SBAA492A September 2021 – April 2022 PCM3120-Q1 , PCM5120-Q1 , PCM6120-Q1 , TLV320ADC3120 , TLV320ADC5120 , TLV320ADC6120
To remove any DC offset that leads to incorrect input level estimates, the AGC algorithm processes the input signal through a high-pass filter. This HPF is exclusive to the AGC, and is different from the second-order HPF filters used by the decimation filters.
The transfer function implemented by the high-pass filter is given by Equation 1.
The HPF is a first-order filter implemented using three coefficients: AGC_HPF_B0, AGC_HPF_B1, and AGC_HPF_A1. The transfer function parameters (N0, N1, and D1) are converted to coefficients using Equation 2, Equation 3, and Equation 4.
These coefficients are user-programmable to set a different cutoff frequency from the default cutoff (-3 dB) of 100 Hz for a 48 kHz sample rate. Increasing the cutoff frequency results in faster settling of signal-level estimates, while decreasing the cutoff frequency improves the accuracy of the signal-level estimate. The default filter coefficients provide a good balance between speed and accuracy, and are suitable for most applications. Table 2-2 shows the coefficient registers. The coefficients are represented in 2s-complement, 32-bit format.
COEFFICIENT | PAGE | REGISTER | RESET VALUE | DESCRIPTION |
---|---|---|---|---|
AGC_HPF_B0 | 0x06 | 0x78 | 0x7F | AGC_HPF_B0 Byte[31:24] |
0x06 | 0x79 | 0x7F | AGC_HPF_B0 Byte[23:16] | |
0x06 | 0x7A | 0xD2 | AGC_HPF_B0 Byte[15:8] | |
0x06 | 0x7B | 0xB4 | AGC_HPF_B0 Byte[7:0] | |
AGC_HPF_B1 | 0x06 | 0x7C | 0x80 | AGC_HPF_B1 Byte[31:24] |
0x06 | 0x7D | 0x80 | AGC_HPF_B1 Byte[23:16] | |
0x06 | 0x7E | 0x2D | AGC_HPF_B1 Byte[15:8] | |
0x06 | 0x7F | 0x4C | AGC_HPF_B1 Byte[7:0] | |
AGC_HPF_A1 | 0x07 | 0x08 | 0x7E | AGC_HPF_A1 Byte[31:24] |
0x07 | 0x09 | 0xFF | AGC_HPF_A1 Byte[23:16] | |
0x07 | 0x0A | 0xA5 | AGC_HPF_A1 Byte[15:8] | |
0x07 | 0x0B | 0x68 | AGC_HPF_A1 Byte[7:0] |