For sensors (ie high bit-depth adc,dacs) running in isolated noise/power supply domains, are optical isolators preferred over magnetic isolators or vice versa? I am interested in noise immunity, more than HV isolation.
For example, comparing datasheets for two similar options for spi,
opto isolation (ACSL-6410-06TE)
https://au.mouser.com/datasheet/2/678/AV02-0235EN_DS_ACSL-6xx0_2018-08-21-1827365.pdf Input-Output Capacitance 0.7pF. at 1MHz.
Low Level Supply Current 4.6mA typ 7.5mA max.
Input Threshold Current 2.7mA typ. 7mA max.
magnetic isolation (ADUM4151)
https://au.mouser.com/datasheet/2/609/ADuM4151_4152_4153-1503566.pdf Capacitance (Input to Output)1 CI- O 1.0pF, f = 1 MHz
Quiescent Side 1 CurrentIDD1 (Q) 2.9 mA
Quiescent Side 2 Current IDD2Q) 6.1 mA
Input Current per Channel 0.01uA typ +1μA max.
The key characteristic of coupling capacitance, appears similar for both (0.7pF vs 1pF at 1MHz).
One thing, that stands out is the higher drive/pulse power for the optocoupler. eg. (pulsed 7mA), which is perhaps more likely to introduce noise in the sensor domain. Although maybe it's not an issue just driving an open-collector pull-up resistor placed right next to the chip.