Shoelast

Post here for help on using FreeCAD's graphical user interface (GUI).
Leatherman
Posts: 88
Joined: Thu Feb 02, 2017 2:49 pm

Re: Shoelast

Postby Leatherman » Thu Mar 09, 2017 4:02 am

kridhond wrote:Nice!! That looks right!

I am working on an instruction of geometric preconditions for an average foot. With these formulas the model would build a last which can be used to make shoes you can actually wear and walk with. For example the tip of the last should be cleared from the ground depending on heel rise and length of the last. If you don't respect these "rules" you end up with shoes impossible to walk with.



We have already lots of Formulas and metrics collected, which we have not been published in this thread so far, because at this point of time it wouldn't be really interesting for other members. One of the next steps is to bring these metrics into action and see where we are really landing up. Up to now, it's more a testing about what could be done and how. From my first experience from this project is, that it would not be too difficult to add the important metrics for the shoe related measurements. The important part would be to see how FreeCad reacts on the "not" defined points or how it's interpolate them. At the very beginning of this thread we have already cut some ribs out of a demo-last with the Part Cross-Section tool, in different ankles and after manual adjusting these ribs, I have made a sweep and the result was very close to the original last in the forefeet, really great, but quite lousy at the heel area, in fact not usable at all. So, the final result will not so much depend on the 10-20 points, which we like to confirm and declare on the final shoe last, it depends more on the other 99.9% of points. Let's just see what happens once all the formulas and metrics are included.
I have quite less doubt, that somehow we will get a reasonable result and with reasonable I mean that it will certainly looks like a shoe last :D , but would it be for the beginning exactly the shoe last we have "thought" off is the still the big question mark? How much more manual work and tweaking must be done after the loft or "blow" is the real big question for myself. I guess we just need to wait and see.
Regards
Andre
User avatar
microelly2
Posts: 2446
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Shoelast

Postby microelly2 » Sun Mar 12, 2017 12:50 pm

After some problems getting the right parameters for the side walls of the shoe
my focus was on the methods to make the inspection of the scanned shoe last point cloud easier.
I see that the shoe must become broader but its hard to see how much.
When I create Bsplines from the point cloud I can compare them easier to our model.

The result of the weekend: I can create BSplines for any cut of the shoe and get the editable BSpline.
So we will have the curves. https://freecadbuch.de/doku.php?id=blog ... _to_spline

The next question now is how to careful reduce the number of poles without destroying the layout.
bp_411.png
bp_411.png (207.28 KiB) Viewed 310 times

At the picture the white is the scanned last, green is the cad model and black are the parametric points.
The sole matches exactly 1 mm. and the silhouette too. :)
kridhond
Posts: 9
Joined: Tue Nov 01, 2016 6:45 pm

Re: Shoelast

Postby kridhond » Mon Mar 13, 2017 9:21 pm

And the crowd goes wild! :) well done once again! Really!

I still needed to share my system to construct lasts. It would be nice if it could help to further develop the model. It's not clear from the pictures but no last is "flat" from left to right (so to speak) if you look at the bottom of a last.

In my two picture attached, I drew by hand 3 lines: black, red and green. For my purpose I need the freedom to design those 3 lines in xy-plane (supposing the ground is xy-plane):
- the black line is what would be called something like 'insole line' (I hope you get the point)
- the green line is medial line to support the longitudinal arch of the foot (sorry for the slang)
- the red line is the deepest last line (at the heel it becomes an area rather than a line). This line should pass more or less under head of 1st metatarsal (behind the big toe) or wherever the user wishes.

In side view, the red line is very important because its change in height (z-axis) is constructed using formulas. Typically you choose heel height (h10) and it changes the ground clearance of the tip + the curving in between. I would need some more time to let you know in detail, but you wouldn't be far off.

Starting from the red line the underside of the last is curved to end on the black line. From the black line it is going to the green line. Both on black and green line you should have discontinuity (rather sharp edges), but not on the red line. (In fact the part of the black line below the green line shouldn't be a sharp edge, but then you would need a 4th line drawn by hand. So it would become too complex.

Anyway, if you construct a last in that way, you get the bottom of a last suitable for shoemaking.

Anything above this bottom part of the last is making sure dimensions like instep, waist, girth, joint are correct according to user input and what the model is doing already looks very nice!! Respect to that!

I hope this maybe useful to add to the model.

I also could give you how the last must look below the ankles (top side of the last) in case of low shoes.

In case interested, I will further elaborate on that.

