9.3 Analog-to-Digital Conversion: Successive Approximation

Successive-approximation converters have been the general-purpose ADCs for many years. Their resolution is typically 10-12 bits and the speed can reach megasamples per second (Msps) although those in the MSP430 are not that fast. For some reason the name is expanded into "successive-approximation register" so that it can be contracted to SAR.

SARs work by homing in on the result using binary chopping, which is a standard way of finding solutions to equations of the form f(x) = 0. The sequence of operations for an input voltage of V _in = 0.4 V _FS to a 4-bit SAR ADC is illustrated in Figure 9.13. Here are the steps:

1. The input voltage V _in is compared with the midpoint V _FS of the full range.

2. In this case V _in < V _FS so the most significant bit (msb) = 0.

3. We now know that the input lies between 0 and V _FS. The input is next compared with the midpoint of this range, V _FS.

4. We find V _in > V _FS so the next bit is 1.

5. Now we know that the input lies between V _FS and V _FS. The input is next compared with the midpoint of this range, ? V _FS.

6. We find V _in > ? V _FS so this bit is 1 again.

7. Now we know...