The Fm Second-Adjacent Protection Rule

COMMENTS ON

PROPOSED CHANGES TO

THE FM SECOND-ADJACENT PROTECTION RULE

1.BACKGROUND

A B-TAC sub-committee has been reviewing the protection requirements for FM stations utilizing 2nd-adjacent channels, with a view to relaxing the current Desired-to Undesired (D/U) ratio of –20 dB to a new value of –40 dB. As co-siting of 2nd-adjacent stations is not on the table at this time, this change would simply allow stations having this frequency relationship to be sited somewhat closer together geographically, while still ensuring that the protected service contour of one station does not enclose the transmitter site of the other.

The stated intent of this proposal is to increase the availability of drop-in allotments in radio markets where new services may be required but where vacant FM allotments are either scarce or non-existent.

2.FM RECEIVER PERFORMANCE IMPROVEMENT: FACT OR MYTH?

In FM broadcasting, the rules governing the geographical spacing of stations operating on adjacent channels are a function of the RF front-end and IF selectivity of “typical” consumer receivers[1]. It has been suggested by some broadcast engineers that the 2nd-adjacent signal rejection characteristics of modern receivers have improved substantially since the original rules were set and that a relaxation of the spacing rules is therefore overdue.

Receiver design changes may not have been all that beneficial

When the receiver improvement hypothesis is examined in detail, it begins to look more like an “urban legend” rather than an engineering fact. When FM radios employed tubes, discrete components and variable-capacitor tuning, they generally had tuned RF front ends that provided some adjacent-channel selectivity. Most low to mid-tier solid-state receivers now on the market have no front-end selectivity at all. Adjacent-channel rejection is provided solely by their double-tuned IF transformers. Moreover, they have AFC circuits that are extremely wide, in order to compensate for the fact that there is no protection against thermal or mechanical drifting by their local oscillators.

FM Receiver performance has in fact been assessed regularly

Those who are promoting relaxation of the 2nd-adjacent protection rules always seem to point out that the current rules were set decades ago, leaving the uninformed to infer that no one has ever re-assessed FM receiver performance in the meantime. In fact, over the past twenty years, considerable efforts have been made to quantify changes in FM receiver performance.

US receiver tests were done in both 1981 and 1999

In 1981, tests on a wide variety of consumer receivers were conducted by the National Radio Systems Committee (NRSC). The published results stated that a 2nd-adjacent D/U ratio of at least –26 dB was necessary for unimpaired reception. They recommended that a minimum protection ratio of –25 dB be utilized, in order to ensure that an audio signal/noise ratio of 50 dB could always be achieved for stereophonic reception[2]. They made this proposal notwithstanding the fact that the FCC rules, even at that time, required a D/U of only –40 dB for 2nd-adjacent stations.

Again in 1999, comprehensive receiver testing was performed by a number of groups in the US, in response to an FCC initiative to increase the number of LPFM assignments by reducing or eliminating certain adjacent-channel spacing requirements. The NAB’s receiver tests were particularly thorough. In its comments under FCC Docket 99-25, the association said:

“Our test data clearly shows that the alleged “vast improvements in receiver technology” cited by the Commission ... are, in fact, a myth.”

The association goes on in its report to state:

“Clearly, there has been no improvement in this aspect of receiver performance because the just completed NAB tests show today’s radios to be interfered with at slightly lower adjacent channel signal levels than the radios tested by the NRSC in 1981[3].”

Industry Canada’s receiver tests are the most current

Most importantly, however, Industry Canada commissioned receiver tests in 2001 (with dual interferors) and 2002 (with a single interferor). These are the results on which the comments in this submission are based.

ASSESSMENT OF TEST RESULTS

The following comments relate to the methodology and test results for the 2001/2002 Industry Canada receiver tests.

Including car radios in static-performance lab tests was a mistake

One of the problems with the Industry Canada tests is that they included a fairly large representation of after-market auto receivers in the sample. The tests that were conducted in 2001 utilized car radios for 21% of the sample, while in the 2002 tests auto radios comprised 31%. This seems to be disproportionately high, especially for the 2002 sample. As the adjacent-channel selectivity of auto radios is known to be far superior to that of all other types, this tends to skew the overall results, suggesting median receiver performance that is better than it would be if a more statistically-valid sample had been used.

While it may well be desirable to test the performance of auto receivers, it is not appropriate to use static lab testing techniques to do this. The high adjacent-channel selectivity that is typical of car receivers has become necessary because moving receivers are subjected to very wide variations in both desired and undesired signals as they travel along roads and highways. The static lab tests that Industry Canada conducted are not indicative of the environment in which these receivers actually operate and so this data should not be lumped in with the results obtained for fixed receivers.

To be meaningful, selectivity tests for auto receivers should be done by simulating what really happens when a receiver is in motion. This requires simultaneous random signal level shifting, within prescribed ranges, of both the desired and undesired signals.

