The R.F. isolator is popularly described as a circulator that is equipped with a suitable load termination. The device has the characteristic of conducting radio frequency power in one direction with minimal attenuation and providing a high loss in the opposite direction. The isolator is designed to operate over a range of frequencies in the manner described and to exhibit a design impedance at its input and output ports.

Single junction isolators provide less than 0.25 dB of attenuation in the forward direction and 35 dB of attenuation in the reverse direction. Design impedance is 50 ohms, with a return loss of 30 dB (VSWR of 1.05:1 or better) at input and output ports. Two standard ranges of power handling are provided; up to 125 watts continuous power handling and up to 250 watts power handling. A range of load terminations rated from 10 to300 watts are available. Dual junction isolators have a nominal forward loss of 0.5dB or lower and isolation (reverse attenuation) exceeding 75 dB. All isolators are thermally compensated for stable operation over a wide ambient temperature range under continuous duty conditions.

Isolators are manufactured for the FM broadcast band of 88-108 MHz and in the 944-960 MHz range for studio – transmitter link (STL) applications. Operationally, the isolator permits transmitter power to be coupled into an antenna system or between transmitter stages with very low signal power loss and to attenuate any in-band signal from being coupled back to the transmitter power amplifier stage by up to 35 dB for a single stage unit or up to 75 dB for a dual stage unit.

In translator system application, the isolator prevents signals from other translators or nearby high power FM transmitters from reaching the P.A. stage of the transmitter where mixing can take place, producing undesired intermodulation products that would otherwise be radiated.

As interstage filters, the typical benefit is to prevent changes in vacuum tube amplifier grid circuit characteristics from affecting the loading of a prior vacuum tube or solid state exciter stage. The exciter stage is thus protected from damage should the P.A. tube fail or degrade in characteristics as well as assuring a constant impedance load for linear exciter performance.

Filters, providing second harmonic or low pass characteristics are provided to insure that conducted or regenerated harmonic energy is suppressed sufficiently to preclude the generation of intermodulation products external to the transmitter stage.

Since the standard design impedance of EMR isolators is 50 ohms, should the system be designed for 72 ohms, impedance transformation devices are available as required.

In the typical translator application, we find two or more translators operating with their antennas in close proximity, or one or more translator transmit antennas in close proximity to a high power F.M. transmitting antenna. Either a single or dual junction isolator equipped with suitable load terminations and a 2nd harmonic or bandpass filter will effectively reduce the coupled power to each translator P.A. stage to reduce generated I.M. to a legal or acceptable level.

The isolator and filter are placed in the feedline between the translator output and its antenna. Special isolator models are available for installation inside the transmitter housing of many models of 10 watt translators.

Dollars and Sense in Isolator Selection

There are several things to consider in making a choice of isolators. These include:

1. The power output of each transmitter that is to feed an isolator.

2. The amount of isolation actually needed between the antenna and each transmitter, or the amount of isolation required between transmitters being, combined in a transmitter combiner.

3. Transmitter maximum duty cycles. If more than 75%, use an isolator rated for continuous duty.

4. The probable worst case reflected power to be returned to the isolator’s output port. This will determine the ratio, of the load termination of the last isolator stage.

5. Ambient heat that may exist in the area in which the isolator is installed.

We find that most isolators on today’s market are honestly rated as to power input capability, however, it has been shown that some makes become marginal in performance or drift under continuous power at their maximum power input capability and at high ambient temperatures. As was stated earlier. the combination of heat due to power lost through insertion loss plus environmental heat resulting from dissipation of other devices such as transmitter power supplies, power amplifier heat sinks. etc. must be considered in the temperature compensation of the ferrite, magnets and other elements that make up the isolator. For these reasons, all of the five considerations above must be taken into account. To provide its assigned tasks, e.g.: maintain a constant input impedance, low insertion loss and required system isolation over the ranges of from zero to 100% transmitter duty cycle and throughout the ranges of ambient temperature and reflected power an isolator must remain electrically stable. This electrical stability can only be provided by careful component choices, design choices and the application of techniques chosen to “tune out” the drifts of parameters of all components used in the device.

