Sign in:

Category: Uncategorized

Here at Modular Robotics we’ve been working hard on an exciting new product that enables people to wirelessly control and reprogram their Cubelets.  We’re getting ready to release that system to the world, but thought we’d give our friends in Boulder a sneak peek.  So!  We’re hosting a Hackathon to give you a chance to play with Cubelets and be the first to try out our new Bluetooth Cubelet.

If you’re in the area on October 4th, swing by our headquarters and join us for an evening of building robots, eating pizza, and maybe even winning a set of Cubelets!  The programming is done with text-based C code, so it’s probably not for the young ones.  12+ is about right.  We have limited space so please RSVP in the comments section below. When: October 4th, 2012, 6:00pm Where: Modular Robotics 3085 Bluff St. Boulder CO 80301 [As you head East on Bluff (from 30th), you’ll see our low brick building on your left.  Pull into the parking lot behind (and just after) the building and you’ll see us.] What to Bring: A Windows laptop with a Bluetooth adapter, and/or an Android device.
As Space Shuttle Endeavor is transported to its retirement, my Facebook and Twitter feeds are flooded with comments and photos of its final flight.  I can’t help but feel like this is, at some level, the end of an era. An era of big dreams and phenomenal achievements. An era of curiosity, exploration and discovery. It’s the end of an era of great things.

In my opinion, the Shuttle program is (symbolically, and perhaps physically) the single largest achievement humans have ever made.  A highly reliable, mostly reusable machine that can transport and sustain human life outside our protective atmosphere, bringing with it complex scientific equipment, and the spirits of a nation. What enabled its success?  A passion for science, the desire for discovery, and the collaboration of thousands of individuals.  Thousands of people working together in teams, collaborating to design and build each of the 1,000,000+ components that comprise one shuttle.  Each of those parts must to do it’s job AND work perfectly with each of the other 999,999 parts. This complexity and collaboration is the motivation behind Cubelets. Just one Cubelet doesn’t do anything useful, just as a single NASA design team couldn’t have built a Shuttle.  But when you put a few Cubelets together, they spring to life.  Working as a team, each of the members of the group accomplishes what none of them could accomplish individually, and amazing interactions result.

