I often hear that size, weight and power are major impediments to the adoption of software defined radio. With the hundreds of thousands of SDRs now in deployment in various markets, I have to wonder – is this still true, or is it just a myth?

To answer this question, I considered three handsets utilizing SDR technology that are available in the market:

• The Apple iPhone 3G, discussed in my last blog <insert hyperlink>, which is 5.2 cubic inches in size, weighs 4.7 ounces, and has a battery life reported of 5 to 10 hours depending on use.
  
• The Motorola APX 7000 radio, which was advertised at the SDR’08 Conference and Product Exposition, is 32.2 cubic inches in size, weighs 18.4 oz. (with battery), and has a battery life of 8 hours under typical use. 
  
• The Harris Falcon III Radio, which was also advertised at the SDR ’08 Conference and Product Exposition, is 42 cubic inches in size (with battery and GPS), weighs 41.6 oz (with battery) and also has a battery life of 8 hours for the AA adaptor configuration. This battery life is similar to other radios in the same class. 
  

The battery life on each of these radios is pretty similar, but the size and weight vary dramatically as we move from commercial to public safety to defence products. I contend that the reason for this has nothing to do with SDR, but rather lies in the economics associated with the radio platform and the requirements levied on the system in their respective markets.

Consider the following: a typical modem architecture consists of a general purpose processor, one or two DSP’s, some interface logic, and maybe an FPGA or application specific device for higher performance operations. Each of these devices requires its own separate power, much of which is dedicated to moving data between chips. Take these same technologies and integrate them onto a system on a chip device, and you instantly gain significant economies in size, weight and power. This is exactly what has happened in the commercial market for “smart phone” devices (again, see our last blog entry).

So why doesn’t everyone create a custom SoC for their radio platforms?

The answer is: it costs too much. Spinning up a device fabrication is really expensive. If you plan to produce hundreds of thousands of devices, then you can amortize this expense across these devices, making the cost per unit associated with chip fabrication reasonable. At lower volumes, as is typical in the defence and public safety markets, the ability to achieve these economies diminish such that it is difficult to justify the per unit cost and so equipment manufacturers are forced to use more general purpose technologies with their associated increase in size and power.    

This problem is exacerbated as customers introduce additional requirements on a radio for use in a specific segment of the market, further limiting their production volumes. Consider the following differences in the three classes of radios presented above:

 

                                                                                                Typical Handset Specification

                                                             Commercial                      Public Safety                       Defense

RF Range                                             824 to 849 MHz                 136 to 174 MHz                 30 to 512 MHz
                                                            869 to 894 MHz                 764 to 776 MHz                 (continuous)
                                                            876 to 916 MHz                 794 to 806 MHz
                                                            921 to 960 MHz                 806 to 825 MHz
                                                        1710 to 1785 MHz                  851 to 870 MHz
                                                        1805 to 1880 MHz       
                                                        1850 to 1910 MHz
                                                        1930 to 1990 MHz

Application Framework                            Custom                             Custom                              SCA 2.2.2

Cryptography                                          Commercial                       FIPS 140-3                         NSA Type-I Certified

Environmental                                         Non-rugged                        MIL-STD 810F                    MIL-STD 810F

Operating temp                                       0 to +35^o C                        -30 to +60^o C                     -30 to +60^o C

Relative humidity                                     5% to 95%                         Submersion to 2m              Submersion to 2m

 

It is clear that the requirements levied on public safety communications are higher than commercial, and that the requirements levied on defence communications are often higher yet again. There is also a difference in life cycle expectations in these three markets: commercial handsets are essentially disposable, whereas radios in the public safety and defence markets are in use for many years. This requirement drives a further increase in size, weight and power (and costs!) as the use of suboptimal components may be required to meet the life cycle target.

 

I contend that these differences, when combined with lower volumes, are the real drivers of size, weight and power in advanced wireless systems, and not the use of SDR technology.
As always, I welcome your feedback on these views: please post your comments here or email me at Lee.Pucker@SDRForum.org.