Filters
In the digital or analog RF environment, there will be interference and other noise to deal with. That is why filters are a very useful tool. They help bring in weak signals or get only the signal you want to hear at that moment. Filters narrow the spectrum the radio is working with. They help focus the signal in one area, or exclude spurious signals that may be created from mixing. One type of filter might be an “AM broadcast” filter. This would drop the signal level in that band to help reduce spurious signals in your radio.
One way to measure a filter is by the bandwidth, which is the amount of signal that passes through the filter. The passband is the frequency allowed to pass through the filter. The cutoff is the frequency at which the outside signal is suppressed. On our AM broadcast filter the cutoff frequency would be between 500 and 1800 kilohertz. Any other signal would be in the passband. This is called a “band stop” filter.
The term shape factor measures a filter’s ability to reject signals occupying an adjacent channel. Imagine you had a signal you wanted to listen to on 1.8 Megahertz. Since that’s right next to the AM band, the shape factor of the AM filter would be an important consideration.
Filters can be created using a variety of components. They include capacitors, inductors, crystals and cavities.
A Chebyshev filter is a type described as having ripple in the passband and a sharp cutoff. Simplistically, ripple is a small loss of signal in the passband.
An elliptical filter has extremely sharp cutoff with one or more notches in the stop band. The stop band is another term for the cutoff area. Sharp cutoff is good to minimize impact to desired signals.
A Pi-network attenuator filter gets its name from how it looks on a schematic, like the Greek letter Pi. The arrangement of components is important in a low-pass filter Pi-network. The capacitors and inductors go between the network’s input and output. That arrangement is a capacitor is connected between the input and ground, another capacitor connected between the output and ground, and an inductor is connected between input and output.
Schematic of a pi-network attenuator filter
Crystals are a common component of filters as well. Using quartz crystals to attenuate undesired frequencies makes a crystal lattice filter. A crystal lattice filter is a filter for low-level signals made using quartz crystals.
Physical crystal lattice filter
A high-pass filter is one that has a passband higher than a particular frequency. It only allows frequencies above the cutoff frequency to pass. A T-network with series capacitors and a parallel shunt inductor has the frequency response of a high-pass filter. Yes, there are low-pass filters too.
There are filters optimized for VHF and UHF applications. What’s the best choice to keep spurious signals out of your 2-meter band repeater duplexer? That would be a cavity filter.
Cavity filter diagram
If you’re not running a repeater, but need some filtering on a VHF and UHF transceiver, consider a helical filter. A helical filter is frequently used in transceivers as a band pass or notch filter.
Digital Signal Processing Filters
Digital signal processing brings its own unique type of filters. How does a DSP generate an SSB signal? One common method is when signals are combined in a quadrature phase relationship. One type of DSP filter for generating an SSB signal is a Hilbert-transform filter. It’s named for David Hilbert, the scientist who discovered this filter in the early 1900s.
If you want to remove unwanted noise from a received SSB signal you might use an adaptive filter. An adaptive filter is capable of adjusting, or adapting, its response via an algorithm. It keeps the desired signal strong while knocking down the interfering noise.
Let’s focus on two types of adaptive filters. They are Finite Impulse Response (FIR) filters and Infinite Impulse Response (IIR) filters. An FIR responds for a fixed period of time, while an IIR responds indefinitely. Many DSP’s chose the FIR filter. Its advantage is that FIR filters can delay all frequency components of the signal by the same amount.
In some DSP filters you will find components called tapped delay lines or “taps.” Taps provide incremental signal delays for filter algorithms. The FIR is just one digital signal processing filter that has taps. If you want to allow your DSP to create a sharper filter response you would add more taps.
Filters may need some troubleshooting. If you have ringing in a filter it means you have undesired oscillations added to the desired signal. Unwanted ringing and audio instability can be prevented in an op-amp audio filter circuit.
