Transmitted TV signals are waves of energy and act like waves in water and/or light waves.  That is, they are reflected when they strike an object.  They can also be diffracted (change direction) when they encounter an edge such a mountaintop or the blade of a wind turbine.

TV signals are strongly reflected by metal surfaces whose dimensions are greater than a critical length of the reflecting surface.  This corresponds to about 18 inches for signals in the UHF TV band (Most TV signals are in this band.) and to about 53 inches for FM radio signals.  (FM radio reception can be affected by extremely strong echoes.)  The steel tower and to a lesser extent, the non-metallic blades of a wind turbine will reflect TV signals.  These waves move at the speed of light.  That is about one sixth of a mile per microsecond (1 one millionth of a second) or six microseconds per mile.  The effect of reflected signals is to create echoes of the TV signal which have traveled further than the distance from the transmitter to the receiver; hence they are delayed and generally weakened with respect to the direct signal from the transmitter.  In some cases, the direct path from Transmit to Receive is blocked by man-made structures or hills.  An illustration of how echoes can be caused by reflections from a wind farm is shown in Figure 1.

                                   

Analog TV signals from the A-town Transmitter of Figure 1 would have suffered a lagging “ghost” (an echo), part way across the screen and to the right of the image formed by the direct path signal.  The longer path reflected signal is much weaker than the direct signal at the viewer’s home.  With analog TV signals, even a ghost image echo 10,000 times weaker than the direct signal was annoying to viewers. [1]

If the direct path were partially blocked by a mountain range or man-made structure, the delayed reflected signal arriving at the receiver may be stronger than the direct transmitted signal.  In this case the weaker signal (the ghost) would lead the stronger signal.

The US upgraded from analog to digital (DTV) in 2009.   Canada will do so soon.  The effect of echoes on DTV reception is much smaller than it was on analog TV reception.

In our example, DTV signals travel 3 miles between the Transmitter and A-town’s home viewer.  These signals also travel 2 miles to the wind farm where they will be partially reflected and then travel an additional 3 miles to A-town: a total path length of 5 miles.  The reflected signals travel 2 miles further than the direct path so they arrive 12 microseconds later than the direct signal, causing a lagging echo.

Figure 2 shows the static echo rejection by early (2004-2005)  US DTV ATSC receivers tested by the Mackenzie Presbyterian University (Brazil).

The 2004 vintage best receiver is shown in blue (ATSC Previous Generation).  It provided excellent rejection of lagging echoes, but very poor performance in rejecting leading echoes.  Shown in red is the best receiver vintage 2005 (ATSC Latest Generation)  Here it appears that the design objective was to improve the echo rejection for leading echoes, but in doing so, the rejection of lagging echoes was compromised.

The Mackenzie Presbyterian University also tested static echo rejection of the best European DTV DVB-T receivers at the same time  Those results are shown in Figure 3.

Here we see that the DVB-T previous generation models (blue) behaved ideally for leading echoes up to – 75 microseconds, and equally well for lagging echoes up to + 75 microseconds.  The DVB-T latest generation performs just as well over this range of echo delays and performs better for leading and lagging echoes > 75microseconds.

The US DTV System (ATSC) relies upon digital signal processing (equalization) in the receiver to remove echoes from the television signal.  There are multiple types, brands and generations of equalizers utilized within the US population of DTV receivers, with each generation of receiver showing significant advancement in its ability to reject multipath interference.

Following the US DTV conversion in 2009, the FCC released a report detailing their performance testing of 102 of the latest generation digital converter boxes [4].  The report states “Multipath performance of converter boxes was quantified using single-static-echo tests and field-ensemble tests.  Single-static-echo performance of all approved converter boxes satisfied the ATSC guidelines within a small margin for measurement error (0.2 dB).  The converter boxes successfully demodulated a median of 39 of the 50 ATSC-recommended field ensembles—with even the poorest performing converter box significantly outperforming 64 percent of FCC-tested receivers that were on the market in 2005 [4]”.

The 102 receivers contained 13 brands of demodulator chips.  As the histogram in Figure 4 shows, the latest receivers were much better equipped to successfully demodulate signals containing multipath interference than DTV receivers manufactured in 2003 through 2005.

