Software-defined radio (SDR) is popular technology these days and is used in a diverse range of applications. An SDR consists of two main components: a configurable radio front-end for transmitting/receiving radio waves and a programmable baseband signal processing board.

• Radio front-end – The front-end of an SDR provides significant flexibility because it’s configurable, sometimes even at run-time. For example, the “spectrum hopping” used in the field of cognitive radio reconfigures receive and transmit frequencies at run-time. An SDR enables you to look at the signal you want in terms of center (carrier) frequency, signal bandwidth, and amplitude. A good SDR front-end (such as the Nutaq Radio420x) lets you configure the following parameters at run-time: center frequency (sets the carrier frequency from the two PLLs shown in the picture), signal bandwidth, and amplitude (using the variable gain amplifiers shown in Figure 1).

• Baseband processing – SDRs are also known for flexibility in their baseband processing. A good SDR baseband board, like the Nutaq Perseus601x board (found in PicoSDR and PicoDigitizer systems), will be equipped with a fully programmable FPGA. On the FPGA you can implement any algorithm you want, from digital up/down conversion to a filtering and automatic gain control algorithm that tracks the amplitude of the digital signal using a moving average. This functionality lets you retrieve useful information from a noisy signal received by the radio front-end.

When combined together, the flexibility of these two components enables you to a) look at whatever RF signal you would like and b) retrieve whatever information you want from it.

Figure 1: Radio420 block diagram

Figure 1: Radio420 block diagram

A great example of programmable baseband board/RF front-end interaction is the implementation of an automatic gain control (AGC) like the one Nutaq provides on in its MIMO-OFDM QAM64 reference design. The algorithm on the baseband board tracks the amplitude of the signal using a moving average and controls, in real-time, the amplifier gain in the receive chain of the RF front-end in order to maintain a decent signal amplitude.

Between the radio front-end and baseband processing board are the ADCs (in the receiver chain) and the DACs in the transmitter chain. These handle the respective analog-to-digital conversion and digital-to-analog conversion of the baseband radio signals.

Clipping vulnerabilies

If we look at the receiver chain, however, we see that the ADCs have a fixed dynamic range, a fixed analog input voltage range that they can receive without saturating or clipping the signal. The same constraint applies to most of the components of the receiver chain.

Whenever the voltage at the input of the ADC (or any component in the receiver chain) is greater than its dynamic range, nothing can be done to fully recover the signal that has been clipped. This is a well know radio vulnerability, one that exists across many applications.

Stronger transmitter scenario

One scenario in which a radio typically suffers from the risk of signal clipping due to the presence of a too strong (too high amplitude) input is when an undesired transmitter is located close to the receiver and transmitting signals with greater amplitude than the one you are trying to receive (the transmitter that you are trying to listen to is placed further away and its signals are received but have less amplitude).

One option is to adjust only the gains of the receiver amplifiers in order to avoid clipping the ADC input. This may however result in the amplifier gains being so low than the signal you are trying to receive (from the transmitter placed further away) is too low, lost under the noise floor.

Signal filtering

To counter this vulnerability, a well-known approach is to place filters at the input of the radio receiver chain to filter out signal coming from undesired transmitter. Assuming the undesired transmitter is slightly out of tune (with regard to the transmitter you are trying to listen to), it is possible to filter out the undesired signal.

Nutaq’s Radio420x is equipped with fourteen RF bandpass preselection filters, selectable by software at run-time. Their location in the receiver chain is shown in Figure 1.

The allowed frequency band (frequencies that can pass through) by each filter is shown in Figure 2. Each filter corresponds to a commonly used band (some are associated with RF standards), defined by the authorities responsible for spectrum allocation.

For example, if your Radio420x receiver is tuned to 915 MHz with an instantaneous bandwidth of 26 MHz (or less), you can select (via software) the ISM Region 2 filter preselection filter. It will provide a strong attenuation for any signal below 902 MHz or above 928 MHz (top part of Figure 2). This prevents clipping in any component in your radio receiver chain (including the ADCs) due to the presence of an unwanted stronger transmitter.

Figure 2: Frequency band filters

Figure 2: Frequency band filters

Bypath pass and custom filters

An RF bypass path is available on the Radio420x in case there isn’t an available filter that corresponds to the frequency band of interest. Nutaq also has the ability to customize the Radio420x (upon request) by adding filters not shown in Figure 2.