Share this project

Done

Share this project

Done
A glove-based input device that provides full mouse/keyboard control built for wearable/mobile computing and handicapped users.
A glove-based input device that provides full mouse/keyboard control built for wearable/mobile computing and handicapped users.
95 backers pledged $12,474 to help bring this project to life.

Keyglove Kit PCB Revision Update, Glove Status, and an Anniversary

Happy Saturday, everyone!

I've been continuing to make progress on the Keyglove, and especially for those of you who are interested in kits, I've got a great update. I've also got some sad news about the potential glove manufacturer I referred to in my last update, but let's start with the good stuff for now.

For the "TL;DR" folks who would love to jump right into oodles of Kit PCB pictures, check out this, this, this, and this photo album on the Keyglove Facebook page.

Kit Board Updates: 3rd time's the charm!

As I mentioned last time, I've been working on a compact, simplified circuit board that will allow people to put together a collection of off-the-shelf parts and prototype modules easily without having to worry about exactly how to connect everything together. My goal is to end up with something that requires only connections from the sensors on the glove itself (clearly labeled, of course), and everything else should be plug-and-play--or perhaps plug-and-program-and-play. The kit board is based first and foremost on the Teensy++ board from PJRC and the 6DOF IMU motion sensor from SparkFun, both of which are easily obtainable and proven, tested components. The first attempt at this board, version 0.1, was a good effort:

But even before the v0.1 board made it back from the fabricator and into my hands, I'd already scrapped it in favor of a new version. First, the circuit was missing a few connections that are required to provide seamless wireless (Bluetooth) connection management. Second, I had just modified the sensor count to include a total of 37 instead of the 34 I had been working with, and this required a significant reworking of the RGB/vibe/buzzer feedback module due to the limited number of pins on the Teensy++ board. Third, this board contains a whole module that won't even be necessary for the kit version of this project. So, I got rid of that, modified the feedback module, laid out the components a little different, and tried again:

This revision, v0.2, is better than v0.1 from a layout standpoint and supports all the right modules and connection, and it pretty much turned out exactly the way I hoped it would. It even has a cozy little space to hold the lithium polymer battery.

But two bad things happened with this board. First, although it worked wonderfully after I first plugged everything in and tested it for a few minutes, the board spontaneously decided to act up anytime I actually wanted to use it. Each component still worked fine on its own, but the moment I plugged them into the board, the Teensy++ would lock up and refuse to blink its pretty little orange LED at me.

Usually.

It would actually work briefly at random times. Intermittent problems are the worst. I still haven't pinpointed the issue, but I suspect the messiness of the circuit traces on the PCB itself may be introducing some capacitance in...actually, I'm not sure. I haven't learned enough to really know what I'm talking about when it comes to that in-depth level of electrical theory. It is honestly more likely that it's a very simple error that I wasn't fortunate enough to accidentally stumble upon.

However, that issue aside, I also discovered a usability problem: although the board was smaller than the solderless breadboard I'd been working with before, once I added those 90-degree header pins on the edges, it became even more unwieldy. Add to this the fact that the PCB is extremely rigid, more so than the solderless breadboard (which has a thin foam pad on the bottom), and my super awesome kit board suddenly felt painfully big and ugly, despite being more convenient from an assembly standpoint. This would not do. So, to attempt to fix the unidentified electrical issue and to drastically shrink the board, I tried again:

All right! Now we're talking! This v0.3 board is considerably smaller, which makes it easier and more comfortable to mount on the back of your hand without feeling the rigidness of the board. Also, it works. All the time. (Well, all the time so far; it hasn't stopped yet, and I've been doing a lot with it.)

The photo above, ladies and gentlemen, demonstrates the latest in cutting-edge Keyglove Kit technology. The sensors work, the wireless module works, the battery charging circuit works, and of course the microcontroller driving it all works as well. Some of the larger modules do hang off the edge a bit, but this causes no electrical or structural problem, and it allows the base PCB to be smaller. There's not much that could be done to make this any more compact without stacking vertically or cutting out a module entirely (both of which are possibilities, but for now I'm sticking with this design).

Not only is this probably something very much like what the eventual Keyglove Kit will use for a foundation, but it also simplifies the development process for me. Connecting sensors to the prototype module for testing is something I've had to do many, many times, and will have to do many more times during the course of development and debugging. With three dozen or so sensors, this gets to be tedious (or worse). With this board, that tedious-or-worse-ness goes away almost entirely.

However, there's still that issue of what to plug into those 37 sensor-ready right-angle headers, which brings us to the fact that the...

Glove Manufacturer Search Continues

The exciting prospect I had last time has not panned out--not for lack of trying or interest on their part, but the main manufacturing facility in China that they usually use didn't feel comfortable attempting the mixture of textile and interwoven electronics that is necessary for the Keyglove's physical component. The management stateside was as disappointed in this as I was, and they said I could contact them again if could provide a very clear design that would be easier for a manufacturer to follow, rather than a conceptual explanation, simple diagrams, and a prototype. Although I was really hoping to have this taken care of by now (or at least be well on the way to), this is a perfectly reasonable request on their part, really.

So, aside from the kit PCB, module testing, and embedded software development, I'm trying to start some conversations with reputable 3D artists who can help me put together something that will be suitable for a glove manufacturing company to use as a reference design. I've got one prospect on that front so far, but if you're reading this and know of someone (or are someone) you wouldn't hesitate to recommend, I'd be happy to hear from you. I've got a high-quality 3dsmax model of a hand to start with, but I don't have the software or knowledge to develop something better with it:

So, that's the story on that. In my mind, this is still by far the biggest difficulty to overcome, and I'm intent on tackling it as soon as possible.

One Last Thing: Your Enduring Support

I honestly had hoped to be considerably farther along on this project by now, having real Keygloves on your hands and a strong market beginning to develop. My eternal optimism and lack of prior experience in many areas made my estimates off-target, and for this I am sorry (both for myself as well as for all of you). It isn't from lack of money--I still have a good deal of Kickstarter funding left from your generous backing, and I've even been able to take care of many recent Keyglove-related investments with my own income in an effort to stretch the Kickstarter money out as far as I can, to better accommodate future purchases that I don't even know I'll need to make yet. Most of the difficulty comes from a lack of time, since this is not my primary source of income, and that has to take precedence. Some of the difficulty also comes from a lack of knowledge, but I see that as a very temporary hurdle since I love learning new things.

In less than a week--April 27th, 2012 to be exact--we will reach the one-year anniversary of the success of the Keyglove Kickstarter campaign. Wow. The truly incredible thing to me though is that in all this time, not one of you has sent me any nasty messages asking where the @#$! your rewards are, and telling me to hurry up. This is relieving and encouraging, but it is also humbling. I have a magnificent set of supporters, and though I've said so before, I want to tell you again that I am inexpressibly grateful for you all. You've given $12,474.80 and so far received very little in return (unless you really enjoy reading these lengthy updates). I promise it will not remain this way forever, and I'm doing all I can to finish the project and put products in (and on) your hands.

So thank you all, again, for your generosity and especially your patience. You will not be disappointed. This thing is going to be awesome when it's done.

Jeff

Comments

Only backers can post comments. Log In
    1. Julianna Yau Yorgan on April 24, 2012

      I've been reading all the project updates and just wanted to say keep up the great work!