[JayCop]

Thanks to PaulRowntrees program warp drive I have resolved most of my inaccuracies. With Pauls warp drive program I was able to find that my machine is about .15 degrees out of being orthogonal.

I have a new problem now that I have not been able to get figured out. I am making custom computer cases and the fan grills that I make are half circular arcs. On one side the archs are perfect every time and on the other side they vary pretty substantially. Attached is a picture of what I am talking about. Left side good, right side bad :tired:

[ger21]

Is the cutter plunging on the right side? Try starting in the middle of the slot.

[JayCop]

Is the cutter plunging on the right side? Try starting in the middle of the slot.

Yes it is. I am unsure how I could change that in my cam program CUT2D. I tried ramping the plunge of the cut over 100mm this last time and it did not seem to help.

I did cut a panel that had the same grill on it 90deg from the other ones and they look significantly better. Could this be backlash on one of the axis? I couldn't imagine that there would be so much backlash in the R&P setup but hey you never know. I have been chasing my tail with this machine.

[PaulRowntree]

Thanks to PaulRowntrees program warp drive I have resolved most of my inaccuracies. With Pauls warp drive program I was able to find that my machine is about .15 degrees out of being orthogonal.

I am glad to hear that it helped you out!

EDIT : Does the warped G-code look correct on the right side of the cut? Did you try a previewer to make sure that the warping isn't messing things up? It never has for me, but some children (and some programs) sometimes misbehave when out on their own ...

Cheers!

[JayCop]

I am glad to hear that it helped you out!

EDIT : Does the warped G-code look correct on the right side of the cut? Did you try a previewer to make sure that the warping isn't messing things up? It never has for me, but some children (and some programs) sometimes misbehave when out on their own ...

Cheers!

I can't seem to get the previewer to install on my laptop. I will have to wait til tomorrow when I get to my shop to try it on one of my systems there.

[kreutz]

Thanks to PaulRowntrees program warp drive I have resolved most of my inaccuracies. With Pauls warp drive program I was able to find that my machine is about .15 degrees out of being orthogonal. .......

I would like to read some details about PaulRowntrees' program, please.

Best regards,

kreutz

[JayCop]

Here is a link to Paul's website where you can learn more about the program and download it WarpDriver - Paul Rowntree

[JayCop]

Yes and I am unsure of how to get my cam software to not start where it does. I use cut2d by vectric software. If anyone knows how to start it some where else I would love to know and will try it asap.

[JayCop]

I looked at the g-code viewer. When I run it it stops at this line
N23930G2X20.6098Y23.0182I0.0625J0.4519
and says "G Code Error Radius to end of arc differs from radius to start"

I have attached the g-code file from this. Any help would be greatly appreciated guys.

[kreutz]

Here is a link to Paul's website where you can learn more about the program and download it WarpDriver - Paul Rowntree

Thank You!!

Best regards,

kreutz

[ger21]

I looked at the g-code viewer. When I run it it stops at this line
N23930G2X20.6098Y23.0182I0.0625J0.4519
and says "G Code Error Radius to end of arc differs from radius to start"

I have attached the g-code file from this. Any help would be greatly appreciated guys.

Change your IJ mode in General Config.

[Khalid]

Seems your Z-axis is not stiff/rigid enough.. Also check the squareness of your Gantry...

[JayCop]

Change your IJ mode in General Config.

I did have the same problem in mach3 before but I did change the IJ mode and do not get this error in mach anymore just in the Kmotioncnc viewer. I haven't been able to track down if there is a IJ mode setting n the Kmotioncnc software.

[JayCop]

Seems your Z-axis is not stiff/rigid enough.. Also check the squareness of your Gantry...

Thanks for the sugestions Khalid. I went and checked on the stiffness. While the Z axis was very rigid (CNCrouterparts High Z plate on a R&P driven Y axis) I noticed while pressing on the z axis while it lowered the x axis was actually moving with only little force applied. I had loosened the R&P on both sides of the X axis when I was trying to figure out why I was loosing/overshooting steps. I did not tighten them back down enough especially the motors to the rubber drive pulley belt. Now that they are tight it takes quite a bit of force to move the x axis a few thousandths now. Every thing does move without binding still too. I will be able to test this out tomorrow or the next day to see if that helps any.

[PaulRowntree]

I looked at the g-code viewer. When I run it it stops at this line
N23930G2X20.6098Y23.0182I0.0625J0.4519
and says "G Code Error Radius to end of arc differs from radius to start"

I have attached the g-code file from this. Any help would be greatly appreciated guys.

@JayCop : is this the code from your CAM or from WarpDriver? I don't think my program should produce G2 or G3 codes at all, since they should be changed into short G1's. Warped arcs aren't arcs anymore.

EDIT : It IS from WarpDriver, but I see that Z=0.000. WarpDriver normally requires the tool to be zero'ed at the top of the work, and it uses -ve Z values to flag cuts. I will look at the code tonight to see if anything else is amiss. Can you post the original g-code as well?

@ger21 : thanks for the tip on IJ mode ... I never would have thought of that, and you saved me hours of looking for a bug in the code that wasn't there (hopefully).
Cheers!