B.R. Kristof
Attachments
ShoelastXY.jpg
ShoelastXY.jpg (532.78 KiB) Viewed 289 times
Shoelastprofile_small.jpg
Shoelastprofile_small.jpg (952.46 KiB) Viewed 289 times
User avatar
microelly2
Posts: 2446
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Shoelast

Postby microelly2 » Mon Mar 13, 2017 10:27 pm

Thank you kridhond,
at the moment I'm on the way to make the model easy to edit.
One task is to convert the last scan to a collection of sketches on predefined locations (your drafts we can use to define these locations and ankles).
The next is to reduce the generated model to a minimum number of parameters.
and then create the gui for these parameters

for fine tuning the sketcher already works
here a picture where the sketcher is active on one rib.
bp_415.png
bp_415.png (43.26 KiB) Viewed 285 times


I can downsize the number of poles from the scan to 12 - 24 by script,
but I see further possible simplifications.
User avatar
microelly2
Posts: 2446
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Shoelast

Postby microelly2 » Tue Mar 14, 2017 11:17 pm

Today the question was
how can I reduce the number of poles in a bspline by a algorithm with minimal tolerance
The result in average distance

('BC-8 6 Start', 0.001)
('BC-15 12 BC-8 6 Start', 0.004)
('BC-15 6 BC-15 12 BC-8 6 Start', 0.012)
('BC-2 6 BC-15 6 BC-15 12 BC-8 6 Start', 0.03)
('BC-4 20 BC-2 6 BC-15 6 BC-15 12 BC-8 6 Start', 0.098)
('BC-4 20 BC-4 6 BC-2 6 BC-15 6 BC-15 12 BC-8 6 Start', 0.638)
('BC-3 16 BC-4 20 BC-4 6 BC-2 6 BC-15 6 BC-15 12 BC-8 6 Start', 0.6)

and a last step with the sketcher by hand ( 5 minutes interaction)
('BC-3 16 BC-4 20 BC-4 6 BC-2 6 BC-15 6 BC-15 12 BC-8 6 Start Sketch', 0.296)
That means we can simplify from 17 to 10 poles with an error of 0.3 mm. :)
User avatar
microelly2
Posts: 2446
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Shoelast

Postby microelly2 » Wed Mar 15, 2017 1:08 pm

I have played to edit multiple ribs in parallel. Switching between the sketches is to slow (for me), so the idea is to give the parameters for any sketch into an dialog and open all dialogs needed.
the feeling is like a conductor in an orchestra :lol:
Image
https://freecadbuch.de/doku.php?id=blog ... _von_daten
https://www.youtube.com/watch?v=cZppPX9zbRQ
Leatherman
Posts: 88
Joined: Thu Feb 02, 2017 2:49 pm

Re: Shoelast

Postby Leatherman » Wed Mar 15, 2017 3:25 pm

This is brilliant! It's looks very simple indeed and is very helpful. Are the weight points 1-10 randomly or are they positioned for each rib at the same position, which would be quite helpful? It would be nice to know where is the weight 1 at the rib located? Not that it is very important with three ribs, but if you need more ribs to adjust then this would be certainly good. Just my five cents.. Really cool stuff 8-)
Andre
User avatar
microelly2
Posts: 2446
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Shoelast

Postby microelly2 » Wed Mar 15, 2017 3:56 pm

The poles are the orange points in the image.
There the weights work: A high value means that the curve is pulled to the pole, a low value makes the curve more straight.
On the picture in the last post the poles are the red centres of the blue circles.
We will label and show the poles when the dialog opens and hide them again when it closes.

I have done such a dynamic labeling in an other context.
Image
Leatherman
Posts: 88
Joined: Thu Feb 02, 2017 2:49 pm

Re: Shoelast

Postby Leatherman » Thu Mar 16, 2017 8:25 am

Great! Found this for clarification. May be useful?
http://geometrie.foretnik.net/files/NURBS-en.swf
If u hover over the control points u can change the weights of each individual, it's nice to play with around.
Regards
Andre
Attachments
Nurbscurve.png
Nurbscurve.png (106.01 KiB) Viewed 194 times
User avatar
microelly2
Posts: 2446
Joined: Tue Nov 12, 2013 4:06 pm
Contact:

Re: Shoelast

Postby microelly2 » Thu Mar 16, 2017 2:57 pm

Now the poles labeled and the connection between the poles and the knots visualized.
https://youtu.be/l1KXbeTOvwE
Attachments
bp_420.png
bp_420.png (76.26 KiB) Viewed 172 times