Advanced Search

I made a small video, sorry it's not a proper screen recording.


It shows the behavior quite well. The mpg count is correctly counting, joint.2. Jog count also but the axis doesn't move each time, it jumps from time to time.
Deadband is quite low, I don't think it come from this

Hallo miteinander,meine bisherige Musteranlage (LinuxCNC → Mesa → Schrittmotoren) hat grundsätzlich sehr gut funktioniert. Allerdings ist und bleibt das Ganze ein reiner Testaufbau.
Jetzt möchte ich etwas größer denken und den Schritt von der Demo- und Testanlage hin zu einem echten Prototypen machen – mit dem Ziel, daraus irgendwann einmal eine wirklich verkaufsfähige Maschine zu entwickeln.In diesem Zusammenhang bin ich auf die Leadshine EL8 gestoßen. Zusätzlich habe ich das Video von talla83 gesehen, was bei mir die Frage aufgeworfen hat, ob ein Umstieg von Mesa auf EtherCAT sinnvoll wäre.Ein mögliches Beispielsystem, wie ich es mir aktuell vorstelle:

• Linux-Rechner
• LinuxCNC
• Leadshine EL8 (Servo-Treiber)
• Beckhoff EK1100
• Beckhoff EL1018 (DI)
• Beckhoff EL2004 (DO)

Was mir aktuell noch schwerfällt einzuschätzen:

• Welches System ist im Alltag wartungsanfälliger?
• Welches ist auf Dauer robuster und stabiler?
• Welches zickt beim Systemstart eher mal herum?
• Gibt es typische Stolpersteine bei LinuxCNC + EtherCAT im Vergleich zu Mesa?

Ich würde mich sehr freuen, wenn Ihr Eure Gedanken, Bedenken oder auch ganz konkrete Erfahrungen dazu teilen könntet – insbesondere im Hinblick auf einen späteren Dauerbetrieb und einen professionelleren Maschinenaufbau.Hier noch die relevanten Links:

• Video talla83:
  
• Leadshine EL8:
  www.leadshine.com/product-detail/EL8-EC400F.html
• Beckhoff EK1100:
  www.beckhoff.com/de-de/produkte/i-o/ethe...-koppler/ek1100.html
• Beckhoff EL1018 (DI):
  www.beckhoff.com/de-de/produkte/i-o/ethe...-eingang/el1018.html
• Beckhoff EL2004 (DO):
  www.beckhoff.com/de-de/produkte/i-o/ethe...-ausgang/el2004.html

Vielen Dank schon mal für Eure Rückmeldungen!Viele Grüße
Mücke

Hi EdWorks, Some of it is documented on this forum. I got stuck at one point in the ethernet comms losing packets but that is resolved (documented on the forum). If you want to have a look, I could show you over video (Telegram/Signal).

In hindsight the setup was pretty clean and simple with a few gotchas along the way.

When I run Goode in the simulator, it asks me with a “tool change request, insert tool #”, to insert the tool before it will proceed. However, when I  run a program on my main install I do not get that same request screen. The program will wait for me to load a tool before running but it never shows me the tool change request screen or any notion that’s it’s waiting on a tool to be loaded. Is this a normal function or have I not set something up correctly?  I am running without ATC.

Just curious, would two separate LinuxCNC instances, running on two physical PC's, one for the motion control, and the other acting as the PLC meet such requirements?
Cheers,
Mark

Hi. This is almost what I was looking for . I need to make a cylindrical dial for a leadscrew with 1/8 inch move per rotation. So it would be 125 ticks, each 5th is slightly bigger, each 10th is bit more bigger. And numbers 0, 10, 20,...110, 120 beside 10th ticks. How I can modify your code?
BTW, other dials, 1/10" per rotation or even metric would very useful. 

Actually I would move to the new and better supported hm2-modbus
as there are some long-standing bugs in mesa-modbus
(the call response  mismatch is a known one)

It'5 days i'm away for work, hopefully tomorrow evening i can come back to my cnc.
Anyeway, to give an answer to tommilight and nwe.
I've 2 phisical lan cards. The 1st (let say A) is the MB's onboard lan card, the 2nd (let say B ) is a separate PCI card.
A is connected directly to mesa card. B is connected to my domestic LAN (then to my router). 2 different phisical cables. This because, if I remember, mesa cards don't love to be connected via router.
Untill 2 years ago, when I worked with my cnc, I used to connect to A (I don t remember if eth0 or eth1), that case my cnc worked perfecly, but i didn't have any access to the web or to my nas.
Then, if I connected to B (in case I need to access the web or my Nas) obviusly the cnc wasn't available.
But I remember I hade 2 different networks (eth0 and eth1).
This year something happened (I didn't do anything but start the pc), so if I connect to the web, i can access internet,but if I connect to mesa (and linux tell me "you're connected to mesa 7i76e) linuxcnc gives me errors.
Now, ok i can try to setup again the network as tommylight suggests, but could it be better to phisically remove the PCI card (B )? I do not need internet or access to my nas whe I use my cnc (that pc is exclusively dedicated  to the cnc machine), I can easily use a normal usb pendrive to transfer the cnc files (I design and program my jobs on another pc),I just care about having my cnc working......


 

Coming from a background in mostly mazak programming and alittle haas the way linux cnc handles tool offsets has been confusing me and maybe it's because I'm not setup with somthing like the carousel macros.

I am currently using axis since it's the only thing that supports the weird display size I accidentally went with. So anyways when I set tool height it moves g54s z value to the tooltip like I was touching off the part not the tool. Currently I have been storing all tool heights too p59 as a work around.

This seems like a weird way of being setup rather then storing it too it's own unique tool height reference for a tool setter. 

I'd be curious too see how the tool setter macros work exactly where do they store the height of the tool setter 

Maybe this is somthing super basic I have never had to setup a tool setter at work before and as far as I fan tell and when manually setting tool lengths mazak seems to base them on a unique height offset.

Thanks for any and all responses!

Light looks nice, very nice.

Per chance, have you written up in some detail how you accomplished this?

Hey Guys,

i need some opinions about the design, i kind of like it but im not quite sure if it is worth the time.

What would you say?

The edited version (Light) or Standart?

Thx for any help

Thanks for reply. I didn't know that as soon as I downloaded latest 2.10 and compiled and installed. The qtdragonhd also got updated. I didn't know how to check it's latest or not. Then I checked files in downloaded repo and system files. Both were same.

I have adjusted the code accordingly.
With a bit of luck, this fixes the reported issue.In addition, I have extended the README with a basic usage guide to better explain the workflow and operation.There is now an integrated grooving cycle, which is written into the program automatically when required.Feedback on whether this resolves the issue would be appreciated.

Hello;

I'm finalizing my retrofit and i struggle on the very last detail: the MPG.
My drivers are run by analog command and its works fine, I manage to get decent repeatability and precision.
However the MPG is a huge mess.
2 issues, the first one is the worst to me 
1. When i send only one pulse, the motor is not moving. When i send the second pulse, the motor moves and eventually catch back the 1 first pulse, no always. The issue is the same if joint.0.jog-vel-mode is 1 or 0.
My mpg-counts signal is straight forward and is working well. the axis and joint jog counts are equal to mpg counts 

2. When jogging, the velocity of my axis is not linear, it goes from 0 to 300 very quickly but by step (maybe 3 step by half a second) and does like this all the way

I attach my config as I'm a little bit lost on what should I do

Thank you