JAJSN92 December 2022 AFE78101 , AFE88101
PRODUCTION DATA
If the AFEx8101 are used in a noisy environment, use the CRC to check the integrity of the SPI data communication between the device and the system controller. This feature is enabled by default and is controlled by the CONFIG.CRC_EN bit. If the CRC is not required in the system, disable frame error checking through the CRC_EN bit, and switch from the default 32-bit frame to the 24-bit frame.
Frame error checking is based on the CRC-8-ATM (HEC) polynomial: x8 + x2 + x + 1 (9'b100000111).
For the output register readback, the AFEx8101 supply the calculated 8-bit CRC for the 24 bits of data provided, as part of the 32-bit frame.
The AFEx8101 decodes 24-bits of the input frame data and the 8-bit CRC to compute the CRC remainder. If no error exists in the frame, the CRC remainder is zero. When the remainder is nonzero (that is, the input frame has single-bit or multiple-bit errors) the ALARM_STATUS.CRC_ERR_CNT bits are incremented. A bad CRC value prevents execution of commands to the device.
When the CRC error counter reaches the limit programmed in CONFIG.CRC_ERR_CNT, the CRC_FLT status bit is set in the ALARM_STATUS register. The fault is reported (as long as the corresponding mask is not set) as an ALARM_IRQ on SDO during the next frame. The ALARM pin asserts low if enabled by the alarm action configuration (see Section 7.3.3.1).
The CRC_ERR status bit (see Figure 7-19) in the SDO frame is not sticky and is only reported for the previous frame. The ALARM_STATUS.CRC_FLT bit is sticky and is only cleared after a successful read of the ALARM_STATUS register. Read the GEN_STATUS or ALARM_STATUS registers to clear any sticky bits that are set.
The sticky status bits are cleared at the start of the readback frame and are latched again at the end of the readback frame. Therefore, if the fault condition previously reported in the status register is no longer present at the end of the readback frame, and the data are received by the microcontroller with the CRC error, the fault information is lost. If a robust monitoring of the status bits is required in a noisy environment, use the IRQ pin in combination with the status mask bits to find out the status of each fault before clearing the status bits. Set the CONFIG.IRQ_LVL bit to monitor the signal level on the IRQ pin, and unmask each status bit one at a time to retrieve the information from the status registers.