Strange movements at start of g-code files

[b57378]

Hi all, have a bit of an issue not too sure if it's a bug or on my side.

I am running a uc400eth on version 2,112, (I realise this is a test version so it could possibly be the problem).

The system has been running flawlessly for as long as I have been using this test release but occasionally at the start of loading a new g-code file and pressing cycle after zeroing, the machine will rapid to random x- y positions and then return to the zero point and start the job as usual.

This problem is not g-code specific as it usually only ever occurs when rezeroing and loading different g-code files.

I am running a custom toolchanging macro but as I stated, this occurs in the beginning of the file, before an m6 or m31 is even called.

I am also running the uccnc atc arcs post processor.

Does this issue seem like a bug that a test release would create or could it be something else going on?

Any help will be greatly appreciated.

[ger21]

Post your g-code, because that's what most likely is doing it.

[b57378]

Will do, but as I've said the problem isn't linked to one specific g-code file and is not repeatable.

For example if it happens with a file and I run it for a second time it will not do it again.

[ger21]

Try 2.113.
You should always try the latest version before making any bug reports, especially when using test versions.

[b57378]

OK noted, just thought someone may have experienced something similar.

[zerodivide]

You very likely have the first absolute XY movement in a block like that:

G0 G54 X12 Y34 S.. M3

The problem appears if the previous WCS was other (ex G55) than G54 In the sample above G54 will not be applied to X12 Y34 movement.
Instead XY will move according to the previous active WCS

I have been burned by this the very first time I used UCCNC.

The fix is to move WCS code out of the XY block:
G54
G0 X12 Y34 S.. M3

[b57378]

Great, thanks for the feedback! I definitely thought that could be a potential problem, will try this fix and let you know.

[b57378]

Would it be best to edit this in the post processor which I'm using?

[zerodivide]

I ended up modifying my post processor:
1) Put G54 and G90/G91 blocks before positioning (X,Y) blocks
2) Removed usage of G28 and instead use G53. G28 too a bug: G28 Z0. moves all axes home, not just Z

Other than that everything else seems OK (or at least as intended).

[b57378]

Thanks, appreciate the info!