Tag Archives: Integrating USB

1:1 Interview with Erik Kettenburg of Digistump (Part 2)

(Part 2 – continued)

digispark-atiny85-avr-atmelTV: Why AVR as the microcontroller? What did you especially like in the development, prototype, production use-cases for AVR 8-Bit Microcontrollers?

EK: The best thing about AVRs is that they have a huge community and it makes sense why – as for microcontroller use, they are pretty easy to use, thanks to all of the tools for them – of course Arduino is a big part of that – but even before that, probably the key component was avr-gcc – being able to write C/C++ code and easily compile it for AVRs with a completely free unrestricted compiler is a huge incentive to develop with it, especially for makers who don’t have the big bucks or the desire to use proprietary compilers. PIC took a long time to catch on to that, and still isn’t nearly as easy to get started with, and the community is not as open and beginner friendly either, in my opinion – and I started with PICs back in the late 90s.

The other great thing about AVRs is that they are pretty robust and there is nothing quite like a device that can source up to 40ma on a pin and survive – that makes it much more beginner friendly!

TV:  Are there any particular microcontrollers or microprocessors you would like to get your hands on more easily in design and prototyping?

EK: There are lots of them – lots of the new silicon MCUs being developed in China would be fun to play with and see what they can do – like the Allwinner ARM SoCs but we have lots of trouble getting them – we find ways though.  Lots of MCUs have a ton of potential for use in startups and IoT but we don’t even think about using them because of all the non-competes, minimum orders, binary blob drivers, etc – it doesn’t make sense to develop open source software with a very closed company.

TV: How do ARM and AVR microcontrollers differ to you, why do you design your first project in AVR?

hello-digispark-avr-tiny-avr-8-bit-attinyEK: Our first project was using an AVR 8-bit Microcontroller because that is what we knew best and that is what our target audience was used to – because of Arduino.

TV: Why did you choose Atmel ARM for DigiX? Development, IDE options, Design, Production, Quality, Supply or any other reasons?

EK: We chose the ARM chip for the DigiX because we do see it as the future of IoT products – the AVR is being pushed to its limits when you stack Wifi, LCD touch solutions, sensors, and more all on top of it – the ARM still has plenty of room to handle more – so it is a natural progression for power users to move towards more powerful hardware.

TV:  Our goal was then to make it as easy to use as possible – we chose the same chip as the Arduino Due because we wanted to add to their support for that chip in Arduino rather than compete with it – we wanted to help make both the DigiX and Due easier to use.

From a design, supply, and production standpoint it has been far more difficult, expensive, and harder to work with – but it is worth it to enable people to access that kind of power in an easy to use format.

TV:  Is it more about low power and smarter hardware design or beyond the core development with Restful cloud based architectures to empower the Smart Connected Thing? What do you find more important, if you had to choose?

EK: Low power and ease of use – many of the best ideas come from people who are in no way experts at the hardware or software involved – the more people are able to use the tech, the more likely a breakthrough will come about. Low power is also important because the more we can move to small and battery powered, the more we’ll see these ideas break free of the home and desk and move into everyday on person use.

TV:  Do you see the Maker Movement and Enterprise coming together in matrimony someday? What are some of the challenges?

EK: I don’t see them at odds – as I mentioned before – I think the enterprises who embrace the maker culture will be met with more success and the rest will be left behind, at least by those embracing the maker culture.

TV: For a designer or developer, are we at the hour of product creation or the hour of connectivity?

EK: Both – we need people doing both to fuel this cultural change!

TV:  If you had to build the perfect smart city integrated with personal networks, take me through your choice of network topologies and protocols most considerably ideal to optimize the connectivity for people and customers. What would your design look like in terms of ideal integrated connectivity (802.11.x, 802.15.14, ZigBee, 6LoWPAN, IPv6, CoAp, MQTT, etc)?

EK: Whether or not it is ideal – I think that it would be met with the most success if it is 802.11.x based for long range, and 802.15.4 for personal area and shorter range links – the short range 802.15.4 protocol would be ideal for devices communicating with other devices around them and wifi is already so ubiquitous that it makes sense to use it for longer distance and high bandwidth connections. Its range and prevalence also allows it to be very successful without worrying nearly as much about infrastructure.

