Modulation monitors are used to assess the performance of the entire broadcast transmission system and should therefore be considered vital rather than expensive and unnecessary. For competitive reasons it is mandatory that performance of the entire broadcast transmission path be maintained as high as possible. Since FCC deregulation of monitors, and even before, it has been tempting to measure modulation with a spectrum analyzer or calibrated tuner and oscilloscope, avoiding the noise and overshoot present in many monitors. However, spectrum analyzers cannot be used to measure deviation since there is no simple relationship between the deviation and RF spectrum under normal program modulation conditions. Spectrum analyzers can be used for reference calibration using the Bessel function null technique for a peak deviation reference. This involves modulation with sine waves of specific frequencies. Furthermore, even the highest performance wideband tuners do not have adequate bandwidth to measure modulation accurately. A true monitor must be used to measure modulation, but it must be designed correctly in order to be useful.

If aggressive, competitive audio processing is used, accurate transient response through the entire transmission path including the modulation monitor is essential. The monitor peak indicator circuitry must also be transient-accurate to assure correct indications. If the monitor has poor transient response, it is useless as a measuring instrument and reference standard. The more aggressive the audio processing, the greater the transient accuracy required from the monitor. Accurate transient response in the monitor allows transient response degradation elsewhere in the system to be detected.

Peak indicator transient accuracy is a specification that was never required for FCC type-approval. Nevertheless, maximum loudness can only be achieved if the monitor permits maximum legal modulation by indicating dynamic peak modulation accurately. If the monitor falsely indicates higher peak modulation than that actually occurring, more aggressive audio processing must be unnecessarily used to achieve a given loudness level in an attempt to reduce peak-to-average ratio. If the transient response is improved, processing levels can go down and loudness can be maintained, or loudness can be increased for a given level of processing.

Audio processing ordinarily includes peak control to prevent overmodulation. Maximum loudness is only obtainable if peaks are controlled tightly and accurately. This is only achieved if the transient response of the entire signal path after the final limiters, including the monitor, and monitor peak indicator circuitry, does not increase the peak level significantly. Many overshoot problems have been blamed on high frequency ringing, when in fact, the overshoot is a result of low frequency tilt. The entire signal path after the final limiters must introduce less than 1% low frequency tilt and/or high frequency ringing, and band-limit the signal to 15 kHz. Total energy above 15 kHz must be better than -40 dB below 100% modulation. Significant energy above 15 kHz will cause crosstalk, producing aliasing distortion which is audibly offensive, since it is not harmonically related to the signals which cause it. If the left and right inputs to a correctly designed stereo generator are band-limited to 15 kHz, the stereo generator will produce an output which is tightly band-limited to 53 kHz. Provided that static stereo separation is good, only good dynamic crosstalk assures good dynamic stereo separation, and impresses your friends.

The FM modulator in the exciter or composite STL, if used, must be free of bounce due to underdamping of the AFC loop. An improperly designed AFC loop can intermodulate with the composite signal affecting low frequency response causing overshoot, reducing loudness capability and audio definition. The AFC loop design is the primary reason for sonic differences between various FM modulators.

In an attempt to assess system overshoot, it is invalid to apply non-band-limited square waves to a composite STL or to the input of a stereo generator which does not include 15 kHz low-pass filters. Such square waves contain harmonics outside the valid frequency range. Any test signals must be band-limited to 15 kHz for stereo generator inputs or 100 kHz for composite STL inputs in order to assess system overshoot. It is similarly invalid to evaluate modulation monitor transient response by exciting the monitor with square waves which have significant energy above 100 kHz. Non band-limited square waves can only be used for the purpose of determining if a system is linear phase. If no ringing exists, or ringing exhibits pre and post overshoot, the system is linear phase. An STL, exciter, or modulation monitor may produce ringing on non-band-limited square waves, but produce ABSOLUTELY NO OVERSHOOT on legitimate baseband signals. Most current-generation audio processing systems provide more than enough dynamic band-limiting to prevent the composite low-pass filter in the monitor or composite STL from overshooting significantly, even when the composite stereo signal is further clipped by several db. However, because they introduce ãsplatterä into the baseband, most current-generation audio processing systems and all composite limiters DO NOT sufficiently dynamically band-limit the audio signal to fully realize the dynamic stereo separation capabilities of the stereo generator or provide sufficient sub-carrier protection.

