John Melton holding a HPSDR Mercury Digital Down-conversion board at OSHUG #5
Last month the Open Source Hardware User Group were treated to a presentation on the High Performance Software Defined Radio (HPSDR) platform. For those new to SDR the introductory text for the event follows.
"Radio spectrum is a finite resource and it should therefore come as no surprise that radio systems are a particularly hot area of research. Whilst ever more advanced schemes for modulation, digital encoding and spectrum access promise increased efficiency, step upgrades more often than not require new hardware. As has been evidenced in the evolution of mobile telephony from analogue to GSM and 2.5G (GPRS) to 3G, and similarly in the evolution of wireless computer networks. A disruptive development in radio technology promises to change this and to bring an unprecedented flexibility to radio systems, and one similar to that which programming brought to the task of machine computation. Despite, or perhaps due to being at the cutting edge there are a number of open source hardware projects concerned with developing software-defined radio (SDR) technology. As with the earliest developments in radio systems radio amateurs are once again at the forefront, and at this month's meeting we will have a presentation on the comprehensive HPSDR platform."
Further details along with a few photographs can be found on the OSHUG website and we hope to soon add a video of the talk and the slides.
HPSDR is a hugely impressive project and serves as a great example of the potential for open source hardware; not only is it concerned with cutting edge radio systems technology but it's a well thought out, highly modular system that draws upon a multitude of skills including multilayer PCB layout, logic design for FPGAs and software development. The best thing is that this is the result of open collaboration and all the design files are made available under an open source licence, and you are free to build your own or a derivative, else simply to reuse elements of the design.
HPSDR is just one example of an open source hardware/software project concerned with software-defined radio and others of note include GNU Radio, USRP and DttSP. In addition to which there is a growing list of proprietary technologies, such as those from Vanu Inc and GE Intelligent Platforms to name but two.
Make no mistake, SDR is not just about niche applications or even maximising the return on investment in infrastructure and extending the useful life of user devices - which is not to say that the potential gains here will not be huge. Moreover, it will also be an enabler for new services and has the potential to drive far more efficient use of spectrum. An example of the former might be the consolidation of fixed radio infrastructure, whereas Cisco's CleanAir technology hints at the potential for unprecedented levels of spectrum efficiency brought about via cognitive radio (something that is much easier realised when signals are being processed in the digital domain).
SDR really is a game-changing technology and it was great to see that non-radio amateurs and indeed people with no background in communications really got this at the OSHUG meeting. It's an area that I've been keeping an eye on for around five years or more, and the pace of innovation continues to increase and with a proliferation of applications in recent years. This trend can only be set to continue as we witness the arrival of low power multi-core computing, software-defined silicon and ever more flexible reconfigurable logic, in addition to advances in digital converter and RF IC technology. My guess is that it can't be that long before we start to see SDR being put to use in more mainstream applications. Definitely a case of watch this space!
-- Andrew Back
Next OSHUG meeting: #6, Thursday 18th November.
Martin Keenan
Interesting stuff. What has previously made me sceptical about the adoption of SDR is the prohibitive entry costs - but recently (last couple of years or so) there have been strides made to make the technology more accessible. The best solutions for lower end applications that I've seen are based on Blackfin processors with or without small FPGA's. Size, power and cost are important in many radio applications and if you need a $1000 FPGA before you get started then this rules many applications out. With a Blackfin and an IF sampling A/D converter you can get started with a low level and scalable application. ADI have published this case study from quite a while back from Aerostream where a Blackfin and an AD9874 combine to give a low cost and low power starting point for SDR.
There are a few open source designs for SDR about now based on Blackfins and smaller FPGA's - check Darrell Harmon's DSPCARD here and there are a few others about
Martin
LStacey
Are you a radio guy too?
There seems to be a few of us about!
Martin Keenan
Nope, not really! I've got an interest in all things electronics and I've got involved in lots of interesting projects as an applications engineer. What I really like seeing though are costs and entry barriers come down so applications are opened up to the masses - not just those with huge budgets
peterjfrancis
An amazing technology with so much potential.
LStacey
I really enjoyed the talk on SDR. Completely opened my eyes to something I never even knew existed... Which is always nice!
For some reason it never occured to me that radio gear could be built in this way... Odd considering I used to be a radio geek and have since become a DSP geek!
I want to see more... Much more of this stuff!