Advanced Search

are you on the soft limit or the hard limit? if the hard limit, then set a home offset so you have a bit of cushion for drift on power cycling. if its a hard limit switch, in probe basic in the side bar tab for "WCS" there is a button for LIIMITS OVERRIDE. this should ignore the limit switch and allow you to jog away. but this isn't really going to fix the issue over all. you should set the home offset with enough distance away from the limit towards the useable axis travel side so it doesn't ever encounter this situation.

hello,

I see the TC3 on your screen .. I did TC3 master to lcnc master configurator but it is not tested with the SoE drives ... maybe it will help you

let me know if it helped

regards
 

kubes wrote:

Additionally, I have found incorrect and inaccurate details regarding the setup of absolute encoders and how LCNC handles them in both document and forum searches.

Yeah search is fuzzy on the forum and utterly useless on google lately.
And this is a rare topic, so not much to go on, but more and more users have similar setups with absolute encoders and homing to index and EtherCAT, mainly due to inrush of cheap china drives with haphazard user manuals leaving users to pretty much figure stuff by trial and error.
-
I did a retrofit with absolute encoders, used the "immediate homing" feature in LinuxCNC as it was simpler for them and more in line with their type of "one off" work.
-
Did you test adding a bit of HOME_OFFSET? Usually negative values, not always.

hello,

Some time ago, Sebastian did the config ..

regards
 

Now that you confirm this feature is supported and should work. Which would have been good to know in the first response. I can dig deeper, I assume this is exposed in the hal, correct? Is there a special condition or hal pin for jog while within soft limits? While, there is tons of ductiontation, it is hard to find it when you don't know what you are looking for. If I had to guess, the issue is that it's not a direct EtherCAT issue (has it just a HAL component), but rather a dependency on the configuration from a configuration wizard.

I do agree a machine should never be beyond the limits, but if the logic is just a straight greater than or less than than I think there could be an edge case that needs better logic. Maybe this has already been addressed. I have 23-bit encoders, so that is > 8.3 million per rev. So even with the translation to machine units there is a very very small "error" factor. Meaning the "actual" location will fluctuate (bad word choice) with no movement by this very small amount. This can make the axis and/or joint appear its beyond a soft-limit. For instance, when I zero the DRO will remain "zero" but the "-" will flash on and off. Again, maybe all this has been addressed.

Additionally, I have found incorrect and inaccurate details regarding the setup of absolute encoders and how LCNC handles them in both document and forum searches. I am more than willing to contribute back, but "guidance" would be more helpful.

Lastly, tommylight  Thanks for all your efforts as a moderator; it's a thankless task.

On my opinion, better not to use USB communication. I had USB joystic connected to CNC, if USB lost communication - need to restart linuxcnc to access it one more time. USB - it's not for device control - only for file exchange - on my opinion.

Linux cnc supports EtherCAT interface. Try to make your device with communication protocol as EtherCAT slave. It will be more professional device in this case.

I have a DIY RouterMill that I've been fighting with for many years. I recently ditched MACH3 and paraports for the MESA 6i25 <=> 7i76u combo.
After flashing the 6i25 with the 5i25/7i76u x 2 bitfile and updating LCNC to 2.94 and configuring with PNCConf, all of the steppers worked as expected and were tuned for optimum movements for my machine. Turning my attention to the INPUTS, based on what I understand from the docs, the INPUTS are sinking so I configured the machine with all INPUTS as follows:

VF(24v) -> switch(NC) -> 7i76.0.0.input_nn

After starting LCNC, I notice that the ESTOP was triggered but, the button was not closed. Checking the INPUT pins with halshow, I could see that the board was showing all INPUTS as TRUE, therefore I changes all of the INPUTS to input-nn-not and restarted LCNC. Now all of the INPUTS seemed to respond correctly. When I attempted a HOME_ALL, the Z moved to positive, so I reconfigured it to home negative with PNCConf. That worked and after more reading and adjusting of some parameters, I got the Z homing correctly backing off the trigger position by .050 in.  Then I directed my attention to X.  However, I had to adjust a few more parameters for the X axis then restarted LCNC. *)#$!! All of the inputs had switched and were now reporting as TRUE. Frustrated I changed all INPUTS back to input-00 and restarted LCNC. With this session I was able to get both the Z and X axes to home correctly, however, Y was now the problem child as it would not back off the trigger position. After several attempts I finally decided to swap DIR-/DIR+ at the microstepper so the Y config was exactly as the X config. (My Y axis is at the back of my machine so I had reversed the motor in the config for correct movement). After making the change and while running PNCConf I confirmed that the Y axis was now responding correctly then proceeded to resume verifying  the HOME ALL function, The attempt failed, so I tweaked the HOME parameters but, when I restarted LCNC all of the INPUTS had again switched their state (*)#$!!,*)#$!!). More bizarre was that now no matter how I set the PROBE input (input-nn or input-nn-not) the halshow remained the same.  After several additional restarts, the input pins again flipped the input state.

At this point I'm half inclined to suspect the 7i76 board as it has one INPUT, input-04, that is always stuck in the same status in halshow much as the PROBE input is now responding.

If anyone has any ideas that I can try, I would greatly appreciated the support.

Regards,
Stephen

I use Gmail just like Todd, and there are no emails in any folder.

Hi -- did you have any luck sorting this out? I am not working with linuxcnc, but I am having a similar error with the akd2g + ethercat in ros2 and was wondering if there was a common thread.

If you unplug the Ethernet cable from the 7I96S, do any lights on the RJ45 connector illuminate?
If so, this indicates ESD damage to the 7I96S's Ethernet port. This can only be repaired by replacing
the Ethernet chip.

To avoid this kind of damage, it important to ground the chassis ground on the 7I96S to earth ground.
 

rtt min/avg/max/mdev = 0.085/0.161/3.281/0.042 ms

is better but still not good

Take a look at this thread, especially the kernel command line (setup with grub)

forum.linuxcnc.org/18-computer/56380-error-finishing-read-sigh
 

There’s a diagram in the previously provided link. No resistor for the reed switch.

I tried a resistor and no luck with the contact. I'll give Tommy's reed relay a try. Does it matter where it is in the lead and should it need a resistor?

Yes, LinuxCNC supports backing off of the limits, be it hard or soft, it has a button for that in the GUI, but :
-you never mention any info about the machine, like it uses EtherCAT
-i have no experience with EtherCAT, so no idea if normal LinuxCNC stuff applies to it
-you keep asking for help, but still fail to provide info, this is frankly agonizing
-you posted twice in the wrong sections, there is EtherCAT section on this forum, but this is not important as we do not mind moving topics around, and i did this once, now i have to do this again as you did not provide the info on the first post, nor the second nor....
-
In general, limits should never be crossed if everything is set up properly, and that is easy to achieve, but this is EtherCAT, so i am moving this there, much more experienced users there so should figure where the issues is.
I would venture a guess at setting some small HOME_OFFSET in the ini file, like a mm or two, so even when you disable motors they could not reach limits. Maybe this is all you need...

So you are saying that Linux CNC supports jogging back into soft limits? 

config attached