SDAA014 November 2025 AM2752-Q1 , AM2754-Q1
Figure 6-1 shows a simple example of how the McASP can use a single internal reference to send and receive audio data across multiple domains. The McASP is operating in asynchronous mode, but because the root clock source is the same for the transmit and receive domain, there is no risk of buffer overrun or underrun (as long as the audio data frame formatting is the same on the input and output).
For this system, the Audio PLL uses HFOSC1 to generate a high frequency audio clock reference. The high speed divider (PLL4_HSDIV0) divides the high frequency reference to 49.152MHz for the AUXCLK input. In this case, the TX and RX domains have AHCLK, ACLK, and AFS all configured to be internally generated.
The audio data frame is four audio channels for a single TDM4 stream and, assuming that the word depth is 32 bits, then the bit clock can be calculated based on the product of 4 channels of 32 bit words that are sampled at 48kHz = 4*32*48,000 = 6.144 MHz.