[Spunk]

Dear

I was testing the manufacturing abilitys in pro-e.
This works out fine but the G02 and G03 commands give the following errors in EMC for a sherline 5410 mill:

...zero radius...
...radius at begin differs from radius at end...

Could this have something to do with the rounded numbers? not enough number after the point?

Hope someone has had the same problem before.

Cheers

Philip

[jpgdesigns]

The I & J coordinates can be either absolute or incremental... check your post-processor settings... check EMC's settings.

I don't use EMC so I can't tell you what it is looking for or if there are settings to adjust...

[Spunk]

The problem occurs at the very first time the G02 command is used.

all the numbers look okay, absolute (like i need them to be)

[Andre' B]

I do not think you can change the arc endpoint error setting in EMC like you can in other controls.
Can you try bumping up the number of decimal places you CAM software puts out.
I remember reading a while back (years) that 6 places was recomended for best results in inch mode, probably get by with 5 in mm.

[Spunk]

So i need to get into the postprocessor in pro-e...

anyone know how?

[Big John T]

If it is the first arc only you are having a problem with then your post is not moving to the proper start position for the arc. Check your post processor.

John

[Spunk]

According to the read-outs the cutter is in the right position to start the curve.

that is why I think that it has something to do with the accuracy...

[Big John T]

Then the arc is defined wrong. Take a look at the Arc Buddy on this page.

http://wiki.linuxcnc.org/cgi-bin/emc...ode_Generators

It will generate the start position and the arc after you put in a couple of things. Then you can check your output to see what is wrong with your post processor.

John

[Spunk]

i think i need ubuntu with EMC2 to do that?
now i have the standard issue EMC debian from sherline...

[cyclestart]

i think i need ubuntu with EMC2 to do that?
now i have the standard issue EMC debian from sherline...

Any linux distribution worth it's salt includes python.

That wiki page has grown since I last looked ! The problem is it assumes people know what python is and how to use it. What percentage of the population might that be ?

Do this:

1) Right click "upload arcbuddy.py" on that wiki page.

2) "Save link as" and save it to your home directory

3) Open a terminal and type the commands

 chmod  a+x  arcbuddy.py

 ./arcbuddy.py

if you get a "bad interpretor" error, open arcbuddy.py in a text editor and modify the top line to read

 #!/usr/bin/python

and rerun the ./arcbuddy.py command in the terminal.

4} That's it. These scripts should also work in Windows with a python interpreter ? Think so anyway.

Btw: There's a quick text editor that can be run right from the terminal.

 nano  arcbuddy.py

The instructions are all listed at the bottom and take the form

"cntrl btn"   "some letter"

very handy and intuitive.

Anyway, just me up much to early and rattling on. Hope that helps in someway.

Edit/ There is a possibility another error might pop up. Depends on the python installed and what the gcode generator calls up. If another error shows up, just paste it here and someone will tell you what to install.

[Spunk]

owkay sure going to try it

but this is going to be a method to check, not to solve though...
i did the math and the places are pretty accurate but perhaps not accurate enough???

no-one knows something about pro-e postprocessors? can't be hard to tell him to print 4 or 5 digits after the point is it?

[cyclestart]

no-one knows something about pro-e postprocessors? can't be hard to tell him to print 4 or 5 digits after the point is it?

ProE has it's own forum here. You could post the question there also.

Another option is telling ProE to use R's instaed of I's and J's. R's allow more fudge factor. I's and J's are the preferred way however.

[Big John T]

Any linux distribution worth it's salt includes python.

That wiki page has grown since I last looked ! The problem is it assumes people know what python is and how to use it. What percentage of the population might that be ?

Actually the page is so big now that you must have missed the last part about how to use the python files with EMC

Do this:

1) Right click "upload arcbuddy.py" on that wiki page.

2) "Save link as" and save it to your home directory

However this part is/was missing...

4} That's it. These scripts should also work in Windows with a python interpreter ? Think so anyway.

Yep just change the file extension to .pyw and install Python.

John

[Big John T]

owkay sure going to try it

but this is going to be a method to check, not to solve though...
i did the math and the places are pretty accurate but perhaps not accurate enough???

no-one knows something about pro-e postprocessors? can't be hard to tell him to print 4 or 5 digits after the point is it?

Can you post the move to the start position and the arc move?

John

[Zig]

I too have see this problem with the G2/G3 codes.

The problem a the time i seem to recal was that the g code was generated on bob cad imperial to 3 decimal places

EMC2 moaned about errors and upon further reading ( somewhere in the fine manual ) it became obvious i needed another decimal place in order to apease the accuracy criterion.

Generaly i work in metric units and my current cad/cam software outputs metric g code to 5 decimal places no problems.

[Big John T]

The amount of decimal places does not matter, what does is it is how close to the proper position you specify with your offsets. Now if your post rounds up to 3 places in inch then most likely you will be off enough to get the error. The confusing thing is if you add more precision your error might go away. For example the arc G3 X1.5 Y1 I-0.5 will run fine if your current location is X1.5 Y1. However if your current location is X1.5005 Y1 the arc will fail.

If the distance from the current point to the center differs from the distance from the end point to the center by more than 0.0002 inch (if inches are being used) or 0.002 millimeter (if millimeters are being used) then you will get the error.

http://www.linuxcnc.org/docview/html...ub:G2,-G3:-Arc

John

[Zig]

John,
thats the refference on error spec I was thinking of.
By increasing the number of decimal places teh data was no longer truncated ( not rounded ) and the error condition disapeared.

[Big John T]

The cool thing is in 2.3 you define a G2/3 360° circle with just the offset you don't need to specify the end point.

John

[skullworks]

So your saying a command like:

G3 J-1.

Will now work as it should?

[BMG]

John stated that will be added in 2.3 - meaning release 2.3 which is expected sometime this year. The current release is 2.2.8 so I reckon that has not been implemented in the current releases.