Laboratory Report
Laboratory Report 97/1
Preliminary Survey of Signals
from Cellular Telephone Base Stations.
(Study by Keith Malcolm and Maurie Daly)
Report compiled by K G Malcolm
11 February 1997
Revised Text 15 May 1997
Review of Cellular Telephone
Site Signal Levels
Background:
The Laboratory has undertaken a preliminary study of RF signal levels in the vicinity of a sample of cellular telephone base station sites.
This preliminary study was undertaken by selecting a sample of cell-phone sites around Canberra and conducting a representative set of measurements as follows:
(a) at a convenient roadside location (usually outside the house nearest to the cell-site), a survey of signal levels across the VHF and UHF bands covering 20 MHz to 1000 MHz was undertaken to identify dominant existing signals and the levels of those signals recorded
(b) the cell-site control carrier was identified and its signal level measured as a function of distance from the base of the antenna support for as many (approximately) radial paths as could be conveniently followed.
The resulting data provided information about:
(i) the absolute level of the cell-site control carrier signal(s)
(ii) the level of cell-site signals compared with signals from other sources
(iii) the distribution or variation of cell-site signal around the cell-site antenna.
Equipment Considerations:
Efficient conduct of survey work of the nature undertaken imposes a number of requirements on the equipment to be used, which include:
portability
wide frequency operating range
large signal-level measurement range.
The equipment used consists primarily of a wide-range signal-level measuring instrument and appropriately calibrated antenna(s). The performance levels required of the test instrumentation are non-trivial and not easily met.
Operating Frequency Range:
for this preliminary survey, the measuring range was constrained to 20 MHz to 1000 MHz. This limitation was accepted on the basis of knowledge about likely sources of signals that are present generally in the community.
In practice, at the present time, signals on frequencies above this range are most likely to be either highly directional in narrow beams or of very low operating power level (and thus not of significant level in areas accessed by the general public). This situation may change over time as new services are introduced.
Below this frequency range, the only signals of significant level are likely to be MF-AM broadcasting stations. There are few, if any, sources of high level HF signals in typical urban areas. However, in particular locations, industrial RF energy sources (such as heat sealers and plastics welders) and amateur radio stations may act as significant localised sources of RF exposure. For this exercise, the MF broadcasting stations were ignored because permissible exposure levels (as defined in AS 2772.1) are much higher than those relevant to the VHF/UHF range. It should be noted, however, that any comprehensive survey work would need to search for MF and HF sources and make an assessment of likely contribution to total exposure in accord with the relevant provisions of AS 2772.1 before proceeding with measurement work.
Within the frequency range of interest, dominant signal sources are FM radio and TV broadcasting stations, and, close to cell-phone base stations, the cell-phone base station signals. Depending on location, pager stations may also be sources of significant signal level. Other signal sources such as taxi radios, CB radio, amateur radio etc are usually only itinerant in operation and relatively low power, so do not make much contribution to total exposure conditions. An exception to this general rule might occur in locations of unusually high density, or occurrence, of transmitting facilities (eg hill-tops used as communal transmitter sites). In a general survey, such sites should be examined over an extended time period so as to be able to assess the aggregate exposure from the large number of sources (particularly during peak traffic periods).
However, even the limited frequency range examined imposes quite difficult requirements on the measuring equipment.
Signal Level Measurement Range:
Even though wide frequency measuring instruments (usually called RF hazard meters or similar) are available, these devices are necessarily relatively insensitive. They do not detect the low signal levels generally expected to exist within the general environment and are intended primarily for use where an actual RF exposure hazard (within the terms as defined by the exposure limits set out in AS 2772.1) may exist. At most of the measuring locations used, even the most sensitive of the commercially available units would have indicated no reading at all and would have given only a minimal indication at even the worst case exposure locations found.
As a cautionary note, it is to be observed that these instruments constitute broad-band, untuned detectors followed by very high-gain instrumentation amplifiers which can lead to the false indication of high exposure level under some measurement conditions even though actual field levels are low.
A consequence of these requirements is that to gather meaningful data about exposure levels it is necessary to use a tuned, high sensitivity receiver of some sort. A spectrum analyser which is capable of providing a broad overview sweep for the purpose of identifying signals for further study and then focussing in on individual signals for measurement is an essential tool. This analyser must be used in conjunction with calibrated antennas if meaningful measurements are to be made.
In practice, it is the calibrated antennas that are the limiting element in the measuring system.
The Laboratory has both a spectrum analyser which operates from 9 kHz to 2200 MHz and a measuring receiver which covers the range 20 MHz to 1000 MHz and calibrated antenna units which cover sub-ranges 20 MHz to 200 MHz and 200 MHz to 1000 MHz.