For example, let us assume that in static tests median car radios can accept a D/U of –60 dB before exhibiting a reduction in desired signal S/N below acceptable levels. Let us further assume that, as vehicles move around within a station’s service secondary area (between the 70dBu and 54 dBu contours), the desired and undesired fields that they encounter will fluctuate independently within a range of ±15 dB[4]. In this case, then, a margin of 30 dB must be subtracted from the nominal static D/U in order to account for vehicle movement. This brings us to a required 2nd-adjacent D/U value of –30 dB for regulatory purposes for auto receivers taken alone.

If this correction is not made, and if static tests on car receivers are included in determining the required D/U ratio for regulatory purposes, the car radio data will skew the overall median receiver values. This can result in erroneous conclusions about the protection levels that are required by the majority of receivers.

The proposition that car radio test data should be treated with extreme caution when deriving service rules is discussed in an academic paper that reviewed four studies on FM receiver adjacent-channel immunity conducted in connection with the FCC’s 1999 rule-making proceeding[5]. In this paper, the authors state:

“All four studies made the mistake of reporting the results of tests of car radios along with other radios. Every report showed that car radios outperform other radios with respect to adjacent channel interference rejection. In fact, car radios need this capability if they are to provide reasonable performance. At the very least, car radios should have been tested and reported separately. Including car radio performance in these studies provides misleading signals about the overall performance of receiving systems.”

Dual interferors must be considered in the real world

If valid decisions are to be made concerning protection levels required in realistic situations, it is appropriate to evaluate 2nd-adjacent interference with signals on both sides, since this situation is likely to occur frequently. This is because the Canadian FM Allotment Plan often contains 4th-adjacent allotments in the same market. Therefore, an adjacent community is highly likely to have been allocated a frequency that is 2nd-adjacent to both of these neighbouring allotments.

The impact of the second interferor cannot be ignored. Appendix 1 compares the median receiver results obtained from the 2001 and 2002 Industry Canada tests, in relation to 2nd-adjacent rejection for each category of receiver[6]. It also compares the overall median results when the Category 4 auto receivers are included and when they are excluded. At the lowest measured desired signal levels (54 dBu), a

–40 dB D/U value for dual interferors can only be supported when auto receivers constitute 31% of the sample. When auto receivers are removed from the sample, the D/U median value drops to –25 dB at the lowest desired signal value. Even at this level, increased interference could be expected for Category 1 and 3 receivers located near the protected contour.

Adopting D/U = -40 dB will reduce interference-free service areas

Considering the above, it becomes apparent that, if the FM rules were modified to allow 2nd-adjacent assignments to protect the 54 dBu service contours of existing stations to a D/U ratio of -40 dB, the net result would be to limit interference-free reception to about the 59 dBu contour of desired stations.

This is demonstrated in Table 1, the entries for which are based on two Class C1 2nd-adjacent stations separated by 100 km. This separation produces a D/U of –40 dB at the protected contour of the desired station.

Table 1.

Appendix 1 says that a D/U ratio of –25 dB is necessary to avoid interference on the median non-auto receiver at the lowest signal level that was tested, in the presence of dual interferors. Row 2 of Table 2 shows this situation occurring when the test case D signal is 59.2 dBu and the U signal is 84.5 dBu.

Therefore, adopting a –40 dB 2nd-adjacent protection rule would be equivalent to reducing the protected contour for existing stations from 54 dBu to 59 dBu[7]. As can be seen from Table 1, this would reduce the radius of protected coverage for Class C1 stations from 86 km to about 74 km[8].

The main beneficiary of the proposed rule change would be smaller stations adjacent to larger markets

Reducing the D/U ratio to –40 dB would likely only benefit lower-class 2nd-adjacent stations located in smaller markets adjacent to existing Class C1 stations in the major markets. Whereas the latter have no “head-room” left to increase radiation towards their neighbours, the smaller stations may be free to do so.

Examples of this can be found in and around Toronto. CICZ-FM Midland is a Class B station that is 2nd-adjacent to CHUM-FM. The Midland station’s 74 dBu contour is now flush against the CHUM-FM protected 86 km radius. If the protection rule were to change to –40 dB, then CICZ-FM may have the opportunity to increase radiation southward toward Toronto. On the other hand, CHUM-FM has no opportunity to make an offsetting increase northwards, since it is already operating at maximum Class C1 parameters.

Other examples that could have the same outcome in the Toronto area include:

CJLF-FM Barrie Ch 262B vs. CKFM-FM Ch 260C1
CKEY-FM Niagara Ch 266B vs. CHIN-FM Ch 264C1
CKMB-FM Barrie Ch 298C1 vs. CILQ-FM Ch 296C1

CKMS-FM Waterloo Ch 262A1 vs. CKFM-FM Ch 260C1
CKGE-FM Oshawa Ch 235B vs. CING-FM Ch 237C1

The rule change would not produce many new drop-in assignments where they are really needed

Empirical evidence suggests strongly that the proposed –40 dB rule would not produce many brand new drop-in channels in locations where they are most needed. Since the 2nd-adjacent channels are usually already occupied in areas abutting the larger markets, the most probable result of the rule change would be to allow stations on the periphery of major markets to improve their penetration of larger adjacent CMAs.

