2D Adaptive leaves whiskers on my parts.

[Anonymous]

I'm having a problem with 2D Adaptive leaving whiskers on my parts. These whiskers are roughly .010 wide and I can see them in simulation. Is there any way to fix this?

I'm not sure how to share my design so that you can see the CAM files. Any help with that would be appreciated.

Thanks,

Chris

[Anonymous]

I'm having a problem with 2D Adaptive leaving whiskers on my parts. These whiskers are roughly .010 wide and I can see them in simulation. Is there any way to fix this?

I'm not sure how to share my design so that you can see the CAM files. Any help with that would be appreciated.

Thanks,

Chris

[HughesTooling]

To export your model go to model space and right click on the top component and pick export from the menu. Save as f3d and attach to your post.

Mark.

[HughesTooling]

To export your model go to model space and right click on the top component and pick export from the menu. Save as f3d and attach to your post.

Mark.

[Anonymous]

OK, I've attached the F360 archive.

[Anonymous]

OK, I've attached the F360 archive.

[fredsi]

Well, can't say for sure without running the code, but if you switch the Stock/Mode in the simulation to 'Fast' you don't see the whiskers. May just be an artifact of the simulation engine.

I seem to remember an earlier discussion of something like this either here or on the Autodesk CAM forum.....maybe Jeff or others here will recall it.

Fred

[fredsi]

Well, can't say for sure without running the code, but if you switch the Stock/Mode in the simulation to 'Fast' you don't see the whiskers. May just be an artifact of the simulation engine.

I seem to remember an earlier discussion of something like this either here or on the Autodesk CAM forum.....maybe Jeff or others here will recall it.

Fred

[Anonymous]

The problem is that in reality I'm seeing the whiskers on my parts. I've measured them at around .010" thickness and ~.125 long at the base. I'm sure that some of it is sloppiness in the machine but there's no reason F360 couldn't overcut that thin section slightly. Maybe I'm taking the wrong approach, but 2D adaptive seems to be the only way to cut that flat area on the part.

[Anonymous]

The problem is that in reality I'm seeing the whiskers on my parts. I've measured them at around .010" thickness and ~.125 long at the base. I'm sure that some of it is sloppiness in the machine but there's no reason F360 couldn't overcut that thin section slightly. Maybe I'm taking the wrong approach, but 2D adaptive seems to be the only way to cut that flat area on the part.

[fredsi]

Chris,

Uncheck Stock Contours on the geometry tab and see what you get (it was checked in the downloaded model). 

In general, I've noticed that with the availability of 3D tool paths, more and more 2D Adaptive is being bypassed for just Adaptive Clearing from the 3d tool paths. Makes programmiing go quicker with fewer selections, settings, etc. 

Fred

[fredsi]

Chris,

Uncheck Stock Contours on the geometry tab and see what you get (it was checked in the downloaded model). 

In general, I've noticed that with the availability of 3D tool paths, more and more 2D Adaptive is being bypassed for just Adaptive Clearing from the 3d tool paths. Makes programmiing go quicker with fewer selections, settings, etc. 

Fred

[martin.dunschen]

Hi Chris

Thanks for taking the time to report this. I just had a look at your part, and you are using a flat mill with a zero corner radius. Can you check the data sheet for the actual mill you are using and make sure that you have not missed a chamfer at the bottom of the tool? Your should model a chamfer with a corner radius for the tool selection in Fusion.

But even if you have a completely 'square' corner on a brand new mill, the corners will wear off very quicky and have the effect of small corner rad, You can estimate this radius or better measure it, and use this in the tool definition for the operation.

Another factor for this slither of left over material can be tool deflection. There is always a small amount of deflection of the tool causing a deviation from the rotational axis and the cutter is leaving more material at the bottom than the top. Deflection becomes bigger for a longer tool and a bigger stepoveror cutter engagement, so you could try with a slightly less aggressive cut.

Please let me know if that improves the results you're getting

Martin Dunschen
Prinicipal Engineer

[martin.dunschen]

Hi Chris

Thanks for taking the time to report this. I just had a look at your part, and you are using a flat mill with a zero corner radius. Can you check the data sheet for the actual mill you are using and make sure that you have not missed a chamfer at the bottom of the tool? Your should model a chamfer with a corner radius for the tool selection in Fusion.

But even if you have a completely 'square' corner on a brand new mill, the corners will wear off very quicky and have the effect of small corner rad, You can estimate this radius or better measure it, and use this in the tool definition for the operation.