[JayCop]

Thanks for looking at it for me Paul. Attached is the original gcode.

Also Paul is it possible to use the table surface as z-zero since the nominal thickness of my material changes a little each time so I prefer to zero off the table if not I will change to zeroing from the material?

[PaulRowntree]

@JayCop;
1) I have run your g-code produced by one of the Vectric programs through WarpDriver, and it ALWAYS generates the correct series of G1 codes from a G2 or G3, as long as the Z value is at or below the WarpThreshold. As far as I can see, you have Threshold set to 0.000, so this shouldn't be the problem, BUT since you are using Z=0 to be the bottom of the work, you should set WarpThreshold to be above the top of the material (see below).

2) The g-code from your Vectric program seems to missing a line [N190 I believe]. Your code is numbered
N180G00G20G17G90G40G49G80
N200T1M06
N210 (End Mill {0.125 inch})
Similarly, it looks like a line has been removed from the Warped Code you sent,
N12G0G20G17G90G40G49G80
( Passed along with only skew+renumbering )
***** Was something here removed ? ***
( Passed along with only skew+renumbering )
N16T1M06
( *** New Segment Start ********* )
( Warping code written by Paul Rowntree )
( Passed along with only skew+renumbering )
N18 (End Mill {0.125 inch})

3) The Z=0 only matters when using the roughing option of WarpDriver. If set to the bottom of the work, WarpDriver will do bad things to your material, things that don't even look like roughing. If you are not roughing (ie Passes=1.0) it shouldn't matter anywhere.

4) Setting the Accuracy to 0.0010 may be excessively small. It just makes the files bigger and the steps smaller; it should affect the G2 -> g1 conversions. I use 0.010"

I would like you to try setting the WarpThreshold to be about 0.020 above the top of your material, set the 'Verbose' flag on, and run this again. Please send the exact input and output files from this test and I will look again. So far, I see nothing wrong, and I am wondering if some round off error is at play.
Cheers!

[JayCop]

@JayCop;
1) I have run your g-code produced by one of the Vectric programs through WarpDriver, and it ALWAYS generates the correct series of G1 codes from a G2 or G3, as long as the Z value is at or below the WarpThreshold. As far as I can see, you have Threshold set to 0.000, so this shouldn't be the problem, BUT since you are using Z=0 to be the bottom of the work, you should set WarpThreshold to be above the top of the material (see below).

2) The g-code from your Vectric program seems to missing a line [N190 I believe]. Your code is numbered
N180G00G20G17G90G40G49G80
N200T1M06
N210 (End Mill {0.125 inch})
Similarly, it looks like a line has been removed from the Warped Code you sent,
N12G0G20G17G90G40G49G80
( Passed along with only skew+renumbering )
***** Was something here removed ? ***
( Passed along with only skew+renumbering )
N16T1M06
( *** New Segment Start ********* )
( Warping code written by Paul Rowntree )
( Passed along with only skew+renumbering )
N18 (End Mill {0.125 inch})

3) The Z=0 only matters when using the roughing option of WarpDriver. If set to the bottom of the work, WarpDriver will do bad things to your material, things that don't even look like roughing. If you are not roughing (ie Passes=1.0) it shouldn't matter anywhere.

4) Setting the Accuracy to 0.0010 may be excessively small. It just makes the files bigger and the steps smaller; it should affect the G2 -> g1 conversions. I use 0.010"

I would like you to try setting the WarpThreshold to be about 0.020 above the top of your material, set the 'Verbose' flag on, and run this again. Please send the exact input and output files from this test and I will look again. So far, I see nothing wrong, and I am wondering if some round off error is at play.
Cheers!

I output from cut2d one with the z=0 table surface and one z=0 top of the material. Both were run through the warpdriver with Accuracy= 0.01, Warpthreshold=0.197, Passes=1, and Skew=0.150

I hope I ran those with the correct parameters. If I did not run them right I will make them up again asap.

Again thank you Paul for all your help and your great program.

[PaulRowntree]

Well, 'not so great' would seem to be a better description. For some reason the version that is uploaded is 1 day older than the version I use at home, and apparently I fixed something that day, but I don't recall the repair at all. My home version handles the G2/G3 correctly, but the posted one does not. Obviously a short between the headphones, as the radio guys used to say.

I will verify the behaviour and upload a newly compiled version tonight.
I apologize for the time that you have wasted on this; I hope the newer version works as advertised and is worth the wait.

The attached image shows what the code should look like when 'Verbose'. You can see how the original G2 codeline in ( ) gets turned into multiple G1 steps.

PS : My Point #4 of the previous post SHOULD have read as

4) Setting the Accuracy to 0.0010 may be excessively small. It just makes the files bigger and the steps smaller; it shouldn't affect the G2 -> g1 conversions. I use 0.010"

Cheers!

[PaulRowntree]

The new WarpDriver package has been uploaded to PaulRowntree.weebly.com. Please let me know how it works (or if it doesn't!)
My best guess is that I grabbed the wrong version when I packed up the zip file back in January, since my changelog at that time says that I found and fixed the bug with skewed arcs ...

Either that or I screwed up.

Cheers!