2022-10-31 08:45 Tags: # Resistor String DACs Possibly the simplest way to make a DAC is a large voltage divider. $2^N$ resistors, switches and control signals.  Likely that the control signals come from a decoder, where the input binary signal is decoded into a unary encoding at the output.  Ths would likely need a buffer at the output.  ## Downsides 1. Very slow due to large capacitance on the output nodes due to the series FETs. It scales by $C_{out}=2^N\cdot C_D$ Series resistance (mid code) can be as large as $R_{max}=2^{N-1}\cdot R$ Thus the worst case time constant will be $\tau=\frac{1}{2}R_{max}\cdot C = \frac{1}{4}R\cdot C_D\cdot 2^{2N}$ (since we would get two parallel paths of half the maximum resistance at the centre of the string). Thus we can't have many bits in this converter, it will just be too slow! ## Advantages Monotonic - The voltage across a resistor is always positive, so the voltages at each node are always strictly reducing. ## Subranging Resistor String DAC One part deals with MSB and another part deals with LSB. Saving in the number of resistors. Need buffers between MSBs and LSBs unless we have very large resistors in the output string (else it will load the primary string).  Decoders also needed:   Overall circuit:  ### Advantages Significantly reduces number of switches and resistors. Significantly increase speed ## Disadvantages 1. Potentially no longer monotonic. If the offset difference in the buffers is comparable to the voltage drop across the MSB resistors, we can lose monotonicity (not as linear).  2. Missing codes - less voltage across the stack as across one resistor. # References 1. [vr-4602-wk07-sc01-dacrstring](../../Spaces/University/ELEC4602%20–%20Microelectronics%20Design%20and%20Technology/Lectures/W7/vr-4602-wk07-sc01-dacrstring.mp4)