Survey Methodology:
A small selection of sites was chosen for examination. The selection was made on no particular basis but included convenience and an attempt to assess a range of types of installation (noting that Telstra, OPTUS and Vodaphone sites are different in detailed implementation).
The first step in the assessment was to set up a spectrum analyser and conduct a sweep of the 20 - 1000 MHz range to identify signals present and to measure the levels of common signals (in practice this meant FM sound and television broadcasting signals and pager signals). In addition, the control channel of the cell-phone sector under study was identified and measured so as to provide a means of comparison of exposure from the local cell-site signal(s) and the background (mainly broadcasting) signals.
This first step was undertaken at a convenient access point, usually in the street outside the houses closest to the cell-site antenna structure.
The second step was to then tune the measuring receiver to the control carrier of the cell-site sector and to measure that signal as a function of (approximately) radial distance starting from a point of closest public access to the antenna structure. In two locations, this was right at the base of the antenna support poles (the antennas being installed in unfenced areas outside a telephone exchange building) while in the other cases, the first measuring point was alongside the site fence closest to the antenna support. Signal level measurements were then made at nominal distances of 10, 20, 30, 40, 50, 100, 150, 200, 250 and 300 metres from the antenna support. The measuring distances are not precise because of the intervention of obstructions such as trees, roads, houses and fences etc.
During this phase of the process, the measuring antenna was held at a height of about 1.5 metre above ground and signal checked using both horizontal and vertical orientation of the antenna.
This step in the process gives information about the variability in signal level around the cell-site, the location of highest signal level and an indication of the decline in signal level with distance from the antenna support structure.
The third step in the process was an attempt to identify the number of cell-phone signals originating from each site so as to permit an estimation of likely maximal exposure.
Measurement Considerations:
Signal Level Variability:
It is a characteristic of radio wave propagation that the signal from a transmitting source falls away rapidly in the immediate vicinity of the source and then declines relatively slowly with distance. In a typical environment, the signal declines in level generally according to either an inverse distance law (twice the distance gives half the signal power) or an inverse distance squared law (twice the distance gives one quarter of the signal power) or something in between. Localised additional attenuation occurs (particularly at higher UHF channels) due to obstruction of the path by trees and houses.
One consequence of this is that for the short distances involved in the detailed assessment of cell-site signal levels, it is neither useful nor necessary to measure the levels of remotely originated signals at every measuring point. At each site observed, in addition to the dominant cell-site signal, the principal signals seen are sound and television broadcasting signals. These signals originated from a distant source such that any variation over the 300 metre measuring range can be expected to be negligible (the 300 metre length of measuring path is of significance only for total paths of less than about 3 km and in all cases examined, the broadcasting site is more than 3 km away).
Cell-Site Total Exposure Levels:
The measurements undertaken have utilised the cell sector control carrier. This carrier operates continuously and at the maximum power assigned to that sector. Accordingly, the exposure data collected applies only to the effect of the control carrier.
To determine the total exposure from the site, it is necessary to take account of the maximum number of channels that can be used in the sector. This information is not readily available nor can it be easily determined by inspection at the site. The total number of channels and the power level in each channel is a function of the traffic demands on the cell and the signal level received from the user terminal(s).
For a GSM site, the control carrier operates continuously and at full power. Traffic is carried on separate carriers each of which is capable of carrying a number of traffic channels (the signal is broken up into 8 time (traffic) slots), hence a low utilisation GSM base station might have only two RF channels. The maximum possible number of RF channels at any base station is a function of the spectrum bandwidth available and the cell planning structure adopted. At the present time, it would seem that each sector could have up to 4 RF channels. If the traffic at a site is sufficient to warrant more channels then they can be provided by "stealing" channels from surrounding base stations.
In the preliminary study undertaken so far, the GSM sites observed appear to have from 2 up to a maximum of 9 (3 overlaid sectors) RF carriers available for use. The total exposure is simply the addition of the powers from each of the available carriers. This is calculated by multiplying the power level observed on the control carrier by the expected maximum number of carriers in the sector.
For an AMPS site, the control carrier similarly runs continually at maximum power. However, traffic channels are activated in direct proportion to the number of calls assigned to the cell and at power levels dependent on the signal level received from the user terminal. This means that the traffic channels occur fleetingly and at unpredictable levels. As for the GSM case it is possible to estimate the maximum number of channels and thus derive an estimate of "maximal exposure" for the site. During the survey, we observed 6 or 8 RF channels in operation at the AMPS sites studied, but it is possible for a very busy AMPS site to have up to 32 carriers per sector.
Some sites have both AMPS and GSM stations. In these cases it is appropriate to simply add the estimates for each service.
