ADC Mysteries Are Finally Explained

Analog-to-digital Converters (ADCs) have both analog and digital functions. Most simply, an ADC can be considered a device that provides an output that digitally represents the input voltage or current level. The analog input is compared to the ADC's analog reference voltage or current. The digital output word denotes what fraction of the reference voltage or current is the input voltage or current. In essence, the ADC is a divider. These definitions hail from the application note, "ABCs of ADCs: Analog-to-Digital Converter Basics" by National Semiconductor's Nicholas Gray.

At 64 pages long, this application note is extensive. For the sake of simplicity, the text assumes an ADC with a binary output. It begins by delving into numerous definitions surrounding ADCs. For example, an ADC's resolution is its number of output bits. Yet resolution also may be defined as the size of the least significant bit or one count. The least and most significant bits are those bits that have the least and most weight, respectively, in a digital word.

Better accuracy can be realized through the use of a higher-resolution converter and/or smaller reference voltage. Yet more bits means higher cost while a reduction in the reference voltage translates into a loss of input dynamic range. The note examines a host of errors relating to ADCs, such as quantization, offset, full-scale offset, and gain errors. The document moves on to define different types of linearity and their corresponding errors. Differential nonlinearity and linearity errors are both used to describe the error in step size. In contrast, integral nonlinearity and integral linearity error describe the maximum deviation from the ideal transfer function.

When using high-speed ADCs, there are many possible sources of problems. Yet many users are not aware of them. As a result, they end up using higher-resolution ADCs than they need because they are trying to get better noise performance or lower distortion. These common design mistakes involve either inadequate attention to noise minimization or the overdriving of any input. To avoid high-frequency coupling, for example, the designer should keep the signal path straight, avoid running analog lines parallel to each other, and keep inductors adequately separated or orthogonal to each other. Finally, care should be taken with resistor packs. This work is a valuable resource for any engineer who uses these mixed-signal devices.

National Semiconductor Corp., 2900 Semiconductor Dr., P.O. Box 58090, Santa Clara, CA 95052-8090; (408) 721-5000, Internet:

TAGS: Technologies
Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.