Heat is the natural enemy of circulators and isolators! The dynamic characteristics of all ferrite materials change as a function of temperature. Without compensation, the resonances of these materials changes in frequency as the ambient temperature ranges increase and from heating of the material resulting from applied R.F. power. The fixed magnets also change in coercive force as temperature changes, generally opposite from the ferrite in detuning effect. Were these effects equal and opposite, the overall performance would be self compensated, but this is never the case since characteristic changes are neither linear nor “mirror imaged” over wide temperature ranges. To accomplish full thermal compensation, careful selection of ferrite types and magnet characteristics, along with studies of dynamic heat migration in all parts of the assemblies must be made along with the application of special materials having, negative or positive magnetic flux conductivity as a function of temperature. Exhaustive laboratory tests and adjustments must be employed to thermally stabilize a given isolator design after its electrical characteristics have been verified. The entire range of ambient temperatures must be addressed along with R.F. duty cycles from zero to continuous power application at power levels from 10% to full rated power of the device. This requires many hours of testing and optimization in order that all possible conditions of field application may be simulated and successfully compensated.

We have dwelled on temperature compensation for a very good reason. In every case drift, in isolators, results in a change in the return loss at all ports. If an isolator drifts due to ambient heating, or heating as the result of conducted power loss, port impedance chances. Unless adequately compensated these shifts in impedance result in higher losses and further heating, eventually causing a runaway condition and failure of the device. Moreover, since insertion losses and isolations will both change, erratic effective system radiated power and decreased I.M. rejection will result. To produce uniform system performance, any isolator must be properly temperature compensated.


Short lengths of line function as linear transformers. A 1/4 wavelength (electrical length) of line translates a mismatch back to it’s other end which is 90 degrees displaced in phase. An inductive phase angle through a 1/4 wave line length will translate to a capacitive angle of equal magnitude. A true 1/2 wave line length will repeat the mismatch, since the wave front rotates 180 degrees along its length. In coordinated system designs, such as transmitter combiners, we use “trick” line lengths, adjusted to correct inductive or capacitive loads to resistive characteristics. The true length of cable must take into account connector lengths, cavity loop lengths, etc., to arrive at a 50 ohm resistive characteristic at the isolator’s port. Although we have developed known good lengths for many applications, it is often necessary to arrive at different loop coupling adjustments in a given cavity may require adjustments to those known line lengths. The point that we wish to make is that just any convenient length of jumper cable is rarely the right length! However, with the use of a device consisting of a simple miniature “PI” network (Impedance Matcher, Z Matcher, Micro Matcher or Line Matcher) which has two variable capacitors and a small inductor with values selected according to the frequency band of its application and matching range, impedance matches between elements of a system can be accomplished, such as between transmitters and amplifiers or cavity resonators and antennas.

Suppose that you cut and terminate a length of RG214/TJ cable that neatly dresses from your transmitter output to the isolator input on a transmitter combiner. Later, you find that the transmitter runs hot or its internal protective circuitry shuts it down before you arrive at its rated output power. The cable might also get warm, with definite “hot spots” along its length and the isolator might also seem to heat up unnecessarily. What to do??? You could make a longer cable and cut it off and re-terminate it a half an inch at a time until things cool down, or, place a Line Matcher right at the transmitter output port.


Due to certain non-linear elements used in ferrite garnet materials, weak harmonic signals can be produced in isolators at popular system power levels. The strength of these harmonics is highest for the 2nd order, weaker for 3rd, 4th, and so on. Since the second harmonic of the transmitter may be only 60 dB below the conducted carrier power at the isolator’s output, it must be suppressed by another 40 dB or more to eliminate it as a factor in I.M. generation. Third and higher order harmonics are down 95 dB or more and are generally disregarded as an I.M. source. When isolators are cascaded, the 2nd harmonics from the first device in a “string” are not conducted through successive isolators. It is, therefore, necessary to provide harmonic filtering only at the last isolator before feeding the antenna system. 2nd harmonic filters typically provide 40+ dB of rejection at the harmonic, with a VSWR of 1.05:1 or better.


Low-pass filters reject all responses above a given band limit and are down 60 dB at 2 times fundamental frequency.


The bandpass cavity resonator consists of a 1/4 wavelength or 3/4 wavelength radiating element positioned in a circular, square, rectangular or irregular shaped enclosure.

The volume of a cavity enclosure (body), overall length, material used, etc. will determine its “Q” or figure of merit for a given frequency of operation. The 1/4 wavelength radiating element may be compared to a typical ground plane 1/4 wavelength antenna with the extremes of the ground plane pulled up into a circular shield. As the diameter of the shield increases, the capacitive loading on the radiating element decreases and “Q” is greater. If the distance from the end of the radiating element to the end cap of the cylinder is 1/4 wavelength or more the capacitive loading effect is minimal.

Where a 3/4 wavelength radiating element is employed, the same relationships apply and circuit “Q” is increased. A fundamental rule in the application of cavity resonators is: The higher the “Q” the better (narrower) the cavity’s selectivity compared to a given coupling loss.

