Pierce Plasma RUR 2000 - Almost got the POST working, need your input

[thova]

Hi there!

Im sorry if my technical terms for POST is probably a bit off 🙂

Our school got a Pierce RUR 2000 GP ITEC Norway APC15PIERCE (PMX105 its own)

And the control unit is a PIERCE15. It operates with absolute for starting a section and then the machine wants incremental positions for the cut job.

Im trying to write a POST script for Fusion360, so we can quit using the OMNIWIN software. Several schools would benefit from this POST getting online.

Ive based this on the WinPC.nc POST file. It was the one closest i found at that time to our machine.

Ive managed to have the output text and G / M codes match so the machine accepts it.

The attached files are:

NestPlan_from_omniwin.cnc    // POST file generated with proprietary OMNIWIN for reference

Pierce_POST_almost_there.cnc    // My output POST file for the Pierce RUR 2000 Machine

PiercePOSTv4_backup.cps    // My adapted POST generator file ive made for the Pierce RUR 2000 Machine

I had to upload the files as .nc instead of .cnc

https://pastebin.com/mLjKLxiR // Generic FANUC.cps // Ive tried to combine this with the PIERCE for getting incremental support, i do think something is still missing.

This is our machine. It only runs using the plasma cutter.

 

This is what the machine wants to cut when i finally managed to format the POST so the file was accepted.

Looks like this is what happens when you only feed the machine absolute values and it wants both absolute and incremental. xD I was sort of flabbergasted for a minute there!

This is what the machine wants to cut when it gets the industrial POST from the omniwin software.

Its made using the same shapes as the Pierce POST got but at different positions.

Just so i can compare the output easier.

This is what the CNC wants to cut with the generated gcode from my adapted POST processor.

The arc on the top of the doorway is the wrong way, what can cause this? I also notice that many rounded edges displays the arcs the other way. Looks like it can do complete circles, because then it will do a complete circle and doesnt care about my fumblings.

 

As you can see above, what the machine now gets is almost correct. The first big square is ok, the big circle is ok.

The G90 G00 start call for the door shape is down and to the left. And the machine codes put the small circle way to the left but at correct heigth.

 

Ive probably spent 12 hours now to get this far, and I will finish this somehow!

Hope anyone of you experts can look over the POST generator and the two generated post codes and give hints on how to correct this behaviour.

I can think of a few things next:

1. Break up arcs into more segments to se if that handles better, or linearize. But i really dont want to linearize.

2. Check if i forgot some code from the "Generic FANUC.cps" thats incremental. I dont quite understand it all yet.

3. Must be something with this onCircular missing code for incremental support

function onCircular(clockwise, cx, cy, cz, x, y, z, feed)

4. How should i attack the G90 G00 POST for the doorway and the small circle beeing totally off position? What could cause this.

That would be very much appericiated!

Kind regards,

Thor

[carl.j.barker]

EDITED   ....

I am unable to test any of this so you'll have to do that.

My thinking is after the machine cuts an arc, the cutter position is now at the end of the arc/circle but the xOutput and yOutput variables still contain the arc centre point.

Try adding -

                                yOutput.format(y);

after the writeBlock at line 1182

Again this is just a guess looking at the POST and I can't test this.

And swap the 3 : 2   for 2 : 3 in both writeblocks in the onCircular Fuction.   

[thova]

Thanks.

Going to test the changes today.

[thova]

Hey, that seems to have done the trick! Thank you! 😊

Now it seems that the last issue is that the absolute coordinates on the doorway and the small circle is off compared to the drawing.

[thova]

The absolute coordinates in the post code is generated wrong. So the CNC shows it the right way.

Ive attached the drawing and the POST processor if anyone could try to generate.

Ive highlited the wrong lines with 2D Profile5 and 2D Profile6

Maybe its caused by something from the incremental coordinates.

(MACHINE= PowerMAX105_MS_45)
(MATERIAL= true)
(THICKNESS= 5)
(% Post program name:Pierce POST v3)
(% Post comment:Nope)
(% Configuration: )
(Machine: )
( vendor: Pierce)
( model: RUR 2000)
( description: Pierce Plasma Cutting Machine)
(% Operation comment: 2D Profile1)
(Plasma cutting)
(tool.jetDiameter = 1.5)
(tool.jetDistance = 0.)
(THROUGH CUTTING)
(QUALITY = 0)
N10 G54
N11 G90 G00 X22.67 Y-32.5
N12 G00
N13 G91
N14 G00 Z-15.
N15 M113 T1
N16 M04
N17 G01 X4.33 Y2.5 F0
N18 G01 Y3.
N19 G01 X376.
N20 G01 Y-276.
N21 G01 X-376.
N22 G01 Y273.
N23 G01 X-4.33 Y2.5
N24 M113 T1
N25 M04
N26 G00 Z15.
N27 M03
N28 G40
(% Operation comment: 2D Profile4)
N29 G90 G00 X52.644 Y-346.383
N30 G91
N31 G00 Z-15.
N32 M113 T1
N33 M04
N34 G01 X4.803 Y-1.389 F0
N35 G02 I119.883 J-124.728
N36 G01 X-1.199 Y4.854
N37 M113 T1
N38 M04
N39 G00 Z15.
N40 M03
N41 G40
(% Operation comment: 2D Profile5)
N42 G90 G00 X-61.248 Y-530.855
N43 G91
N44 G00 Z-15.
N45 M113 T1
N46 M04
N47 G01 X4.33 Y-2.5 F0
N48 G01 Y-171.227
N49 G03 X3. Y-3. I3. J0.
N50 G01 X225.
N51 G03 X1.342 Y0.317 I0. J3.
N52 G03 Y330.366 I-82.592 J165.183
N53 G03 X-1.342 Y0.317 I-1.342 J-2.683
N54 G01 X-225.
N55 G03 X-3. Y-3. I0. J-3.
N56 G01 Y-153.773
N57 G01 X-4.33 Y-2.5
N58 M113 T1
N59 M04
N60 G00 Z15.
N61 M03
N62 G40
(% Operation comment: 2D Profile6)
N63 G90 G00 X274.905 Y205.911
N64 G91
N65 G00 Z-15.
N66 M113 T1
N67 M04
N68 G01 X4.486 Y2.209 F0
N69 G02 I27.939 J-1.847
N70 G01 X-4.156 Y2.78
N71 M113 T1
N72 M04
N73 G00 Z15.
N74 M03
N75 G40
N76 M03
N77 G40
(PlanFooterCommand)
N78 M02

