Redefining Software Defined Radio
BitWave’s SDR for the Masses

Software Defined Radio (SDR) is certainly one of the hottest buzzword technologies being bantered about these days. Although most current SDR development was started in the military with initiatives like the US government’s Joint Tactical Radio System (JTRS), the advantages of the approach extend far beyond military use, and the technology is now center stage for consideration in many commercial embedded applications as well. SDR offers flexibility and field-upgradeability to RF systems that simply cannot be matched by traditional analog-based RF design techniques.

The typical architectural approach to SDR locates a wideband analog-to-digital converter (ADC) very near the antenna of an RF system in order to sample and digitize incoming RF waveforms. The channel modulation scheme is therefore implemented completely in software, so down-conversion and demodulation happen entirely in the digital domain, typically on a digital signal processor (DSP) or other general purpose processor. This setup is extremely versatile, of course, because the software that defines the scheme can be swapped on the fly, so one hardware platform can implement an almost infinite variety of radio standards without modification.

An obvious commercial application that would benefit from this approach are mobile phones, where a plethora of competing standards has burdened the industry with building a large number of variants of each handset in order to support service providers and standards around the world. The problems with stuffing SDR into a cell phone, however, boil down to our old friends - price, performance, and power. That wideband ADC is an expensive function in any implementation and only gets more expensive in the frequency ranges used by mobile phone carriers. Right behind the ADC, the power consumption, price, and silicon area of DSP processors that can handle the problem are all well above the budgets of even the most high-end cell phone projects. It has always been cheaper to implement multiple transceivers in a single phone than to try to slide into an SDR solution.

BitWave Semiconductor, a Massachusets-based startup, has recently announced a new technology called Softransceiver that could change that balance, however. BitWave’s single-chip radio frequency IC (RFIC) takes a more conventional radio architecture and adapts it to SDR-like configurability instead of implementing a full-tilt SDR scheme. Rather than running the whole show in software, BitWave started with a traditional transceiver design, applying software control to configure components like the low-noise amplifier (LNA), mixer, voltage-controlled oscillator (VCO), and corresponding transmitter parts. The result is a system that doesn’t require the extremely expensive wideband ADC and high performance processing of traditional SDR, but still delivers the same SDR flexibility under software control.

The use of digitally controlled analog components may cause some purists to question the SDR label, but the benefits are essentially the same. Most interestingly, the technology promises to offer a single transceiver capable of handling the wide variety of current (and hopefully future) radio standards used in mobile phones. Additionally, Softransceiver will allow carriers to add services and features to phones by simple reconfiguration of the devices, paving the way for a number of revenue-generating add-on service options. BitWave isn’t announcing who their partners and early teacher-customers are yet, but you can bet that they’ve captured the attention of handset manufacturers looking to lower their cost of design and deployment into the many large global handset markets.

BitWave says that their tunable parameters on analog components include input match, output match, center carrier frequency (Fc), bandwidth, second- and third-order intercept points (IP2 and IP3), noise figure (NF), effective number of bits (Enob) and many others. Digital states are used to control these parameters, and they can be altered or reconfigured in real time. According to BitWave, this design yields much better performance with lower power in less silicon area than traditional SDR.

BitWave will be offering the transceiver fabricated in a regular, garden-variety digital CMOS process, so a future product enabling the same functionality within a hard IP block for highly-integrated System-on-Chip designs seems a possibility. For now, however, BitWave is focused on delivering the first samples and shipments of this product to their customers over the coming months. Their first samples are scheduled for delivery in summer 2006.

Behind the scenes, BitWave obviously tackled a highly complex, mixed-signal, system-on-chip design to realize this technology. The analog component alone would send most SoC design teams running for cover. Rather than betting their bankroll on a single-pass, hope for first-time-success tapeout, BitWave broke the process down into a sequence of controlled stages, taping out a new IC for each one. By incrementally iterating on their design as they pulled together the layers of integration, they are able to debug each set of samples that come back with far greater precision and productivity. This approach is important because there are no reasonable approaches to prototype such a system that would give sufficient verification visibility to allow a good chance at first-silicon success.

The specs on the new silicon include an RF tuning range from 700MHz to 4.2GHz, bandwidths from 200kHz to 20MHz, and full support for a number of existing standards including UMTS (GSM, GPRS, EDGE, W-CDMA, HSDPA), CDMA2K (IS-95B, 1xRTT, EV-DO), iDen, and 802.11 b/g. They also claim “partial support” of AMPS, PHS, IS-136, Bluetooth, 802.16d/e (WiMax), GPS, DVB-H, DVB-T, DTV, ISDB-T, and DAB. Their modem interface is 3G DigRF plus extensions and their front-end module (FEM) interfaces include both multi-port, single band, and single-port multi-band.

BitWave is also working to provide tools and services to help customers integrate the new device. Obviously having an almost infinitely reconfigurable transceiver is an awesome feature unless you are left to your own devices figuring out how to reconfigure it. Support for many of the popular standards is built-in, and additional support is available for upgrading and adding additional standards over time.

The proliferation of SDR technology into cell phones and other handheld devices continues the trend of pushing programmability and configuration options as far out in the production chain as possible, often even into the hands of the end customer. Programmability ultimately gives systems suppliers the ability to get products to market faster with greater ranges of features, and keep them in the sweet spot of the market longer with low-cost upgrades allowing electronic devices to ride the wave of new innovations and customer demands much longer than traditional fixed-architecture systems.

Kevin Morris, Embedded Technology Journal

November 29, 2005

[back to top]

Comments on this article? Send them to comments@embeddedtechjournal.com