Adding a function in Autodesk post processor class without using TCP

[Anonymous]

A 5 axis simultaneous tool path made on Fusion 360, post processed without TCP mode, is resulting in a jerky motion of the machine. But at the same time, with TCP mode ON, the tool path is smooth but the controller requires an in-built TCP mode to perform smoothly. And a lot of economic controllers in the market does not have this feature. 

I would like to request for the addition of a function to Autodesk Post Processor Class, so that any controller which does not have a TCP mode can be used to run 5-axis simultaneous G-code. 

A similar function, as below, could be added or the same can be updated to include an option for using controllers without TCP mode, so that the tool-path will not be jerky.

optimizeMachineAnglesByMachine ( MachineConfiguration machine, Integer tcp )

 0: Maintain tool tip position (TCPM).  In this function we have only 3 options

1. 1: Map tool tip position.
2. 2: Map tool tip position for machine axes in table only.

Kindly look into this issue.

Hoping to hear a positive response from your end.

Amit Kumar.

[Laurens-3DTechDraw]

Your TCP and NON-TCP path is jerky.

Because I'm sure the toolpath is jerky in the CAM.

The issue is not the post.

NON-TCP post is correct and should work fine.

[Anonymous]

But TCP is working smoothly not jerky motion, but coordinate system is rotating in TCP mode. 
and i tried dry run of TCP mode toolpath, there is no jerk. but cutting in wrong place. and can you send an example of toolpath file in fusion 360 , so i can verify what am i missing.

[Laurens-3DTechDraw]

My solidworks part I used for the .tap files is in the google drive folder too.

[Anonymous]

@Laurens-3DTechDraw
I tried making a simple contour on hemisphere. And generating gcode for both. Again i am getting less accurate in non tcp and accurate in tcp. As my controller is showing and backplotter is showing both same . I think conversion of tool coordinate system to machine coordinate system is not accurate and not same. Please correct this algorithm in fusion post processor library.

[Laurens-3DTechDraw]

@Anonymous

I am not working for autodesk.

But I'm sure the code is the same for TCP and NON TCP.

[Anonymous]

With all due respect sir, it is not coming as same for TCP and non TCP but still ill check if there is a mistake on my part.

[Laurens-3DTechDraw]

I've shown you two NC files one with TCP and one with NON-TCP.

They give the exact same output if you backplot them correctly.

So if you post a good toolpath you get a good output. TCP or NON-TCP.

I've given you the post processors and the part file.

So if you can't get them to work that is not due to the post, but due to something else.

I mean how can you simulate TCP on the control when it doesn't have TCP?

[Anonymous]

@Laurens-3DTechDraw it will show in toolpath preview window and its visible that tcp toolpath is accurate i sent you screenshots of my controller.

[Laurens-3DTechDraw]

@Anonymous

So it will do the correct preview just not run the program?

[Anonymous]

@Laurens-3DTechDraw i tried running the program, i am getting same jerk for non tcp and wrong place cutting in TCP.

[richpicture]

Hi Thought I would join in - I am running Mach 3 and tried non TCP and TCP with lots of different cps files. Understandably TCP on will not work in Mach but paths generated in non TCP are also wrong for multi axis. There is definitely a path translation error for multi axis non TCP, with incorrect paths when looked at in Mach. The code paths with  TCP on  when viewed in Mach look cleaner, but again Mach cannot run them. 

So many users are hitting this problem and failing to get resolution. I have found that Multi 5 axis does work with the correct paths in Mach with hand crafted 5axis code and low feeds.

The majority of the Post Processors are authored by AutoCAD as they know their product. Non TCP multiaxis code should be a walk in the park for them.  The Fusion  Team really need to talk to ArtSoft and get heads together, get an actual copy of Mach to get a fix to this and soon.

This is bad press and users will go and soon out of frustration!

Fusion itself is great  - well done team but this final coat of paint urgently needs Fusion team attention       

[Laurens-3DTechDraw]

So TCP paths look good but run wrong?

Non-tcp paths look rubbish but do the correct moves but a little stuttering?

I'm pretty sure people are running Mach3 5-axis with Fusion and HSM. So there must be other things in play here.

[Anonymous]

@richpicture @Laurens-3DTechDraw What other things are in play here?

and yes @richpicture its true they have to talk to artsoft about TCP enabling.

[Laurens-3DTechDraw]

@Anonymous

How can Autodesk talk to Artsoft about enabling TCP. They have no connection and Autodesk has no power over Artsoft. I believe people have made Mach4 with TCP.

But the other things could be settings in MACH3.

You have full control over all the settings and Autodesk cannot know how you setup the machine.

Especially when you guys say the TCP code is simulated correct on the machine, it should also execute it right. If it doesn't there is a problem on the control not with the code.

[Anonymous]

@Laurens-3DTechDraw 

I tried all controllers available in market , but all without TCP option. I have one suggestion, when post processor converts tool vector to table coordinates the toolpath get distorted in that conversion. i just want to say that there is smoothening algorithem available in researches we can implement in our fusion post processor also. In tcp all conversion is done by controller but in non tcp conversion is done by post processor.

[Laurens-3DTechDraw]

@Anonymous

How can it get distorted?

It's the same path just in a different notation.

There isn't any change to the toolpath.

But to get everything clear.

If you use the post without TCP the machine runs the code, and produces the correct shape just not very smooth?

Or doesn't is produces the correct movements and shape?

[Anonymous]

@Laurens-3DTechDraw produces the correct shape just not very smooth?  the shape i am getting is completely jerky, there should be an smoothening after conversion of tool vector to machine angles. I am having that algorithem just want to know how to implement it using fusion 360 post processor.

[Laurens-3DTechDraw]

@Anonymous

I think your problem is much more in the points created in Fusion than the conversion.

If you have the algorithm you can see if it helps when you apply it after fusion produced the code.

But if you can send me the file, the fusion file, I can check for you if that code is jerky or if there is an issue in the so called conversion.

[Anonymous]

@Laurens-3DTechDraw i have reduced the jerk with that algorithem that i made in python, i just need your help in implementing it. So it can help other users who faces the same problem which i faced in 5 axis toolpath.

[Laurens-3DTechDraw]

@Anonymous

So it changes to points in that algorithm or what?