Intermediate positions in linear movement during adaptive toolpath

[stefanoprato8]

Greetings,

I'm a beginner and I'm working on a DIY machine running on mach3. 

I was testing an end mill with an adaptive toolpath and I noticed that in some points there was a sort of micro-stalling (like a lost step in the stepper motor) in the linear motion and I thought it was a stiffness problem. After running the contour toolpath for finishing, which ended up having a width of cut comparable to the roughing pass due to my concern about stiffness, I noticed that this stalling was absent. I looked at the toolpath and noticed that in the adaptive strategy in single axis movement there were intermediate positioning command. Basically suppose I'm in A in order to go to D the adaptive toolpath commands to go to B and then to C and finally to D; contour instead simply commands to go to D (which  is equivalent in the example). 

I'm confident that the micro-stalling in the machine is not related to stiffness but that, possibly for some problematic settings on the mach3 side, the machine stalls when it reaches this intermediate position and then it starts again after few milliseconds. 

In the video you can see the stalling event with no load. The code is from another run but for example lines 81, 82 and 83 seem useless to me.

I will investigate on the mach3 side but for what concerns fusion is there a way to set the adaptive toolpath so that it avoids placing this intermediate positions?

Thanks. Sorry for any English problem.
Stefano Prato.

[seth.madore]

Set your tolerance to a reasonable value (I prefer .001") and then turn smoothing on. What's happening is that Mach 3 is choking on the massive amount of code being thrown at it. 3200 lines of code for what the part is, that's pretty hefty. Curious, could you share your file?
File > Export > Save to local folder, return to thread and attach the .f3d file in your reply

[stefanoprato8]

Thanks for the answer,

The file is just a simple block as the goal was to test an end mill. When simulating the adaptive toolpath the dots are visible in the simulation. Adjusting tolerance doesn't help. I tried to simulate a code where I manually eliminated those intermediate dots and, as I thought, they actually seem to be useless.

I also attach the file for the real piece I have to make. Alse here in the adaptive strategy I use to remove the stock on the sides the are intermediate dots. 

[seth.madore]

"Intermediate dots" are actually positional moves. If these short movements are posing an issue for your machine, I suggest staying away from the 3D toolpaths, as that's essentially how most of them work

[stefanoprato8]

Isn't there an optimization to eliminate useless ones. 

Something like:

IF (current line has only one axis movement AND current line +1 has only one axis movement) 
THEN eliminate current line

Take this lines of code: 

0    X14.356 Y43.199
1    X52.147 Y43.2
2    X102.534
3    X152.921
4    X200.159
5    X201.195 Y43.027
6    X202.144 Y42.579

Lines 2 and 3 are useless.
If I'm in a line that commands a movement only in X and the next line is also commanding a movement in that exact same line then my current command is useless. 

Do you thing I can write down a line of code to edit my .nc files to make this correction or is there something that I'm missing?

Non Adaptive toolpath are much slower. 

[seth.madore]

Repetitive linear outputs are just something we accept. Smoothing only converts line segments into arcs where possible.

Is the issue that the Y isn't being output as well? I must admit, I don't think I've seen another Mach3 user have this same issue.

[stefanoprato8]

The adaptive toolpath creates intermediate points only for linear movement that are longer then a certain number of units so in the test, because of the short length, no extra points have been created in Y. If I create other simulations the adaptive toolpath creates intermediate positions also in Y. 
Anyway I have come to the conclusion that possibly my machine has a stalling with every line of gcode and that I was able to notice this in linear movement only because it is more visible than in arcs. 
I have seen few people complain about this but all of them were running mach3 on a PC via parallel port, with all the problem this implies, while I'm running it on a smoothstepper motioncontroller which should be a more high end solution, and less pc specs dependent.
Seems that CAM side investigation is over, I still don't understand the reason for the existence of those intermediate lines which only seems to make the code longer and heavier but they should be easy to eliminate with a simple text manipulation program. However if the problem is present with every line I will have to find solution on the machine side too.

Thanks for the help.