ATSC document A/74: ATSC Recommended Practice: Receiver Performance Guidelines could be the basis for National Telecommunications and Information Administration performance standards for the 10 million federally subsidized digital-to-analog converters.
That document was never intended for such purposes. It appears to be a compendium of the performances of commercially available DTV receivers in the 2003-04 era, which has little bearing on what it takes to reliably receive DTV signals as envisioned by the FCC in its planning factors for DTV, (see OET Bulletin No. 69, revised).
Specifically, A/74 states the dynamic range of DTV signals is -83 to -8 dBm. However, the FCC used as planning factors a signal-to-noise ratio of 7 dB at UHF and 10 dB for both VHF bands. Based on the 7 dB receiver noise figure, the minimum usable received DTV power is -84 dBm, not -83 dBm. By inference, the ATSC document indicates that DTV receivers may not achieve the 7 dB noise figure.
Where DTV reception is noise-limited, a low noise pre-amplifier mounted on the rooftop directional antenna could more than make up for the weak signal in UHF and high VHF, but probably not in low VHF, where manmade noise picked up by the antenna dominates receiver-generated noise.
Alas, the cost of such a low-noise receiver exceeds the subsidy, so its benefits do not extend to folks in weak signal areas, unless these DTV converters are tested under weak signal conditions not anticipated by A/74. That document lists three desired signal powers: -68, -53 and -28 dBm, all of which are well within the dynamic range of terrestrial broadcast DTV signals and over which digital-to-analog converters should work.
The weak signal power, -68 dBm is 16 dB above the low end of the range. A DTV converter with an effective noise figure of 20 dB would work at -68 dBm, but not much below it, nowhere near -84 dBm.
DYNAMIC RANGE
A simple test to find the lower end of a receiver's dynamic range is to feed a DTV signal at 0 dBm into a calibrated attenuator set to -85 dB and then to the receiving device under test. Unless the device under test has a noise figure below 6 dB, it will not decode the data. Decrease the attenuation in 0.5 dB steps until it does decode. You have found the lower limit of its dynamic range.
The other end of the dynamic range of DTV signal power is limited by receiver overload. This is much more important. A/74, in Appendix D, Table D.1, gives the computed received power at an urban site in Southern Florida for all stations in that area. It is based on the FCC planning factors for DTV.
While the text of A/74 suggests the maximum received power is -8 dBm, Table D.1 gives higher values for several DTV signals, up to -5 dBm. This is average power. The text notes that multiple strong undesired signals may be present, which Table D.1 clearly illustrates. However, A/74 does not suggest how to calculate the combined effect of such multiple DTV signals.
DTV-to-DTV interference results from nonlinearity in the signal path. The distortion that results is related to the peak voltage seen by the nonlinear devices. When two or more DTV signals are present, it is their combined peak envelope power, not average power, that counts.
Convert power to RMS voltage squared across 75 ohms. Convert this to peak volts squared and take the square root (peak voltage per signal). Add these peak voltages, square the sum and divide by 75 ohms to get peak envelope power.
One strong DTV signal can overload a receiver. When overloaded, third-order intermodulation (IM3) products are generated. It is well known that some of these fall in the first adjacent channels. The rest of the IM3 is within the DTV channel masked by the DTV signal. These IM3 raise the noise floor under the desired signal. If the noise floor is 15.2 dB below the desired signal, reception fails.
Dr. Oded Bendov has shown the desired signal is subject to cross-modulation that results from the same third-order nonlinearity that produces IM3. Where there is IM3, there will also be cross-modulation.
The upper end of the dynamic range of DTV signals that a given receiver can handle is set by the combined IM3 and cross-modulation in the channel. Dr. Bendov provided such data in "Transactions on Broadcasting," Vol. 51, March 2005. He will also be presenting a DTV interference paper at the IEEE Broadcast Symposium Sept. 28-30 in Washington, D.C. His paper will be published in the December issue of "Transactions."