I don’t see IoT world being built on closed protocols – industry is trying to do that, but they are ending up with expensive devices that people actively resist while we openly welcome more open devices like WiFi into our lives. WiFi is extremely prevalent and the (finally) emerging widespread us of IPv6 opens up many new doors. Technologies like 6LoWPAN, or simply put IP over low power radios has the potential to increase standards between low power 802.15.4 devices. Additionally, by essentially moving the internet into the realm of low power wireless devices the transition between personal area networks and WiFi/standard IP networks becomes far more seamless, and therefore much more user friendly. These technologies will reduce complexity, barrier to entry, and part count – which should allow for IoT to spread freely and at a much lower cost. These technologies are still working there way into the Maker culture, and the more open they are the more quickly they will be adopted.

TV: How can Internet of Things and Internet of Everything come to more realization when it comes to design choices and connectivity designs? What design aspects would like to share with the community or similar-minded engineers, makers, hackers, and designers of tomorrow?

EK: As creators we all need to focus on removing the barriers to other makers and consumers using the technology – this means price and ease of use. Maker culture remains rather elitist, mostly due to price – a simple, elegant, and affordable design will take us farther towards mainstream culture then all the features that can be possibly packed into it. digispark-avr-8-bit-mcu

We also need to keep those designs accessible, and our communities friendly – to all, all people, all types, all skills, nationalities, sexes, etc – the open source hardware community has a long way to go to shed the elitist mens club image – it is getting there, but we can always work to be more welcoming.

TV: How can designs that you bring to the community help change the product landscape? Any last thoughts you would love to share with the Embedded Design and Makers community?

EK: I hope the products we bring to the community change the landscape – and I think in general products made by makers can do that by accessibility (a common theme with me) and doing new things with old tech, or old things with new tech – basically looking at things a different way. Since the Digispark came out we’ve seen many clones/derivatives taking advantage of using V-USB and the ATtiny – which has been awesome! We like to see when makers share their methods and inspire new applications through taking a different approach.

Last thoughts: Keep making, I am firm believer that the more of us make, and the more things we make, the better idea will come around – so lets all make our ideas reality, regardless of whether or not other people think it is possible! My other thought would be, open up your process – the more open we’ve become as a company the better we’ve done – I love the popular quote that goes something like “Don’t be afraid someone will steal your idea, you’re not that important, and if it is that good of an idea you’ll make money off it anyway” – we’ve found that to be very true and will keep pursuing that strategy.

(This concludes the interview, part one can be found here).

1:1 Interview with Erik Kettenburg of Digistump (Part 1)


 1:1 Interview with Erik Kettenburg co-founder of Digistump (Part 1)

Tom Vu: Why did you create Digistump? Did you see a strong need to do this primarily for the movement of connectivity hardware and embedded software going too slowly or too conventional? Why do this yourself instead of use what is available?

Erik Kettenburg: Digistump was created from the success of our first project launch on Kickstarter. We launched the Digispark hoping to sell 500 units on Kickstarter and ended up selling 25,000 – in the course of that campaign I started to design several add-on boards (shields) for the Digispark and when it was all complete we had orders nearly 100,000 units total across all products. Turning that momentum into a company was a natural choice and in doing so we wanted to stay true to our original goal of making open source hardware cheaper and more accessible.

The creation of the Digispark was fueled by a feeling that existing open source development products were too expensive to leave in all the little projects that I had built. The Digispark was created to be as minimal as possible while still being very user friendly – thus enabling it to also be much cheaper – which is something we didn’t see in the open source market.

TV: Walk the reader through your two successful Kickstarter projects. What do they do and who should use these products?


EK: Our first Kickstarter, as mentioned above was for the Digispark, and ended up raising just over $330,000. The Digispark is an Atmel AVR 8-bit Microcontroller ATtiny85 based development board with on-board USB. The ATtiny85 doesn’t have hardware USB and the AVRs that do or an external solution (as is used on the Arduino) would have made it too expensive, so we built off of the V-USB project which is essentially a hack that does USB communications from regular i/o pins. The Digispark has 6 i/o pins (we disabled the reset pin to get an extra), a 500ma voltage regulator for using with external power, it can be powered over USB and has an integrated USB connector, a bootloader that allows it to be programmed over USB with our slightly modified version of the Arduino IDE (or by command line with any hex file), and two LEDs – one for power, and one connected to Pin 1 for use as a status indicator. It is about the size of a quarter. The Digispark now retails for $9.95 and has over a dozen shields kits available for it that allow you to quickly add everything from a motor drive or real time clock to a RGB LED or protoboard. It makes both a great first board to get started for a low price and also a great board to leave in projects, or use in projects where cost and size are a concern.