Another factor for this slither of left over material can be tool deflection. There is always a small amount of deflection of the tool causing a deviation from the rotational axis and the cutter is leaving more material at the bottom than the top. Deflection becomes bigger for a longer tool and a bigger stepoveror cutter engagement, so you could try with a slightly less aggressive cut.

Please let me know if that improves the results you're getting

Martin Dunschen
Prinicipal Engineer

[Anonymous]

Hi Martin,

I'm using a brand new end mill with square corners. I can assure you that taking a .010" wide cut in the middle of a face is not causing the mill to deflect appreciably. 

If you run a simulation on the 2D adaptive operations in my CAM you can actually see that F360 is not cutting far enough into the center of the part. You can see it in the screen shot I attached to my first post. I've tried changing the width of cut to see if there is a magic number that will catch these whiskers, but so far I haven't had any luck. 

I think it has to do with the shape of the part. I'm trying to rough and finish a face that has a vertical wall. I tried using the same operation on a truly flat face and the last pass of the tool goes right down the center of the face. That's what I would expect it to do.

I'll try contacting the CAM group at Autodesk and see if they can give me some recommendations on how to fix it.

Chris

[Anonymous]

Hi Martin,

I'm using a brand new end mill with square corners. I can assure you that taking a .010" wide cut in the middle of a face is not causing the mill to deflect appreciably. 

If you run a simulation on the 2D adaptive operations in my CAM you can actually see that F360 is not cutting far enough into the center of the part. You can see it in the screen shot I attached to my first post. I've tried changing the width of cut to see if there is a magic number that will catch these whiskers, but so far I haven't had any luck. 

I think it has to do with the shape of the part. I'm trying to rough and finish a face that has a vertical wall. I tried using the same operation on a truly flat face and the last pass of the tool goes right down the center of the face. That's what I would expect it to do.

I'll try contacting the CAM group at Autodesk and see if they can give me some recommendations on how to fix it.

Chris

[martin.dunschen]

Hi Chris

Thanks again for this information. I have logged this as a problem report in our ticket system and will investigate further. There might well be a bug and there is a bit of remaining stock that the algorithm has missed. I'll keep you posted here.

Martin Dunschen

[martin.dunschen]

Hi Chris

Thanks again for this information. I have logged this as a problem report in our ticket system and will investigate further. There might well be a bug and there is a bit of remaining stock that the algorithm has missed. I'll keep you posted here.

Martin Dunschen

[martin.dunschen]

Just a quick update: When I use my stock simulation with a very high resolution I can't detect any remaining stock, so I think the toolpath calculation is correct; the reason for the probem must be found elsewhere, not in the toolpath itself. 

[martin.dunschen]

Just a quick update: When I use my stock simulation with a very high resolution I can't detect any remaining stock, so I think the toolpath calculation is correct; the reason for the probem must be found elsewhere, not in the toolpath itself. 

[kb9ydn]

Leftover slivers is a fairly common issue with 2D adaptive; not just in Fusion but also HSMWorks and Inventor HSM.  So far in my experience it's always been related to tool corner radius (including general wear), tool deflection, or machine/setup rigidity.  But it really would be better if the toolpath calculations took this into account and applied some sort of overlap to eliminate slivers and cusps.  To this point, it is an issue that is being looked at and there is a thread over on the CAM forum about it.

https://camforum.autodesk.com/index.php?topic=7444.0

The simulation really should show that all material has been removed so there may be something else going on here as well, but I at least wanted to point out that the leftover material issue is being looked at.

C|

[kb9ydn]

Leftover slivers is a fairly common issue with 2D adaptive; not just in Fusion but also HSMWorks and Inventor HSM.  So far in my experience it's always been related to tool corner radius (including general wear), tool deflection, or machine/setup rigidity.  But it really would be better if the toolpath calculations took this into account and applied some sort of overlap to eliminate slivers and cusps.  To this point, it is an issue that is being looked at and there is a thread over on the CAM forum about it.

https://camforum.autodesk.com/index.php?topic=7444.0

The simulation really should show that all material has been removed so there may be something else going on here as well, but I at least wanted to point out that the leftover material issue is being looked at.

C|

[HughesTooling]

Here's a backplot done with Rhino and the curves offset, it shows all the toolpaths should overlap and not leave any material behind. It could be down to constant contouring in your control smoothing the path. That said there's not much of an overlap.

Edit. Updated picture. Original curves 1,2,3 at the top Offset curves at the bottom

Mark.

[HughesTooling]

Here's a backplot done with Rhino and the curves offset, it shows all the toolpaths should overlap and not leave any material behind. It could be down to constant contouring in your control smoothing the path. That said there's not much of an overlap.

