We found a great manufacturer to partner with on this project, and they sublimation printed and sewed a test article for us with the graphic we designed. The fabric is super comfortable and the print looks amazing! Now it’s time to stretch its limits.

A piece of clothing so comfortable and awesome that you want to wear it all the time needs to be able to hold up to any activity you decide to do in it. Within the diverse realms of bouldering, road cycling, yoga, and dancing, we’ve been very happy so far with these leggings.

• Climbing: the polyester/spandex blend rebounds to its original shape even after extended wear. It was also the most-discussed item we’ve ever tested in a gym – no shoe prototype has ever started so many conversations! Maybe unusual shoes just are harder to talk to a stranger about than saying, “I love your leggings!”
• Cycling: the fabric weight was perfect for a coastal California ride.
• Yoga/dance: the leggings stretch gently with your motion and stay opaque when sized properly.

We’re excited to continue testing these SFTights to learn how they perform over time.

All of the design, minus the webbing, is in CAD now, so the prototype scheduled to be built this week will have lasercut patterns. Shipped the materials to the laser cutter today; plan to finish the webbing CAD while the parts are being cut and shipped back. After this prototype is built (aiming for Friday), it'll be time to start reliability testing.

Last week, I built the first full shoe with rubber on Dan's scaled last. This shoe also pioneered the first CAD patterns: after modelling the fabric on the computer, I used software to generate flat patterns from the 3D form, printed out the patterns, and sewed a shoe from them. It took a couple of hours to optimize the patterns, since the inputs for algorithm that flattens the surface have to be tuned to minimize distortion.

This means looking at a curvature plot of the surface to identify high curvature regions (which correspond to areas where the fabric would need to stretch a great deal), and then tweaking relief cuts on the surface to decrease the maximum stretch the fabric experiences. Patterns that equate to less stretch in the real world can be flattened by the software with the fewest errors and distortion.

Once I had hacked together a file with mostly optimized patterns, I started all over in a new file. Why? In honor of two talented engineers at my previous job, I wanted to create a parametric pattern file. Also because this will save me effort and time when building shoes to order in arbitrary sizes: a parametric model means that all the patterns should automatically update if I change the last scale.

In order to do that, the CAD file has to be structured very carefully, so I wanted to start from a blank slate with no distractions. I'd done all of the steps to create the patterns in the hacky file, but they were out of order. This gave me a clear idea of how to get from a last to a pattern; I just had to put all the steps in an order such that changing the scale (a step in the middle) wouldn't break any steps that came after it. This can be fairly simple if you use some best practices, like only using universal references (references that rarely have to be changed, and are usually some of the first five steps in a model). However, a last is a completely organic shape with basically no obvious references.

However, due to a quirk of 3D modelling, the ends of my lasts each have a small flat surface. CAD programs have difficulties computing models that have singularities - places where all the nodes of a surface converge on a point, like the pole of a sphere - so it's best to model a small area around these poles as flat, in my case - like the top of a hot air balloon.

Since these flat surfaces already exist in my last model, I created universal references on them. Now, all my sketches refer to those origin points, so when the references move, the sketches that generate the patterns move accordingly.

With this parametric model, creating Dan's shoe patterns was a matter of changing 3 scale constants and printing the patterns. The first prototype built this way had some problems with its rubber patterns (which are not yet in CAD), so I tore off its sole and modified it yesterday. Dan tried it and thought that the new sole offered better range of motion and fewer pressure points, but the lacing system over the tongue has to be redone. Since the lacing runs under the rand rubber, I'll have to build a new shoe. My goal is to finish the new shoe, incorporating all design modifications, by tomorrow at 7 pm, when we're going climbing with friends.

The most recent prototype made on Last 8 felt excellent during climb testing last week and passed the majority of tests I use to assess a prototype design, so I committed to grading Last 8 to Dan's size. Using scale factors derived from the slope of best fit lines from Army foot data (example graph below) and from Dan's foot measurements, I scaled up the CAD file, lasercut the new parts, and assembled Dan's last.

Now, I'm assessing methods for efficiently creating the graded patterns in 3D CAD and flattening them accurately.

