Choose Converters with Care

In the world of electronics, most measurements involve standard analog-to-digital converter (ADC) circuits. An understanding of how three types of popular converters work will help interpret specifications for equipment such as a data-acquisition system or PC-based measurement cards.

Convert in a Flash

As its name implies, a Flash ADC performs a conversion almost instantly. A chain of resistors establishes a reference voltage for each comparator, which also receives the unknown voltage signal. By simultaneously comparing this signal against the reference voltages, a Flash converter quickly produces a digital output. This type of converter requires 2-1 comparators for a resolution of n bits, so an 8-bit converter requires 255 comparators. Adding another bit of resolution more than doubles the number of comparators and increases the complexity of the encoding logic.

Because the complexity of a Flash ADC increases rapidly for each added bit, manufacturers provide ADC resolutions of eight or fewer bits. Generally they offer high throughput, but a low number of effective bits.

Approximate the Result

A successive-approximation ADC applies an unknown voltage to a comparator and uses the output of an internal digital-to-analog converter (DAC) to home in on that signal.

At the start of an 8-bit conversion, the ADC first compares the unknown voltage to the voltage produced by setting the DAC's most-significant bit (MSB) to a logic 1. If this voltage output does not exceed the unknown voltage, the circuit maintains the test voltage and tries the voltage associated with the next least-significant bit. If voltage added by this bit exceeds the unknown voltage, the converter resets this bit to logic zero (no added voltage), and moves on to the next least-significant bit, until it tries the voltage associated with each bit. The test voltages decrease by half in successive bit tests.

Because this type of converter applies one test per bit, a 10-bit converter requires only two additional tests, yet it offers a four-fold increase in resolution over an 8-bit ADC. Resolutions for successive-approximation ADCs range from 8 to 16 or more bits. In most cases, circuits that employ a successive-approximation ADC require a sample-and-hold amplifier.

Add the Difference

A first-order sigma delta converter (or delta-sigma converter) relies on a modulator that produces a stream of serial information about an unknown signal. The modulator's negative feedback circuit causes the average value of the DAC's output to equal the unknown signal.

As the unknown input signal becomes more positive, the number of 1's in the serial output increases, and the number of 0's decreases. Likewise, as the input signal moves toward a more negative voltage, the number of 1's decreases, and the number of 0's increases. Thus the ratio of the 1's to the total number of samples in the same interval — the 1's density — is proportional to the value of the unknown input signal.

Because a sigma-delta converter oversamples a signal, the analog front end does not require a complex anti-aliasing filter. This type of converter circuit finds application in audio circuits and other devices that require conversion with 10- to 24-bit resolution and a high effective number of bits. Sample rates can exceed 10 MSamples/sec.