#### Robert_

##### Newbie level 3

I'm currently working on a projet for fun and to learn about signal synthesis.

My setup is a digital to analog converter of 8bits (home made R2R) driven by an FPGA.

I'd like to generate different waves forms (square, sawtooth, triangle, sine) with a frequency of 1 to ~20kHz.

Let's take for example the sine wave :

How to calculate the number of sample I need to get a good reconstruction ?

I know Shannon theory that tells frequency sampling must be at least twice of the max frequency.

What is the link between the number of bit in my DAC and the complexity of the signal I'd like to produce ?

I mean if you have a 1bit DAC you won't be able to produce a sine wave but square wave is ok.

Let's say I have fmax= 20kHz I take as sampling frequency fs = 200Khz.

Given that I have 256 possible states and assuming 0 of my sine wave is at 128. It is useless to take more than 128 dots per half period.

I'm lost ,I would be grateful if somebody could help me

Robert