Milled dome-like shape seemingly has low-poly count

[Anonymous]

Hi everyone! Quick bit of background I'm fairly new to Fusion and milling in general, however have quite a bit of experience using Sketchup and 3d-printing so I'm very familiar with translation of 3d-model to gcode and such.

I'm attempting to mill a mold out of aluminum for a part that's been designed from the ground up in fusion (none of it is imported from non-fusion files or anything). I've attached a picture of the first attempt I made at this. As you can see in the picture, what is supposed to be a smooth sphere-like shape all the way around is clearly made up of lots of line segments. In Sketchup, I would have just increase the number of segments that defined the circle, but I can't figure out how to do an equivalent operation in fusion since fusion natively knows what a circle is (unlike Sketchup which converts everything to lines; one of the reasons I'm wanting to move away from it).

I originally thought that the "tolerance" parameter would have directly impacted how close the lines in the gcode approximated the original circle. For this attempt, I set that parameter to 0.001 mm. The mill I have actually has a worse resolution than the 0.001mm, so I figured this quality would be as good as the machine could possibly do, but clearly it leaves quite a bit to be desired.

Is there some other parameter or setting somewhere that would be impacting this?

I'm thinking this is in fusion because the smaller circles for alignment pegs look like perfect circles (also pictured), so I firmly believe the mill itself is capable of better so I have to assume it's the gcode instructions it's getting from Fusion that are resulting in this. I was super concerned about having ridges from each "layer" it stepdown to, but that seems to have been great in terms of resolution (when I run my finger over it, it feels completely smooth).

Settings I used for this:

Roughing step: Adaptive clearing; set to leave 0.5 mm of stock

Finishing step: Contour, tolerance: 0.001mm, maximum stepdown: 0.05mm

Both used a 1/4" ball end mill.

[Laurens-3DTechDraw]

What machine do you have?

And have you tried the smoothing setting in the operation?

[Anonymous]

Looking at the pictures, I see a lot of chatter, potential tool instability, feed rate and / or rpm too high.

If you can step up to larger tool size for rigidity, also look at the difference between Contour and Scallop tool path.

Contour doesn't do well on transition between steep and shallow curves while Scallop has a lot more uniform stepover thru entire surface.

[Anonymous]

The machine is a 3-axis mill from Microkinetics (https://www.microkinetics.com/index.php?page=cncmill); it's really a Sherline mill with some additional drivers and such.

I did try the smoothing option as well, however that had the same issue. during that it seemed like the machine would move, pause, move, pause, etc. and showed the same thing (I was originally trying in wood prior to getting the aluminum blocks). I assumed this was due to the machine having to spend more time doing the circular interpolation instead of the linear and those minor pauses (say 1/4s) were causing the mill to dig further into the piece there, causing this pattern. Now that I'm thinking about it though, I don't think that was the case since I see the same thing happening without smoothing on when those pauses generally don't happen.  

[Anonymous]

---

@Anonymous wrote:

Looking at the pictures, I see a lot of chatter, potential tool instability, feed rate and / or rpm too high.

Could you elaborate on what makes you say this? Like what specific things in the picture(s) leads you to suggest this? (I believe you, I'm just trying to learn what those things look like)

I did do this very slow, so there were parts of the operation where it was literally cutting off nothing. I'm assuming that's certainly what can cause some of the scuffing and such that's on there, but I guess I would expect that to show up as a fairly random pattern/texture across the whole part as opposed to the very geometric pattern that I'm trying to solve. Am I on the right track there or could those issues you pointed out be causing this too?

---

@Anonymous wrote:

Contour doesn't do well on transition between steep and shallow curves while Scallop has a lot more uniform stepover thru entire surface.

---

 Would this be causing the issue I'm seeing in the x/y plane though? I was super concerned about that step-over in the z-axis (hence the 0.05 stepover), but that direction is smooth enough for the purpose. This is what's got me so perplexed is that I just don't see how that geodesic-like pattern is appearing.

Thanks for your suggestions!

[Anonymous]

So, let's play detective and try to decode what's in the pictures. You say you used Contour to finish and yet I see vertical lines suggesting tool path was working from top to bottom then back to top before taking another pass.

If I zoom in to few spots I see texture similar to fish scale or saw tooth, that's result of tool vibration while cutting.

It looks to me like finish cut never touched the surface actually. 