Edit. Updated picture. Original curves 1,2,3 at the top Offset curves at the bottom

Mark.

[martin.dunschen]

Hi

Yes, that'll be me looking at it. 😊

Martin

[martin.dunschen]

Hi

Yes, that'll be me looking at it. 😊

Martin

[Anonymous]

Thanks for the link to the camforum thread. I subscribed to it.

[Anonymous]

Thanks for the link to the camforum thread. I subscribed to it.

[Anonymous]

Hi Martin,

I don't understand why the toolpath generation routine can't overcut that last bit of material. Setting a radius on the tool is a workaround, not a solution. I also got an email response from CAM support. His response was to increase the Optimal Load to .040" from .0375". That seemed to work in simulation, but maybe increasing the load by 6% won't work in every situation. Please let us know if you make any headway on this issue.

Thanks,

Chris

[Anonymous]

Hi Martin,

I don't understand why the toolpath generation routine can't overcut that last bit of material. Setting a radius on the tool is a workaround, not a solution. I also got an email response from CAM support. His response was to increase the Optimal Load to .040" from .0375". That seemed to work in simulation, but maybe increasing the load by 6% won't work in every situation. Please let us know if you make any headway on this issue.

Thanks,

Chris

[jeff.walters]

The suggestion was meant as a workaround until a better plan is developed. “I don't understand why the toolpath generation routine can't overcut that last bit of material” the simple answer is it can. The hard question is how much is a bit? There has to be some logic to that amount. This is getting some serous discussion by the developers and a solution will be implemented as soon as they can.  

[jeff.walters]

The suggestion was meant as a workaround until a better plan is developed. “I don't understand why the toolpath generation routine can't overcut that last bit of material” the simple answer is it can. The hard question is how much is a bit? There has to be some logic to that amount. This is getting some serous discussion by the developers and a solution will be implemented as soon as they can.  

[Anonymous]

Thanks for the update Jeff.

[Anonymous]

Thanks for the update Jeff.

[kb9ydn]

---

@jeff.walters wrote:

The suggestion was meant as a workaround until a better plan is developed. “I don't understand why the toolpath generation routine can't overcut that last bit of material” the simple answer is it can. The hard question is how much is a bit? There has to be some logic to that amount. This is getting some serous discussion by the developers and a solution will be implemented as soon as they can.  

---

Nice!  Do you have any insight on the potential solutions being looked at?  Maybe an overlap setting by percentage of tool diameter?  Or maybe always try to maintain the set engagement angle (even if there is not enough material left) unless it will hit a boundary?

C|

[kb9ydn]

---

@jeff.walters wrote:

The suggestion was meant as a workaround until a better plan is developed. “I don't understand why the toolpath generation routine can't overcut that last bit of material” the simple answer is it can. The hard question is how much is a bit? There has to be some logic to that amount. This is getting some serous discussion by the developers and a solution will be implemented as soon as they can.  

---

Nice!  Do you have any insight on the potential solutions being looked at?  Maybe an overlap setting by percentage of tool diameter?  Or maybe always try to maintain the set engagement angle (even if there is not enough material left) unless it will hit a boundary?

C|

[jeff.walters]

I'm not sure 100% when it will be out but it’s on the top on the list and close to being released. The issue isn’t just tool defluxion as the Adaptive developer said that would be simple for them to calculate. You also have the material deflecting or pushing away from the tool as well. So the solution sill most likely be a mix of adding to the final stepover and or adding an additional pass depending on the geometry involved.

[jeff.walters]

I'm not sure 100% when it will be out but it’s on the top on the list and close to being released. The issue isn’t just tool defluxion as the Adaptive developer said that would be simple for them to calculate. You also have the material deflecting or pushing away from the tool as well. So the solution sill most likely be a mix of adding to the final stepover and or adding an additional pass depending on the geometry involved.

[onlinemailaccount]

I have the exact same problem as the original poster. Fusion is leaving many tall thin whiskers on my part when using 3D adaptive, or 2D adaptive. It even detects them in the simulation, and results in a crash in the simulation to the whisker.

I see this thread is 5 years old, and the Fusion team was working on a solution then. Did a solution ever become available?

Thank you.

[onlinemailaccount]

I have the exact same problem as the original poster. Fusion is leaving many tall thin whiskers on my part when using 3D adaptive, or 2D adaptive. It even detects them in the simulation, and results in a crash in the simulation to the whisker.

I see this thread is 5 years old, and the Fusion team was working on a solution then. Did a solution ever become available?

Thank you.