Frage Random movement and phantom circles

I have had some problems with the movement of my stepcraft 600. When running the G code I occasionally get random circles at the corner radii of some cutouts (cutting inside the shape to make lightening holes in wood with corner radii of about 5 to 10mm). These circles do not occur on all the corners and do not appear on the toolpath viewer and there is no clue where they will appear. However, they always occur inside the shape so do not affect the actual shape (so far). The circles appear when running the code in demo mode (no machine connected) so its not the step craft as far as I can tell. Feed rates etc are 500 - 600 mm/min, spindle speed is high 20,000, cut depth is 0.5mm with ramping.

I also had some files where the tool path suddenly jumps or skips some lines / paths / parts of paths. the tool path starts of progressing as expected then suddenly it will jump or miss 1 or 2 cut passes on a shape, then jump position, and often miss shapes out altogether. Again this is happening in demo mode with no machine connected so it seems to be a corrupt G code file?

I created the drawings in Turbocad on my Mac, converted to DXF and imported to Cut 2d, made sure all vectors were closed, created and previewed the tool paths with no indication of a problem. everything that i can see seems OK up to then point of running the code in UCCNC. I have used the exact same DXF files to create the same tool paths in Cut 2d 3 or 4 times and get different behaviour each time. I don't think is the size of the files, I have had issues with G code files of 630 line but ones with 1500 lines work OK.

I think I have my laptop set up OK, internet is disabled, no unneeded software, antivirus switched off, and I think I have got it optimised OK. Files are transferred from Mac to laptop with a USB flash drive but I do not run the files from the flash. I copy them to the laptop hard drive.

Tools used:
MaC Desktop OSX 10.8.5
Turbocad Deluxe 2d/3d v6
Lenovo laptop (brand new with plenty of memory and processor speed).
Cut 2d only installed a 2 months ago
UCCNC
UC100
Stepcraft 600

Could it still be the virus software??

Cheers Dave.
Proxxon spindle

Check your post processor?

At the point where you save the toolpath check the post processor selected

If the Demo in cut2D is correct and the view in UCCNC is not then its the post processor. UCCNC will only do what its told from the Gcode.


you should see exactly what you see in Cut2D - on the UCCNC


you want the Mach2/3 arcs mm post processor.




if your program moves position and starts again somewhere else (randomly) - and if the toolpath is correct - then this is mist likely due to the machine stalling mid program and then starting again.Make sure she is well tuned.




let us know if you find what post processor you are using to cause this issue.

Rory.

Thanks for your responses.

The post processor is definitely set to Mach 2/3 Arcs (mm)(*txt). I do not see how it can be the machine as the errors occur when I run the files with UCCNC in demo mode without any connection to the machine at all, I can run the files in UCCNC sat at my kitchen table and the errors occur. The circles do not appear on the blue tool path trace in UCCNC but you can sit and watch the tool point do the circle. when the jumps / position errors occur the pointer will jump to a new location and the skipped tool paths will turn yellow as if they have been instantly completed. I do not see how it can be a machine problem. I have also been very careful to set the machine up the best I can. I already found the video you attached on u you tube. I also took great care to make sure the alignment of the axis was correct and when it cuts it does a great job so far, easily accurate enough for my needs, no bad noises and smooth running.

I will attache a couple of files later when I get a chance.

Thanks Dave

okay - you've covered most points - please send some toolpaths to see can we reproduce what you are seeing.

Rory, all.
Here is a video of a tool path that shows the random behaviour. I have confined the video to the bits of interest. The code was running in demo mode, no connection to the step craft. I have also uploaded the DXF, Cut 2d and output tool path text files.

UCCNC Random tool paths









Cheers Dave.

Cheers Dave - interesting video. I've just ran the toolpath text file and it worked perfectly here.

What is your PC setup? There is something funny going on there.

and what version of UCCNC?

Rory
UCCNC details

software version 1.0031
firmware version 1.0000
hardware version 0.0000
API version 1.5150

PC details in attachment below.







Cheers Dave

Rory, all.
I think I may have found some clue to my issue with random circles and movement. I disabled my backlash compensation and the problems seem to have gone away. Why the backlash is a problem I do not know, perhaps I was trying to set things up wrong or too accurately / precisely?

I tried to attach my step craft 600 profile so that you could try it and see if you could reproduce the problem but for some reason the forum won't load.


Cheers Dave

Dave,

I took the code and simulated it in NCPlot.

It does not show the random movements.



I loaded it in Mach3 as well and it looks OK.

I see some strange invalid arc movements of UCCNC, maybe a SW bug. I would contact them directly.

