The success of Mobile TV is dependent upon service reliability. Generally speaking, people have been trained to accept poor cellular voice service and intermittency. Dropped calls are not a big deal since the information can be repeated on another hopefully successful attempt. Video is much different. Gaps in reception result in a lost experience and will not be acceptable to the viewer.
Over the last four years, extensive testing has been conducted to quantify the benefits of transmitting circular polarization (CP) to a linearly polarized mobile handheld.
In 2007, it was concluded that small handheld devices are limited to linear polarization; there was defined margin improvement; and tests in a controlled environment showed, on average, that transmitting circular polarization provides 5dB of margin improvement over linear polarization. In 2008, tests in a controlled environment showed that, compared to CP, elliptical polarization with a 66-percent Hpol/33-percent Vpol split provides the highest margin improvement in heavy scatter environments. Also that year, it was discovered that adding a separate vertically polarized antenna to an existing horizontally polarized antenna provides 2dB of margin improvement.
The next year, in 2009, outdoor field tests showed that, on average, transmitting circular polarization provides 5dB of margin improvement over horizontal polarization and 7.5dB of margin improvement over vertical polarization. And, in 2010, tests showed transmitting VHF circular polarization to a small, linearly-polarized handheld provided 3.5dB of margin improvement over horizontal polarization. Also, mobile UHF had a 15dB gain margin advantage over VHF. BER testing confirmed all of those results in 2011.
Even though the testing quantified the margin improvement when using circular polarization, there was still concern over the benefits since all measurements were based on signal strength. Many believed that to truly prove the margin benefit gained by circular polarization for digital transmission, the measurements must be based on bit error rate (BER). In March of 2011, in a joint effort with the West Central Florida Group, the opportunity to conduct measurements based on BER became possible.
BER vs. SNR curve
BER is the number of bit errors divided by the total number of bits. The relationship between BER and SNR is inversely related by a waterfall curve — general examples of which can be found in most communication text books. Two BER measurements can be converted to an expected margin improvement. (See Figure 1.)
These general curves are typically published for free space and static conditions, neither of which is true for real-life conditions. In order to be able to relate true, expected SNR based on BER measurements, the curve must be adjusted for multipath fading and the modulation scheme used within the communication channel. In mobile handheld situations, a fading channel is best represented by a Rayleigh distribution where there is typically no dominate line of site signal. The equipment used in the experiment was an ICOM LMR 450MHz system, which is based on non-coherent 4 level frequency shift keying (4FSK). The probability of signal, in a Rayleigh fading environment when using non-coherent 4FSK, can be determined as shown in Figure 2.
The BER vs. SNR curve can now be adjusted for the experiment situation and used for determining margin improvement from measured average BERs. It should be noted this adjustment highlights the need for higher SNR in real-life conditions in order to reduce the BER.
A circularly polarized antenna was placed next to a vertically polarized antenna at 800ft on the ATC broadcast tower in Riverview, FL. Also, a mobile unit simulating a linearly polarized mobile handheld was constructed. A motorized dipole was used to measure in parallel mode as if the user was holding the handheld in the upright position and perpendicular mode as if the user was holding the handheld horizontal to the ground. While continuously measuring all orientations, the mobile unit was moved over a long run in location. At the base station, a logging program was used to continuously sample the BER and GPS location.
In order to ensure a fair comparison, the circularly polarized antenna was designed to provide similar coverage on the main beam as the vertically polarized antenna. In doing so, the estimated field strengths per polarization are calculated to be equal from 3mi out in free space. For this reason, all measurements were taken within 60 degrees of the main azimuthal beam and no closer than 4mi from the tower. The data was collected in three different environments: outdoor, indoor and driving. Multiple experiments were conducted in each of the three environments, including: open areas, city, residential, mall, office complex, and inside and outside of a vehicle.
A large number of samples were recorded, and the average BER in each run was calculated for both the circular polarization and linear polarization. The margin improvement was then figured by transposing average BERs onto the BER vs. SNR curve.
When averaging all of the outdoor tests, the average margin improvement of circular polarization was 8dB in the parallel mode and 7.9dB in the perpendicular mode. It may appear strange that the average margin improvement of the perpendicular mode is not higher than the parallel mode since, in the perpendicular mode, the receive dipole is held horizontal to the ground and should be completely depolarized from the vertical signal. The answer is that small-scale fading has created as much vertical component in the horizontal plane as there is in the vertical plane. Multipath has completely depolarized the signals. If this is the case, then logically: If the vertically polarized signals are so depolarized, then the received signals should be independent of orientation and location. So, why does circular polarization provide 8dB of margin improvement?
That answer is because circular polarization is made up of two orthogonal polarizations time-shifted by 90 degrees. The odds of both polarizations destructively interfering at the same time and same location is much less than a single polarization.
For the indoor cases, the average margin improvement of circular polarization was found to be 6.8dB in the parallel mode and 8.3dB in the perpendicular mode. Note that both the indoor and outdoor measurements produced similar results for both the parallel and perpendicular cases with an overall average margin improvement of 7.5dB. This is explained by understanding circular polarization primarily helps mitigate the effects of small-scale fading that is present both indoors and outdoors. Large-scale fading, such as attenuation through structures, tends to only shift the mean signal strength. As that strength decreases, BER increases. But, the margin improvement gap remains the same. This is due to the effect of Rayleigh fading has flattened out the SNR vs. BER curve in the region of usable operation. It can now be said the benefits of circular polarization hold true both indoors and outdoors.
The next tests were performed both inside and outside of a moving vehicle during long drives of 25 to 60 miles. The first test was inside of a hatchback car. On average, it was found that vertical polarization actually provided 0.5dB more margin improvement than circular polarization. The second test was with a small monopole on top of the car. Results showed circular polarization started to provide a benefit with margin improvement being 1.5dB over the vertical polarization. The third test was with a larger monopole farther above the top surface of the vehicle. Margin improvement here was 2.5dB.
The results provide an interesting insight into why circular polarization does not provide any benefit inside a vehicle, and only starts to provide benefit when raised off the surface of the top of the car. The answer lies in a boundary condition commonly used to solve Maxwell's equations. It states: “The E-field tangent to a ground plane is zero.”
Inside a vehicle, there is basically a ground plane above and below the linearly polarized antenna of a mobile handheld. Therefore, most of the horizontally polarized signal is filtered out, leaving only the vertically polarized component of the incoming circularly-polarized signal. When the antenna is placed on top of the vehicle, there is only a ground plane below it. As the antenna is raised higher above the ground plane, the circularly polarized signal begins to retain shape. This concept can be demonstrated using HFSS modeling software by launching a circularly polarized wave at a low grazing angle onto a ground plane as shown. (See Figure 3.) Note that only the vertical component exists near the ground plane, but, a few wavelengths above the ground plane, the circularly polarized signal is fully intact.
It must be mentioned here that this situation should not discourage the use of circular polarization for mobile applications. Who needs margin? It's not the fireman or policeman inside a vehicle. They have the option to use larger, more efficient external antennas in conjunction with much higher-power radios. The users that need margin are the ones that are carrying small, inefficient, low-power handheld devices, and this is where circular polarization provides a significant advantage in reliable connectivity over liner polarization.
In summary, margin equals reliability, and BER testing confirms that transmitting circular polarization to linearly polarized handheld devices provides necessary margin that will be imperative to the success of Mobile TV.
(Note: The author would like to thank the West Central Florida Group, especially acknowledging Ed Allen and Paul Toth for their support in conducting these tests.)
John L. Schadler is director advanced antenna systems development for SPX Communication Technologies.