If you left stock in roughing operation then followed by 3D contour, tool would start at the top and work its way down in circular pattern, with step over ridges showing in horizontal or spiral fashion, like those in pictures I uploaded earlier.

Speaking of my pictures, look at the bottom of the test model, difference in density of Contour passes compared to  Scallop is huge. So what I am suggesting is that you have to pick finishing strategy that produces better uniformity given shape of the surface.

Having f3d file to see what you are actually doing would be a lot better way to find potential cause of undesirable effects and avoid a lot of "detective work". 

[Anonymous]

Here are your pictures suggesting finish toolpath never touched surface after roughing operation, or something else is terribly wrong so I'd like to see your F3d file, just curious how you ended up with this.

[Anonymous]

---

@Anonymous wrote:

You say you used Contour to finish and yet I see vertical lines suggesting tool path was working from top to bottom then back to top before taking another pass.

---

You'd think that right?! That's what has me so befuddled. It didn't start at the top and go down, but instead started at the bottom (I told it to start there instead of top) and worked its way, painstakingly slowly, up the entire surface in a more or less spiral; the operation took like 8 hours to do, most of which I sat through watching it so I'm very confident that it went around in circles as expected.

---

@Anonymous wrote:

If you left stock in roughing operation then followed by 3D contour, tool would start at the top and work its way down in circular pattern, with step over ridges showing in horizontal or spiral fashion, like those in pictures I uploaded earlier.

---

Yeah, that exact issue is what I was trying hard to prevent (visible step-over ridges), which I was able to avoid, but somehow had this other issue.

---

@Anonymous wrote:

If I zoom in to few spots I see texture similar to fish scale or saw tooth, that's result of tool vibration while cutting.

It looks to me like finish cut never touched the surface actually. 

---

Ah, thanks for that. I didn't really notice that earlier, but I see what you mean. 

I exported and attached the f3z that was used to machine it so you can see exactly the settings and such used.

[Anonymous]

Simulating the file now, I see that bottom is flat and ball end mill has a problem there, you should finish that using 2D contour or pocket and bullnose end mill then follow with ball end mill working from top to bottom.

In screenshot, you can see that tool picks up increasingly larger cut as it works towards the steep part of the curve, that overload is causing it to deflect from cut and bounce back as do heavy ridges left by roughing, indicated by two short red lines.

Tool reacts to varied thickness of stock while it works backwards and that is not benefitting finish surface.

You need more uniform stock left by rougher because .25 diameter tool is week and needs more stable load to produce better finish surface.

I'd say add one larger tool to take balk of the stock out in larger step overs, follow with semi finish pass leaving about .005 radial stock then run finish pass always driving each tool from top to bottom.

If you set step over to .014 of an inch value for .25 ball end mill, your theoretical surface finish will be 62 microinches, providing there is no chatter caused by tool vibration.

I'll see if I can work  out better option.

  

[Anonymous]

So, I didn't know that there was flat at the bottom of middle hole. I used 8 mm bullnose with 1 mm corner radius to rough and finish.

I deleted some of the clutter to concentrate on hole and what I am showing to you is gradual removal of stock so that finish pass has uniform stock which contributes to finish surface consistency, and there is no benefit in driving tool from bottom up as shown in my last post, if you do that , problem gets worse as tool gets closer to top because of increasingly more vertical wall that increases footprint of tool engagement in cut.

Simulation shows texture but that's just resolution, it should come out clean in machine.

[Anonymous]

I don't think it's just due to the resolution that's causing those lines to show up since they're the same lines that I'm seeing on the part. Here's the simulation after applying the settings you suggested, you can clearly see where it's going to cut too deep in vertical lines.

[Anonymous]

We are looking at magnified computer simulation where graphics and resolution are exaggerated that's why I say it will look better in machine.

I am not sure how to optimise toolpath for best outcome with smoothing but reducing stock for finish pass and driving tool from top down should result in a lot better part then your original. 

[HughesTooling]

From the pictures and what's been said in this thread I'd be looking at problems with the machine.

First check backlash in X and Y. Next test for an axis sticking, if you use a clock and move an axis slowly one way then the other does it move smoothly or does it stick and then move? Last you need to check the rigidity of the head and column. 

The way you have the bands in the finish that make the 3d contour toolpath look more like a parallel toolpath point to a mechanical problem. When you have problems caused by tolerance accuracy you tend to start seeing a mesh in the finish that would be radial more like your picture of the simulation not something along an axis.

Mark