Crafting chocolate and cases
We’ve made a lot of progress in development in the last few weeks. (More on that in a bit.) We also crammed in a visit to our friends at Dandelion Chocolate. Since we’re all about making stuff these days, I want to give you a peek at how they make stuff, and how they’ll use Twine.
Todd and Alice were nice enough to show us their new bean-to-bar chocolate factory, one of only a dozen places in the US that make chocolate from beginning to end. Dandelion travels around the world to find the right cacao beans, then brings them back to San Francisco to turn into fine chocolate, adding only sugar. We tasted it at each step, and it is magic to get from that bitter bean to a uniquely tasty bar. Like us, they’re really small, and have improvised a lot of their processes. As you can imagine, the machinery needed to do this is very specialized (ever heard of a bean guillotine?) and almost as hard to get as Oompa Loompas. So they started with roasters made for coffee beans, and a grinder powered by a hand drill.
Because they make small quantities with personal attention, Dandelion wants to use data to keep the quality consistent and learn from each batch of chocolate. Even the industrial machines they have don’t accurately measure critical things like temperature and viscosity. So they plan on using Twines with custom high-temperature probes to log all data into a computer, where they can not only monitor the progress (some steps can take a day), but also go back months later and learn what made that batch so good. This is the sort of thing that we would’ve never imagined Twine being used for, and we’re gratified that it’s helpful.
Thanks to Dandelion for the tour and tastes! It’s amazing to realize how many steps there are to making our chocolate — finding, sorting, shelling, roasting, grinding, more grinding, mixing, tempering, molding and cooling — and seeing it all being done by one small company reminds me to savor each piece.
We’re also grinding away. Assembling the first pre-production boards, stringing together the software all the way up to the web interface, and iterating through case prototypes. We’ve got some extra hands now but it’s a lot to do, and do well — especially when we got 15 times the response we were expecting! As other Kickstarters have found before us, that requires an adjustment in plans.
For example, instead of casting each case by hand, we’re now going to produce them using injection molding. The design of an injection molded part is a complex process that requires many rounds of iteration to get all of the details exactly right. Injection molding requires very expensive steel molds with several interlocking pieces that must come apart cleanly and automatically to form a case that meets our quality standards. Before we commit tens of thousands of dollars to that, we’ve had to prototype the process. Here’s a peek:
Sucking bubbles out of urethane using a vacuum pot.
Pouring the urethane into a syringe to inject into the prototype mold (in the clamp).
A prototype curing in a 3D-printed acrylic mold, next to an earlier cast prototype.
But these are good problems to have. Your amazing support is giving us the opportunity to make Twine much more than we had planned, and David and I are working double shifts to make sure you’ll love it. This project is our lives now (apologies to our significant others). We’ll be back soon with an update to cover some of the improvements.