Nice summary of these interesting boards. I have been using an Odroid board for a while, and spotted a few small mistakes in your description. First, the Odroid pictured is the X model, while the model described is the X2. Unfortunately, the X2 sells for $135, not the $89 listed. The Odroid U2, the version I have been using, does sell for $89 and differs substantially from the X2 only in that it has two (instead of six) USB ports. I have gotten around this limitation without issue by using a small powered USB hub.
The Odroid U2 serves as my main Linux machine (Ubuntu 13.04, no overclocking, classic gnome desktop, 1920x1200 resolution using the board's HDMI port). While I do not set out to stress the system (and have not done any benchmarking), I have yet to run any application that showed any noticeable lag. Indeed, I recently installed the Accellera (OSCI) SystemC Simulator on this system for fun, and the performance is reasonable.
I also run Android (Jelly Bean 4.1.2) from an micro SD card, primarily to use XBMC, which runs quite well on this machine under Android (but not yet under Ubuntu)
The Odroid user community (accessible on the Hardkernel site), while not as extensive as that of the Raspberry Pi, is active and helpful.
I have previously owned a Raspberry Pi and a Hackberry A10, and found that both were too underpowered for everyday general use.
Turns out that Hardkernel is getting set to release in September a new board, the XU, which has a big.LITTLE architecture (Exynos5 Octa Cortex™-A15 1.6Ghz quad core and Cortex™-A7 quad core CPUs). Thus, while the U2 has more than enough power to serve as my everyday Linux system, I have my preorder in for an XU Board. One can never have enough speed!
Many of the boards shown in this article are Evaluation Boards (EVM) for the microprocessor manufacturers and as such they are made in small numbers, often at a great loss to the manufacturer in order to entice customers to buy their Micrcoprocessor chipsets.
The other components on the board, and 'their' manufacturer such as the Ethernet PHY, DRAM, Flash and/or Power management end up getting a 'free ride' for their silicon being on the board since many people copy EVMs precisely.
Of these boards the manufacturers place a 'cap' on the number of boards that can be purchased by a single company or individual to stop them using it in a product due to the losses made in their production.
The Rasberry Pi board is made by a non-profit charity, with private funding. Real Companies have to make a profit to stay in business, so this board is undermining the industry as a whole.
I don't really see the relevance of this article. If you want to use these boards in a product you hope to make for a long period of time, then be warned, you'll have endless issues in a couple/few years time when these boards no longer exist.
Yeah, that is a really good point. Rick Merritt keeps asking me how prototyping boards fit into the industry, and I really think the ultimate answer is "they don't". They are their own little industry.
However, if you're using one of these to build something for yourself, you probably dont' really have to worry about future supply.
These proto boards don't fit into a full production product but they are a great way to get a quick demo/proof of concept for a design team. Afterwards, however you are faced with migration of the design into something that you can sell at a profit.
The current rage in laptops is super-slim, bordering on supermodel anorexic. Heck, some even refuse to include standard Ethernet connectors because they make the case too "fat"... you have to use a special adapter cable. meanwhile, your dev board (cubieboard, Pi, beaglebone) have adapter boards plugged in on topdoing exotic things.
I have a basic calculator, using one of these dev boards. To make it practical I'd have convert it to clamshell, throw in a battery, design a case and translate most of the "guts" in a 4 (6?) layer board to a new home. That's a big ask for a student/hobbyist, I reckon my amateur layout will stay put.
When I started university in the mid-1970s we worked on both analog and digital computers. The main digital computer wa sin its own building -- we created out programs on punched cards and/or paper tapes. My mind boggles when I think that each of the SBCs discussed in this column could probably out-compute that room-sized beast...
@Susan: I bet the Apollo program could have used a few of these boards.
I'm sure you will be amazed to hear that I have a story about this. Konrad Zuse was an engineer in Germany who had a fully mechanical binary floating-point computer called the Z1 working in his parent's front room in Germany by 1938 ... way ahead of anyone else in the world. During the war he went on to create relay-based and vacuum tube-based computers.
Meanwhile, Wernher Magnus Maximilian, Freiherr von Braun was a German rocket scientist, aerospace engineer, space architect, and one of the leading figures in the development of rocket technology in Nazi Germany during World War II and, subsequently, in the United States. After the war, von Bron moved to Huntsville, Alabama (where I now hang my hat), where he lead the development of the Mercury, Gemini, and Apollo missions.