The FCC data also show the variability among the ATSC receivers with regard to their ability to successfully process echo rejection.

The European DTV system (DVB-T) does not rely on an equalizer in the receiver to reject multipath.  DVB-T signals are modulated by a technique called Coded Orthogonal Frequency Domain Multiplexing (COFDM)  Nearly 8,000 carriers are equally spaced across the channel bandwidth (e.g. 8MHz in the United Kingdom).  Each carrier is digitally modulated at a very low bit rate. Carriers are about 1,000 Hz apart, so each symbol period is 1,000 microseconds.  It is well known that nearly all signal reflections (echoes) arrive at the receiver within +/– 60 microseconds of the arrival of the same symbol.  The beginning of each symbol period is reserved as a “guard interval” (typically 60us) during which receivers ignore the signal, so they do not “hear” echoes within this interval.  This has the practical implication that the echo rejection of DVB-T receivers of different makes and model years are very similar among different manufacturers.

The DVB-T Standard provides the flexibility needed by the various governments in Europe to customize their broadcast signal to meet local propagation requirements.  In a mountainous country such as Switzerland, the government chose a large guard interval to eliminate the effects of echoes.  In Holland, a smaller guard interval suffices.

The United Kingdom has 8 MHz wide TV channels and Germany has 7 MHz wide signals.  These variations are all encompassed by the DVB-T Standard.  That is to say, every country in Europe has the flexibility to tailor its DVB-T system to accommodate its terrain and proximity to broadcasters in neighboring countries.

A static echo is generated by a TV signal reflected from the steel tower of a wind power generator.  If the non-metallic blades are not rotating, they may cause some of the incident DTV signal to also be reflected to reach receivers, but echoes from today’s non-metallic wind turbine blades are very weak compared to reflections from the tower.  The shape of these blades is optimized to capture as much energy from the wind as possible, that is, they are aerodynamically shaped; with tapered curved surfaces.  To date, models to calculate the wind turbine blades’ reflected signal power have used rudimentary approximations.  Results generated by these models have not been able to consistently predict actual field measurements [ 3 ].

Although echoes from wind turbine blades are weak compared to reflections from the tower, dynamic echoes from rotating blades may be much more troublesome to DTV reception than echoes from static blades.  Modeling of dynamic echoes is difficult because the Doppler frequency shift for signals reflected from the fast moving tips of rotating blades is much higher than the Doppler shift from the slower moving centers of the blades.  European researchers have concentrated their studies on blades, but they also reported that the signal power reflected from the massive steel tower is significant. [ 3 ]  While static reflected signals from the tower can be predicted with reasonable accuracy, more research is needed on predicting the dynamic echoes that are generated from moving blades.

Pre 2005 ATSC receivers had good static lagging echo suppression, but poor static leading echo suppression (Figure 2).  They also had good dynamic echo suppression.  However, early experience with DTV broadcasting indicated that static leading echoes were more frequent causes of interference than dynamic echoes from rotating turbine blades.  Consequently after 2005 the emphasis in ATSC equalizer design shifted toward better leading echo suppression, but at the expense of dynamic echo rejection.  Figure 5 shows the dynamic echo rejection for a lagging echo of the best available ATSC receivers before and after 2005.

Figure 5: Single Dynamic Echo +8 µs delay ATSC

Good performance is indicated in Figure 5 when the trace is near the top of this chart.  Static echo rejection is shown for zero Doppler frequency (top, center).  The earlier ATSC receivers (blue trace) degraded slightly in dynamic echo rejection as the Doppler Frequency increases from zero to +/- 200 Hertz Doppler.  The best of the latest ATSC receivers had very poor dynamic echo rejection as the red trace falls sharply at +/- 40 Hz. Doppler.  In Figure 6, the same data is presented for the best DVB-T receiver.

Here the red trace shows the latest receiver performs much better to +/- 80 Hz Doppler.  The earlier best DVB-T receiver was just as good, but only to +/- 35 Hz Doppler.  These Doppler Frequency numbers are related to the maximum vehicular speed for autos.  Doppler frequencies for echoes from wind farms depend on other variables.  However good echo rejection (trace near the top of the chart) is always desirable and the DVB-T system provides this result over a wide range of Doppler frequencies for all receivers.

