Advanced Search

What the difference between AI and the crazy guy yelling at the sky on a street corner, the advice appears the same.

I would change these in the ini file
from:

DIRSETUP = 100
DIRHOLD = 1000
STEPLEN = 2500
STEPSPACE = 2500

to:

DIRSETUP = 10000
DIRHOLD = 10000
STEPLEN = 5000
STEPSPACE = 5000

also:

P = 0.0
I = 0.0
D = 0.0
FF0 = 1.0
FF1 = 0.0
FF2 = 0.0
BIAS = 0.0
DEADBAND = 0.0

should be:

P = 1000
I = 0.0
D = 0.0
FF0 = 0.0
FF1 = 1.0
FF2 = 0.0
BIAS = 0.0
DEADBAND = 0.0 

These changes apply to all joints.
 

I've went through the pncconf wizard to set up a 7i76eu board. I can get the x-axis to job in Linuxcnc but the motors are not moving.  All wiring is connected correctly to the board and driver.  I know this because it was moving at one point and then i made the mistake of trying to swap from Axis to Probe Basic and ChatGpt was telling me to do some stuff that obviously messed it up.  So my files here are the result of going through the set up wizard a second time.  It's not the originals.  I'm essentially trying to get back to where i was.  I have communication with the board.  DMM Dyn4 drivers has a solid green light on the front of each.  The only thing of note worthy is the Z-Axis pins from the driver are connected to the Step 0 pins on TB2, X axis on Step 1 pins, and Y axis on Step 2 pins.  This is due to short cables and i don't want to move them.  I made these changes in the wizard too.Can anyone take a look at my Hal and Ini files to see any issues?  I'm attaching the files and TXT files for reading.

If those are AI suggestions, they are pure nonsense...

To reduce network latency:

1. Make sure basic host latency setup is done (disable power management etc)

2. If you have an intel MAC on the host PC, disable IRQ coalescing.

3. Use isolcpus to run LinuxCNC on the last CPU

4.  (in conjunction with #3) Pin the Ethernet IRQ to the last CPU
(and set irqbalance to one-shot mode so it doesn't mess with the IRQ pinning)

To check maximum network latency:

sudo chrt 99 ping -i .001 -q -c 60000 10.10.10.10

( Assuming 10.10.10.10 is the 7I76E IP address ) 

This will run for 1 minute and print statistics

 

I did the 
ping between mesa is less then 0,3ms and it is per to per connection

I did not try the custom firmware 
goal is -
resualt is no change at all ... 

So you tried stuff AI has, what have you tried and what were the exact results ? To save repeating stuff.

Hello gentelmen,

How to reduce the read-all timig from 0,8ms to something below please? Ping I have beloe 0,3ms and latency at servothread below 800ns 

I am using 7i76e+7i77 at P1 with 3 axis setup..

I have tried everything what AI told me but without any success ... 

thanks

Guten morgen Tommylight
Danke für den beitrag

ich hab in der tat vergessen das min
3 Encoder eingänge
gebraucht werden. 1 im Bedienpult zum einrichten und 2 an den Spindeln zum Synchronisieren

4 Servos davon 3 0-10V für 2 Angetriebene Werkzeuge und 1 Für die Spindel 2 und -10 bis +10 für die Hauptspindel

7i96S
5 Step-Dir
11 iso Eingang
6 iso Ausgang
1 Encoder

7i83
6 -10 bis +10 Servo
6 5-24V Ausgänge

7i84S
Wie kann ich das verstehen vierkanal Encoder A/B/Z/
(Die Beschreibung ist für nicht Profies wie sie etwas schwer zu durchblicken bisher hatte ich es nur mit 7i76E zu tun)
und 4x2 Step/Dir oder PWM oder 8x 5V Ausgänge
Hier stellt sich mir die frage wie Bekomme ich den DB25 dann an die 7i96s bzw die 7i83? Es haben ja alle 3 karten verschiedene anschlüsse.

danke für die hilfe und schöne grüße


 

The mouse issue sounds like EMI so likely a grounding/line filter issue with the servo drives

Ethernet is much more EMI resistant than USB so that may be a latency issue.

