I suppose you mean a shunt regulator not a series one

Hallo Asok,

I didn't mean a shunt regulator. Efficiency would not be very good.
I meant a series regulator consisting of zenerdiode at the base, input voltage at collector and output voltage at emitter of NPN transistor. Regulation characteristics is rather poor, compared with modern three pin voltage regulator chips, but voltage drop of one pn junction can be saved, which results in a slightly reduced minimum differential voltage between input and output.

If you have a look at internal circuitry of a standard three pin voltage regulator, you will mostly see a NPN darlington at series transistor. So, allways about 1.4V, the voltage drop of two pn junctions, is needed between input and output.

A word to LDO regulators:
Common LDO voltage regulators do only contain a PNP transistor as series transistor, which results in a drastically reduced voltage drop. But PNP design is unstable and needs output capacitor with it's equivalent series resistance as compensation. Compensation fails, if series resistance falls outside a highly defined range. So, applications needing multiple power supply decoupling capacitors can become problematic. Also, when temperature changes drastically or when used electrolytic capacitors are becoming old and no longer show predictable series resistance.
Also, because only a single transistor is used, instead of NPN darlington, quiescent current can become rather high. Think of decreased current amplification of series transistor.

So, LDO voltage regulators require much lower differential voltage between input and output than any NPN design, but application is much more problematic.

Kai