Our second Kickstarter was for the DigiX and just ended a few months ago. The DigiX aims to be the opposite of the Digispark – it is an Atmel ARM SAM3X8E based development board that aims to have everything you could ever need. It is also compatible with the Arduino IDE. The main feature is that it has built in, easy to use, client and server capable low power WiFi, allowing it to be an IoT device without purchasing anything else! It also has a mesh networking module, microSD reader, EEPROM, status LEDs, switching voltage regulator, audio output, and 99 i/o pins. It is compatible with all Arduino Due compatible shields and we also sell a level converter shield along side of it that makes it electrically compatible with all classic Arduino shields. The DigiX is the ultimate Arduino compatible board for the ultimate power user – or anyone looking to get a WiFi enabled Arduino platform for an affordable price. Currently it is on presale (expected to ship late October) for $59, with the level shifter also on presale for $15. lessThan2cm_digispark-avr-atmel-tinyavr

TV: What challenges do you see in the maturation of a embedded design project especially when dealing with operations, prototyping, production, licensing, distributors, supply, etc? Any tips to share?

EK: We have had many challenges with both of our products. The Digispark was a very simple design but used standard parts in non-standard ways. This presented some challenges when it was handed over to a factory for production. They were not able to program the Atmel 8-bit AVR ATtiny85s successfully because they had never before programmed them in that way, it took several weeks to get it worked out and eventually we sent them the programmer we were using and the exact commands and they were able to duplicate our setup and then transfer that knowledge to their automated programming machines.

digispark-avr-attiny-atmelThe other major obstacle with the Digispark was that we had to scale very quickly, as I mentioned before we expected to, at best, sell 500 units – we sold 100,000 across all products and 25,000 Digisparks alone – we had to scale everything, production, packaging, kitting, shipping, bank accounts, etc. It was good problem to have but added significant delays. The result of that scaling, though, was that we now have a full company set up and running, and the Digispark continues to be very successful, as well as our other products we have introduced since.

From a distribution standpoint we were surprised to find that the distributors found us, and so that hasn’t been too hard. Micro Center stores, MCM, and RobotShop, to name a few, carry our products and just today we have shipped our first order to MakerShed for them to carry them as well. We’d love to see even more distributors carry them, but I would say the biggest problem we’ve had is finding the time – my wife (the other Founder, and my partner in the business) Jenni works on Digistump whenever possible as her main job (in addition to taking care of our 2 month old daughter Maple, and our 18 animals on our small farm) – I still work a day job (as CTO for Portland based Vacasa) and split the rest of my time between Digistump and family as well. We hope that someday I’ll be able to work on Digistump full time.

With the DigiX – which is still in production – our issues have been more with supply – for instance we needed 2,000 SAM3X8E chips right away to start production, that meant we had to buy from about 6 different suppliers to get enough of them. The WiFi modules had to be made to order, and generally having such a complex and cutting edge board has meant the parts can be harder to find.

Some of the challenges we’ve experienced have made us plan ahead more and other have made us give serious thought to moving production in house. Our next product the Digispark Pro (to hear when it comes out join our mailing list at digistump.com) will be made entirely in house on a pick and place machine we recently installed.

TV: Is DigiStump sort of like the composite utility belt or multi-purpose digital rivet for USB based hardware design/development?

EK: I hope that Digistump will grow to be a sort of smaller version of Sparkfun or Adafruit – focused on in-house products, hopefully built in house as well. Where Digistump differs already and will continue to is that price will remain a major focus. We want to keep things accessible to everyone, folks in other countries, students, kids, retirees, etc. We believe that anyone should be able to get into the electronics hobby and we’ve already seen this approach pay off with many schools using Digisparks to not just teach children, but allowing them to take them home and continue to use them.

Digistump is also expanding beyond hardware – in an effort to enable makers we will also be adding a low cost prototype PCB service and laser cutting service.

TV: Give an outlook of what there is to come – hypothetical and speculative products, empowerment of individual makers, enterprise designers, and tinkerers? Convergence?

EK: The momentum and mainstream acceptance of the maker movement is something I never could have dreamed would happen. When I was a kid and very much into computers and electronics in the 90s, it was not cool or encouraged at all, now we see children being encouraged by schools, parents, and more importantly each other to invent and innovate. I only hope the I (and Digistump) and continue to find ways to enable their creations – I think that should be the number one goal of the entire open source hardware community – to enable the next generation of makers. digispark-avr-atiny85-8-bit

With this mainstream acceptance we are also seeing the culture move into enterprise situations, which some may think is the death of the movement, but I think is an excellent step – when a enterprise tries to be “disruptive” it is far better than it simply writing off the movement and continuing to make marginal improvements – we should welcome all disruption and all innovating, regardless of the source.

