Very cool! I was struggling with this earlier today. I was increasing the stock to leave by .002 on the rest machine tools to eliminate the whispers. Thanks for explaining why they happen.
Phenomenally useful video. I wouldn't have known where to even start diagnosing this issue, let alone digging deep to figure it all out. And you top it all off with a super easy solution. I really appreciate the hard work and thought that went into this, John.
It would behoove the Fusion/HSM CAM team to solve this problem automatically and not force us to dink with spreadsheets to figure out stuff our $$$ CAM package should do for it.
Watched it numerous times. Unfortunately I have the problem with whisper cuts in the first 3D Adaptive, not in the next ones. Sometimes it's okay, sometimes it's so bad, that the toolpath almost isn't usable. Maybe someone got an idea to it.
What exactly was this project for? It looks like a gearbox housing for a Baja SAE team. I'm designing one of those right now and I'm getting pretty excited that I get to do the entire CAD, CAM, and machining!
Are whisper cuts those little bits of stock that Fusion seems leave and refuse to generate toolpaths for even though you told it not to leave any stock (because it's meant to be your final finishing pass?
Is there a reason not to lower the initial tolerances in the first op so the STL that is created has more resolution straight away? Thanks for the videos. I always learn a lot.
The whisper cuts had been bugging me for a while. This took care of them all! Cycle time for the one pallet dropped from 23 minutes to 21.5 minutes (24 parts on one pallet)
I get whisper cuts at a higher Z level when I'm using 3D adaptive. I don't think this really helps me with those. It happens on sidewalls parallel to the tool, so it makes no sense to me.
John this is good info. But you are most likely (though Autodesk would have to confirm) not totally correct in the relationship between the CAM sim output STL and the tool path generation. It appears to me that (as someone who has studied OpenGL) that the STL output model is a result of a tessellation shader and a collision intersection shader between the tool and the tessellated model at simulation. You can confirm this to a degree by running the CAM sim on a high power CPU but low power GPU computer. Where you will typically find that the tool path generation has no real problem, but the CAM sim is as bogged down as a HSS 4 flute in Al with no coolant starting to chip gual. However, there is still probably a high correlation between the tessellation size and the tolerance in the tool paths. But what you're mostly seeing in the output STL is just how the intersection shader is setup. For instance, if you try upping the smoothing in the tool path as shown by Haas and Autodesk (th-cam.com/video/vt7rVVRiiKk/w-d-xo.html) your STL tessellation is going to have to make a lot more faces even though the number of finite jumps in the tool path has decreased. I don't know what algorithm Autodesk is using to figure out the adaptive (tricordial ) path but it is most likely an optimization calculation based on keeping the angle of engagement constant. For example see here citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.590.8130&rep=rep1&type=pdf, and the following vid from Sandvik Coromant th-cam.com/video/Zt4NKgodCzY/w-d-xo.html
I have always wondered if there was a way to get fusion to not do some random moves during 3d clearing, might be good for another video? I have found they creep in now and again, where it has roughed out a section, then goes back and practically redoes the whole or part of a sequence. it's not a rest machining issue as it's usually in the first big hog out, but sometimes it can add ages to machining time. another thing that could be a cool video would be optimising depth of cut, as a few times I have found it hogs out hard once or twice and then does a third time with only a tiny removal off z. it's easy to tweak the roughing Stepdown to half the total height to fix it, but I have forgotten a few times and stood watching it run mostly cutting air for the last pass.
Won't this result in the part not being exactly the size of the model? Or are you assuming a 3rd finishing operation would then be added with a Stock to Leave unchecked? I'm a newbie...
Very cool! I was struggling with this earlier today. I was increasing the stock to leave by .002 on the rest machine tools to eliminate the whispers. Thanks for explaining why they happen.
Phenomenally useful video. I wouldn't have known where to even start diagnosing this issue, let alone digging deep to figure it all out. And you top it all off with a super easy solution. I really appreciate the hard work and thought that went into this, John.
