Imagine the chaos that would arise if my smartphone could pick up any other nearby phone signal. Have you wondered how a smartphone knows which particular signal it needs to communicate with the local cell tower? In addition to multiplexing, your phone and plenty of other RF devices will make judicious use of filtering in order to communicate at the right frequency.
RF devices might seem esoteric, but designing and analyzing critical components for a number of RF devices is not any more difficult than working with other components. With the right analysis package, you can easily simulate an RF bandpass filter and analyze all aspects of your RF filter.
Why RF Filters?
Technically, any type of filter can be used at RF frequencies when the circuit includes the right components. The goal is simply to set the relevant natural frequency and bandwidth to the desired values in order to ensure your filter exhibits the right behavior.
The goal in RF filter design is to maximize the signal strength that propagates into a circuit at a particular frequency while excluding other frequencies. Ideally, the antenna and receiver should be impedance matched in order to suppress signal reflections and ringing, which may require an impedance matching network at the source or load end of the interconnect.
In practical applications, rather than using first order filters, higher order filters are typically used for RF applications. This is because receive/transmit channels in RF devices have narrow bandwidth, and the filters involved need to have steep rolloff at the edge of a channel. This ensures that the RF signal itself propagates into a circuit while noise or other signals that lie outside the desired bandwidth are excluded.
In GHz systems with carrier aggregation, especially in upcoming 5G systems, the level of required filtering in each channel is quite important to suppress nonlinear effects like passive intermodulation. Nonlinearities in poorly manufactured components can create sidebands in neighboring channels due to nonlinear frequency mixing. Using a higher order filter with steep rolloff at the desired receive/transmit frequency will help suppress sidebands to the point where they do not produce appreciable bit errors.
Modulation and multiplexing are extremely important in modern telecommunications. The return loss and insertion loss spectra for your filter circuit will need to be matched to the same spectra for your antenna and your intended receive/transmit band so that power in different bands is not inadvertently removed from a multiplexed signal. A higher order filter is generally used to provide the bandwidth and steep rolloff required to work with multiplexed signals that span over a receive/transmit band in telecommunications systems.
Passive RF Bandpass Filter Simulation
With RF/wireless devices, the goal is to make the return loss seen by the propagating signal as low as possible, meaning that the scattering parameter S11 should be as negative as possible when measured in dB. Ideally, an RF bandpass filter simulation will help you design the filter so that the return loss spectrum nicely overlaps the carrier signal and the modulation signals (in frequency modulation). Similarly, you want to bring insertion loss (S21) as close to 0 dB as possible.
With antennas, you want to pass as much of the carrier signal to the antenna as possible, bringing the voltage standing wave ratio (VSWR) as close to 1 as possible. Commercially available antennas that include filtration/impedance matching circuits typically quote VSWR values between 1 and 2 for these components.
Just like any filter, signals that fall within the pass band and are output from an RF bandpass filter will acquire a phase shift. Filters also produce a group delay, but this is generally ignored when working with a purely sinusoidal signal of constant amplitude. With multiplexed amplitude-modulated signals (for example, quadrature amplitude modulation), the group delay becomes very important as this represents the delay in the time-dependent amplitude of specific AC frequencies.
Since you are often working a multiplexed envelope of frequencies, different frequencies in the envelope will experience different levels of group delay. The example below shows the output from a 10th order Chebyshev filter with passive components. Here, you’ll see that the passband (9 to 11 GHz) is very narrow and only provides minor ripples in the insertion loss spectrum. The phase spectrum has a number of peaks, which may be undesirable in some applications, or may require compensation at the PCB level. A Bessel-Thomson bandpass filter generally provides flatter group delay spectra.
Output from an RF bandpass filter simulation for an example 10th order Chebyshev filter
Determining Losses, Group Delay, and Phase Delay
If you can calculate all of these parameters from your filter circuit, then you have all the information you need to evaluate your design. With any filter, the response can be determined using frequency sweeps at particular sinusoidal frequencies. Once you have a measurement of the output voltage and the input voltage at each frequency, you can determine the phase difference between the two signals. You can use the ratio of these voltages to determine insertion loss and return loss. If you are working with an antenna, you’ll also want to calculate the VSWR from the return loss:
Insertion loss, return loss, and VSWR equations
The group delay spectrum is equal to the derivative of the phase shift spectrum with respect to frequency, multiplied by -1. Note that you may need to convert the phase between degrees and radians when calculating group delay. This tells you all the information you need about your RF bandpass filter.
With newer wireless devices running at ever higher frequencies, you’ll need tools that help you design RF antennas, receivers, and components. A powerful SPICE package can help you design an RF bandpass filter simulation for use in complex PCB designs. The OrCAD PSpice simulator package from Cadence can help you perform many other tasks that are important for designing and analyzing analog circuits for any application.
If you’re looking to learn more about the solutions Cadence has for you, talk to us and our team of experts.