Correctly scaling and biasing a microphone preamp input to nRF51822

[sd]

I'm using the MAX9814 electret pre-amp's output as an input to an analog pin of the nRF51822. The MAX9814 has a 1.2V swing at 1.2V DC (so it goes from 0 to 2.4V). VDD for the nRF51822 and the MAX9814 is 3V.

I believe that a direct connection (MAX9814 MICOUT -> nRF51822 An) should work with a 1/1 scaling and VDD as reference, although I will lose 0.4V of the range (2.6--3V). Any potential issues with this setup?

If I wanted to use a DC block capacitor (MAX9814 MICOUT -> CAP_DC -> nRF51822 An), what is the best biasing scheme to use in order to get a full resolution? I looked through the datasheet, but the scaling doesn't quite work in my case unless I add another reference voltage, which isn't ideal. Any advice? Any reference resources? (I prefer to avoid using a voltage divided for AREF, as the board space is very tight).

[Stefan Birnir Sverrisson]

Hi

To sum up the possibilities:

• You have amplifier that outputs 0V - 2.4V. Is that constant or is it relative to the 3V battery voltage? If it is relative to the battery voltage, choosing VDD as reference would be a good choice, otherwise, choose the VBG voltage.
• If the amplifier output is relative to the battery voltage, then choosing 1/3 prescaling on the input and VDD/3 as reference would give the best resolution for you signal, i.e. 2.4/3.0*1024=819 for 10-bit ADC resolution setting.
• If the amplifier output is 0V - 2.4V independent of the supply voltage, then only option would be to use 1/3 prescaler on the input with VBG as fixed reference. The resolution would be 2.4/3.6*1024=682. You could also use external reference AREF with input of 0.83 (the minimum) to have resolution of (2.4/3)/0.83=987

[sd]

There is a battery, but it's regulated to 3V, which supplies the pre-amp and nRF51822. I think that your second options is indeed the best way to go: no Aref divider/connection, no DC blocking cap...