Fillets and chamfers suck

[inno]

just for the sake of understanding why Fillet Component resists so much to working properly, I did this kind of thing where I start form the un-chamfered curves (I reverted the chamfers by hand in Rhino and re-internalized the closed non-manifold curves)

then got their corner points to create a sort of tiling, then compared the center point of each tile to each QR-curve: if a point is contained odd times then its tile is good, if it’s contained even times then discard the tile:

after region union of all the remaining tiles, applying the Fillet to them works for each and all corners:

2351_MOD_QR-Code_unchamfered.gh (38.8 KB)

Faith in standard Fillet component -sort of- restored for sure a bit too much fragile…

[osuire]

Wow. I see what you did here !
You make use of the checkerboard nature of QR codes, and that’s great !
Anyone wanting to carve a QR-code can now use your tool

…but in general cases, much like Rhino fails miserably to fillet rather simple cases of polysurface edges, the fillet and offset commands (at least in V6) will trip over the slightest geometric difficulty and output garbage, or just do nothing.

I’ve been endlessly exporting solids to Solidworks and now Fusion 360 just to fellet edges for years now.
If V8 can solve this, that alone is worth the price of the upgrade.

[inno]

I just had chance to explore the new Clipper2, the non-manifold-curves version looked like an interesting benchmark easy as pie

2351_MOD_QR-Code_unchamfered_Clipper2_OMG.gh (34.9 KB)

[saeed_hasan_zadeh (saeed)]

@inno , May i ask how it goes on pure source curves?

[osuire]

@inno

easy as pie

Yeay, Clipper saves the day !
Funny how robust these QR codes are. Even with these rounded edges, my phone will still send me to the Xylotica website.

[inno]

this is how it looks on the source curves (chamfered to avoid non manifold)
there are still tiny fillets on the chamfers, but much smaller than the fillet radius

this is with the non-chamfered curves:

I noticed there are tiny areas that are behaving a bit differently, like this one, where I would have expected curves to move as sketched in yellow:

GH file with both curve-types internalized if you want to experiment:
Both_curves_internalized.gh (60.7 KB)

[saeed_hasan_zadeh (saeed)]

@inno thank you so much

[anon39580149]

You can collapse the tiny chamfered corners

qrcode_fillet.gh (34.1 KB)

[osuire]

Yes, that’s a good workaround.
But again, a proper tool should deal with these small edges, leaving the long edges untouched and gobbling up these critters with the Fillet.

[anon39580149]

This is another way using explode, negative extend and connect curves

This one with Qrcode plugin: QrCode | Food4Rhino

And for the one with chamfered corners you need to remove the tiny curves:

qrcode_fillet.gh (40.2 KB)

[Ashley_Cichocki (Ash)]

@anon39580149 Love this solution!

[inno]

@anon39580149

image1682×637 78.5 KB

mindblowing, that’s it https://www.youtube.com/watch?v=9CS7j5I6aOc

[anon39580149]

Use this code to replace smooth corners by arcs

[anon39580149]

I notice that Fillet component works as expected with Qrcode result.

qrcode_new.gh (11.1 KB)

Try the new update with level input

QrCode