One side of pocket is "fuzzy"

So, I’ve been experimenting with a few things, trying to get a feel for what I can and can’t do with the Nomad. And stumbled across something a bit odd.
I was cutting a rather complex STLfile with “machine whole stock” selected, so the Nomad cuts a rectangular “pocket” around the actual geometry.
The odd bit is the pocket: three sides are smooth as glass, but the fourth (the X+ end of the pocket) is “fuzzy.” Not in a way that suggests any odd motions of the cutter, but rather in a way that suggests the cutter was “ripping” rather than cutting during those passes. I was using the #101 and #102 0.125" mills, with fairly conservative settings (I was trying to keep the noise down) of 25ipm and 0.001" stepdown for the roughing and fine passes (10ipm for the pencil pass, but I doubt that pass touched the pocket walls). The material is “craft” plywood that I suspect is basswood (the label doesn’t say).
The only possible explanation I’ve been able to think of so far is that, on that side of the pocket, the Nomad might have been moving in a direction that meant the outer wall was being “normal” machined rather than “climb” machined (or vice versa?). Does MeshCAM control for this?
For additional detail: I had “X&Y” selected for the Fine pass. The “stock boundary” setting in MechCAM was 0.25", so there was more than enough gap between the outer edges of the geometry and the inner walls of the pocket that there should not have been any issues with the cutter “kissing” the outer wall while cutting the geometry. There were no signs at all that the stock might have slipped, or an axis might have slipped, during th cut.
One really strange bit: the fuzzy edge? Has an undercut, I kid you not. About halfway down the “cliff” the wall bulges inward, for roughly 1/4 of the height of the wall, then sucks back in beneath the bulge. The bulge is maybe .010" high (I don’t have a good way to measure it, but it’s very obvious to the touch). The “fuzzy” part of the cut is above the bulge, the bulge itself feels a bit rough textured, and below the bulge feels pretty smooth (it’s hard to get a fingertip in there). I’ve held the ball-end mill up against this wall, and if I rotate it just right, its profile seems to fit the shape of the bulge and the flat areas above/below it. But that doesn’t make much sense, since at 5000RPM the diameter variations in the mill should have been irrelevant. The flat end mill doesn’t fit nearly as well, although I would assume that the pocket would have been machined with the flat mill, given the floor/wall relationship is perpendicular and wouldn’t need the Fine passes with the ball mill.
Curiouser and curiouser…

So, some more experiments. And the “fuzzy” side is consistent… except that it flipped sides. Now it’s on all the X- facing verticals, rather than the X+ sides.

This time, though, I was watching (part of the time – it was an 8hr cut). The fuzziness seems to have been there from the beginning, rather than a smooth cut initial cut that got “kissed” later.

I also noticed that MeshCAM is definitely flipping between “normal” and “climb” cutting all over the place in this model. It also is making zig-zag cuts when there appears to be no reason to, and some of the long “climb” cuts are making really nasty noise (oddly enough, though, that doesn’t seem to be related to the “fuzzy” sides).

I took some videos of these events. This time, I was cutting in solid basswood, rather than any sort of ply. To see if I could decrease my cut time, the roughing pass is using a 2-flute, 0.250" flat end mill (my first try with the Big Cutter) at 25ipm and 7500RPM.

So! Further research – looks like I can select for the Climb/Conventional direction by un-checking the “Use Parallel Path” option. Have to try that next.

Hm… I had “X and Y” selected for the Fine pass, based on some issues I’d had with another piece of artwork earlier. I wonder if that might be causing the “zig zag” I’ve been seeing?

The Climb/Conventional setting only applies to roughing, SkyeFire. MeshCAM seems to always attempt to use climb milling on waterline passes, and of course parallel X/Y finishing is back-and-forth “lawnmowing” so alternate passes are climb and conventional. For 2.5D work pencil appears to always be climb also.

Can we get some pictures of said “Fuzzy wall” you’re experiencing? You’ve got me curious what it looks like :relaxed:

Yeah climb vs. conventional definitely does make a difference in 1) cleanliness of finish, 2) noise during the operation and 3) the risk of chipping out material or causing the stock to come loose if it’s taped down.

Generally I have found that with wood, the harder the wood/smaller the grain the better it does with conventional cutting, and softer woods do better with climb cutting because it doesn’t “peel” the fibers apart as much as it chops them cleanly with the fresh engagement of the cutter on each “bite” with the flute every revolution.

-Jonathan

Well, it’s a bit hard to make out all the details, but if you look at the left edge of the “pocket” around the cut (the big one, ignore that tiny Millenium Falcon in the lower left corner), you can see some of the fuzziness I’m talking about. The undercut isn’t visible, but it’s there as well. If you look at the far wall of the pocket, you can see that it’s glass-smooth. Possibly burnished? I’ve been deliberately running my feed rates low, probably sub-optimally so, but since both walls would have been done by the same cutter with the same parameters, I don’t really know what could have caused this. Not to mention, the undercut is a complete mystery. The dark layer just below the fuzzy band is the bulk of the “bulge” – it sticks out further than the rest of the pocket wall both above and below it

SkyeFire, this is a mystery to me. Does the height of the lower smooth area correspond to the length of the flutes on your cutter?

Also here is a blowup of part of your photo.

The rightward-facing walls of the interior geometry are smooth all the way down. But on them the X axis was in constant motion rather than being static as on the perimeter wall. That makes me think of some kind of resonance in the X axis when it was being held steady.

Did you really mean that you were using a .001" stepdown? That is probably much smaller than the width of the wood fibers in the plywood. I can imagine the tip of the cutter ripping out fibers rather than the side of the cutter severing the fibers. Have you tried a larger stepdown like .020" or .030"?

With plywood “climb” vs “conventional” probably loses most of its meaning. It would be rare for the wood fibers of each layer to be lined exactly up with the X or Y axes (much less perfectly straight), and most probably are at more or less of an angle.

Do you have a piece of blue building foam or such that is grainless that you could do a test cut on for comparison?

'tis a pozzlement! :smile:

(didn’t realize I had enough resolution to support that much zoom. Handy)

Yeah, I’m kind of weirded out. The in/out pattern does seem to fit the flutes of my cutter, but only at one particular rotation. And at 5000RPM, and ~25ipm, I can’t see any sort of pattern there.

I had another project where the torn fibers were on the opposite side of the X axis. So far the only common element is that it’s always on the X axis (or, more precisely, on the sides of the pocket perpendicular to the X axis). On this cut, the opposing wall is just as smooth as the “back” wall.

Yeah, that one really was 0.001" step-down – did I mention I was being reall conservative? One of the things I was trying to do was keep the cutting quiet, since my Nomad is actually sitting in my living room and I have to be careful not to interfere with people’s TV watching or sleeping (bedrooms with adjoining walls).

I’ve got a couple chunks of machining wax. I’ll give them a go.

I’m thinking that if your roughing passes are “along X” then it may be because of how it’s kissing the walls on the passes. If you don’t mind trying to do the roughing with “use parallel path” unchecked, and see what you get from that, it might also help with the sleuthing!