Given the potential impact that the proposed rule change could have on existing stations, it would be inappropriate for Industry Canada to proceed with the change, based solely on an assumption that shortening the required 2nd-adjacent spacing would automatically liberate useful new FM frequencies in places where they are needed. This would have to be tested by examining a number of test markets and determining where new frequencies could actually be added.

Most areas of frequency congestion in Canada have already been studied extensively by a number of very experienced engineering consultants. It is probably unreasonable to expect that any new comprehensive study would be fruitful, especially if the analysis takes into account the interference limitations that any new drop-ins would be obliged to accept from distant co-channel and 1st-adjacent stations.

Moreover, while it may be possible to drop some small channels into the plan without affecting related stations, the practicability of these frequencies would be severely constrained by incoming interference.

Although experienced broadcasters usually understand the effect that these limitations can impose on their future operations, the practicability factor is often ignored or overlooked by smaller community, campus and ethnic licensees. Industry Canada may like to think that the “caveat emptor” principle applies in such cases. But in reality, the end result of licensing very marginal channels is often an endless litany of complaints from the inexperienced broadcasters who end up occupying them, as well as from the listeners they are supposed to be serving. This is especially a problem when new licences are issued for niche services that are supposed to provide programming unavailable elsewhere (e.g. smaller ethnic groups that are not served by the mainline ethnic broadcasters). Often these audiences are widely dispersed within CMAs and may not be well served by frequencies that have severely constrained coverage capabilities.

Canada’s situation cannot be compared with that of the US

Some observers have also suggested that Canada should adopt a –40 dB D/U value for 2nd-adjacent protections because that is the value specified by the FCC for domestic use in the USA.

In the US, the majority of FM stations operate in large urban agglomerations with high population densities. Given the limited number of available FM channels, it may have been appropriate for the FCC to restrict interference-free service to a radius of 65-72 km[9].

However, due to the lower population densities and smaller urban agglomerations in Canada, there is a greater requirement in this country to ensure that FM services can be received without interference up to the point where the signal strength drops below that required to produce a reasonable signal-to-noise ratio for stereo reception. In other words, we have a greater need in Canada to noise-limit our FM stations, as opposed to interference-limiting them, as the US does.

It is also useful to remember that the US -40 dB 2nd-adjacent protection rule for commercial stations has been in place for many years and radio listening patterns have developed in the presence of whatever extra interference this produces. When existing services become subject to new interferences, there is a much higher probability of listener protests than is the case when a desired service has experienced such interference from the day it first went on the air.

Any benefit of a rule change would be diluted by US action in border areas

As well, if Canada were to adopt a D/U ratio of –40 dB, consistent with the US domestic rules, this would inevitably prompt the FCC to seek an amendment to the Canada-US FM Agreement to reflect this change. After all, there would be little sense in retaining a D/U criterion of –20 dB in the Canada-US Agreement if both countries are using a value of –40 dB domestically. Assuming US stations would then be free to radiate up to 20 dB more power toward Canadian 2nd-adjacent protections, it is reasonable to expect that some of the benefit of this Canadian change in border areas would actually accrue to US stations, rather than Canadian licensees.

SUMMARY AND CONCLUSIONS

The conclusions that are developed in this paper are as follows:

Receiver tests performed in the past two decades have not demonstrated that FM receiver selectivity has improved substantially since the original allotment rules were established.

In today’s congested radio environment, it is reasonable to assume that FM stations will be subjected to both upper and lower 2nd-adjacent interference.

Selectivity test results for auto receivers, obtained during the recent Industry Canada tests, should be excluded from receiver performance results unless they are corrected to account for the dynamic changes in D and U signals levels that that occur when vehicles are in motion.

All recent Canadian and US receiver tests suggest that a 2nd-adjacent D/U ratio of –40 dB is not supportable. At best, a D/U of about –25 dB could be justified and even this would result in increased interference to Category 1 and 3 receivers near the protected contour.

If a –40 dB rule is adopted, there will be an erosion of the protected contours of full-parameter stations in major markets, such that interference will be increased between the 59 dBu and 54 dBu contours of desired stations.

The main beneficiaries of a relaxed 2nd-adjacent rule would be smaller stations located in markets adjacent to large CMAs, which might be able to increase their radiation toward the neighbouring major markets.

Very few brand new drop-in assignments would result from the rule change, in locations where frequencies are in short supply. Where such frequencies can be found, they will often be subject to severe incoming interference limitations and may not be suitable for serving specialized (e.g. ethnic) audiences that are often spread over a wide area.

Changing Canada’s domestic rule on 2nd-adjacent protections will result in pressure from the FCC to modify the Canada-US FM Agreement accordingly. Loosening the protection rules will allow border US stations more flexibility to increase their own parameters, thus diluting any salutary benefit of the change in Canada.

W.A. Stacey, P.Eng.

CAB Technical Advisor

3 June 2002

APPENDIX 1 Median Data for Charts (All Receivers)

Median Data for Charts (Receivers by Category)