Digital-analog Converter - Part 2
As the name indicates, the task of a digital-analog converter is the conversion of digital signals into analog signals. They are being used in various aspects of our daily life.
Areas of Application
Digital-analog converters are installed in music players and PC sound cards for playback of digital signals on CDs or MP3 files via speakers.
In order to be able to display digital signals on a (PC) monitor, they must be converted by circuits connected to the work memory (RAMDAC).
Generation of transmission signals in mobile communication devices requires very fast conversion of digital to analog signals.
Digital-analog converters are also used for control of many technical devices with electromechanical or electrochemical actors and as digital potentiometers or multipliers, for example for volume control on TVs.
Function of a Digital-analog Converter
A digital-analog converter is not a reversed analog-digital converter. A once quantized time- and value-discrete digital signal can never again be converted back into the original analog signal.
The reason for this is that an analog signal is time- and value-continuous, so that the amplitude can be measured at any time of the signal. On the other hand, a digital signal is reduced to specific scanning points: Accordingly, statements concerning the condition of the signal can be made at the respective point in time, but not about the condition of the signal between the individual scanning points.
An input register keeps the value at output of the signal from one scanning point to the next. This creates various courses of an analog, i.e. time- and value-continuous signal. As only the digital scanning points are available for the process of the digital-analog converter, different courses can result. Undesirable frequencies in the high range can be eliminated with anti-aliasing filters applied to the analog signal.
Quantization steps, i.e. the division of the signal into scanning points, also cause distortion of the signal. This is caused by the phenomenon of enveloping quantity curves, the so-called sampling or cardinal curves, the Sinc function. As the deviations caused by this are often rather in the low frequency ranges, they can be reduced or eliminated with the aid of low-pass filters. Depending on the type of signal, a quantization step can include one or more scanning points.
However, when the distortion is below the filter range, the well-known quantization noise is caused.
Specific causes for measuring value errors
In addition to the zero point error, the amplification error, and the non-linearity error, some further specific causes for a deviation from ideal and reality can be named for the digital-analog converter:
When the input values increase step by step, a reduction of the output values may occur depending on the implementation method being used. These are called step errors. The probability for occurrence of this error increases with the number of binary digits.
Time fluctuations, so-called jitter, have a considerable influence onto the output signal.
Classification of digital-analog converters according to the implementation method
A voltage divider has exactly 1 resistor for each quantization step. All resistors are weighted equally and are allocated to their quantization step by means of a so-called 1-of-n switch.
Although this method can be realized fastest, the effort increases strongly with increasing signal resolution. A known example is the 8-bit converter with its 256 resistors and 272 switches.
A digital signal logically is composed of binary digits. The number of resistors with the parallel method corresponds to the number of binary digits of the signal. With this, the individual resistors are weighted according to the significance of their associated binary digit. Each bit in the representation of the digital signal requires a switch. The different-value currents are switched to a bus line or directed away, depending on whether they are 0 or 1. All connected currents are converted to voltage by means of an operational amplifier.
With this implementation of the digital-analog converter, the signal is generated with the aid of time steps. Here, the total number of time steps corresponds to the number of quantization steps. Only a single switch is used, which is switched on or off by the digital signal, and this specifies the scanning frequency. Accordingly, the final signal is the arithmetic means of the voltage switched on and off in this way.
As the process requires counting of the time steps as well as forming of the front, this means it is time-consuming, but can be realized rather easily and cost-effectively. It is often used as an integrated circuit for microprocessors.
Classification of Digital-analog converters by wiring
Another type of classification is the way how the digital signal is directed into the digital-analog converter. This is done either parallel, with one bit per connection line, or serial, with only one data line. In this case, the input signal in most cases is an electrical voltage with standardized representation. An additional control line is used to verify the validity of the supplied data.
The output of the analog signal can also be used for classification of digital-analog converters. Thus, the signal can be provided either as current (voltage-output DAC) or as voltage (current-output DAC). In most cases, an amplifier circuit is required after processing for processing of the converted signal.