There’s one way to ensure that the end of one era is succeeded by great things in the next:  by educating and encouraging kids to do great things.  Children possess such amazing creative capabilities.  Every group of students we’ve given a set of Cubelets to has combined them in a new way to make a different robot that is completely unique from anything we’ve seen before.  It’s incredible! Dream it and build it – just like NASA.  Let’s inspire the next generation of dreamers!
We’re hiring for three positions right now: an Education Coordinator, a Supply Chain Manager, and a VP of Finance. I posted the three positions on Friday and then promptly left for a week vacation mountain biking in Idaho. I’m back now and starting to parse all 1200 resumes. There are some amazing-sounding candidates in the mix, so all of us at Modular Robotics are looking forward to starting interviews soon. Unfortunately, though, I’m tossing the bulk of the emails into the trash. I know that there are all sorts of evil robots out there throwing unfit resumes at every single job posting on the internet. But on the optimistic assumption that some of these people are actually real, qualified humans who are interested in working with us, I thought I’d share my quick 6-step pre-screen algorithm. If your email ended up in the trash and it shouldn’t have, please make a quick change or two to pass our little pre-screen and email us again. This isn’t meant to be a “how to apply for a job” post, I just thought making the pre-screen transparent would be a good idea. It’s simple:
  1. Is there a cover letter? A traditional cover letter isn’t necessary, but a couple of sentences about why you’re applying or why you think Modular Robotics is a good fit for you or why you like robot toys is necessary. It seems to me like the best place for this would be in the body of the email, not as an attachment with a cover letter for the cover letter in the email body. Just a resume without a cover letter? To the trash.
  2. Does the cover letter have a glaring typo? Or two? To the trash.
  3. Is the cover letter a generic copy/paste? Does it mention “your organization” and “your product” instead of “Modular Robotics” and “Cubelets”? To the trash.
  4. Is the cover letter just an explanation of how, although the person doesn’t meet any of the requirements for the job, they’ll be a great candidate and a perfect fit? To the trash.
  5. Does the cover letter direct us to do things without using the magic word? “Call me to discuss this opportunity.”, for example? To the trash.
  6. Does the resume have a glaring typo? To the trash.
OK! Back to the jobs@ inbox!
I’m happy to announce that we just scored a big fat venture capital investment from our friends at the Foundry Group. We’re pretty excited — this deal will give us the cash to do so much more than we’re able to do now. For the last two years, we’ve been building Modular Robotics from an academic project into a toy company. A project like this, of course, requires cash. Normally, companies like us get a VC investment, selling part of the company in exchange for cash. But if you’ve talked with me about this in the past, you’ll know that I’ve been an adamant bootstrapper. I’ve seen all sorts of companies get pressured by investors to release crappy products, and I didn’t want that to happen to us. We had the freedom to bootstrap because our patient customers were willing to pre-order kits that they didn’t receive for months, and because the National Science Foundation supported us with us with three SBIR grants. Grant funding and pre-sales allowed us to grow to 18 people, to build and ship almost 20,000 Cubelets, and to carry on adamantly pulling ourselves up by our bootstraps. We didn’t need any business guys in pleated pants telling us what to do! But a couple of months ago, I changed my my mind. Why? Spreadsheets. With all of these employees and all of these tiny robots, my job has changed quite a bit. I spend a lot more time in front of a spreadsheet now than I spend in front of the laser cutter. And when I started to make projections, to see where we might be in one, two, five years, there was no question that a cash infusion now would let us scale up to make a tremendously bigger impact in the future than we could by bootstrapping. Bootstrapping can work, and it’s worked swimmingly for us so far. I mean, here we are, 18 people in a cool lab with a ping-pong table and a production line sending thousands and thousands of tiny robots all around the world! But we can do more. Scaling up now, building out our factory in Colorado, and hiring a few great people will let us get Cubelets out to a far greater number of kids than if we stuck to our previous path. Another reason for taking on this investment has to do with speed. Since we’re just a bunch of kids in a robot lab, we’re building our business based on agility. We think we can get cool new hardware built, tested, and onto store shelves before the big toy companies have any idea what’s going on. But lately, we’ve been doing a lot of waiting. Waiting for parts because we can only afford to order small quantities at a time. And waiting for tooling changes because we’re not first priority to some of our vendors. We realized that there was a better way. We went straight to the best VC firm in the country. Seriously: they wrote the book. The Foundry Group has had amazing success with tech companies and they also happen to invest in all of our friends like Orbotix, Sifteo, and MakerBot. We called managing director Brad Feld and he immediately understood what we were trying to accomplish. And so here we are! I’m not sure that anyone reading this really cares about our financial strategy; I’m pretty sure that our audience consists mostly of kids, educators, technologists, alpha-geeks, engineers, and artists. Not people who read techcrunch. I thought I’d write about this here, though, because it means that we’ll be able to make Cubelets (and you should see the other robots coming down the pipeline!) significantly faster and better than we even dreamed about yesterday.
I just gave a talk about tiny robots at the TEDx Front Range event. Excitement! Bright lights! As I began, my slides didn’t appear. There was confusion, audience banter. The AV guy was giving me one of these. So I proceeded, waved my arms around a lot, and talked for ten minutes about how I think tiny robots are going to help our kids become smarter than we are. I won’t lie: I prepared to give this presentation with a lot of imagery, and when it all of a sudden didn’t appear, I got stressed. Spontaneous challenge! But everything was fine: tiny robots speak for themselves. For those of you who were there, I’m sorry the images didn’t pop up; the talk would have been better if they had. For those of you who weren’t, here are the slides!