Coupling loops are used to induce R.F. power into the cavity and to couple power out of the cavity. These loops may be made larger or smaller, or made rotatable to determine the depth of coupling. The larger the loop and its orientation to the transverse electromagnetic mode wave propagation within the cavity the greater or lesser the depth of coupling. Also, the loop size and shape will determine its characteristic impedance. Figure 2 provides a means to compare the resultant selectivity of cavities of various diameters and loop coupling setting. Note that square, rectangular or irregularly shaped cavity resonator performance may be roughly compared by comparing the volume within the cavity itself.

The materials from which the cavity is constructed will have significant influence on its effective Q, those materials having lower skin effect losses yielding higher values of Q. As was suggested earlier , the cross-sectional area of a cavity is a factor in it’s figure of merit, “Q”. With this in mind it can be seen that the area of a circle 8″ in diameter (50.265 square inches) is quite close to the area of a 7″ square (49 square inches). Given other factors which effect “Q” such as material and length, the two shapes are functionally equivalent.


This type consists of two or more band pass cavity resonators with suitable connecting cables in the transmit and receive branches to secure desired rejection of transmitter carrier and noise powers as required for full duplex operation. The two branches are connected to two ports of a three way “tee” connector using selected cable lengths. The antenna is connected thorough its coaxial cable feed line to the third port of the “tee” connector.

The size and number of the cavity resonators, coupling loop design and adjustment and cable lengths between the cavities are determined by, the operating frequency band and the frequency spacing between the transmitter and receiver frequencies. Where multi-frequency, or multicoupled receivers. multi-frequency or combined transmitters are to be served by a single duplexer, special coupling and tuning methods along with critical cable lengths permit duplexing of expanded pass bands within certain limits.

The characteristics of the band pass duplexers are such that sufficient rejection of transmitter wide band noise, harmonic and spurious emissions are provided along with narrow pass band responses to protect the receiver not only from its associated transmitter emissions but from all other undesired signals. Transmitter spurious and harmonic radiations are also controlled by the highly selective band pass characteristics. Branch losses are generally higher than other diplexer types since each cavity, must be set for at least 0.5 dB of coupling factor to secure required branch selectivities. The minimum T to R spacing that can be used is determined by cavity “Q”. More cavity elements are required for a given T-R spacing than for the other duplexer types, resulting in higher costs to manufacture.


This duplexer type employs one or more notch type cavity resonators in each branch, depending on T-R frequency spacing, operating frequency band and transmitter power level. In this design, a single loop is used to excite each cavltv. using the resonant response of the cavity to define a rather broad selectivity characteristic, centered on the transmit and receive frequencies. The coupling loop is resonated by a series capacitance to form a notch element. The notch in the receiver branch is tuned to reject transmitter carrier power sufficient to prevent desensitizing (overloading) the receiver front end stages and the notch in the transmitter branch is used to trap out residual transmitter wide band noise existing at or near the receiving frequency.

With a given cavity size, the spacing between transmit and receive frequencies decreases, the notch depth becomes less and throughput losses become greater. As T and R frequencies are brought closer together more cavities are required to provide needed rejection of transmitter noise and carrier power as needed to yield interference free duplex isolation and resulting branch losses increase. The benefits of this duplexer type include lowest possible loss in each branch, use of smaller cavity resonator dimensions, smallest overall duplexer size and lowest cost to manufacture. The disadvantages are that the pass responses are very broad, providing very little protection against the radiation of harmonics or spurious emissions present in the transmitter’ s output and little or no improvement in the receiver’s “front end” selectivity.


This duplexer design is similar to the straight band reject type except that tighter coupling into the cavity field is provided by smaller loop geometries and placements. When property designed, above and below resonant loop notch responses are used to wield a “psuedo band pass” effect. Excellent isolation notch depths are provided along with notches above and below the desired pass bands.

Although a slightly higher branch loss results, compared with the strict band reject type, the pass bands of both branches provide valuable selectivities to help attenuate the radiation of transmitter noise, harmonics and spurs and improve effective receiver front end selectivity. For application at moderately and highly populated sites, this duplexer provides the desired aspects of the band reject type in terms of small size, low loss and low cost to manufacture with some of the benefits of the bandpass type in terms of selectivity.


Many translator installations evolve into radiating the signals of two or more transmitters in a common service area or direction from a high altitude site. Where antenna mounting space or height is restricted, the multiple antenna costs and problems arising from antenna-to-antenna coupling make the combining of two or more transmitters to a common antenna arrays viable.

