Contour tool path offset in the negative X direction?

[Anonymous]

I am having a very strange issue I have not been able to resolve.  I created a contour tool path to cut out 2 parts (in the same program) but it does not appear to be following the contour correctly, it seems to be cutting somewhere around 0.020" short (X-negative) in the X axis.  The parts are pistol grips with 20^ slopes at the top and bottom of each grip, this weird offset makes the top of the left grip and bottom of the right grip too short and the opposite ends too long.  See below, the vertical face on both grips should be the same height.

 
At least I am pretty sure this is what is happening ...  looking at the pictures, I have the machining done left to right along X - I tried/tested a set with the grips oriented in the Y direction.  Same issue - the red circled ends cut too short and the green cut too long.

Pretty much out of ideas here,  I have deleted and recreated the tool path a couple of different time,  even created separate paths for each grip ...  the machine is fine as it reproduces the behavior when the work is reoriented. 

The other weird part about this is that in the same nc program generated there are 8 boring operations done with the same tool....  they are in perfect alignment and do not show the weird offset. 

Here are my tool path settings:

[seth.madore]

Can you share your file?
Right click on file in Data Panel and select the "Share Public Link" option

[Anonymous]

Sure thing - https://a360.co/2S0sECm

[seth.madore]

I'm not seeing this issue in the toolpaths. I do question why you have 6 bodies selected in your Setup, when there are only 2 being machined, but that's most likely NOT the cause of the issue.

I do see that the two parts are ever so slightly off of the same plane, but it's such a small value, shouldn't be causing this.

What is the machine and control? Have you tried running it at a much reduced feedrate?

[Anonymous]

No, I've been over the tool paths repeatedly and can't find an issue either deleting them and recreating them shows the same issue ... 
6 bodies, yea well I'm pretty new to the software, I did have one of the grips about .020" below the plane, but I though I fixed that (that was causing machine issues as well) 

Machine is a Sienci Longmill (basically a home brew open source machine) using the UGS Platform for sending the nc. 30x30 cutting area... pretty nice in fact. 

I've cut probably 30 sets of grips and this issue only showed up after I added the 20^ slope to the top side of the grips (the bottom slope was always there) not sure it could be the model though .. on the ends that get cut short you can still see the scallop from the last ball nose tool pass - so the entire path ~just~ for those paths is definitely getting shifted in the -X direction 

I can certainly try a slower feed rate, but not sure how that would change things?? 

[Anonymous]

OK - tried a few things ... 

- cut the parts on a different part of the table

- broke the end contouring into 4 tool paths 

- left one of the offset outline paths at the same feed rate

- changed the second offset path to half - a painful 20in/min 

No change to the grip machined at 40in/min and a very slight improvement to the one done at 20in/min. So while the improvement is there its still not viable as its still wrong AND really slow.

I don't understand why the feed rate should be making a difference in this case.  

[seth.madore]

The feedrate test was to rule out machine issues.

Are you changing tools manually, and how do you establish the length offsets?

[Anonymous]

Sorry on the late reply .... seems like I am not getting notifications for this thread! 

I am changing tools manually - establishing the length using a probe, which has been reliable. 

I set my XY using a 1/4" stainless pin - just to avoid accuracy issues with fluted tools, then set the Z for each individual tool.

While I am pretty careful changing tools I'm not sure that could be the issue - I would expect some variation in the amount it was off in the X axis... That being said, loosening the collet ~would~ put exert force in the direction the tool appears to be offset to.  

I guess it would be trivial to check the machine's position before and after a tool change.

[seth.madore]

Do you set a G54 work offset, and does that post out on every tool? Do you post out a program for each tool, or is there just an M0 for manual tool change? I'm wondering if it's calling a different position at that tool, for some reason. 

I'm NOT seeing the issue in the CAM file. @daniel_lyall you have more time on the CNC Routers, what do you think?

[Anonymous]

@daniel_lyall  will probably need the new public link?  https://a360.co/2GMk6tx

Yes, I am creating an nc program for each tool - I don't need the M0, I just have to load each tool program (the M0 actually stops my machine) 
I had gone through the tool library (several times) specifically to see if there was some sort of offset or setting related to the tool (as it does not do this on the boring jobs)  and no  I did not intentionally set a G54 offset ... 

~where~ would you specify such a thing? 

I had checked the tools in my cloud library - not the ones saved in the project itself ... 

[daniel_lyall]

It could be something to do with minimum cutter radius I will have a look.

[daniel_lyall]

Being out by.02 that pretty much is the same amount the one I replied to yesterday what you are looking at Seth, I did the toolpath with minimum cutter radius at 0 with the offset set to in computer the files were different with minimum cutter radius at 0.02 it is a circle, not a corner that definitely should not happen. 

3D contour was within 0.002 with minimum cutter radius on or off.

In this case, it is not but 0.02.

[Anonymous]

This is happening on a straight edge, not a radius... 

[seth.madore]

What toolpath and where are you applying Min Cutting Radius?

[daniel_lyall]

It is your model shows that, have you double checked everything on the machine to make sure it works as it should?

[Anonymous]

"What toolpath and where are you applying Min Cutting Radius?"

the cutting radius is set to 0 - I think all the settings are in the screenshot  I don't understand what you mean by " It is your model shows that" 

I have tried cutting the parts at several different locations and orientations on the machine - all the same,  I have also been over it (several times) mechanically to make sure all is as it should. 

Though I would expect that since the holes are all bored in the same NC program using the same tool and they are all correct, the issue must be with the contour path(s) and/or the model itself ~ somehow. 

[daniel_lyall]

This one https://forums.autodesk.com/t5/fusion-360-manufacture/bug-2d-contour-quot-minimum-cutting-radius-quo...

It does not look like it is the problem in this post it may just be coincidences that it is 0.02 in this post I added 0.02 to the minimum cutter radius and it made no differences to the 3rd toolpath from the end on this post. 

[Anonymous]

I think I ran across that post - and a few like it ...  what is really strange that it is -.02 on the right side of the part and +.02 on the left side.   If I change the stock orientation along the Y axis,  the -.02 is at the bottom of the part and +.02 at the top.  So the same error in both orientations.  
I'm honestly completely baffled. 

[daniel_lyall]

To work out where it is coming from you have to rule out mechanical issues, the control and the G code from the cam program It is not minimum cutter radius making this happen, you said the machine is fine you tried reducing the feed rate.

This leaves double-checking the tool size and what you use to set X and Y 

Doing an air cut and measuring the start position of the axis and end this helps for double check that there is no or very little backlash, you can have one bad axis that stuffs other axes up.

Make sure the stock or bed of the machine is not moving.

When different post have different problems but the error is pretty much the same it makes it hard to find what is wrong.

[Anonymous]

Been over that last night - very carefully checking X/Y positioning, can it reliably find zero again, air cutting etc.   My table is built like a tank .. 🙂  it is not moving 😉   when I got the table installed I swept it with a dial micrometer, unbelievably  it was out at most .015 (from level) over it's 30x30" cutting area.  A leveling bit took care of that. 
I have to test the wheel tension and backlash nuts before starting anything (because I am paranoid) - they usually need only minor adjustment if at all... 

So now you got me wondering about the tool .... if it were to be bent I would expect the cut to be over on both sides, if it were under or oversize... ? the cut should still be under or over on both sides.   I can definitely (will) try with a different/larger tool and see. 

The only other thing I can think of there is maybe the router is not square to the table.  Though I would think that should be evident in the holes being bored as well.  (and also show the error on the perpendicular sides when the stock orientation is changed)