Need to improve surface quality
You have Smoothing turned on, which can affect the surface quality - depending on the geometry.
My guess is that this what you are seeing, and that you can improve the results and performance by turning Smoothing off, or tightening the tolerance quite a bit on that one.
Note also, that if you use the Fast (3-axis only) stock simulation mode, it's often possible to see this kind of toolpath effect before running it on the machine. It requires a bit of fiddling, but you can see stuff like this:
Note that to get these images I simplified the part and the stock to "zoom in" on a sub-section of the part that should be representative of the surface quality. This is a 50x50mm section around the area where your photo appears to be from.
The simulation also has a "Zoom stock" option in the right-click menu, but I'm afraid that it's a bit hard to control. Testing out finishing quality on a sub-section of the part also means that it's quite fast to regenerate toolpaths.
I hope this helps.
Best regards (og med venlig hilsen!)
/Mark
Tha
@porsbym wrote:You have Smoothing turned on, which can affect the surface quality - depending on the geometry.
My guess is that this what you are seeing, and that you can improve the results and performance by turning Smoothing off, or tightening the tolerance quite a bit on that one.
Note also, that if you use the Fast (3-axis only) stock simulation mode, it's often possible to see this kind of toolpath effect before running it on the machine. It requires a bit of fiddling, but you can see stuff like this:
Note that to get these images I simplified the part and the stock to "zoom in" on a sub-section of the part that should be representative of the surface quality. This is a 50x50mm section around the area where your photo appears to be from.
The simulation also has a "Zoom stock" option in the right-click menu, but I'm afraid that it's a bit hard to control. Testing out finishing quality on a sub-section of the part also means that it's quite fast to regenerate toolpaths.
I hope this helps.
Best regards (og med venlig hilsen!)
/Mark
Thanks 🙂
I'm doing a third run on the finish right now, we'll see after that 😉
I thought the 'smoothing' thing was supposed to improve surface-quality, but I guess not 😄
About the simulation-graphic-representation-quality-thing….
This is what I see when I use the 'fast'-mode...?
Don't see a lot of details in that?
So....
Tolerance 0,001mm , no smoothing, stepover 0,2mm, parallel finish didn't work.
I tried a scallop instead. same parameters (just as much to verify that it's not the machine)
Some areas look great (apart from the 150´ corners), some look like crap. The toolpaths are not aligned properly?...
Don't know what to do...… Could get a better result with an anglegrinder... 😞
Hello,
I'm not sure if the issue you are getting is with the small join between the two surfaces, or just the finish overall.. I've taken a look at what is causing the obvious join and can see a few things that could possibly cause this.
To inspect the toolpath, it is always good to have a larger stepover so you can easily view the actual lines. I changed the stepover to about 20mm so it's easy to view the cutting moves. When simulating, I noticed there are two small arcs in the toolpath, right where you are getting this bump:
What tolerance are you working to? These arcs seem to get smoothed out with a looser tolerance and smoothing tolerance.
Also, do you know what software this was modelled in? I loaded it in the render environment and can see the join as well:
I have a strong feeling that it is actually the geometry causing this, and when it is followed with such a fine tolerance, you will get the result you are seeing. I created a plane through the center of the model and projected the intersection of the curve and you can see a very fine step in where the surfaces join.
I'm a bit confused when you said you ran 14 parts and half were bad? To me, that says the toolpath is correct but the cutting parameters are wrong (speeds / feeds) or perhaps the tool has worn? Have you tested with different speeds?
Hi George 🙂
Thanks for taking a look at this 🙂
The thing is, the transition between the two surfaces is what it is (I didn't design it 😉 but the errors in the other parts of the 'smooth' areas are really killing me. They feel like being 3-4/100 deep, and they don't appear in the same place, if I do the finishing in different ways.
This is the same part/area done with parallel finish (0,2 step 0,05mm/tooth, 10.000rpm's)
And with Scallop (0,2 step 0,05mm/tooth, 10.000rpm's).
So depending on which way I run the finish, I can 'move' the bad areas around.
There may be some design-related error between the two surfaces, but the rest should be smooth?...
When I said I had done 14 parts and half was good, that's related to these errors (marked in blue). Some parts are allmost free from errors, others need a lot of polishing.
Right now, I doing finish-run no.6, to get the parts done. I tilted a parallel finish 20 degrees (had to try something) we'll se in 2 hours ;)...
The thing that bugs me (and please remember that I'm very new to 3d and cam) is that I cant seem to get a consistent result from part to part, even though they are very much alike, and I use the same settings, same machine, same stepovers etc...