Ive changed lead in and lead outs, and then the G00 absolute coordinates changes. Suddenly the doorway is correct Z, but wrong y. Small circle is closer but not yet there. Ive attached this generated Pierce POST v5 Gcode as well.

[carl.j.barker]

OK sort of expected.  can you move the yOutput.format(y) line into the writeblock above it. Make sure to move not copy.

writeBlock(gMotionModal.format(clockwise ? 2 : 3), xOutput.format(x), yOutput.format(y), iOutput.format(cx - start.x, 0), jOutput.format(cy - start.y, 0), feedOutput.format(feed));

[thova]

Do you mean remove the yOutput.format(y) on line 1190 and add it into writeblock on line 1182?

[carl.j.barker]

Ah looking at your latest you already did that. forget that then.

[thova]

I just copied, do you think i should i remove yOutput.format(y) from line 1190 and retest?

[carl.j.barker]

Just a thought but what happens if you manually remove all G90's from the output ?

[thova]

The machine operates with absolute on G00 codes, and incremental on the rest.

Without those G90, coordinates will stretch out everything to a long incremental line as in my first post.

With the G90 G00, the four figures are drawn correctly. But initial G00 is off on the last two.

What im wondering is why the autocad feeds out the wrong G00 absolute coordinates to the gcode.


I will try to remove lead in/out and see what that does.

[carl.j.barker]

Try this.

Change the code at line 770 on wards to .

  } else {

        // RUNS AT OTHER SECTIONS BEGINNING

        // writeln("initial90blockposition");

        writeBlock(

            // NOPE

            gAbsIncModal.format(90),

            gMotionModal.format(0),

            "X",

            initialPosition.x,

            "Y",

            initialPosition.y

        );

    }

I'm thinking that the output Vars are still set to incremental type. this will force the actual absolute value.

If it works we can tidy it up.

[thova]

That did not change anything.

[carl.j.barker]

It should only change the rapid, the shapes would be wrong after, The incremental will still be tied to the position before the rapid. can you send the outputted .cnc file ?

[thova]

Ok, i got further. Bu comparing my post to the FANUC, i added some handling for the incremental codes.

This resulted in the correct placement and drawing of the last three sections. Almost there!

But the first rectangle is now in the wrong location.

And the machine head zero position it coming down from up by the middle. Should not this be down at origin from autocad?

Ive attached the code.

[carl.j.barker]

Can you run this nc file and show me the picture ?

[thova]

Heres the picture from the CNC.

Did you check the latest v5 POST to v8 gcode?

[carl.j.barker]

 So, got this (Tested in a simulator) but I had to put any G90 on a separate line before G00 ? which is odd.

[thova]

Still some issues.

Machine pressed down into the plate with the nozzle.

The machine starts a homing position when our post scripts are run.

The gcode machine zero point is off, see the attached image.

Also the figures moved again, the two last ones this time.

Ive added the new Fusion360 scetch if you could test.

The below Gcode, is generated from omniwin:

(MACHINE= PowerMAX105_MS_45)
(MATERIAL= S235JR)
(THICKNESS= 5)
G00 X350.85 Y-57.52
G41
M113 T1
M04
G01 X0.00 Y-7.70
G01 X0.00 Y-370.00
G01 X-270.00 Y0.00
G01 X0.00 Y370.00
G01 X270.00 Y0.00
G03 X1.80 Y1.80 I0.00 J1.80
M03
G40
G00 X218.21 Y-135.82
G41
M113 T1
M04
G01 X-5.44 Y-5.44
G02 X7.32 Y-17.68 I-17.68 J-17.68
G02 X-25.00 Y-25.00 I-25.00 J-0.00
G02 X-25.00 Y25.00 I0.00 J25.00
G02 X25.00 Y25.00 I25.00 J0.00
G02 X17.68 Y-7.32 I-0.00 J-25.00
G03 X2.55 Y0.00 I1.27 J1.27
M03
G40
G00 X-215.61 Y-329.72
G41
M113 T1
M04
G01 X-5.44 Y-5.44
G02 X49.79 Y-120.21 I-120.21 J-120.21
G02 X-170.00 Y-170.00 I-170.00 J-0.00
G02 X-170.00 Y170.00 I0.00 J170.00
G02 X170.00 Y170.00 I170.00 J-0.00
G02 X120.21 Y-49.79 I-0.00 J-170.00
G03 X2.55 Y0.00 I1.27 J1.27
M03
G40
G00 X444.30 Y-20.20
G41
M113 T1
M04
G01 X0.00 Y-7.70
G01 X0.00 Y-225.00
G01 X-325.00 Y0.00
G01 X0.00 Y225.00
G02 X325.00 Y-0.00 I162.50 J-81.25
G03 X2.41 Y-0.80 I1.61 J0.80
M03
G40
(PlanFooterCommand)
M02

[thova]

I got no further on this.