TV: What is the difference between products built from Startup DNA vs Enterprise DNA? Are there recognizable traits or differentiators? Is there a special need for “Made in Makers” seal to advocate authenticity, consideration, and brand?

EK: I think Enterprise DNA is moving towards the maker culture, but 99% of companies are still stuck in the marketing and/or marginal improvements approach. A very recent example of this would be Intel – they just released a new processor aimed at IoT but they are pushing it towards existing industrial customers and making reference boards available through select channels – if Intel really wants to compete with ARM producers like Atmel on the IoT platform they need to get those in the hands of makers and tinkerers, not just career design engineers. A company like Intel needs to go to a company like Arduino, Sparkfun, or even Digistump and say “We want you to make a cheap dev board for this chip, we’ll even help you market it” that would be disruptive (mostly to their internal structure) and far more effective for being part of the future of IoT then millions of dollars in marketing. I guess what I see is that there is still a wall that makes no sense – Makers can’t get access to many of the newest parts and Enterprise can’t get anyone to make something truly innovative with them, so they aren’t capturing the attention of our increasingly maker obsessed culture, and equally important, the new companies that these makers spawn aren’t using their tech.

Because of this divide I don’t think we need a “Made by Makers” seal – it is obvious what is made by Makers – and if a enterprise makes something disruptive or innovative enough to seem like it fits in maker culture then they deserve to be called makers as much as I do.

TV: What’s your definition of disruption? Who successful/innovative in industry do you look to as a mentor? Do you draw forth similar way of thinking?

EK: To me disruptive is overused – a technology or company can be disruptive when they launch, but years later when millions are using them they are no longer disruptive, the next thing is, and it only is if it challenges the current norms. Stripe was disruptive to the merchant payment industry, now even Paypal follows their lead (and actually Paypal was very disruptive when it started). Disruptive seems to often, today, be self applied when it doesn’t fit – I look at companies and individuals who change markets or even better cultures as disruptive – Arduino, RepRapMakerbot, Adapteva and the Parallela, Laen as and OSHPark to name a few. I wouldn’t dare call Digistump disruptive, we built on an existing community (Arduino) and technology (Arduino, V-USB, and AVR). I hope someday to come up with an idea that is disruptive.  neopixel-digispark-avr-tinyavr

As far as those I look to in this industry for inspiration – Sparkfun and Nathan Seidle, Arduino and Mossimo Banzi, Ubuntu and Mark Shuttleworth

TV: What are some of the challenges of the day for Makers and Startups (Kickstarter, Indiegogo)?

EK: Since the Digispark campaign we’ve seen the crowdsourcing arena get very competitive – and dare I say more corporate. While there is still a fair amount of organic success on Kickstarter and the like – there is also a strong segment of companies, startups but well funded ones, with big production budgets, slick movies, etc – and those are getting harder and harder to compete with. Additionally we’re seeing established companies (even some who call themselves open source) latching on to new ideas and using their considerable resources to put a competing product in the market by the time the new project ships. This sort of competition is not just a challenge but is hurting innovation that the community has worked so hard to foster. That isn’t to say all companies are doing that – SparkFun, SeeedStudio, Atmel, lots of companies have been good friends to the small startups – it is, after all, very much in their benefit to do so. Atmel chips run the majority of new open source projects because Atmel gets it – Atmel gets that they need to support Makers, starting with Arduino and now expanding to many other startups like Digistump. Our Atmel reps (Mike as Cascade Tech) treat us like we are big customers even if we only buy a fraction of the chips the big guys buy. Other companies’ reps have treated us very badly, and made it nearly impossible to get advice, samples, or help because we aren’t huge – I think companies like that will find themselves left behind in the maker revolutiondigispark-rgb-shield-tinyavr-avr-atmel

TV: What do you plan on doing with DigiX and Digispark, its quite noticeably public especially in its origins and funding from Kickstarter. What is the eventual roadmap for Digistump LLC?

EK: As I mentioned previously, we’ll continue to build development boards, and other electronics to enable makers – while keeping these devices as accessible and affordable to everyone as possible. That doesn’t stop with our products either, where most companies want $30-$100 to ship internationally we are shipping some products for as little as $3 – and most for under $10 – we really want to make sure that everyone can get into electronics – so much so that we’ve yet to take any profit – we just keep investing everything back into products, efficiency, and keeping the prices low. That is part of the reason I still need a day job!

