My colleague Linley Gumm and I recently conducted a series of measurements of the threshold protection (desired/undesired (D/U)) ratios required for modern ATSC receivers to reject interference from LTE signals radiated by base stations in the 600 MHz Band. We tested 21 receivers and forwarded our results to the International Telecommunications Union through Dr. Charles Einolf, a U.S. member of the ITU working group dealing with such interference issues. Our paper described how we conducted these tests and our results. This report is available from the ITU website www.itu.int/pub/R-REP-BR.2215-6-2016, entitled “Measurement of Protection Ratios and Overload Threshold for Broadcast TV Receivers.” Our contribution is Annex 3F.
Our problem was to find a source of “real, live LTE signals” for testing purposes. We found two such LTE signals “on-the-air” in the 700 MHz band with strong signals, each 10 MHz wide centered at 739 MHz and 751 MHz. These we heterodyned down into the 600 MHz band. Such signals in the 600 MHz band would consist of two 5 MHz blocks of contiguous frequencies, so we were emulating what we believe will be common practice in the 600 MHz band after the 600 MHz band is repacked following the spectrum auctions.
By now, it is generally accepted that ATSC reception may be jammed by pairs of undesired signals on Channels N+K and N+2K with the desired signal on Channel N or N+3 if third-order intermodulation (IM3) products are being generated in the frontend of the affected receivers.
Gumm proposed to heterodyne our LTE signals into Channels 18 MHz above and below the desired ATSC signal on Channel N. None of the third-order distortion products fall within the desired Channel N. The desired ATSC signal was being jammed. That is what we reported in Table 3 on page 8 of what today is “Annex 3F” to the ITU Report BT-2215. This data is given in the column marked No “IM3.” Without any IM3 in desired Channel N, we found that the threshold D/U ratio was, for 90 percent of the receivers, to perform –32 dB. Two LTE signals 18 MHz above and below the desired ATSC signal power cannot be more than 32 dB greater than the ATSC signal power. That is a long way from the –60 dB D/U threshold, which the FCC assumes for planning purposes. When our two LTE signals were centered on Channels N+K and N+2K, the undesired threshold was –28 dB, only 4 dB worse. This is shown for FCC Scenarios 7, 10 and 12 in the abovementioned table in Annex 3F. Even with a perfectly linear receiver front-end (which would not generate IM3), interference was only slightly reduced.
What could explain why we saw interference, which could not have been generated in the receiver? Well, we saw desensitization. All modern DTV receivers have wideband RF automatic gain control (AGC). Strong undesired signals on Channels N+/–3 forced the wideband RF AGC system to reduce the gain of the RF amplifier to protect the mixer from being overloaded. Table 1 summarizes our test results for the 21 modern ATSC receivers detailed in Annex 3F of the ITU Report.
Table 1: Threshold D/U (dB) for 21 modern ATSC Receivers from LTE signals
So the FCC is in error to believe that consumer DTV receivers, engineered to be linear at least up to –8 dBm, would eliminate ISIX (Inter-Service Interference). The only solution I know of is to provide RF selectivity before the mixer with a tracking filter. TV receivers used to have such a tracking filter between the RF amplifier and the mixer, but that was before tuners were fabricated as an IC (aka “tuner on a chip”). That means a tracking filter would have to be developed and I know of no such effort presently underway for the consumer DTV receiver market.
Solid-state or integrated circuit ATSC receiver front-ends were first described in 2007. By 2012 they were found in consumer TV receivers and are now found in all receivers currently in production, so the development of tracking filters for consumer TV receivers would probably take at least five years. However, there is no obvious need to develop such tracking filters at this time.
The FCC has not yet acknowledged that there will be interference to ATSC reception by base station emissions. Our test results suggest otherwise. The FCC claimed that the protection ratio for ATSC-ATSC interference would be –60 dB. That assumption was made in 1995 based upon laboratory data, which involved the only the prototype ATSC receiver available.
ANOTHER FINDING IN THE ITU REPORT
We noted that four of the 21 receivers tested behaved quite differently from the others. Tables A 2–9 show this difference. Receivers 35, 42, 45 and 49 stand out as having protection ratios as low as –28.8 dB as shown in Table 2. Similar observations have been reported in Europe. We believe these four receivers are adversely affected by the “dynamics” of LTE signals in actual use. Our tests were conducted with over-the-air LTE signals actually carrying traffic between base stations and subscribers. Tests by other laboratories have been conducted with a laboratory LTE signal generator. Their LTE signals do not provide for dynamic testing, so their test results are less representative of real-world conditions.
Table 2: The average threshold D/U for the four worst performing receivers
The LTE standard provides a feature called Automatic Power Control (APC). When a cell phone initiates a call, it uses its maximum allowed power and the responding base station replies at its maximum allowed power. The base station then determines the minimum power for both the calling cell phone and itself to continue communicating. APC minimizes interference between cell phones and base stations, but it causes heavy interference to some DTV receivers. Table 2 gives the threshold D/U (dB) for the worst four receivers. The dynamic range over which APC varies LTE signal power is up to 20 dB. ISIX occurs in random bursts, which makes it easier to identify, but sometimes these bursts are more than 20 seconds apart.
Our tests were unique in that we used real 10 MHz LTE signals carrying traffic, not LTE signal generators used in earlier tests. Laboratory LTE signal generators do not simulate the operation of APC. Perhaps that will change.
A single 5 MHz LTE signal is comparable to an ATSC signal as an undesired signal as the FCC has said. Our work comes far closer to what I expect is actual practice where multiple LTE 5 MHz blocks are deployed in many markets, DTV reception will be subject to unanticipated ISIX interference from nearby base stations when the 600 MHz band carries LTE signals. Such interference will occur where the receiver is receiving weak DTV signals and is near one or more base stations.
In short, I doubt that receiver manufacturers will design new receivers for the small market that would benefit from such redesigns. Over-the-air free television will become history for some people. Where and when TV translators vanish, an even larger number of people will also be forced to seek other media. Stay tuned.
Charles Rhodes is a consultant in the field of television broadcast technologies and planning. He can be reached via email at firstname.lastname@example.org.