Frank.
Thanks. I have contacted CNCDrive directly. I will see what they say.

Thanks Dave.

Hi Dave - thanks for further information on this - so you have modified the profile from the profile we supplied to you?

Can you check that the original profile works okay? And also - just emailed you link to 1.0034 (latest version - up from 1.00033 as supplied in the installer pack).

All the profiles we have setup (and supply) do not have backlash enabled.

The Gcode you supplied is perfect - and has run 20+ times here in the workshop without a hiccup on the same profile and version as supplied. it also runs perfectly on 1.0034 here.

The issue your side has manifested itself when the UCCNC interprets the Gcode. If you have changed / re-configured the backlash settings this is most likely where this issue has come from - you are running on a custom profile. The UCCNC is a relatively new software and as such there may well be some bugs that pop up. That said - we only supply versions (and profiles) that have been extensively tested in house.

There may well be an issue with the backlash algorithm/operation - we haven't tested this yet.

The next version of the UCCNC will be V 1.100XX And there are a lot of major changes - including changes to backlash operation.

What I would say is that any identified bug that is spotted gets deal with extremely quickly by CNC drive.

Let us know how you get on

Rory
thanks of the feedback. I contacted CNCdrive and they had the same results as you - it runs OK on 1.0034 even with my profile loaded. I am going to load that version and give it a go later.

Looking at the amount of backlash I had set it could be well below the resolution of the machine. I will try with it on and off.

One of the reasons I used a custom profile was that I had a slight problem with homing. Using the homing speeds in the default 600 profile the axis homed oK but moved into the limit switches so fast they would jam up and then not move after homing was completed. It seems that the balance between free play in my system and the movement speeds were not compatible. I played around with it for a few days trying to get the axis free play right but still had this issue or ended up with the axis too tight and ran into other stalling issues. The end result was that I kept the axis moving smoothly without play and reduced the homing speeds.

Anyway I will let you know how it goes.

cheers. Dave.

Don`t have any solutions, but couple of ideas. The following information could be completely irrelevant in this case.

I see you have a laptop with the UC100, but does the simulation mode get signals from it or the simulation is run by the laptop itself? I haven`t used UCCNC, but mach3 and laptops don`t always get along without some tuning. For example my old laptop (Dell d600 with parallel port) initially failed the Mach3 driver test. I changed my laptop ACPI to standard PC driver and it passed the test after that. If you want to change the driver I suggest doing some reading first because you might lose some laptop functions and having to insert some installation disks for certain software.

I have seen weird tool movements on an industrial CNC lathe before, also in simulation mode. The work itself was done correctly. The reason in this case was that most of the program was done without tool nose radius compensation and the compensation was switched on in couple of places when cutting arcs. The tool reference point suddenly changed when the compensation was switched on and this caused those weird movements, because the software wanted to move the new reference point back to the simulation cutting line. This did not affect the tool movement on the machine when it was cutting. Maybe something similar happens in those inner corners. The program want`s to cut an arc but the reference point suddenly jumps. And in some cases the new position is such that continuing cutting the arc takes the tool back to the cutting line.

david55 schrieb:
One of the reasons I used a custom profile was that I had a slight problem with homing. Using the homing speeds in the default 600 profile the axis homed oK but moved into the limit switches so fast they would jam up and then not move after homing was completed. It seems that the balance between free play in my system and the movement speeds were not compatible. I played around with it for a few days trying to get the axis free play right but still had this issue or ended up with the axis too tight and ran into other stalling issues. The end result was that I kept the axis moving smoothly without play and reduced the homing speeds.

Anyway I will let you know how it goes.

cheers. Dave.

Did you complete the "Verification of the linear axes" procedure described in "First Steps" documentation available on Stepcraft homepage? Even if your system components move smoothly alone they could jam at the end positions after the machine has been assembled unless you do the verification procedure. It happened to me too. The Y axis moving assembly got stuck at the extreme end for a moment and the stepper motor together with it`s drive pulley continued to move while the belt was standing still. Wasn`t a pleasant sound.

tikka schrieb: Don`t have any solutions, but couple of ideas. The following information could be completely irrelevant in this case.

I see you have a laptop with the UC100, but does the simulation mode get signals from it or the simulation is run by the laptop itself? I haven`t used UCCNC, but mach3 and laptops don`t always get along without some tuning. For example my old laptop (Dell d600 with parallel port) initially failed the Mach3 driver test. I changed my laptop ACPI to standard PC driver and it passed the test after that. If you want to change the driver I suggest doing some reading first because you might lose some laptop functions and having to insert some installation disks for certain software.