Most FM modulation monitors use a pulse-count FM demodulator, since it is intrinsically linear without tuned circuits. The output of such a demodulator contains a pulse train whose duty cycle changes in proportion to the FM modulation. To recover this modulation, this pulse train must be passed through a low-pass filter to remove the pulses while retaining the recovered composite baseband. Since the typical pulse frequency is around 700 kHz and the composite baseband extends to 100 kHz, the filter must be reasonably aggressive to provide flat, linear-phase response to 100 kHz while simultaneously suppressing the 700 kHz and associated FM sideband components below the noise level of the demodulator. Noise and adjacencies can cause interference just above the composite baseband region, affecting peak indicator accuracy. For example, with a 700 kHz IF, an upper second adjacency can cause a 300 kHz [700 kHz (IF) - 400 kHz (difference between monitored channel and second adjacency)] component to appear in the baseband. It is therefore desirable to use a filter which has a cutoff frequency as low as feasible given the 0 - 100 kHz linear-phase requirements. The use of phase equalization permits the filter to be substantially more selective without compromising constant group delay performance in the passband. If the cutoff frequency of the filter is increased in an attempt to place the 0 - 53 kHz stereo baseband within the linear-phase region of the filter, without phase equalization, the noise rejection will suffer accordingly. Peak indicator accuracy can be degraded, especially under remote monitoring conditions. Most current-generation monitors employ phase equalization, although not all. Assuming everything else in the system is ideal the stereo separation performance of the pulse-count low-pass filter may be calculated from the expression:




Where:
S = Separation in dB
A = SUB/MAIN = (L-R)/(L+R) = Amplitude Ratio
Ø = Phase Error in Degrees Deviation from Linear Phase

Remote monitoring of the FM signal requires the use of a receiving antenna. Since this is where the received signal originates, the signal must be clean and multipath-free for high accuracy monitoring. Because multipath distortion depends on the ratio between the desired signal and undesired reflected signal, multipath distortion can exist regardless of proximity of the receiving antenna to the transmitting antenna. It is therefore desirable to use a directional receiving antenna, even under strong signal conditions. VHF rabbit-ears will not suffice. Folded dipoles will sometimes work, but beware of a dipole that can be folded. The output impedance of the antenna should match the transmission line impedance as closely as possible across the frequency range to be monitored. Many receiving antennas have extremely poor impedance characteristics and can compromise monitor accuracy. The transmission line from the antenna must match the input impedance of the succeeding RF input stages of the monitor. Maintaining low VSWR in the line is just as important in receiving a signal as in transmitting one. Low VSWR will minimize received distortion, preserve transmitted stereo separation, and maintain subcarrier performance.

Remote monitoring of the FM signal also requires using an RF pre-selector. The bandwidth of a properly designed pre-selector IF is sufficiently wide to cause NO ADDITIONAL OVERSHOOT of the recovered composite signal, provided that the received signal is multipath-free and does not contain any spurious signals. Spurious signals can be received through an IF which is too wide or can be generated within an incorrectly designed pre-selector. Either case can degrade peak indicator accuracy and can contaminate the monitor audio output with unpleasant sounds. There is a trade-off between pre-selector accuracy and adjacent channel rejection. It is impossible for a pre-selector to provide significant rejection of first adjacencies while monitoring the desired channel accurately. This is the very reason that high quality tuners cannot be used for certain high accuracy tests. Tuner performance is optimized toward adjacent channel rejection to reduce interference. The most accurate measurements are made by eliminating any pre-selector or bandwidth-limiting, if possible, and connecting the monitor directly to the transmitter sample. Not all monitors permit bypassing of bandwidth limiting elements, even for high level RF monitoring.

