You need to understand that reducing an input value that in essence can exist as an infinite number of discrete values to one of four thousand and ninety six discrete values (12 bits resolution) will attract noise being the difference the real life value and the nearest value that the digitisation process can resolve.
Going the other way and trying reproduce an infinite number of discrete values from a very well defined and limited values will present the same problems.
Using more bits in the DAC or ADC only serves to delay the problem but does reduce the errors voltage or noise level.
Not forgetting the noise environment immediately around the Pi could/will become quite significant when boosted by a factor of 30 so some care will be needed with power supply decoupling and include both high value electrolytic capacitors and low inductance ceramic capacitors.
(14)
Going the other way and trying reproduce an infinite number of discrete values from a very well defined and limited values will present the same problems.
Using more bits in the DAC or ADC only serves to delay the problem but does reduce the errors voltage or noise level.
Not forgetting the noise environment immediately around the Pi could/will become quite significant when boosted by a factor of 30 so some care will be needed with power supply decoupling and include both high value electrolytic capacitors and low inductance ceramic capacitors.
(14)
Statistics: Posted by RaspISteve — Mon Dec 29, 2025 4:56 pm