I'm not entirely sure about this either.
At least G10 only affects the currently selected coordinate system. As soon as you switch, your shift disappears.
It could be the same with G43, since you're applying the length correction to G53.

Just try it out.

github.com/LinuxCNC/linuxcnc/actions/run...artifacts/6673490803
is uspace preempt-rt

I was under the impression that a 'G0 G53 Xnnn' line doesn't change the active WCS to the machine coordinates, it just causes LCNC to move to a machine coordinate for that line only.

Second, if a change in WCS causes the tool offsets to be removed everyone using LCNC would have to write programs with G43 after every WCS change.

More importantly, the order of G53-G43-G54 is what is output by the Fusion 360 standard LCNC turning post.  I can't be the only one using that post; if that order was wrong everyone who uses the post would be having the same problem.

I will certainly try out your suggestion this evening even if I'm dubious.  And thank you for giving it some thought.

I see
G53...
G43...
G54...
G43 currently refers to the machine coordinates. Then you select coordinate system 1.
That might be why G43 becomes ineffective.
Write G54 before G43
G53...
G54...
G43...
In MIDI, by the way, the last coordinate system is always active, so G54.

I hope this helps

regards

All those dodgy Chinese routers marshaled into a botnet. I don't use TP-link but the ISP's insist on using them.
One example.  www.malwarebytes.com/blog/news/2025/09/t...crosoft-365-accounts .
So far they've not found my home network.

 If so then you weren't the only one. I was trying to buy something yesterday myself , Everything was going 
as per usual until it came time to pay then all I got was an error code. Last night I found out why.
Apparently this is a worldwide outage.

I guess I'm not very tech-savvy. I couldn't find a link to the Bookworm DEB package at the bottom of the page (and I need to figure out whether it's the real-time kernel version or the userspace version).

5 years using mesa cards and this is the first time i had issues.  The cards i'm ordering via cnc.prom.ua (the same price ±, like on the official site but not need to pay customs and tax) Usually they came already opened (the sticker is damaged), maybe they testing it, or just puts their logo.
I found only DS26LS32CM,  AM26LS32AC  , AM26LS32ACDR,  26LS32AC  (which can be used?)

So, the used 7i96s card, which was on a lathe. Signal wires on one side of the El.box, power wires on the other side of the El.box - nearly one year worked well:
ch-a - working
ch-b - not working
ch-i - not working
   

The new 7i96s (rev E was not used and not tested after arrival) :
ch-a - working
ch-b - not working
ch-i - working
     

yeah, I put EtherCAT logo on my device already, and I am now one of the owners of linuxcnc-ethercat, so we should be good, I take responsibility, please share, or if you prefer mail it to me luca at aitalmac.com

github.com/LinuxCNC/linuxcnc/actions/runs/25025393650?pr=3973

look here at the bottom of the page

Here's a program snip from the last round of unsuccessful testing:



As I mentioned earlier, I've also tried the G43 on a separate line, with and without the Hn word.  I added the M1 just to stop the pgm before the spindle ramps up during testing.

The X-values are all negative based my post processor (gang tool slant bed back-lathe, X+ is 'up', and T2 is on the 'negative' side of things).  The motion is correct for my particular setup.

If I start the above program with T2 loaded and offsets applied (via MDI or GUI cmd), everything works...until the next tool change.

I've also attached my remap.py and stdglue.py just in case there's something in there messing things up.

Thanks for having a look.

Thank you so much for your support. You mentioned updating that on the main branch; I imagine you mean linuxcnc. How can I access those changes?
And if you could help me so I could make some modifications myself, I would appreciate it.

Surmetall wrote:

grandixximo wrote:

Would love to have a logo for linuxcnc-ethercat to match
 
I made one, and to be honest, I’m actually quite proud of it . However, I’m a bit hesitant to publish it. As far as I know, the EtherCAT logo is protected.

Yes, I wondered about that. It would be able to be used by someone like grandixximo who is a member of the ETG (Ethercat Technology Group) but not OK for linuxcnc to put on its web site so I think it could be shared here for those users who have the membership

Hi endian,

I spent quite a few hours on this again today and tried to narrow it down properly instead of guessing.
On the full 13-slave bus, all other slaves go OP but the AX5125 at slave 7 still drops to PREOP + ERROR. The relevant lines are

:ethercat slaves:
7 0:7 PREOP E AX5125-0000-0216 EtherCAT Drive (SoE, 1 Ch.)

dmesg:
EtherCAT WARNING 0-7: Slave does not support changing the PDO mapping!
Currently mapped PDO entries: 0x0086:00/16 0x002F:00/32. Entries to map: 0x0086:01/16 0x002F:01/32 0x0024:01/32 0x0050:01/16 0x81E4:01/32 0x81E5:01/32

EtherCAT WARNING 0-7: Slave does not support changing the PDO mapping!
Currently mapped PDO entries: 0x0087:00/16 0x0033:00/32 0x00BD:00/32. Entries to map: 0x0087:01/16 0x0033:01/32 0x00BD:01/32 0x0054:01/16 0x0028:01/32 0x81CC:01/16 0x81CD:01/16
EtherCAT ERROR 0-7: Failed to set SAFEOP state, slave refused state change (PREOP + ERROR).
EtherCAT ERROR 0-7: AL status message 0x0045: "MBX_SOE".

I also made an AX-only test again. Earlier the AX could reach OP by itself, but I still could not get the motor to actually rotate from LinuxCNC. I tried both a direct velocity-command approach and then a stepgen/position-command approach. The command chain seems to work up to the AX command pin, but the motor still does not move.

From the AX-only spin test:
halcmd getp lcec.0.Antrieb-1-AX5-0.slave-oper
FALSE
halcmd getp lcec.0.Antrieb-1-AX5-0.slave-state-op
FALSE
halcmd getp lcec.0.Antrieb-1-AX5-0.Drive-status-wo-6
40960
halcmd getp spindle.0.speed-cmd-rps
8.333333
halcmd getp stepgen.0.velocity-cmd
8.333333
halcmd getp stepgen.0.position-fb
157.122
halcmd getp sercos.0.pos-cmd
157.1553
halcmd getp lcec.0.Antrieb-1-AX5-0.Position-comman-1
164824253

ethercat slaves:
0 0:0 SAFEOP E AX5125-0000-0216 EtherCAT Drive (SoE, 1 Ch.)
dmesg:
EtherCAT ERROR 0-0: Failed to set OP state, slave refused state change (SAFEOP + ERROR).
EtherCAT ERROR 0-0: AL status message 0x001A: "Synchronization error".

I also tested removing IDNs 32972, 32969, 32971 and 32970 once from initCmds, but that changed nothing, so I restored the original file again.
So at this point I have two clear observations:
on the full bus the AX still fails with MBX_SOE,
and on the AX-only LinuxCNC spindle test I can generate command values but still cannot get the drive to rotate.

What would you check next? I am little bit out of ideas. Thanks for your time! 

greetings 

Tom