Gcode strange behavior

[Anonymous]

Hello everybody!

Im new to CNC and Fusion360. I have build a simple cnc machine called Most Printed CNC (MPCNC) that uses the same structure as Reprap 3d printers.

I have made a model on Blender, and exported as .OBJ. I did all the process to cut a 2D contour, exported the gcode, and tested it with a pencil.

The problem is: It makes strange movement at the corners. At the image bellow you can see that the pathway at the left (made using a pencil) is not 90 degrees as it should be ( at the right), and i can see the X motors moving while they should be still.

i ran the gcode on CAMotics ah it generated a perfect pathway:

To test the machine I built a similar structure on fusion and the pathway this time was right as image bellow ( the motor was still this time):

Does someone has any idea why is it happening?

Im asking it because i use blender as my main modeling software and i would like to use fusion360 to export the gcode to my CNC machine.

thanks 

[LibertyMachine]

A few preliminary questions:

1) Can you export and share the Fusion file or your first attempt? There could be something going on with the model or perhaps your toolpath selections.

2) What post processor are you using?

3) Have you ever cut something with this machine using another piece of software? What was it and how were the results?

[Anonymous]

thanks for your repply.

1) yes sure.

http://a360.co/2jJHaMd  

About the toolpath, i had to select , at the contour selection, edge by edge because i imported the model and i converted it to BRep.

2) I tried 2 post process, the only 2 that exports as .gcode. Generic Tiny G and generic Othermill

3)Yes, i tried Estlcam, but not with this toolpath. 

The strange thing is that if i use a toolpath made from a model made on Fusion360 it works flawlessly. But from a similar model imported and converted to BRep, it seens to work on the simulation and on another CAM simulators, but at the machine it doesnt.

After import and converting, the chaining selection doest work as it should. Thats why i had to select edge by edge.

Berthil

[LibertyMachine · Accepted solution]

After doing some searching of my own over on the MPCNC forums, I also found a post processor that someone had written for the Martin control. Is that what you have, or is there another controller you are using?

At any rate, it's producing almost the same quality of G-Code, in that the profile is the same. There is a free G-code viewer that works quite well in determining if the code is valid: NC-Corrector

As you can see in this screenshot, the area that you were having issues with looks proper (enough) The code/profile was the same with the TinyG and the post I found.

There is an inordinate amount of small linear moves in the corner, resulting in a large amount of code, so that "might" be contributing to your error. However, I honestly suspect that your machine might need motor and behavior tuning. 

Have you cut anything with other software that yielded good results?

That last bit about the .gcode format; that's just a file extension that the post sticks on the end of a file. Any posted code can be "saved-as" any file type you want, for testing purposes.

One last piece of unsolicited advice; Are you able to do greater things easier with Blender? The modeling side of Fusion is rather robust for many applications, have you considered doing all your work in one application? Fusion360 gallery

[Anonymous]

Hello Seth,

Thank you for your time!

The post processor you sent me on the link (MPCNC) did the job. It is now working perfectly. 

Ill use fusion to make the models next time i start it from scratch. It is indeed a great tool.

Thanks once again

Berthil