Weighted-resistor and R-2R are the two DAC architectures in the list — option 2.

The weighted-resistor DAC feeds each bit through a resistor of value \(2^{n}R\) into the virtual earth of a summing amplifier, so the currents add in binary proportion:

\(V_{o}=-V_{ref}R_{f}\left(\dfrac{b_{0}}{2^{0}R}+\dfrac{b_{1}}{2^{1}R}+\cdots\right)\)

Simple in principle, but for 12 bits the resistor values span a range of 4096 to 1, and the largest must be accurate to better than the smallest step — impossible to make on a chip.

The R-2R ladder solves exactly that problem by using only two resistor values in a repeating network. Looking into any node the ladder always presents the same impedance, so the current halves at each stage and the binary weighting appears from the topology rather than from the resistor values. Since only the ratio 2:1 need be accurate, and matched ratios are easy in integrated form, essentially every practical DAC uses this structure.

Why (i) is the tempting wrong answer. A successive-approximation converter goes the other way — analogue in, digital out — but it works by containing a DAC: the register guesses a code, the internal DAC turns it into a voltage, a comparator judges it against the input, and the guess is refined one bit at a time, giving a conversion in n clock periods. So a DAC is a building block of it, not a description of it.

Why (iv) is not a converter at all. A multiplexer routes one of several inputs to a single output. Nothing changes representation — digital in, digital out. Its role near converters is to let one ADC serve many channels in turn, not to perform conversion.

Hence, the D-to-A converters are (ii) and (iii).