In this series:

This blog is the fifth part in a series presenting an overview of the theories and practices related to analog-to-digital conversion. In previous blogs, we discussed:

  • the basic principles used by analog-to-digital converters (ADCs) to convert continuous analog signals into a sequence of digital numbers.
  • how we can determine and control the frequency bandwidth of the incoming analog signal.how we can optimize the ADC sampling rate according to this bandwidth.
  • In this blog, we’ll examine how to adequately prepare (or condition) an analog signal prior to submitting it to an ADC for digitization.

The Analog Front End

Most analog-to-digital systems include an analog front end that serves as the interface between the sensor/input analog signals and the ADCs. Incoming signals are very rarely in a state where they can be sent directly to the ADC and usually require some sort of transformation or adaptation to ensure that they will be digitized under optimal conditions.

The main elements commonly found in analog front ends are illustrated in the figure below.

Analog front ends

 

This figure doesn’t represent a typical configuration and is only intended to present the various building blocks that are usually found in acquisition system analog front-ends. An actual implementation is very application dependent and will consist of a multi-stage “assembly-line” configuration, where different instances of these blocks will be combined to provide an optimal analog processing chain between the input signals and the ADCs.

The location of the analog front end is also very application specific, but will necessarily be one of three possible configurations: near or integrated with the sensor; integrated in a dedicated unit located between the sensor and the ADC; and near or integrated with the ADC. The ideal location for the front end is normally determined by a series of electrical, mechanical and economic considerations, which are outside the scope of this discussion. We’ll leave that subject open for other discussions in the future.

 

Conditioning Modules

We will briefly examine one by one each of the analog front-end blocks presented in the previous figure. We’ve selected the most common modules for simplicity’s sake, but other modules could also be used to address more specific processing requirements if needed.

Isolation Block

An isolation stage is sometimes (but rarely) required to ensure that the sensor/signal is isolated from the rest of the acquisition system, and that there isn’t any direct electrical connection between them. This module usually consists of just a simple transformer, but can also be implemented using an active isolation amplifier if the bandwidth of the signal is relatively small. Depending on the configuration and requirements, it is possible for both the transformer and isolation amplifier to also provide some gain (amplification) to the incoming analog signal.

Amplifiers

  • The presence of an amplifier is almost mandatory in any analog front end, and there are usually more than one gain stages in any given configuration. Amplifiers are fundamental building blocks that can be used for many different purposes such as:
  • simple buffering stages to isolate sections of a design so that they don’t interfere with each other
  • amplifying the signal to a high enough level so that the impact of the noise inherently present in subsequent stages will be minimized
  • bringing the signal to the optimal amplitude required by the following processing stage
  • dynamically changing the amplitude of the signal (either automatically or under software control) at different location along the front-end’s internal signal path (requiring that the amplifiers have variable gain capabilities)

Mixers

We briefly examined the purpose of mixers in the previous post in this series. Mixers are used to shift a signal up or down in frequency by multiplying it with a sine wave signal generated by a local oscillator (LO). The multiplication operation creates two copies of the signal, one shifted down in frequency (and mirrored) by FLO and the other shifted up in frequency by FLO. A lowpass (or bandpass) filter is typically used to remove the unwanted frequency components resulting from the multiplication process.

Mixers are typically used to displace the frequency spectrum of a signal to a fixed and controlled location that’s better suited to the operation of the selected ADC.

It’s not uncommon to see configurations where multiple mixers are used to divide the frequency shifting process into separate and optimized stages, even to the point where both upshifting and downshifting are combined.

Filters

Filters are fundamental building blocks of analog front ends, and as with amplifiers and mixers, it’s standard practice in most configurations to use more than one filter. Filters are used to remove unwanted frequency components from the signal at strategic locations in the front end’s internal signal path, or prior to the digitization process. The order of the filter (characterized by the number of internal components or stages) determines the amount of rejection and the rate at which the transition from passband to stopband occurs. The most common filters are:

 

Lowpass filtersThese filters remove frequency components above their natural corner frequency. They are typically used to remove undesired frequency components from an input signal, to limit the bandwidth of a signal to prevent aliasing, or to remove signal images created by the mixing process.
Bandpass filtersThese filters remove frequency components on each of side of their natural center frequency. They are typically used to isolate a very limited portion of the frequency spectrum and reject everything else outside that region.
Highpass filtersThese filters remove frequency components below their natural corner frequency. They are used to remove unwanted low frequency components from a signal, such as its DC level (fixed static offset), which usually is of no interest, or to prevent that signal from being properly amplified.
Notch filtersThese filters remove frequency components located at their natural center frequency. They are used to isolate and attenuate known static frequency components that are present in the signal, but that are of no interest or prevent that signal from being properly amplified (such as an induced 60 Hz component picked up from power line sources).

Output Driver

An output driver is an optional buffering stage for the ADC. It can consist of a simple amplifier or, if the analog front-end is located at a certain distance from the ADC, it can also provide differential driving capabilities to better immunize the signal from unwanted interfering signals that could be picked up along the way. The advantages of differentially driving a signal will be discussed in a subsequent blog post in this series.

 

An Analog Front-End Example

The following diagram presents an example of a simple ADC analog front end for a digital radio receiver.

ADC analog front end for a digital radio receiver

The various building blocks are used to amplify the incoming RF signal, to remove unwanted frequency components at various stages of processing, to shift the signal at a lower frequency so that a low Nyquist frequency could be set, and to amplify the output signal so that the maximum range of the ADC will be used.

The signal present at the input consists of the signal to be digitized (shown in blue) adjacent to an undesired stronger interfering signal (shown in red).
  

 

An amplifier stage is used to buffer the signal present at the input (antenna) from the front-end and to bring it to an optimal level for the following mixer stage.
  

 

The mixer stage creates two copies of the input signal: one shifted down (and inverted) in frequency, and another one shifted up in frequency.
  

 

A lowpass (or bandpass) filter stage is used to remove all unwanted components and only keep the desired signal contents.
  

 

A final amplifier stage is used to bring the amplitude of the signal to an optimal level to drive the ADC.

Conclusion

In this blog post, we’ve briefly discussed why and how analog front ends are used to condition and optimize signals prior to the analog-to-digital process of acquisition systems. We’ve examined the most common building blocks used in the elaboration of analog front ends, and we’ve proposed an example of a simple and typical configuration to illustrate how various building blocks can be combined to create a processing path adapted to the specific needs of an application.