As I also mentioned, we are moving into service – including PCB prototyping (with a US fab partner) and laser cutting. We’ve been slow to roll these out because we are automating them almost entirely so that we are able to offer then at a much lower price than the existing services.

What else is in store? Well Digistump also does lesser-known contract work, everything from manufacturing to design, development, testing, consulting, and prototyping. I’m not an EE – I have a degree in Economics actually – but I have a lot of business experience in both the web development world and now the hardware world – I hope to find more contract work as it allows me to improve my skill set for Digistump while also helping companies look at technology in a less corporate way – of course, I can’t talk much about what I’m working on, but I’ve been able to show some companies how to develop awesome products for a fraction of what they had been paying traditional product development teams, and give some tips on marketing, tech infrastructure, etc on the way. That has been a lot of fun, and I think it will be awesome if we can keep that a part of Digistump as well as we grow.

TV: What needs and requirements were set forth to have custom firmware and bootloader in the DigiSpark?

EK: The bootloader for the Digispark does a lot. The ATtiny85 does not have a reserved bootloader section in its flash – which means it really isn’t setup for having a bootloader.

TinySafeBoot (http://jtxp.org/tech/tinysafeboot_en.htm) came up with serial solution some time ago and Embedded Creations did it with USB support – but those needed to be further refined into something smaller than Embedded Creation’s bootloader that was also more reliable, and ready for production use. After the failures of some of our experimental work on it, Jenna Fox, a very talented individual stepped forward and offered to help and started the Micronucleus open source bootloader  which Digistump sponsored, and supports in every way possible. Digisparks run that bootloader and it provides protection of the bootloader and upload over usb, all using the internal oscillator (with automatic calibration). To date it is the smallest, robust USB bootloader for ATtiny chips – and certainly the best tested ATtiny USB bootloader as well. It also remains truly open source.

Interested in reading more? Tune into Part 2 of Atmel’s 1:1 Interview with Erik Kettenburg co-founder of Digistump


News from the Gallery

News from the Gallery

By Joerg Bertholdt, Director of Marketing, Tools and Software, Atmel Corporation

We launched Atmel Gallery less than 6 months ago, the first app store of its kind to deliver integrated embedded tools and embedded software straight into a development environment, specifically for Atmel Studio 6.

We were excited to get into uncharted territory, but also anxious about the types of reaction we’d get from our customers. It’s time to see how we did.

With over 25,000 developers – and counting — signed up for a Gallery account, our expectations have by far been exceeded. The servers handled the big demands very well, proving the infrastructure design and scalability.

 Keil MDK-ARM Toolchain from Keil enables Atmel Studio to use its highly optimizing ARM compiler

Keil MDK-ARM Toolchain from Keil enables Atmel Studio to use its highly optimizing ARM compiler

With the February update of the XDK, the Extension Developer’s Kit, partners have been able to easily integrate embedded software and package integrations as projects that install directly into Atmel Studio. For example, developers who are interested in exploring commercial real-time operating systems, now have access to Micrium’s uC/OS and Segger’s embOS. Trial versions of these RTOSes are available as ready-to-run example projects for Atmel’s ARM Cortex-M4 based SAM4S devices.

Two popular extensions are compiler plug-ins. The Keil MDK-ARM Toolchain from Keil enables Atmel Studio to use its highly optimizing ARM compilerCodeVisionAVR allows the use of HP Infotec’s AVR compiler, which also includes CodeWizardAVR, a graphical peripheral configuration tool for AVR MCUs.

Besides development tools and embedded software extensions, training modules such as Integrating USB In Your Design have also been made available through the Gallery. With the XDK supporting the development of hardware extension boards for Atmel’s Xplained Pro kits, the drivers for the first Xplained Pro partner board will soon be delivered as an Atmel Studio project.

We’re six months in and happy to see the masses have adopted Atmel Gallery.

Are you a Gallery user? Let us know what you like, what you’d like to see improved or if there are additional extensions you would want to see. Just comment on this blog or send me an email to joerg.bertholdt@atmel.com.

Haven’t checked out the Gallery yet, just take a look at gallery.atmel.com. Don’t have Atmel Studio 6? It’s free, you can download it from atmel.com.

Are you an independent software vendor or developer and want to be part of the growing Atmel Studio ecosystem, join the Gallery as a developer; the XDK makes it easy to participate.