Andersen JB, Mogensen PE, Pedersen GF. Power variations of wireless communication systems. Bioelectromagnetics. Jan 28, 2010. Ahead of print.
The present article discusses power variations in not only mobile phones but also in other digital wireless devices such as hearing aids, data transferring devices. Such new devices represent new type of exposure considering time variations of power along with intermittent use which decreases the time-averaged exposure. Power may also vary due to movement of the user, handovers between base stations, and traffic. In the communications industry the power spectrum is the measure of the occupied spectrum around the carrier, but the two measures are different (the power spectrum is the Fourier transform of the autocorrelation of the signal, while the spectrum of the power is the Fourier transform of the square of the signal). The zero frequency value of the spectrum is simply the average power, proportional to SAR. It is useful to study spectrum of power fluctuations as a convenient and compact way of describing the very complex behaviour of wireless devices.
There are different time scales for the time variations. At the longest time scale, on the order of seconds, the power varies due to the system power control that compensates for the variation in the radio connection. The time scale depends on the carrier frequency since the variations follow the spatial scale of wavelengths; the higher the frequency and the faster the movement, the faster the variations. However, this represents a broad spectrum of frequencies.
The next group of frequencies lie within the Hz region and these are narrow spectra, like spectral lines. Examples are the 2 and 8Hz components of the GSM system. A significant group of frequencies lie within the kHz region, like the 217 Hz and harmonics for GSM due to the power bursting. For WCDMA, the relevant frequencies are 1500 Hz due to fast power control and 15 kHz pilot signals. For a steady channel the basic frequency is 750 Hz. For the high-speed packet access (HSPA) system a new frequency of 500 Hz appears due to the choice of 2ms time frame. The wireless LAN system studied has a broad spectrum in this frequency range. For DECT, which is widely used system in the home and office environment, the basic frequency is 100 Hz.
For most systems the data signals are not modulated with a constant envelope, as is the case with GSM and DECT. This means that the data signals, often in the microsecond range, will appear as power variations in the MHz region. A good example is the WCDMA chip frequency of 3.84 MHz, which appears as a spectral line in the transmit power. The wireless LAN systems also have spectral lines in the MHz region.
The authors concluded that time variations observed in the wireless signals may not have biological effects, but they are part of many experimental studies, and therefore, it is important to do further experiments to explore power variations.