To check network latency, run this:

sudo chrt 99 ping -i .001 -q -c 60000 [card_ip_addr]

Where [card_ip_addr] is either 192.168.1.121 or 10.10.10.10 depending on the 7I76E settings

This will run for 1 minute and print timing statistics
 

Is everything on the machine and the PC grounded?
Is that grounding good? This is very hard to figure out, but give it a try.

Not a lot of info to go on.
If your PC uses a realtek NIC I would search the forums for what to do. That would be a good starting point.

Next thing would be to ensure no "noisy" wires are crossing signal wires.

Next thing would be to ensure that the PSU that are powering the Mesa baord are giving a clean voltage. A good quality PSU, such as Meanwell or Weidmuller are good choices. Ensure the servo drivers are wired according to the manuals, don't skip the recommendations fro any filtering, not a cheap exercise, but well worth worth it.

Keep power wiring away from signal wiring. Don't expect a "rat's nest" to operate trouble free.

Don't wire every at once, wire one section at a time and test, trouble shoot and fix as you go. If something stops performing as it should it makes trouble shooting easier as you can start at the last addition you made.

Where ever possible terminate the ends of bare wires with "bootlace ferrules" makes inserting wiring into screw connectors easier and prevents start strands of wires floating in the breeze.

If you have implemented all these measures and have "noise" issues you will need to provide far more information. PCW may chime in and ask for you to do some "ping tests" and such, as this is his domain I shall leave it to him to guide you through that part.

Hello everyone, 
I am using mesa board 7I76E to control 5 axis cnc with servo motors 
I installed the linuxcnc on a laptop when i run linuxcnc with the mesa board on the laptop everything works well i got some wired interference like loosing Ethernet communication with the board or the mouse pad will be crashed all the time but i was ignore them because they didn't really influence on my project.
After that i installed the same version of linuxcnc on a desktop computer in fact i used the same ssd so it was the same software, the same configuration at this moment the noise issue became more than usually now i lost the communication every 5 min around ana that's a big problem. Do you have an idea what is the exact issues happened to me , how could i decrease those noise 
Thanks in advance 

Hi all!

first of all, great work! I managed to get everything running just by reading all the great posts here! All the nice HowTo´s and so on... 

But now i got stuck...

what do i have?

i´m running a RPi 5 with LinuxCNC (latest Preempt Version) 
a Mesa 7i76EU (everything working)
a Pendant WHB04-6 (everything working)
and some Beckhoff EtherCat modules (everything working)
and last i got my self an EtherCat Module for the Omron MX2 (partial working)

I can´t figure out how to drive the Omron MX2, there is some mismatch in the type of the controlword
my ethercat xml has only the Modulelist, no special entries like syncmaster and so on... 

the Device shows up in the HalShow under lcec, all "Pins" are viseble, ethercat slaves shows the device... so i think its just a small config problem in the xml

maybe someone out here who has it running...

Thanks for looking / helping...

Cheers Einar 

 

Hello all... I am looking at retrofitting a Prototrak DPMV5 mill, its supposed to have electronic handwheels as well as full CNC mode.
I am making new electronic handwheels- instead of using the 100ppr units you can get cheap, i want to use a 1024ppr encoder and custom made wheels. Bringing me to the main question: Can i use the 7i95T Mesa boards 4,5,6th axis encoder inputs as MPGs? Arguably with real good encoder counts? Or am i limited to using the 8 inputs as MPGs for the normal 100ppr units?
The machine I have is electrically stripped, so imagination is the limitation (okay maybe wallet too).

ps: I have retrofit a VMC and converted a manual lathe to CNC using 360ppr encoders on the handwheels to make a similar manual/cnc mode. Sometimes it misses counts when i spin fast using a 7i76e MPG inputs.

Thanks!
 

Thanks for checking things over PCW!

I can't setp the value since it's linked to a signal, but when gcode runs M67 E0 Q0, GPIO 017 does go false.

'setp hm2_7i76e.0.pwmgen.00.scale 100' has been added, this may yield better results. Since I didn't set the scale, all values (1-100) were sending full power. This tip right here may be the biggest sanity check, thanks PCW!