The remote FM monitor must have sufficient rejection of second and higher adjacencies and must not reduce significant FM sidebands. Remote monitoring can be inaccurate if the pre-selector bandwidth is so narrow that significant RF sideband truncation occurs, or is so wide that the monitor suffers from interference from adjacent channels. Although various references disagree about the required bandwidth, it is reasonable to state that an IF bandwidth of at least 565 kHz is required for highest quality professional monitoring, given the 100 kHz baseband limit. This is given by the expression:

B if = 2(Fd + 2Fm)

Where:
B if = Bandwidth of IF
Fd = Peak Deviation of System
Fm = Highest Baseband Frequency

Unfortunately this is an excessively high value to use where the spectrum is crowded. In this case, compromises are inevitable for the purposes of selectivity.

The ideal IF filter has constant group delay in the passband and infinitely steep skirts. If the IF filter is phase-equalized, then a sharper cutoff can be used to reduce noise and interference from adjacent channels. The phase equalization will provide a constant group delay characteristic, thereby achieving good distortion, stereo separation, and subcarrier performance. Conventional L-C filters must be phase equalized to approximate this performance. Other technologies such as surface acoustic wave (SAW) devices may be superior, since they can be designed with intrinsically constant group delay without the addition of phase equalizers.

Since the selectivity compromise renders the demodulator susceptible to noise and interference from adjacencies, these adjacencies can produce noise in the demodulated output. This noise can be removed by appropriate design of the pulse-count demodulator low-pass filter, in essence providing additional selectivity by means of composite filtering rather than IF filtering. Filtering after FM demodulation cannot cause IF spectrum truncation. Because FM demodulation is non-linear, IF and composite filtering are not equivalent and the composite filter will not remove the low frequency intermodulation distortion products caused by the adjacencies. The composite low-pass filter therefore supplements the IF filtering instead of replacing it.

To be accurate, an FM modulation monitor must be designed with due consideration to all these factors. Neglect of any of these factors can compromise performance, as can be witnessed by the performance of many commercially available FM modulation monitors. As the competitive ãloudness warä marches on, the first place to start is with HIGH ACCURACY MONITORING.

The Modulation Index modification program for some of the lower-accuracy monitors creates monitors with transient accuracy of 1% or better even on the most aggressively processed composite waveforms. This modification replaces the original demodulator low-pass filter with a computer-optimized linear-phase design. This filter improves rejection of noise and interference from adjacent channels, since the use of phase equalization allows a lower cutoff frequency.

Accurate low frequency pulse response is obtained through the use of an integrator feedback loop. Monitor-induced overshoot is essentially eliminated, permitting accurate reading of complex peak modulation. This is achieved without covering up poor monitor design with add-on circuits to stop spikes from firing the peak indicators. Such circuits can cause inaccurate, potentially illegal modulation measurement. If the waveform at the input to such a circuit has been distorted by inaccuracies earlier in the monitor, the peak indicator will be forced to ignore the overshoots caused by such inaccuracies. Since this circuit cannot differentiate between overshoot produced in the monitor and that produced elsewhere in the transmission system, the monitor is insensitive to real overshoots produced elsewhere in the transmission system, thus defeating one of the most important monitor functions; measurement accuracy. The monitor can therefore ignore true overmodulation and not protect against potential FCC citations.

In addition, the modification adds temperature compensation to the peak indicator reference. Stereo demodulator performance is maximized by assuring essentially constant group delay throughout the frequency range of the stereo baseband.

Because the FCC has deregulated modulation monitors, these modifications DO NOT VIOLATE FCC RULES. Prior to deregulation, modifications and technological advances which improved performance could not be incorporated in type-approved equipment without re-type-approval. This included any type of add-on circuitry.

The Modulation Index modifications permit more accurate compliance with FCC peak modulation rules. The modification yields performance far superior to that formerly required by the FCC. Baseband frequency response is accurate to 100 kHz, making the modified monitor fully compatible with deregulated subcarrier services. This modification takes the guesswork out of modulation measurement. It allows efficient use of the legal FM modulation limit, allowing MAXIMUM LOUDNESS!!