There have been a number of attempts to predict the interaction of DTV signals with wind farms in Europe for the DVB-T signal.  Results to date have not been satisfactory but this is only a work-in-process, largely in Europe [ 3 ].

Our ATSC signal was designed to maximize coverage for a given radiated power.  In Europe the DVB-T standard was designed to allow neighboring countries to customize their signals and to optimize multipath rejection.  The DVB-T system exchanges a loss in signal coverage capabilities for better echo rejection.  Further research is needed on the development of predictive models for both standards.

It is possible to measure the relative echo rejection capabilities of a range of different ATSC and DVB-T  receivers in a well equipped laboratory.  The Charles Rhodes laboratory has 28 ATSC receivers of recent vintage.  All these are fed Laboratory generated ATSC signals with various echo power and delays.  The Rohde-Schwarz model SFE in this laboratory can generate either ATSC or DVB-T signals and it can simulate multiple echoes.

As mentioned above, there are about 13 different designs by as many firms of ATSC decoder IC devices.  These designs offer different trade-offs between:

a)     Leading echo vs. lagging echo rejection.

b)     The range of echo delays that can be rejected.

c)     The range of “Doppler Rates” that echo rejection works.

d)     How strong an echo can be rejected vs. the echo delay.

In the A-town wind farm scenario of Figure 1, leading echoes do not exist where there is a direct path from transmit to receive sites.  On the other hand, a direct path may not exist due to a man-made structure blocking the direct path (i.e. a multi-family residential structure or an office building).  Under this scenario, lagging echoes will exist and the path length difference will be measured in miles, (6 microseconds of delay per mile of path difference).

With US DTV, an echo 100 times weaker than the direct DTV signal has no effect on the picture quality, for a single static echo.  Put another way, the echo can be 100 times stronger with digital transmission than with analog transmission.  We have measured the “degradation to reception” of  the desired signal power due to a single static echo at 25% of the desired signal level for a range of 2008 vintage DTV receivers.  These results are shown in Figure 7.

Nine ATSC receivers were tested.  These are remarkably similar for delays below 30 microseconds.  That echo would have killed analog TV reception.  The average degradation to reception for an echo delay less than 30 microseconds is about 3 dB. [ 2 ]

The signal degradation was, on average, about 3 dB, which except for the fringe area should be inconsequential.  It is worth noting that all of the tested receivers behaved in the same way for echo delays under 30 microseconds, and some behaved approximately in the same way for even greater echo delays.  These laboratory findings are directly applicable to the challenge of predicting interference to DTV reception by wind farms.

A wind farm will generate a large number of nearly equal power lagging echoes with slightly different delays.  There would be no leading echoes.  This is far from the typical echo ensembles used heretofore for testing in prototype receivers.  Devising a suitable echo ensemble representing a wind farm would be a highly valuable addition to the literature available to both wind farm developers and receiver designers.

The DVB-T and ATSC systems were designed to achieve the objectives envisioned by their respective standards committees.  The ATSC standard was designed to maximize signal coverage at a given power level and has seen several generations of demodulation circuitry improving upon the receivers’ ability to reject multipath interference.  The DVB-T standard was designed specifically to give European Broadcasters the ability to customize their individual countries’ standards to optimize signal propagation within their markets and to minimize the effects of echoes upon reliable reception.

The data shows that the gap between the two standards’ ability to reject multipath interference has narrowed significantly since 2005.  However, we would suggest that more field testing is required to determine the comparative performance of the latest generation of ATSC receivers with echo cancellation within a wind farm setting.  Additional research and development is needed to create models to predict static and dynamic echoes within these settings.  An outgrowth of this research would be to identify today’s best performing receivers and to use this information to help set receiver design, performance and manufacturing standards for echo rejection going forward.