Critical path:

Validate "relaxing" shoe concept & finalize the last shape - scale lasts to prove grading ability - reliability testing - next round of user feedback.

Each milestone is broken down below.

Shoe concept: Originally, our designs were built on flat lasts. This meant that they had to tighten to transform into their aggressive state. As a side effect, we had tight control over the shape of the shoe in the flat state (since this was directly molded from the last shape), but getting the right shoe shape in the aggressive state was a matter of much painful trial-and-error.

After our first official round of user feedback, it occurred to us that the state we really care about is the aggressive shape. The shoe's relaxed state shape is less critical since most of our use-cases in that state are low-performance, i.e. warming up, walking around, belaying, etc. Therefore, the last shape should reflect the aggressive state of the shoe, and then it can relax from there. This change in approach is what I'm calling the "relaxing" shoe.

Lasts:

With seven last designs in the rearview mirror, I've been able to debug issues ranging from scaling problems (just how undersized should a last be for the shoe to fit snugly?) to appropriate heel shape (finally getting that suction-fit around the heel and Achilles tendon which lets me heel hook). Last number 8, which I'll laser cut the parts for this weekend, should finalize the toe box shape.

Grading and other manufacturing issues:

Once the entire last shape is refined, it'll be time to grade the lasts. Grading means producing CAD files (in this case) of a range of last sizes. Unfortunately this is not as simple as scaling the whole last by some amount. I've been messing around in CAD to find a suitable workflow for the grading and have a rough draft of Last No. 8 in Dan's size. If Last No. 8 doesn't need any tweaks, then I can proceed to cut it in Dan's size as well.

Additionally, over Thanksgiving, Dan and I visited Bishop, CA and visited The Rubber Room. Tony, Dan, and the rest of the guys there were amazing - letting me look over their shoulders while they worked on resoling shoes and sharing their insight into climbing shoe construction. Before I left, they looked over a shoe I'd built and gave me a couple of ideas on how to improve the rubber application process.

Reliability testing: The testing plan will need to fleshed out when this milestone draws closer, but for now, at least the following tests will be necessary:

• Abrasion on seams and on adjustment mechanism.
• Slip tests on locking mechanisms.
• Cycle testing between shoe states.
• Flexural cycling - cyclic loading.
• Flexural cycling - don/doff.
• Thermal cycling.

Slip-testing results which compare the off-the-shelf ladder lock (shown in gold in the image) to the custom ladder lock are documented in the graph below.

Since the last post, I've simultaneously been drafting our patent application and modifying the Beta prototypes that fit me. Dan was able to take some nice shots of the Beta prototypes in action while I was testing them at Indian Rock in Berkeley this weekend. I'm proud to say that this pair of Betas is now passing 14 of the 17 tests which I've been using as a metric for prototype performance. This is up from the original 8 tests it passed when I first tested them a week ago.

Next steps:

When we have an official filing date from the USPTO, I plan to put up more detailed photos of SFT's prototype shoes to satisfy everyone's curiosity.

Once I've optimized the Betas that fit me, I'll modify the pair that fits Mak. Then I'll get in touch with everyone who's offered to try out prototypes and organize opportunities for some user feedback! This first round will necessarily be limited to volunteers who fit the two pairs of Betas available, so we'll play Cinderella. If you want to help out (and haven't already let me know) and think you'll fit the glass slipper, email sftclimbing@gmail.com with your contact information and foot size.

From a manufacturing standpoint, that would mean two less parts (+1 DFM), but would require a tool to create the recess in the rubber (-1 DFM) and a tool to mold the grommet in (net 0 DFM since a separate grommet also needs a tool). However, the in-place grommets will likely need to be cast before rubber is applied to the shoe, which means that the bond between grommet and rubber will have to withstand flexing and temperature cycling during assembly. Therefore, durability of the cast material and bond adhesion will probably determine how we proceed. Laura's pair of shoes will be used to test the cast-in-place grommets for durability side-by-side with the pre-cast ones.

While Ray is building Mak's shoes, work continues apace at the California office to get a patent for the shoe design. We've found an attorney and are working with him to conduct a rigorous prior art search to help scope our final patent application. Since prior art searches are less than photogenic, here's a consolation picture of the Continental Divide from Boulder Mountain, CO: