The circuit was modified so that power is supplied simultaneously according to the power sequence recommended by ADI.
When the green signal is measured in the changed circuit, peak noise occurs every 25 samples.
From the FIFO, 25 samples of each Green/IR/Red are taken every 0.25 seconds, and the peak continues to occur in the first sample.
(51th sample (=25*2+1), 76th sample (=25*3+1), 101, 126, …)
It does not occur in IR and Red.
This problem occurred after modifying the circuit.
FIFO uses I2C communication.
Please refer to the Excel file for details and used registers.
I look forward to your quick response.