Moin,
mich würde interessieren wie du deine A6 getuned hast.
Ich hab noch immer leichtes fiepen an manchen Stellen und kriegs einfach nicht weg.

vikingj
Can you post your INI and hal files.

Hmmmm
as said: it has never left the office desk and has never been near a “large machine” in real life.
If damaged it can only be because of the failed attempts to connect to LinuxCNC.

Found this useful thread after posting:
forum.linuxcnc.org/qtvcp/39423-qtdragon-...a-manual-tool-change

Looks like someone wrote something similar and had it working for qtdragon and gmoccapy for 2.8.x. Will start here and update this thread.

Missing grounds between a large machine and the controller can easily damage
analog and encoder pins, that is, unless a permanent ground exists between
the machine and the control electronics. there can easily be enough potential
difference to damage electronic connections.

Hi everyone,
I’ve recently been trying to set up a IndraDrive from Bosch Rexroth with LinuxCNC, and I’ve run into some difficulties.
After quite a bit of research here on the forum, I’ve narrowed it down to two possible causes:

1. The IndraDrive supports EoE in addition to CoE. Could EoE be interfering with PDO configuration? Is there a recommended way to disable EoE on the drive? I’m using the latest LinuxCNC version and followed the EtherCAT tutorial from the forum.
2. I’ve seen that some people define initcmds in the Ethercat-conf.xml, and I believe those may include PDO configuration. What would be the best approach here? In my XML I already have configPDOs="true", so I assumed initcmds wouldn’t be necessary.

I can configure the PDOs exactly as I want using IO Engineering (Bosch’s equivalent to TwinCAT), but once I switch back to LinuxCNC, the PDOs are either missing or appear randomly most of the time. They are dynamically allocated from what I understand.

Any guidance would be greatly appreciated. Thanks!

I wonder how this could be damaged?
it has never been connected to anything else but 5 volt, Ethernet, one encoder and one analog out? It has never been installed and has never left the table.

MRx, nice! what hardware are you using?

and i'm still working on the new platform, also with an ethernet interface... now i can have all supported mitsubishi protocols running simultaneously on a single zynq 7010 core ok, 5 of them - parallel bus (MR-S), 2 copper serial (SSCNET I, II, MELDAS 5/6), 2 fiber (SSCNET III, III/H, MELDAS 7, . as all of them share the same multiple-of-222us cycle time...

still have the second processor core and a spare PTPv2-capable 1Gb ethernet port to try the CC-Link IE/TSN thing to complete the picture...

If you set all PID values to 0 and you still get 1V outputs  from the 7I97T analog outputs
,the 7I97T is likely damaged

Sorry, mismatch between advertised clock and real clock
so the baud rates were all off...

 Denser configurations like this have a slower clock to meet timing
and this is the first time I have made on of these configs, and I had
not quite gotten al the ducks in a row...

This _should_ be better

 
 

Hi PCW,Thanks for that. There's a problem with the sserial port; with the previous firmware (attached), the system as it's currently wired inc. the hal line:
loadrt hm2_eth board_ip="10.10.10.10" config="num_encoders=6 num_pwmgens=6 sserial_port_0=0000000" works and allows us to reference the 7i84 e.g.: hm2_7i97.0.7i84.0.3.output-00

Changing only the firmware to the one you posted above, lcnc fails to start with error:
twopass: Error in file ./morb01_IO_7i84.hal:
    Pin 'hm2_7i97.0.7i84.0.3.output-00' does not exist

this is with all wiring identical. close checking of the readhmid output and manuals suggests that the sserial port settings should remain the same - is this expected? 


 

Hi there
I am retrofitting a giant 7 meter and 7 ton Homag BMF 411.
When it left the factory in 2005 it was equipped with Indramat servos (today Rexroth) and Rexroth DKC 03.3 drives with sercos I communication. 
With a great deal of willingness from our local Rexroth technicians i managed to reconfigure the drives to analog com and now the drives are updated with new firmware ready to be re-installed in the machine.
I have managed so far to connect the 0 joint -x axis-  to my mesa 7i97t ethernet card and i can see the DRO increasing/decreasing respectively when I rotate the motor by hand. So all in all a first table setup with one motor no limit switches, motor brake and enable signals are still not coming from linux but directly from UPS.
Here are my problems:
1  When I enable the drive the mesa card throws out 1 volt to the analog pin and the motor starts turning at a speed that looks OK like 10% of speed max.
2. When I jog the motor in continuous mode the x axis DRO counts up 0,18 mm when pressing and another 0,18 when releasing the tab - same result if i use the jog arrows in axis.
Any ideas?
I am not a linux specialist - in fact this is my first time putting fingers to a linux pc but i have been working with CNCs for 3 decades 
  

I have a 3 axis mill/router that I primarily use for woodworking/plastics. It's an older machine but upgraded/rewired and running great with LinuxCnc 2.9.7 and QtDragon. It has a manual tool change spindle that I'm happy with but I'd like to automate the process of changing tools a little. Currently, I output a separate nc file for each tool that I run on a job and I manually adjust the z-height after each tool change and reset the G54 z-zero before starting the next nc file/toolpath. It works but it's a little slow and error prone.

I have purchased a tool setter switch, mounted to my machine bed, wired it up to my 7I96s and it seems to work (electrically). However, I'm a little confused on what the best workflow is. Ideally, I'd like to make this as idiot proof as possible with some simple macros. Before I get started I figured I'd ask if anyone else has some macros/ideas to share.

For each job:

• Manually install the starting tool and zero it to the machine bed or work piece top
• Start the job
• Automatically move to the xy position of the tool setter to measure the z offset of this first bit
• For each M6 tool change:
  • Automatically move to the xy position for tool changes (front center)
  • Manually change the tool and confirm change complete
  • Automatically move to the xy position of the tool setter to measure the z offset of the next bit
  • Automatically calculate the difference in height between the new tool and the 1st tool and update G54 z zero
  • Automatically continue job 

I'm not sure if this is more of a configuration or QtDragon workflow question. I also realize that to everyone with a tool carousel, this seems like caveman operation. 

Bonjour, 
Je construis une CNC multifonction avec 5 axes , et 2 axes rotatif fraiseuse et pas a pas.
J ai actuellement 2 cartes chinoises mach3 des NEMA des DM556 . Un pc DELL optiplex 7010 .
Le pc n' a pas de port parallèle, j' ai donc commandé une PCIe 2 ports puce ASIX 9865 , impossible de la faire fonctionner, elle n' ai même pas reconnue sûr le slot pcie. Sûr le slot bleu PCI+PCIe elle est reconnue "lspci -v" mais , je me retrouve avec quatre adresses.
J ai essayé toutes les combinaisons possibles dans le "Hall" .
J en ai conclu que il y avait un problème de compatibilité.
Ma question est donc :: que dois-je choisir comme système, continuer a chercher des cartes PCI ou PCIe/parallèle ou une autre solution tel que les cartes MESA ? Pour les quelles je n' ai aucune connaissance.

Merci de vos réponses 
Florent 

 

Okay, will try that when problems retturn. Thanks!