I've been milling for 20 years and I thought that this was going to be fun, but I'm not sure anymore… 😄
I've done 3d before, but only with function in mind, not finish.
This job has to be done with a finish Ra3,2 or better (in some areas) and when I do 3d directly on the machine, it can finish to Ra1,0. The results I'm getting with F360 (in the bad areas) are closer to Ra6,3.....
When I go through the posted data, I get 3 decimals, which should be more than enough to get a smooth surface.
I just don't understand?...……..
@MikkelStenbaek wrote:
Hi. I'm pretty new to F360 and to cam in general, but I have a job to do 🙂
I've done some die's for a customer, but I keep getting these little 'errors' in the surface. (see pictures)
I've done 14 parts now, and I only have the errors in half of the parts.
Would be great with some help 🙂
-Mikkel.
I noticed you have one finishing op set to one direction, are these the good ones? My spindle doesn't really run fast enough and I find 2 direction parallel finishing tears in the conventional direction so always use one way. Also think if I was doing this job I'd rotate the finish cuts 90° starting both cutting from the middle of the stock out. Not sure I'd use feed optimisation either but I'd semi finish the cavities with a parallel cut set to 1mm rather than the scallop you've used.
Start both from the middle like this.
Mark
Mark Hughes
Owner, Hughes Tooling
Did you find this post helpful? Feel free to Like this post.
Did your question get successfully answered? Then click on the ACCEPT SOLUTION button.
Hi Mark 🙂
The cavaties aren't important, they are just there to give space for epoxy to escape 😉
I tried both 0 and 90 degrees finishing with parallel, starting in the middle. Same kind of errors in the surface….
I'm doing a run now, with everything tilted 20 degrees, just to see the difference ( pictures in 20 min)
I tried both 'one way' and 'both ways' milling, and 'one way' does create a better (more shiny) surface, but the errors are the same.
To me, it seems like the F360 miscalculates the toolpaths, as you can actually see them in the simulation.
Here are some pics of the other parts I've done:
First pic. is the vertical plane, came out beautiful in the first try (all parts in one stock). Actually same geometri.
Second pic is the 4 big wings (l: apr. 800mm) Took several attempts to get a so-so surface.
third pic is a detail view of one of the 4 big wings (same geometri as the vertical wings we talked about)
4 and 5 are the latest pictures of the part in question. 20 degrees parallel, same bad result 😞
I normally get better finishes than that using 0.01 for tolerance and 0.01 for smoothing, on small parts might set to 0.005 so what you're seeing is odd. I wonder if it's because the curvature is so small.
Something you could experiment with is setting your tolerances a bit bigger then use compare and edit to set the the Surface Triangulation to a very fine tolerance. Just think having the tolerance bigger would be easier on the control, less blocks per second and the fine surface triangulation should stop too much deviation. Maybe set the surface triangulation to 0.0001mm and tolerance to 0.01, leave smoothing off.
Mark Hughes
Owner, Hughes Tooling
Did you find this post helpful? Feel free to Like this post.
Did your question get successfully answered? Then click on the ACCEPT SOLUTION button.
Hi Mark.
I tried what you suggested last night, but it doesn't seem to do much better...
It's another part but with the same design-ish. It's one of the 4 wings, so everything is just bigger 😉
I set the max radius to 10 and did a parallel finish with almost the same settings, but I still get those faulty lines.
I also do a 'pencil' along one of the curves, to clean it up, which look almost polished! It's so beautyfull, and if the rest of the part could look like this, I'd be dancing right now 🙂
Thanks for the files. I looked a bit more at the toolpath, and there are a couple of differences between the pencil and the parallel, either of which I suspect could explain the difference seen physically on the machine:
- The pencil is only removing 0.05 mm of material vs. 0.2 mm for the parallel (since the areas of the pencil toolpath has already been finished once by the parallel).
- The pencil is machining in one direction only, whereas the parallel is set up for both ways. You did mention that you tried this before, so that's probably not it.
- The pencil toolpath has much denser point distribution. This should not matter, though, as both toolpaths are inside of the tolerance, and if you compare the following stock simulation screenshots you can see that the toolpath should provide similar finishes:
If you disregard the graphics artifact (black lines etc.) and pay attention to the reflections (jagged brown bands), you will notice that the quality is similar between pencil and Parallel. This should be a pretty accurate representation of the kind of finish you would get with a "perfectly rigid" machine, tools that can't bend etc. - in other words ignoring the physical world (which simulation does).
I am sorry that I cannot say something more definitive.
/Mark
