AN1798

Calibration and linearization technique

7/16

3

Calibration and linearization technique

3.1

Principle

ADC calibration is done by software, it consists of a new calculation of the gain, the quantum

and the offset each time the application is executed.

Initially, the you have to obtain the ADC conversion digital result for 0 V and 2.5 V by

applying two stable voltage external sources (0 V and 2.5 V) to the ADC input in order to

calculate the ADC gain and the quantum.

Next, you need to calculate the average of these two values by reading each value n times

and dividing by n to reach the average values.

After that, calculate the unsigned values of ADC_DATA(2.5V) and ADC_DATA(0V) (these

values are two’s complement).

Then, calculate the ADC range or the sweep of conversion in count / volt:

sweep = Unsigned[ADC_DATA(2.5V)] - Unsigned[ADC_DATA(0V)].

Finally, calculate the gain and the quantum:

G = sweep / 2.5 V

q = 2.5 V / sweep

The principle of linearization is to shift the half-lines ‘A’ and ‘B’ respectively down and up to

obtain one segment: ‘C’. Then, subtract the offset as shown in the following figure.

Figure 4.

Linearization

These segments are linearized by software by adding 0x800 if the reading value is positive

(the most significant bit is set to 0). If this value is negative (the most significant bit is set to

1) subtract its two’s complement from 0x800. Then, subtract the offset from this value. The

offset value is equal to Unsigned[ADC_DATA(0V)].

0x800

0xFFF

VCM=1.25V

2.5 V

ADC_DATA(0V)

ADC_DATA(2.5V)

Offset

Vin

ADC_DATA

0x800

0xFFF

VCM=1.25 V

2.5 V

Offset

Vin

ADC_DATA

Final segment

Counts off

Linearize the segments

Subtract the offset

then

subtract the

offset

Offset

Full scale

A

B

C

C