I met Konrad Zuse's son Horst, who told me about a chance meeting between his father and von Bron when both were being evacuated from Berlin -- Konrad with his computers and von Bron with his rockets -- in an attempt to sav etheir work from the allied bombing missions. I think they met in an inn just for one night.
I can imagine the conversation "What do you do?" and Konrad describing his large heavy computers and von Bron describing his rockets. The one thing von Bron was desperate for was a light, accurate control system. Neither of them would have envisaged putting a computer (big, heavy) on a rocket...
...these little single-board computers will motivate more young people to learn programming like the Commodore C64 has motivated me in the 1980s. Yes, it is not only possible to play games with computers. ;-)
Rick, I think they really help the SOM market. For example , the beagle bone helps create a great dev ecosystem , and then these guys start offering a SOM with a processor from the same family(less capable), at an attractive price point($27/1K)for 275MHZ,128MB flash/ram, etc.
SOM = System-on-Module also known as Computer-on-Module (COM). A SOM is a very small card, usually the size of a DIMM and often using a DIMM connector. A SOM makes it easy to incorporate a computer + DRAM + Flash, all with high-density interconnects, into a custom base-board which provides external I/O interfaces and can get by with low-density interconnect.
I think this is very much the equivalent of what the mini-ITX was some years ago for embedded PC based DYI type projects, or PC104 in th 90's. Some of them will go into volume, but very often it's used only for early stage proof-of-concept development work, which is great. For real commercial volume projects though, I think most of them end up with either a Computer-on-Module and a custom carrier, or a making a customdesign with the SOC on a single board. Having said that, if you can solve all of your application requirements with only one off-the-shelf board that is manufactured in volume, that will always be the most cost-efficient alternative, the same way you have a embedded applications where an embedded PC motherboard or SBC has all the functionality onboard that's needed.
Hi Rick, the open source nature of BeagleBone Black is actually what makes it great for commercial applications. The Sitara AM335x processor that powers the board is available for purchase from low quantities to high quantities, so we see hobbyists using the board for DIY projects as well as developers taking their Beagle-based projects to Kickstarter. We're excited about how open source boards and software eases the development process for both groups of developers. Check out the projects page on BeagleBoard.org to check out what people are developing: http://beagleboard.org/.
– Alejandro Erives, Texas Instruments, member of BeagleBoard.org
Hey, nice piece Cabe! Some great tech for next to nothing. Speaking of which, I have an Artigo 1000 kit (http://www.via.com.tw/en/products/embedded/artigo/a1000/) that I got from DesignWest (more on DesignWest in the coming days, standby). I've been sitting on it for a few years. I went to set it up as a simply webserver the other day only to find it's way outdated and no longer supported by Via Technologies. I could mess with it, but just don't have the time for some odd reason.
Anyone want it? It's a great little package and I can't just throw it out.... If you want it, and are based in the US, I can mail it to you: no charge (how's that for a deal?:). Just ping me at firstname.lastname@example.org.
All I ask is that you share with us what you do with it and how you did it. Deal? Just some photos and a description.
And the winner is.. Daniel Winder from Decagon Devices who'll be re-engineering his garden monitoring and control systems. Looking forward to seeing those photos, and maybe a fresh tomato or two in the mail. Good luck, Daniel.
I really like a number of these boards in terms of their features and price, but so far I cannot justify their use in released products in my Industry (Industrial Control). I say this, as all of these and other generic controllers lack suitable industrial hardened power supplies, or transient protection/fusing, etc that are needed for long term reliability... the result is that additional boards, etc would be needed and then one has to determine if it would be cheaper to design a custom single board solution, as opposed to having multiple assemblies (the off the shelf board and one or more custom boards) and associated cabling. Secondly, how does one mitigate long term support risk? A number of the boards use devices that are not widely available and/or are consumer oriented and so design changes, etc are a real possiblity... the result is that support costs increase in order to support the different flavors of the boards. Indeed, some of the boards have released design files so that one "could" remanufacture their own board, but this is also not a simple undertaking, as some parts may be specialized or have high minimum order quantities. In my view I see these boards as excellent learning platforms, proof of concept vehicles, or even as reference designs that firms can use... but certainly nothing that I could use for my production needs.