Last weekend, five members of Team Modular Robotics competed in SparkFun Electronics’ soldering competition in Longmont, Colorado, USA. The event was held on the patio of Oskar Blues brewpub and the weather for the event couldn’t have been better. Game on. There were approximately 70 people participating in the adult competition (SparkFun also had a heat for the young soldering enthusiasts in the crowd). The competition was composed of three different rounds. In each round the competitors had to build one of SparkFun’s soldering kits as quickly and accurately as possible. There were only two metrics on which the competition was judged: quality and time. Your kit had to work and you had to be fast. The first round had three different heats of about 24 competitors each. The top times from these three heats advanced to the second round of sixteen. Each competitor was working on the Simon Says kit from SparkFun. There were some crazy fast times placing and soldering the 15 electronic components and the remaining mechanical assembly. The fastest time from Modular Robotics on this round belonged to Chris at just over seven minutes; this turned out to be the second fastest time of all competitors. Nick was close behind Chris with another jaw dropping time of 7:30.
Jon, in stripes, might have built your Cubelets.
We were happy and proud to have three of our team move on to the second round of the competition. The kit used for round two was the BigTime Watch kit. This round was significantly more challenging than the first round with most times being twice as long as those in the first round. The soldering level of this kit was similar but the mechanical assembly of the kit proved to be difficult for some of the competitors.
Emily, Chris, and Nick, clearly about to announce a trip to Disneyland.
The third round consisted of the top eight individuals from the BigTime Watch build including both Chris and myself. SparkFun had announced the first two kits prior to the event so many individuals came ready with notes or at least some idea of what they would be building. The third round remained a secret until the top eight sat down at the final table. The kit for the final round was a Mr. Roboto Kit. When you purchase this kit from SparkFun, you receive a pre-populated circuit board with surface mount components. The competitors did not receive such a luxury. Each competitor had to hand solder fifteen surface mount components. This was when the provided flux pen and solder wick really came in handy (at least for me, with multiple bridges on my atmega microcontroller). The first two finishers were lightning quick each finishing in less than 26 minutes. It was really a race for third place at the end. I was the third person to finish the kit but the display on my board didn’t work. It took me a few minutes (seemed like 40 but was actually more like four) of troubleshooting before my full board came to life in about 30 minutes of total work time which was good enough for me to place third.
The final seconds.
SparkFun put together an incredible event with many wonderful people. One of the best parts was chatting with other electronics and soldering enthusiasts between rounds. Some I knew, some I met for the first time but all of them came from interesting places and backgrounds. Although my account of the event may seem as though it was cutthroat right from the start, this was not the case. It was a very welcoming environment for participants from all backgrounds. The tone was friendly and the SparkFun employees were quick and eager to help everyone troubleshoot their boards.  We think that the public is finally ready for soldering competitions.  Olympics 2016, anyone?
Yesterday was pretty exciting at Modular Robotics. As you may know, we’ve soldered all of the Cubelets printed circuit boards to date using electric skillets. Thirty thousand tiny PCBs. This might shock an electrical engineer, but the method that Sparkfun published a few years ago works astonishingly well. Astonishingly well for a few thousand boards using a $20 piece of kitchen equipment, but it was time for us to upgrade. This enormous truck arrived first thing in the morning. We rented a forklift for the day to lift our new twelve foot long reflow oven off the truck and bring it into our shop. Matt drove. This beauty is an Electrovert Omniflo 5. We bought it used from a broker in Washington. It’s got 5 heating zones and is programmed with specific temperature curves to solder our Cubelets boards perfectly. Since the old skillet method worked so well, we weren’t really sure that we were going to need a real reflow oven, but it turned out that the boards that got burned or a little undercooked had a nefarious way of making their way further down the assembly process and requiring a lot of time to fix later. So we’re reflowing. 480 PCBs so far and they’re all beautiful! I never thought I’d be so excited about a piece of industrial equipment.
What better way is there to evaluate the reliability of a new Cubelet design iteration than by pitting it against its future comrades? It may seem strangely cynical, but on a practical level it saved an enormous amount of time designing and building a drive block torture test jig. And it’s modular to boot. Cute recursion aside, it’s important business to make sure we’re sending out quality products to our customers, and an all-star design team is no substitute for some good ‘ol torture testing. What you see here is a multi-purpose jig that will repeatedly stall the drive wheels or provide resistance as the drive block runs through a simulated play session velocity profile – for hours and hours. Or days. And all the cubelets in the jig are essentially stock, besides the gutted friction-wheel mounting cubelet and the plug-in power cubelet to run tests overnight. We easily reprogrammed a knob block to output a velocity profile based on the knob position, and I hacked together a cam that snaps onto a rotate block to raise and lower the friction wheel. After the first few hours of listening to cubelets torture, it’s time for another project: build a sound isolation chamber out of Cubelets!
Last week Modular Robotics took Cubelets to the Consumer Electronics Show (CES) in Las Vegas.  Crazy!  For four days our booth was mobbed by retailers, resellers, manufacturers, reporters, engineers, editors, starry-eyed startups and jaded serial entrepreneurs, videographers, bloggers, big-box buyers, business advisors, geeks, and venture capitalists.   It was exhausting and fun.  We were in the ‘Eureka Park’ section (thanks, National Science Foundation, for sponsoring us), surrounded by other new companies with emerging technologies, and we always had a crowd at our table playing with Cubelets.   We did a lot of video interviews, answered a lot of questions, and we got a lot of free advice.   And everyone seems to know a kid who needs Cubelets! Excelsior!