A10 ADC/DAC Converters and Digital Signal Processing

Topic

Converting a real-world analog signal into digital data requires respecting, first of all, a fundamental principle: the frequency at which the signal is sampled must be at least twice the highest frequency contained in it. If it is sampled at a lower rate, false frequencies appear in the digital result that did not exist in the original signal—aliasing—a phenomenon that cannot be corrected after the fact, because once the false frequency is mixed with the real ones, there is no way to distinguish which was which.

Every analog-digital converter also has a resolution limited by the number of discrete levels that it can distinguish, which introduces a theoretical limit to the relationship between signal and quantization noise that this converter can achieve, a limit that grows the more bits of resolution the converter has; In practice, the real performance of a converter is always somewhat worse than this theoretical limit due to imperfections in the circuit itself—returning to this same noise limitation already described by the corresponding model in this catalog—so an "effective" resolution is also measured that summarizes its real performance, more honest than the nominal resolution.

Once the signal is in digital format, digital filters allow it to be processed with great flexibility: some, with a finite response, guarantee that all frequencies are delayed exactly the same - a property highly valued in audio and communications, analogous to what the Bessel approximation was already seeking among analog filters - although they usually require more calculation resources than the infinite response alternative to achieve an equivalent specification.

That is why this module, which opens with a sampling condition that admits no exceptions, closes by returning to this same fast Fourier transform that already described the frequency analysis model: it is the central tool that allows calculating the frequency content of an already digitized signal, and supports applications as different from each other as digital audio, modern communications, radar and measuring instruments.

ID:4068