Two methods of combining are available. The choice of which method is determined from the relationship of the frequencies to be combined. acceptable combining losses and physical space available for the combining apparatus, in the equipment location.

Where the spread between transmitter frequencies to be combined is 500 kHz or more, it is practical to combine up to four transmitters with combining losses of 1.5 dB or lower. Often the combining loss may be recovered through multiple antenna arrays, yielding an equal or greater effective radiated signal than present with a separate antenna for each transmitter. From a cost standpoint, the combining cost is often very close to the cost of good quality antennas, feedline and mounting hardware for the displaced antennas.

Further, complete isolation is provided between all of the combined transmitters. Also, the effects of improper loading on the output stages of all transmitters due to antenna changes, ice loading, etc. are nullified.

With 500 kHz or more of frequency spacings, the method of combining employed includes the use of high Q cavity resonators and isolators, known as “filter-ferrite” combining where transmitters operating at less than 500 kHz of frequency separation are involved, another method known as “hybrid-ferrite” combining is required. In this case, a hybrid coupler is used with an isolator in each input circuit. Unfortunately, the circuit loss is higher, on the order of 3.3 to 3.4 dB of insertion loss per combined path for a two transmitter combiner. The benefits are that identical radiation patterns are provided for both transmitters and more than 60 dB of isolation is provided between transmitters. Cost per channel for this combining method are somewhat lower than in the filter-ferrite method.


It is currently found that the cost of high quality conditioned broadcast quality remote lines has become unacceptable. Accordingly, the use of ultra high frequency studio to transmitter links has become quite popular.

Two systems are found to be available for the stereo transmission of program material. One system stereo encodes the modulation of a single transmitter, which is then decoded at the transmitter site to feed the separate left-right modulation inputs. The second method uses separate left and right encoded transmitter carriers with discrete receivers to derive the stereo inputs to the transmitter.

Both methods are vulnerable to interference from other STL’S, land mobile and radio paging signal interference and multipath signal distortion. The discrete left-right, two signal system is also vulnerable to problems arising from signals from one stereo channel providing phase distortion and other effects on the other channel.

Some STL discrete path equipment manufacturers recommend using separate transmitting and receiving antennas either separated by some distance from each other or using opposed polarization. Others provide simple hybrid only combiners for the two transmitters that often lack sufficient isolation between transmitters, resulting in phase distortion or loss of stereo separation in the transmitter audio feeds.


Each multicoupler includes a custom tuned preselector, a high performance amplifier with choice of power supply and suitable arrangements of signal power dividers. Multicouplers for from 2 to 8 receiver feeds are assembled on a single 3.50″ or 5.25″ high panel-chassis combination suitable for mounting on 19″ wide EIA cabinet or rack mounting. For multicouplers with 12 to 64 receiver feeds, a 4 or 8 way divider is mounted on the main chassis with auxiliary 3.50″ high panels providing feeds for 12, 16,.24, 32, 48 or 64 receivers according to test equipment arrangement at your site.

Operating Theory

The transmission line from the system receiving antenna connects to the input of the preselector. Preselector output is routed to a high performance amplifier, the output of which feeds the signal power divider. A high quality,- regulated and filtered power supply, inverter or converter supplies operating power for the amplifier.


The preselector is basically a bandpass filter. Its purpose is to pass the desired range of frequencies of the receivers to be employed with as near a flat response as possible and as low an insertion loss as possible. At the same time, rejection of all other frequencies is desirable. Preselectors vary from single or multiple cavity resonator combinations to multi-stage inductive, aperature coupled or comb filters. Special preselector assemblies may include one or more of these devices in combination, multiple pass and stop band characteristics, special notch elements, etc. according to the complexity of site requirements.


The preamplifiers used are of the highest quality and dependability to be found. Models are available from 66 MHz to 1.3 GHz and having gains from 15 to 32 dB (dependent on system requirements), noise figures from 2.0 to 3.7 dB, 3rd order intercept ratings of up to +40 dBm and 1 dB compression points of up to 24 dBm. The choice of amplifier to be used in a given multicoupler is determined by frequency band, bandwidth, nature of the install location and other relative factors. Most amplifiers are optimized for best noise figure and linearity at a fixed gain. Coaxial T-pads are used to attenuate excess gain according to the number of power division splits and other considerations.

Power Dividers

The power dividers split the amplifier output power into discrete matched impedance receiver feeds. The hybrid designs used in the power dividing process provides 25 dB or more of isolation between all receivers fed by a common multicoupler. Each time the signal power is split, 3 dB of loss results plus small conducted losses of 0.1 to 0.2 dB.

Courtesy EMR Corporation