1.2104 skip randomly gcode

[shad]

I am notice that UCCNC do not execute any gcode line. Looks like program make jump between gcode lines.
Short video, profile file and cutting program attaching.
UC300ETH-5LPT controller.
This program run in the 1.2047 without this problem.

https://www.dropbox.com/s/t3topfjvjwzoac1/Video_2018-04-09_143345.wmv?dl=0

[CT63]

Looks like its working ok to me. The software is following the gcode. Not sure what you mean by skipping lines?

[dezsoe]

Watch the video: there's a jump at 0:29~0:30. (The code runs fine by me too.)

[CT63]

Watched it again its really hard to see the "skip"... Its not affecting the motion output.

[shad]

Yes, real motion skipped too.
On my machine this issue randomly and in the different points of the cutting trajectory.
This is video from my customer - jump was on 1 min 18 second:
https://www.dropbox.com/s/iu7mt49mi9fc06o/video_2018-04-10_14-45-48.mp4?dl=0

[Robertspark]

Andrew, I see the skips, but they appear to occur at different positions on both video clips / machines. {bit difficult to replicate then}

When re-running the code on both machines do they skip in the same place?

I've been through the tap file {scanned line by line} and nothing jumps out as an apparent error that cause the issue.

I've run the file with your profile and I cannot see any jump on this end.

There is a bit of a catch that you may need to consider..... Neuron always runs a old screenset file, this means that some of the fields that are available in the newer version of UCCNC are not available to Neuron / UCCNC, because if the field is not within the screenset file it cannot be used by UCCNC.
Most of the time I do not expect this to be a problem.

However in the latest development version of UCCNC a couple of fields were omitted and were redistributed to all 6 axis. Hence I am not sure if these fields continue to be used if they can / may affect motion operation even though you are using the newer development release of UCCNC.

Field 96 and Field 275, and these were replaced by fields 2606 >> 2617

These may not be the problem, but although you are running the latest development release the screenset "may" have effects on other things

It may also be related to this mysterious issue too:
http://www.forum.cncdrive.com/viewtopic.php?f=4&t=1118
Unresolved at present / at least not closed out on the forum with an explanation.








The following code that occurs in a few places is a bit odd...

Code: Select all

N16780 X90.5683 Y36.0760 I2.7287 J-1.0529
N16790 M05
N16800 G00 X73.8326 Y68.5818
N16810 M205 (THCON)
N16820 M206 (THCOFF)
N16830 M205
N16840 M03
N16850 G02 X73.6670 Y68.8082 I117.0682 J85.7858

N16790, torch off
N16800, rapid to new location
N16810, THC ON (NEURON AVC ON)
N16820, THC OFF (NEURON AVC OFF)
N16830, THC ON (NEURON AVC ON)
N16840, torch on {including neuron touch off, torch fire etc}
N16850, cut motion begin.

Looking at the NeuronThcUccncPost.scpost from the dropbox folder, its not apparent where the N16810 & N16820 lines are being inserted

I can see where N16830 & N16840 come from, as they are inserted as part of the "OnPenDown()"function

Code: Select all

function OnPenDown()
   post.Text (" M205\n")
   post.Text (" M03\n")
end

code snippet maybe?


Different subject>>> any chance of doing a tidy up of the Neuron dropbox folder? it's not clear what is the stable and what are the testing releases as there seems to be a number of versions.

Suggestion... maybe save the updates in order of reverse date within folders....

[shad]

HI Rob!
This problem issue randomly in the different places of the program.
My russian version of the UCCNC screen adapted to the new futures for all axis and all settings are correct.
M205 duplicate because we need to enable THC before M3 and then code snipped insert another M205.
It's still a big problem to use M205 for AVC on/off control

We make testing for this issue full day - was cut ~260 parts. Any parts was cut fine, any with this problem. 1.2047 version exact solve problem - 50 parts step by step - no problem.
I am very sure this is a bug and hope Balazs can find this and solve. Just run program 1-2-3-4-5-6-7-8-9 times and you will catch him.
Important - we use only UC300ETH-5LPT controller. No chance to check with another controllers now.

Will try to upload new screenset file in the Dropbox

[shad]

Hi Balasz!
Please check this issue.
Already the second table that runs on this version has similar problems.
Return to version 1.2047

Thank you!

[shad]

sadly

[ger21]

This is a test version, where bugs are to be expected.....