SUPERMECHANICAL.BLOG

Note: If you’re a Kickstarter backer, you can update your shipping address at the Spool website we set up, under “Account”. Some of you haven’t yet signed up there - if you don’t have the registration link we sent to the email address you use for Kickstarter, send us a message.

Here’s a material chip of the custom Twine Aqua color formulated for Twine’s shell. (Accuracy subject to the camera and your screen, of course.) A lot of nitpicking went into this — it’s not as easy as picking a color in Photoshop. We couldn’t find the plastics we’re using in a color that we liked, so we sent in one of our original prototypes to get color matched properly in PolyOne’s lab. They got it just right, in the polycarbonate blend and Santoprene we’re using for the inner frame and outer sleeve, respectively.

The UPS man delivered several thousand 3-foot cables for external sensors, custom made for us! Our sensor cables use four-conductor 3.5mm jacks, just like an iPhone’s headphones. While similar audio cables may work, we designed our own cables with better shielding — your ears might not notice a little noise in your music, but sensor data needs to be more accurate, and shielding maintains signal integrity over longer distances. We also took the opportunity to find a nice flexible white cover.

Our intern Tim has been fine-tuning Twine’s sensors. We humans implicitly understand when something is shaking or sitting still, but to an electronic device, it’s just a rapid stream of numbers that represent force. So we have to figure out exactly what pattern of numbers represents what a reasonable person would say is a knock, for example. Here’s a video of him testing the vibration sensing code. You can see it’s sensitive enough to pick up the vibration from a song’s vocals playing on a laptop. You’ll be able to set the sensitivity level to trigger an action when you want it.

As promised, we’re doing shorter, more frequent updates. There’s a lot of progress going on in general, but it’s just a lot of coding, talking with suppliers, and beta testing — stuff that doesn’t photograph well. We’ll be back in about a week with another report!

Since our last update we’ve really gotten into a productive groove here in Austin.  Maybe it’s the sunshine, maybe it’s the much cheaper beer — whatever it is, it’s really been working. We finally have our own workspace, and a hardworking new intern named Tim who’s been improving the sensor software and helping with testing. Here’s where we are:

Case

Twine’s body is designed for injection molding and quick, screwless assembly. The case has two parts: a hard inner shell that snaps onto the main circuit board, and a soft outer shell that provides shock protection and a nice surface feel.  Our brand new injection molds have started cranking out both cases and things are mostly looking pretty good.  The hard inner case fits nicely over Twine’s circuitboard (though we do need to make the board a few thousandths of an inch bigger to eliminate rattle).  However, the first run of outer sleeves was not quite as good. It’s made out of soft and squishy Santoprene, whose properties are very difficult to prototype accurately — sometimes you just have to try it with the real (and expensive) process. The mold had some flow problems that caused distortion and a bad knit line (as you can see here), which is the result of the hot material meeting up from opposite sides of the mold. The sleeve’s fit was too tight as well.

This isn’t good enough, and John’s worked with our injection molder to design a new mold that should give us much better results. We’ll see the new parts soon.

Hardware

The electronics have been looking great in beta testing. Power consumption has met our expectations and the sensors and external port are working as designed. We’ve found only minor edits to make, including better protection against frying Twine when you accidentally put the batteries in backwards. Once these tweaks are finalized we’ll commit to the full production run where there’s little room for error.

Firmware

This is the software that runs directly on Twine, and has the task of managing sensor readings, checking when your rules have been triggered, and talking over the Internet. We all take it for granted with our computing devices, but the last bit is especially hard to do in low-power, low-cost devices like Twine, and David’s had to write a lot of this from scratch. Lately, he’s been fixing the bugs that beta testing shook out. Twine will keep getting better even after it ships thanks to over-the-air updates, but that depends on first having rock-solid communication.

Web software

We’ve made some great strides on the web side of things lately. Our rules creation engine is fully integrated and our beta testers are busy creating and testing all sorts of rules.  They’re doing a good job too — John has been super busy refining the interface and fixing bugs based on beta testers’ feedback.

Our beta testers have been keeping us busy enough that we haven’t yet had to expand our beta to everyone who graciously volunteered, but we’ll send out more units when we have more features to test. We’re working with our supply chain to make an ETA of August for the production Twines.

Some backers have asked whether they could have access to Spool ahead of time. The rules editor won’t work without a live connection to a real Twine — we’ll look at separating them, but we don’t want to lose focus on our primary task of shipping Twine. The outlook in Austin is sunny and things are happening fast now, so as requested by several of you, we’ll be updating in shorter weekly posts from now on. See you next week!