Babar,

forget 16 bit with R-2R.

Typical R-2R with 1% accuracy can give 7 bits
and more expensive (0.1%) up to 10 bits.

10 bit means a maximum error of 1/1024 = 0.1%
E.g. for 10 bit and on using a CMOS driver 4049 with
1kOhm on resistance your R-2R must be made
with minimal 1MOhm resistors.
Also the following amplifier must have an
input resistance of minimal 1GOhm !

Thus, why above 8 bits integrated DAC are
better and cheaper.

There are also some other methods for DACs
like PWM or switched capacitors, which
can give higher accuracy on slower speed.
But its also not easy to implement it discrete.

Audio DACs, which used in CD players, give you
high resolution for low cost.


Peter