FM reception is more robust to multipath than analog TV or ATSC digital DTV reception.  Analog TV and ATSC DTV signals are amplitude modulated (ATSC DTV modulation is 8-level vestigial sideband (8-VSB), which encodes a binary stream as 8 amplitude levels of a carrier.)  Multipath interference affects signal amplitude so analog TV and 8-VSB DTV demodulators are sensitive to it.  Analog FM signals, on the other hand, which are frequency modulated, are unaffected by amplitude interference caused by multipath – up to a point.  The FM demodulator locks on to the strongest signal in its capture range.  As long as the multipath signal is sufficiently lower than the direct signal the FM receiver will ignore the multipath signal. 

FM stereophonic signals are more susceptible to multipath distortion than are monaural signals, but this susceptibility is rarely if ever problematical to broadcasters.  Stereo is enabled by in-band signals above the baseband modulation at reduced modulation amplitude.  The baseband signal, which contains the sum of left plus right (L + R) audio channels, is single-side-band modulated with FM deviation bandwidth of 15kHz.  The difference of left minus right (L – R) audio channel is double-side-band modulated with suppressed sub-carrier at 38kHz and FM deviation bandwidth  +/-  15kHz.  Between the top of the L+R channel at 15kHz and the bottom of the L – R channel at 23 MHz is a pilot tone at 19 kHz, exactly half the suppressed carrier frequency of 38kHz, which is used to regenerate the 38kHz subcarrier for the L – R demodulator.  Because the pilot is at reduced amplitude(8% -10% modulation) relative to the baseband signal it is more susceptible than the baseband signal is to multipath interference.  In practice, however, multipath interference to the stereo pilot tone is rarely a problem.  When the pilot tone is lost, receivers automatically revert to monaural reception (L + R baseband only).  Due to the limited dynamic range of modern, heavily processed FM Radio audio, most listeners cannot distinguish the loss of stereo.

The use of in-band subcarriers has been extended beyond audio to provide various data services.  The most prominent is the Radio Data Broadcast System (RDBS), which uses a low-amplitude subcarrier at 57kHz.  RDBS is a very narrow band system (1187.5 bit/sec.) suitable mainly for text.  Although the low amplitude makes the RDBS subcarrier more susceptible to noise interference than the baseband FM signal, theoretically the low bit-rate would tend to at least partially offset this susceptibility to noise, including multipath.  At the present time there is insufficient data available to predict the robustness of the RDBS signal to multipath from wind turbines.

Other in-band data services above 57kHz, which are referred to as Subsidiary Communications Authorization (SCA), have been and are being used.  These include reading services for the blind, private data services, commercial music services and others.  Little data is available regarding the robustness of SCA signals to multipath interference from wind turbines. 

Digital radio is now being transmitted by many FM broadcasters.  The FCC has authorized a hybrid system branded as “HD radio” (proprietary to iBiquity) in which digital signals are included in-band-on-channel (IBOC) with analog FM transmissions.  The use of audio compression enables low data rate (typically 90–100 kb/s), which does not stress the available bandwidth and tends toward good signal-to-noise.  The digital IBOC signals are low amplitude (10-20 dB below the analog carrier) but because the modulation method is OFDM (like European DVB-T), theoretically HD radio should be robust to multipath interference including that caused by wind turbines.  Since digital radio is still new there is not sufficient data on which to confirm digital radio’s robustness to windmill multipath. 

[ 1 ]         Influence of Echoes on Television Transmission, Pierre Mertz, Bell Telephone Laboratories, published by the Society of Motion Picture and Television Engineers, May 1953.

[ 2 ]         Data from  experiments in our laboratory dated October, 9th, 2010.

[ 3 ]         [An Empirical Comparative Study of Prediction Methods for Estimating Multipath Due to Signal Scattering from Wind Turbines on Digital TV Services, by I. Angulo et al, IEEE Transactions on Broadcasting, June, 2011, Volume 57, Number 2, page 195.

[ 4 ]         DTV Converter Box Program – Results and Lessons Learned, FCC Technical Research Branch, October, 2009

Charles W. Rhodes is a consultant in the field of television broadcast technologies and planning and is regular contributor to TV Technology Magazine.  A number of television papers by Mr. Rhodes have been published by the IEEE, many of which have received awards for excellence.