51.6.18 AFEC Transfer Function

The first operation of the AFE is a sampling function relative to V_DAC. V_DAC is generated by an internal DAC0 or DAC1. All operations after the Sample-and-Hold are differential relative to an internal common mode voltage V_CM= V_VREFP/2.

In Differential mode, the Sample-and-Hold common mode voltage is equal to V_DAC= V_VREFP/2 (set by software DAC0 and DAC1 to code 512).

In Single-ended mode, V_DAC is the common mode voltage. V_DAC is the output of DAC0 or DAC1 voltage. All operations after the Sample-and-Hold are differential, including those in Single-ended mode.

For the formula example, the internal DAC0 or DAC1 is set for the code 512.

The DATA code in AFEC_CDR is up to 16-bit positive integer or two’s complement (signed integer).

The code does not exceed 4095 when the field AFEC_EMR.RES=0 (12-bit mode, no averaging).

Differential Mode (12-bit mode)

A differential input voltage V_IN= V_INP- V_INN can be applied between two selected differential pins, e.g. AFE0_AD0 and AFE0_AD1.The ideal code C_i is calculated by using the following formula and rounding the result to the nearest positive integer.

$C_{i} = \frac{2048}{V_{VREFP}} \times V_{IN} \times Gain + (2047)$

For the other resolution defined by RES, the code C_i is extended to the corresponding resolution.

The table below is a computation example for the above formula, where V_VREFP = 3V.

Table 51-7. Input Voltage Values in Differential Mode, Nonsigned Output
C_i		Gain
Signed	Nonsigned	1	2	4
-2048	0	-3	-1.5	-0.75
0	2047	0	0	0
2047	4095	3	1.5	0.75

Single-ended Mode (12-bit mode)

A single input voltage V_IN can be applied to selected pins, e.g. AFE0_AD0 or AFE0_AD1. The ideal code C_i is calculated using the following formula and rounding the result to the nearest positive integer.

The single-ended ideal code conversion formula is:

$C_{i} = \frac{2048}{V_{VREFP}} \times (V_{IN} - V_{DAC}) \times Gain$

For the other resolution defined by RES, the code C_i is extended to the corresponding resolution.

The table below is a computation example for the above formula, where V_VREFP = 3V:

Table 51-8. Input Voltage Values in Single-ended Mode
C_i		Gain
Signed	Nonsigned	1	2	4
-2048	0	0	0.75	1.125
0	2047	1.5	1.5	1.5
2047	4095	3	2.25	1.875

Example of LSB Computation

The LSB is relative to the analog scale V_VREFP.

The term LSB expresses the quantization step in volts, also used for one AFE code variation.

Single-ended (SE) (ex: V_VREFP= 3.0V)
- Gain = 1, LSB = (3.0V / 4096) = 732 μV
- Gain = 2, LSB = (1.5V / 4096) = 366 μV
- Gain = 4, LSB = (750 mV / 4096) = 183 μV
Differential (DIFF) (ex: V_VREFP= 3.0V)
- Gain = 1, LSB = (6.0V / 4096) = 1465 μV
- Gain = 2, LSB = (3.0V / 4096) = 732 μV
- Gain = 4, LSB = (1.5V / 4096) = 366 μV

The data include the AFE performances, as the PGA and AFE core cannot be separated. The temperature and voltage dependency are given as separate parameters.

Gain and Offset Errors

For:

a given gain error: E_G (%)
a given ideal code (C_i)
a given offset error: E_O (LSB of 12 bits)

in 12-bit mode, the actual code (C_A) is calculated using the following formula

$C_{A} = (1 + \frac{E_{G}}{100}) \times (C_{i} - 2047) + 2047 + E_{O}$

For higher resolutions, the code can be extended to the corresponding resolution defined by RES.

Differential Mode

In Differential mode, the offset is defined when the differential input voltage is zero.

where:

Full-scale error E_FS=(E_FS+)-(E_FS-), unit is LSB code
Offset error E_O is the offset error measured for V_IN=0V
Gain error E_G=100 × E_FS/4096, unit in %

The error values in the tables below include the sample-and-hold error as well as the PGA gain error.

Single-ended Mode

The figure below illustrates the AFE output code relative to an input voltage V_IN between 0V (Ground) and V_VREFP. The AFE is configured in Single-ended mode by connecting internally the negative differential input to V_VREFP/2. As the AFE continues to work internally in Differential mode, the offset is measured at V_VREFP/2. The offset at V_INP=0 can be computed using the transfer function and the corresponding E_G and E_O.