As it is not possible to know the detailed operation of any given site (even if one knew the actual channel allocations at the site, the actual utilisation is determined by the traffic demands which are highly time variable) it is not sensible to attempt to determine the actual exposure from the site. However, by examination of things such as the number of control carriers and the occurrence of additional traffic channels, it is possible to estimate a "maximal exposure" that would occur if all channels were to be fully occupied all of the time. This is a "conservative" process as the actual exposure over time will be less than the calculated maximal value (the systems are constructed to meet the anticipated "busy time" demand so average usage is well below this).
Occurrence of Peak Exposure:
Prior to commencement of the study, it was not known where the peak exposure might occur in relationship to distance from the antenna mast.
The base-stations use antennas that have highly directional (narrow beam) vertical plane radiation patterns as a means of achieving high antenna gain over a broad azimuth. These high gain antennas are used both to reduce the actual power requirements of the transmitters and to provide some ability to achieve more-or-less uniform signal levels over the intended coverage area. This latter also helps to minimise exposures close to the antenna installation by "directing" the signal over the nearby surrounds of the station.
The location of peak exposure (in locations accessible to the general public) is determined by aspects including the shape of the antenna pattern, terrain profile (falling ground tends toward lower exposure, while rising ground tends toward higher exposure particularly if such that buildings protrude into the main lobe of the antenna beam close to the antenna) and the extent of any overlap of coverage from sectoral antennas (most sites studied use antenna arrays designed to provide 120 degree arcs of coverage so there will be some overlap at the edges of the arcs. In practice, the signal level at the edge of the sector will have fallen from its peak value (due to the shape of the antenna beam) and there is typically more than 20 dB separation in level between adjacent sectors (contribution from adjacent sector being less than 1%) such that, unless sectors are intentionally designed to overlap each other (an unlikely situation), the aggregate exposure will not increase above that from a single sector alone.
The occurrence of trees and buildings will strongly influence actual exposure levels as they both absorb (and thus attenuate) the radio signal. Occupants of buildings obtain even greater protection because the building fabric causes significant reduction in the level of signals passing through walls and windows.
For one site observed, the maximum exposure occurred very close to the site fence, but for most sites studied the peak exposure occurred within the range 50 to 150 metres. In one case, where the ground rose rapidly in the direction away from the base station, the peak exposure occurred at 250 metres.
Study Results:
The results of this preliminary study are shown in the attached diagrams.
The first group of diagrams (Fig 1, Fig 2a, Fig 3a, Fig 4a1, Fig 4a2, Fig 4a3) show the actual measured exposures from broadcasting and paging signals and the cell-phone control carrier(s) at the selected control or reference location (as noted above) within the sector under observation.
The second group of diagrams (Fig 1, Fig 2b, Fig 3b, Fig 4b) provide estimates of the likely "maximal exposure" at each site. These estimates have been derived by bringing together the measured data for the "environmental" or "background" signal levels from broadcasting and paging services (in all cases originating from distant locations) and the peak exposure level measured from the cell-site signal within the sector multiplied by the estimated maximum number of cell-site RF carriers available for use in the sector. Where relevant, both AMPS and GSM data is included.
Note that for both AMPS and GSM stations, the control carrier operates continuously and at the maximum power level assigned to the site whereas the occurrence and level of traffic carriers depends on the traffic loading and levels of incoming signals. This means that it is generally impractical to directly measure the signal levels or the numbers of active traffic channels. For the purposes of this study, it is considered appropriate to use the control channel as a "proxy" for the traffic channels (the frequencies are such that practical differences in signal levels across the range of channels should be negligible). Derivation of "maximal exposure" then requires some means of estimating the likely maximum number of channels assigned to the cell or sector. Consideration of the frequency allotments available to GSM at present and observation of sample sites suggests that a typical cell or sector can have up to 4 RF carriers (control plus 3 additional traffic carriers). Observation of AMPS sites suggests a typical loading of about 6 or 8 RF carriers. These values have been used in deriving the estimates of maximal exposure shown in Figures 2b, 3b and 4b. Figure 4b is associated with a chart that provides an outline of the data used to derive the diagram.
Note that Figure 1 as measured appears to accurately reflect the actual "maximal exposure" case for this site as two carriers were measured and only two carriers have ever been seen as being active at this site.
The final diagram (Figure 5) gives an indication of the fluctuations in control carrier signal with distance as measured in each sector studied.
Comments and Observations:
Acknowledgements:
Thanks are given to Jack Rowley, Telstra Research Labs for review of, and helpful comment on, the draft text and to Helen Anderson and Maria Mandic, Telecommunications Industry Division DoCA, for comment and suggestion regarding the graphical presentation of data.
-//-
Report 97/01 Figures 1, 2a, 2b, 3a, 3b, 4a1, 4a2, 4a3, 4b & 5 as an attachment.