@Garcia: I believe that the real significative point here is that all this Biggest-Little revolution is driven by ARM processors!!
How is that going -- I heard a lot about the Big-Little concept when it came out (and I thought to myself "that's an interesting idea") -- but more recently I sort of recall hearing that it wasn't being adopted as widely as had originally been hoped ... so is it soaring, limping along, or crashing and burning?
Actually there are few more companies licensed ARM big.little other than Samsung and Mediatek. For example, The Chinese design house AllWinner released it's quad-core MPU A31 for tablet half a year ago, followed by an version for phablet two months ago, called A31s. I think they are the first one licensed ARM Cortex A7, which is the latest from ARM. Samsung and Metiatec haven't licensed Cortex A7 (they licensed A9, which is the old version). Now you can get a complete A31s based quad-core tablet with HD IPS screen with builtin 3G for about $150 (for example. the uPlay Q7s). It can also make phone calls. Thare are hugh applications for a device like this waiting for people to expore.
I don't think those single board computers listed here are so significant, since those are just board, they still need display and input device. In now days you can actually get a complete tablet computer running Android, soon to be able to run Windows 8 as well, for just $69 (like uPlay Tablet C70).
I would put put the Lillypad and other Arduino compatible systems in a different category. They don't have nearly the capability of something like a Beaglebone. However, the Arduino has been perhaps even more important to date. I think it's done more to expand the accessibility of MCU-based devices than anything else beyond the MCU itself.
The original Beagleboard has been just as valuable in the inexpensive full SPC arena. From what I've seen that board really started the movement that is written about in this article. There were certainly small SBCs before the Beagleboard, but making it open source (as is the Arduino) was brilliant marketing and set the Beagleboard far above anything before it.
@Duanebenson Thanks Duane, agree about the Beaglebone. And the folks at Atmel would be pleased to hear your remarks about the Arduino, as they wondered why it was not included. Some say, though, that it isn't an SBC, although one of the newest Arduino-based boards does run a version of Linux.
I completely agree with Duane. The Arduino is fantastic, but doesn't belong in this list. These are full computers you can plug a display, keyboard, and storage into and begin developing. They are a complete Single Board Computer.
The Arduino is a fantastic prototyping tool that allows many people to get amazing projects done, but is not a complete single board computer. You would have to add "shields" to get video out, keyboard in, physical storage, etc.
I know the RasPi has a video input connector, although the software only supports the Foundation-sourced camera accessory.
Do any of the others have video in?
Also, saying an ARM quad-core is four times as powerful as a RasPi is like saying it's 10 times as powerful as my desktop workstation... maybe true if you're comparing against the fastest ARM processor in any of the peripherals (I think there's one in the SSD), but clearly missing the point. The RasPi ARM core represents only a tiny fraction of the compute power in the Broadcom VideoCore chip.
The BeagleBoneBlack I've also evaluated, and saying it has "quite robust OS support" is misleading. Several important peripherals, including the SGX PowerVR GPU, still have no drivers for any of the available OS distributions.
rbv - The Raspberry, Beaglebone and Beaglebone both support USB video input. I've brought in 1080p resolution cameras on both. The Raspberry can get a little laggy at that resolution, but it does work reasonably well.
This is a very informative slide show with collection of all the different OSBs(Open Source Boards). A detailed comparison wiki is still under evolving state but this really gives very good comparison of the possibilities today. The different boards are providing a good combination of interfaces, may a collective usage and experience with the hardware will enable the user to mix the interfaces and come up with his own design. But overall this slide show gave a good reference.
Thanks for taking the time to discuss this, I feel strongly about it and love learning more on this topic. CAD Drawing Services If possible, as you gain expertise, would you mind updating your blog with more information? It is extremely helpful for me.
What are the engineering and design challenges in creating successful IoT devices? These devices are usually small, resource-constrained electronics designed to sense, collect, send, and/or interpret data. Some of the devices need to be smart enough to act upon data in real time, 24/7. Are the design challenges the same as with embedded systems, but with a little developer- and IT-skills added in? What do engineers need to know? Rick Merritt talks with two experts about the tools and best options for designing IoT devices in 2016. Specifically the guests will discuss sensors, security, and lessons from IoT deployments.