A simple test also will determine the upper limit of a receiver's dynamic range. A DTV signal at 0 dBm is fed through a calibrated wideband attenuator to the receiving device under test. With 0 dB of attenuation, the receiver sees 0 dBm and will probably not be able to decode the DTV data stream.
Increase the attenuation in 0.5 dB steps until the receiver is able to decode the DTV data stream, and the upper limit is established. Table D.1 indicates the strongest DTV received signal power is -5 dBm. DTV receiving devices should be able to handle a signal up to -5 dBm. I believe the dynamic range of DTV receiving devices, including DTV converters, should be -84 dBm to at least -5 dBm. Over this entire range, such devices should work unless the TV service is interference-limited within the noise-limited coverage area of the station.
CHANNEL INTERFERENCE
While we usually think of co-channel interference only near the edge of a station's noise-limited coverage area, there are two exceptions.
In the first, co-channel interference may result within a station's coverage area if it employs multiple transmitters operating synchronously, as in distributed transmission. If the adaptive channel equalizer of the receiving device cannot treat the weaker signals as ghosts, they act as a noise source within the channel. By varying the delay between the stronger and weaker versions of the same DTV signal, one can quickly find the range over which its dynamic channel equalizer works.
The second situation is where a strong DTV signal overloads the receiver, generating enough distortion products within the channel that the data cannot be decoded.
It is now well accepted that adjacent channel DTV interference is due to receiver nonlinearity. Under moderate to strong received signal power, the receiver tuner is overloaded.
This was first published by Gary Sgrignoli in 2003 and elaborated upon by Dr. Bendov and myself in our recent papers published in IEEE Transactions on Broadcasting.
Sgrignoli, Bendov and I reported that the total sideband splatter radiated into each first adjacent channel by a legal DTV transmitter is 44.5 dB below the ERP radiated within its channel. This received sideband splatter is attenuated by about 2 dB in the receiver so it is 46.5 dB below the undesired signal being received.
This is noise within the desired channel. As Sgrignoli reported, DTV reception fails when the desired-to-undesired ratio is -31.3 dB, because the received splatter increases the in-channel noise, lowering the signal-to-noise ratio to 15.2 dB.
The FCC established a desired-to-undesired ratio for DTV-to-DTV adjacent channel interference to -27:+/-1 dB in 1998. There isn't a lot of headroom between -31.3 dB and -27 dB for any additional noise within the desired channel. In fact, where undesired, or U = -5 dBm, the desired (D) signal must exceed -36 dBm for a perfectly linear receiving device. At D = -36 dBm, the receiving device would have to have a third-order intercept power (IP3) above 24 dBm which I don't think is feasible. Fortunately at D = -30 dBm, IP3 of 15.7 dBm suffices, and this is a value that can and should be provided.
The amount of additional in-channel noise that can be tolerated depends on the strength of the received desired signal and the undesired DTV signal power on one or both adjacent channels. While the FCC has one D/U ratio for lower and one for upper adjacent channel interference, it still does not recognize that there is a maximum U for each D level.
A/74 gives two values for D/U for ACI between DTV signals for D=-68, -53 and -28 dBm. These are the D levels at which the ATSC directed testing to be carried out in 1995. Those tests were designed to choose between competing modulation schemes. As each modulation scheme was tested with its own proprietary demodulator, we tested at signal levels well above where those differences in tuners such as noise figure and voltage standing wave ratio could mask the results.
Testing at -68 dBm cannot indicate performance between -84 and -69 dBm, as below -70 dBm, the effect of receiver-generated noise comes into play. Testing at -28 dBm will not provide useful information about receiving device performance between -27 dBm and -5 dBm.
Manufacturers would welcome such restricted testing, but this is not in the public interest. It is in the interest of broadcasters that the market is not flooded with DTV receiving devices that do not operate over the wide dynamic range unique to terrestrial DTV broadcasting. It is time for the broadcasters to speak up.
The NTIA deadline for filing comments on its notice is Sept. 22. There is much yet to be said about testing DTV converters using A/74. Next month, that will be my topic and I hope the NTIA will extend its deadline for comments so this critical issue can be treated as it deserves.