It would behoove the Fusion/HSM CAM team to solve this problem automatically and not force us to dink with spreadsheets to figure out stuff our $$$ CAM package should do for it.
what about when you get these whisper cuts in a first, non-rest machining operation?
fine stepover, set it like roughing stepover if you dont have a 3d part that requires ball mill finishing surfaces
Watched it numerous times. Unfortunately I have the problem with whisper cuts in the first 3D Adaptive, not in the next ones.
Sometimes it's okay, sometimes it's so bad, that the toolpath almost isn't usable.
Maybe someone got an idea to it.
What exactly was this project for? It looks like a gearbox housing for a Baja SAE team. I'm designing one of those right now and I'm getting pretty excited that I get to do the entire CAD, CAM, and machining!
Are whisper cuts those little bits of stock that Fusion seems leave and refuse to generate toolpaths for even though you told it not to leave any stock (because it's meant to be your final finishing pass?
Explained it quite nicely....
Is there a reason not to lower the initial tolerances in the first op so the STL that is created has more resolution straight away? Thanks for the videos. I always learn a lot.
Tighter tolerance means more triangles. so.. Slower to calculate tool path due to more calculations, larger G code file due to shorter move segments.
This is very relevant to my needs.
The whisper cuts had been bugging me for a while. This took care of them all! Cycle time for the one pallet dropped from 23 minutes to 21.5 minutes (24 parts on one pallet)
I get whisper cuts at a higher Z level when I'm using 3D adaptive. I don't think this really helps me with those. It happens on sidewalls parallel to the tool, so it makes no sense to me.
Great video, John!
Thanks Eddie!
John this is good info. But you are most likely (though Autodesk would have to confirm) not totally correct in the relationship between the CAM sim output STL and the tool path generation. It appears to me that (as someone who has studied OpenGL) that the STL output model is a result of a tessellation shader and a collision intersection shader between the tool and the tessellated model at simulation. You can confirm this to a degree by running the CAM sim on a high power CPU but low power GPU computer. Where you will typically find that the tool path generation has no real problem, but the CAM sim is as bogged down as a HSS 4 flute in Al with no coolant starting to chip gual. However, there is still probably a high correlation between the tessellation size and the tolerance in the tool paths. But what you're mostly seeing in the output STL is just how the intersection shader is setup. For instance, if you try upping the smoothing in the tool path as shown by Haas and Autodesk (th-cam.com/video/vt7rVVRiiKk/w-d-xo.html) your STL tessellation is going to have to make a lot more faces even though the number of finite jumps in the tool path has decreased. I don't know what algorithm Autodesk is using to figure out the adaptive (tricordial ) path but it is most likely an optimization calculation based on keeping the angle of engagement constant. For example see here citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.590.8130&rep=rep1&type=pdf, and the following vid from
Sandvik Coromant
th-cam.com/video/Zt4NKgodCzY/w-d-xo.html
I have always wondered if there was a way to get fusion to not do some random moves during 3d clearing, might be good for another video? I have found they creep in now and again, where it has roughed out a section, then goes back and practically redoes the whole or part of a sequence. it's not a rest machining issue as it's usually in the first big hog out, but sometimes it can add ages to machining time. another thing that could be a cool video would be optimising depth of cut, as a few times I have found it hogs out hard once or twice and then does a third time with only a tiny removal off z. it's easy to tweak the roughing Stepdown to half the total height to fix it, but I have forgotten a few times and stood watching it run mostly cutting air for the last pass.
Won't this result in the part not being exactly the size of the model? Or are you assuming a 3rd finishing operation would then be added with a Stock to Leave unchecked? I'm a newbie...
These would be roughing operations, so doesn't matter. If you need a tighter tolerance part, you have to do a separate finishing op.
Very informative!
THAT FILE WORKS IN METRIC??
>1st...couldnt help myself. Great information John.
Come on, Jody ;) LOL thanks!
Holy sheep shyte batman! Now it makes sense! I will swear less going forward.
Muy buen aporte
1st!
KESTREL
Best! 😉
Peter Riis best?
1st?