I have seen weird tool movements on an industrial CNC lathe before, also in simulation mode. The work itself was done correctly. The reason in this case was that most of the program was done without tool nose radius compensation and the compensation was switched on in couple of places when cutting arcs. The tool reference point suddenly changed when the compensation was switched on and this caused those weird movements, because the software wanted to move the new reference point back to the simulation cutting line. This did not affect the tool movement on the machine when it was cutting. Maybe something similar happens in those inner corners. The program want`s to cut an arc but the reference point suddenly jumps. And in some cases the new position is such that continuing cutting the arc takes the tool back to the cutting line.

Good points raised here - but there is no Driver test requred as the UC100 overcomes this. We have production machine running from 250euro laptop. the UC100 uses the USB to communicate so this is not an issue - we would have seen this before. The driver test issue is related to using a laptop with a parallel card - this is no longer relevant.


Good suggestion on the "real life performance" - it would be interesting to see the machine response to this? Just cutting fresh air?

tikka schrieb:

david55 schrieb:
One of the reasons I used a custom profile was that I had a slight problem with homing. Using the homing speeds in the default 600 profile the axis homed oK but moved into the limit switches so fast they would jam up and then not move after homing was completed. It seems that the balance between free play in my system and the movement speeds were not compatible. I played around with it for a few days trying to get the axis free play right but still had this issue or ended up with the axis too tight and ran into other stalling issues. The end result was that I kept the axis moving smoothly without play and reduced the homing speeds.

Anyway I will let you know how it goes.

cheers. Dave.

Did you complete the "Verification of the linear axes" procedure described in "First Steps" documentation available on Stepcraft homepage? Even if your system components move smoothly alone they could jam at the end positions after the machine has been assembled unless you do the verification procedure. It happened to me too. The Y axis moving assembly got stuck at the extreme end for a moment and the stepper motor together with it`s drive pulley continued to move while the belt was standing still. Wasn`t a pleasant sound.

Another good point - very important to verify the free movement of the machine in all areas of travel - but in this case the machine is working 100% and the demonstrated performance happens in demo mode when not connected to the machine.

I had tried the suspect files cutting into "air" and they had the same random behaviour.

I have just done some cutting with UCCNC version 1.0031 and without backlash compensation. Everything worked OK so I suspect that it was the backlash. I tried it with version 1.0034 and no backlash and that worked OK as well.

Therefore it looks like the backlash is the problem.

Thanks for the help everyone.

Cheers Dave.

Let me quote some more.

Rory schrieb: Another good point - very important to verify the free movement of the machine in all areas of travel - but in this case the machine is working 100% and the demonstrated performance happens in demo mode when not connected to the machine.

david55 schrieb:
One of the reasons I used a custom profile was that I had a slight problem with homing. Using the homing speeds in the default 600 profile the axis homed oK but moved into the limit switches so fast they would jam up and then not move after homing was completed. It seems that the balance between free play in my system and the movement speeds were not compatible. I played around with it for a few days trying to get the axis free play right but still had this issue or ended up with the axis too tight and ran into other stalling issues. The end result was that I kept the axis moving smoothly without play and reduced the homing speeds.

Anyway I will let you know how it goes.

cheers. Dave.

tikka schrieb: Let me quote some more.

Rory schrieb: Another good point - very important to verify the free movement of the machine in all areas of travel - but in this case the machine is working 100% and the demonstrated performance happens in demo mode when not connected to the machine.

david55 schrieb:
One of the reasons I used a custom profile was that I had a slight problem with homing. Using the homing speeds in the default 600 profile the axis homed oK but moved into the limit switches so fast they would jam up and then not move after homing was completed. It seems that the balance between free play in my system and the movement speeds were not compatible. I played around with it for a few days trying to get the axis free play right but still had this issue or ended up with the axis too tight and ran into other stalling issues. The end result was that I kept the axis moving smoothly without play and reduced the homing speeds.

Anyway I will let you know how it goes.

cheers. Dave.

Need to read more sorry Tikka - you were of course correct

david55 schrieb: I had tried the suspect files cutting into "air" and they had the same random behaviour.

I have just done some cutting with UCCNC version 1.0031 and without backlash compensation. Everything worked OK so I suspect that it was the backlash. I tried it with version 1.0034 and no backlash and that worked OK as well.

Therefore it looks like the backlash is the problem.

Thanks for the help everyone.

Cheers Dave.

Good news your back in action.

Maybe try the backlash if you have time on 1.0034 and see can you replicate it?