RF filters

November 1, 2007

Sometimes it seems as though radio-engineering work is very involved with filters of one kind or another. Actually this is not so far-fetched. When you think about our radio equipment, the widespread use of filters is easily understood. The science of broadcasting depends on the correct passage of various frequencies through differing pieces of equipment. Some frequencies are in the audio range and others are in the RF domain.

The physically smaller, and usually lower power, audio frequency (AF) filter immediately comes to mind — especially to the audio aficionado. But mention filters to an RF engineer and he immediately thinks in terms of kilowatts and much larger physical size. And that dreaded word "interference."

As a matter of fact, the topic of can become quite esoteric when the whole gamut of the field is considered. Filter consistency and construction can vary from a small collection of capacitors and inductances through quartz crystals and electromechanical combinations of garnet and yttrium such as are used in microwave operation, to tuned cavity filters that are often used in VHF and UHF operations. The major concern of most engineers in the radio broadcast field is with simple L, Tee and P networks and sometimes with acceptor or drain series networks. The Ls and tees are usually the last filters in the station's RF network and are used to match the transmitter output to the antenna.

Silver-plated wire

When a radio engineer speaks of an RF filter, many of us tend to visualize a sturdy coil of silver-plated copper tubing and a few physically large capacitors.

This is especially true of engineers who are fortunate enough to be involved with transmitters. In this area of radio engineering, where heavy currents are frequently handled and skin effect losses are important, it is usual to silver-plate the comparatively large diameter copper tubing used. To further reduce losses, as well as improve stability, this process is frequently extended to silver-plating the flat copper straps used for interconnections in antenna tuning units (ATU) and phasors, it also reduces I2R losses.

At the other end of the scale we find that used in such devices as radio receivers and similar low-power devices tend to use smaller gauge insulated copper wires and quite small physically, molded capacitors. In some cases, although less frequently these days, multi-stranded, very small gauge, insulated Litzendraht wire is used in inductors to reduce losses. In the audio field we can find an interesting similarity between radio tuners and loudspeakers.

Tuners and loudspeakers

Filter characteristics of a Bessel low-pass (top), Chebyshev group delay (middle), and Butterworth (bottom) filter.

The combination of the loudspeaker cone, surround, voice coil and centering spider form the mechanical equivalent of a parallel resonant tuned circuit. At the bass resonance frequency of the combination, the impedance of the voice coil rises to maximum and a VTVM connected across the voice coil peaks at the resonant frequency.

Looking at a schematic diagram of a receiver with a variable capacitor connected in parallel with an inductor, the similarity to an ATU may not be immediately apparent. However further consideration reveals that a parallel resonant circuit of an inductor and capacitor presents a high impedance circuit across which, at resonance, maximum voltage is being developed. This of course constitutes the tuning device in many radio receivers.

In the case of the ATU the combinations of L and C develop a network of impedances that matches the tower base operating impedance to the transmission line. In each case the filter operates because it is tuned to the frequency of interest and it only performs as designed at this one frequency. In effect it becomes a tuned transformer.

As the science of radio developed, the field of filters expanded considerably and a number of brilliant engineers gave their names to the specific type of filter that they designed. Some names associated with filters include, Bessel, Butterworth and Chebyshev (characteristics shown at left).

High- and low-pass

In the radio broadcast field, where the required pass or reject bandwidths are narrower and measured in kilohertz and megahertz, we generally find that the simple high-pass, low-pass and bandpass filters take care of most of our needs.

Sometimes an additional type of filter is required and is often referred to as a stop-band filter. This function is occasionally necessary when a strong signal on a critical frequency is received from a close-by transmitter. Internal cross modulation occurs in a final stage because the unwanted signal enters via the antenna and output coupling circuits and becomes a modulating signal in the final tube or transistor. Once again the simple parallel resonant circuit provides high impedance to the unwanted signal at resonance. The desired transmitter carrier is passed without loss and a series resonant drain circuit bypasses the unwanted cross modulating signal to ground through its very low impedance path.

In the case of a multiplex antenna system driven by more than one transmitter with different frequencies, drain and reject filters are required for each individual transmitter frequency.

Most station engineers are confronted by whose purpose and functions are clearly known and somewhat obvious. Occasionally a piece of equipment is found in an area, or circuit, without any apparent reason for its being there and lacking paperwork and information. Absent the availability of information concerning the purpose of such a filter it is often possible to obtain an idea of its purpose by drawing out the circuit and trying to determine its function based on filter knowledge.

Sometimes it's possible to determine if the filter is low or high-pass by remembering the effect of inductor and capacitor characteristics. For example a filter in which an inductor is in series with the input with a capacity to ground the filter will pass more low frequencies. This is based on the fact that as frequency increases, inductor impedance increases, so lower frequencies are passed. This is known as a low-pass filter. In the same way, if there is a capacitor in the input circuit and an inductor to ground, higher frequencies will be passed on more easily than lower frequencies. The resulting characteristic is known as a high-pass filter.

In brief, these filters take advantage of the frequency characteristic of capacitors and inductances. A capacitor passes higher frequencies and tends to attenuate lower ones as its impedance decreases with frequency. That's one reason larger capacitors are used for inter-stage audio coupling. An inductance has the reverse effect.

Contrarily, if there is an inductance in the input to the filter with a capacity to ground, the L and C characteristics will pass lower frequencies while shunting high frequencies to ground. Such a filter is known as a low-pass filter.

E-mail Battison at batcom@ohio.net.