450) and I will not try to fix this old Controller.
The machine has servos with feedback and Maxsine EP100 AC Servo Driver so I think it is not a big deal to change the Controller to LinuxCNC.
I also have the electrical drawings of the machine.


I would use TCP/IP for connection so I think the Mesa 7i76e is fine for me, but this card is not available. So one alternative is the 7i96.
But this card has less in and output and I think I need a lot...

Spindel: CW + CCW + Analog
every Axis 3 switches (end, end, rev)
Lub level
Start/Stop
Drive X, Y, Z on
Cooling Pump
and much more....

So what do you think of this project.
How to start and what hardware I have to use?

Have a nice day
Martin