Because our last model is parametric (its geometry is determined by a stored set of equations and relationships), it can adapt to accommodate different variations of foot proportions. This video shows how the last shape changes for a 5% increase in length and a 22% increase in width across the ball of the foot.

In light of Kevin's suggestions, I've started planning a new CAD model which will combine the best points of my current lasts and the things I learned from visiting Leahy Custom Hiking Boots. 

By comparing how much weight (based on the volume of water) swatches could hold, I determined which adhesive was best for the SFT shoes and the optimal procedure for curing the adhesive.

Dan was finally happy with how the toebox shape felt - a symptom of getting the rubber patterns and application process dialed in. I believe his words were, "Dang son, I want a pair of these," mumbled as he pulled the shoe on the second time. Three of the gym's staff members were kind enough to try on this prototype; their feedback was also positive, citing toebox shape as well.

The biggest hurdle overcome in this prototype was proving out a new bearing design. The adjustment strap of the shoe currently runs from the instep side of the shoe to a "bearing" under the toebox sole, and then to the ladderlock on the outstep of the shoe. Previous bearings tended to become misaligned from the force vector of the adjustment strap, resulting in high stresses in the materials holding the bearings. In Dan's first shoe, these stresses (which are proportional to the climber's weight) were high enough to tear the bearing retention material.

Having seen what happens when a misaligned bearing is subjected to high loading, I sought to design a bearing that would remain aligned or self-align. Since this part was critical, I didn't want to finish the all-lasercut prototype with out it, but that pushed back the completion of the prototype by about a week. A lesson in supply-chain management was learned (most of the delay came from waiting for endmills of the appropriate size to arrive in the mail), and the bearing performed flawlessly last night.

Additionally, here are some snapshots of rejected concepts as well as the evolution of this mark: 

On the proof-of-concept alpha prototypes, we had problems with friction in the tension system. Combinations of high friction coefficients and too many lacing points meant that adjusting the shoe into the aggressive state almost always left slack somewhere in the tension system. While climbing, this slack would redistribute and the whole shoe would relax. So we minimized the lacing points and spent a great deal of time making bearings to provide a low-friction surface for the tension system to run through.

On the beta prototypes, the friction in the system has been decreased enough that tightening and loosening the shoe is quick and the tension is distributed evenly. This is great progress. However, only once the system had been optimized for low friction, did it become clear that there may be an issue with how the shoe is locked into a certain state. We've been using ladder locks on webbing to secure the shoe state, and the first time I climbed in the beta prototypes, the locks slipped, allowing the shoe to relax. Same failure as with the alpha prototypes, but a different cause.

After inspecting the ladder locks in different states on the shoe, it became clear that because of the curvature of the shoe side, tightening the tension system put the ladder lock on the heel, resulting in a self-unlocking position.

I shortened the travel of the ladder lock so that it remained on the side of the shoe and this appeared to resolve the slippage issue. However, after testing these prototypes on longer duration climbs, it became apparent that there was still a very slow slippage rate. Estimating the force carried by the tension system for a person of my proportions and weight yielded a loading scenario of about 2 kN. The company which manufactured the ladder locks and webbing we're using rates them for 440 lbs before they slip, which converts to about 1.96 kN.

I am a small person, so my characteristic load should be lower than most of the population. Therefore, load forces are almost always going to exceed the rated load for these ladder locks.

But this does not mean total failure! Note that the load forces should exceed the rating for these ladder locks. In fact, this is an excellent development, since now I have a metric to spec the locking mechanism for this tension system. And if there are no ladder locks which are small enough with high enough ratings, then I can always deliberately reintroduce friction into other parts of the system.

If I had super thin circuits,

film super capacitors,

a distributed sensing (or a brain-computer interface),

and digitally reconfigurable materials,

then SFT shoes could adjust themselves.

Let's unpack my wishlist. Super thin circuits are on their way, but why do I want them? In a couple projects I've seen, flex cables and flexible circuit boards were a blessing and a curse. They're the only system that can enable certain designs, but their bend radii (at least 20 times the material thickness for dynamic applications) can start to drive a design. So from this perspective, thinner electronics are great, provided that the bend radius-to-thickness ratio remains similar or improves.

Physical design is gated by power storage solutions even more than it is by circuitry. My basic rage at batteries is due to the fact that their power density is pitiful, they have to be protected like precious eggs, and their form-factors are not nearly varied enough. Definitely looking forward to when thin-film supercapacitors, like this transparent one, can store useful amounts of energy and are cost-effective. So let's pretend we've got that.

A brain-computer interface might be asking too much (though Ms. Jepson disagrees), especially since it would be much slicker to let the shoes sense what state they should be in, rather than the wearer needing to consciously direct them. Distributed sensing could mean a simple GPS and some microprocessors or the ability to pair with a phone. Both could draw data from a user-populated app like Strava that would tell the shoe, "you're in Yosemite, and you're not gaining a lot of elevation, so you're probably bouldering." On the other end of the spectrum, the shoe might be laced with strain gauges and EMG sensors, which would provide data to gently relieve strain on foot muscles during a climb.

Which brings us to the last item on the wishlist: digitally reconfigurable materials. Everything described above is useless if the shoe cannot change in the real world and in real time. There is lots of potential with materials like piezoelectrics or Nitinol, but since I have to wait for further development on energy storage systems, digitally reconfigurable materials better make some advances in the intervening time too.

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:

While I was gluing rands onto shoes last night, Dan started machining a grommet with the manual mill. Mak's finishing it now so that our prototypes can keep rolling. The adhesive cure time (12 hours) means we have to work fast on tasks in series with the glue-up and then do other things in parallel while the shoes dry. If we stick to schedule, we'll be able to climb test in the next set of protos tomorrow.

Since we've been able to cut down the time to sew a prototype from about 10 hours to 5 with our sewing machine, there has been a flurry of new shoes; my current favorite is the "Vomit Comet"... named for its color.

The "Vomit Comet" offers more smearing area than anything we have but our old 5.10 lace-up Coyotes, and it edges as well as Mak's Miuras. The down-side: it's hard to tell if you've tightened it down too far until you get your feet on the rock. I guess there are worse problems to have than a prototype working too well...