From the schematic, this regulator has both MOS and bipolar transistors.
Can this be realized using CMOS process?
I mean the MOS without the oxide layer would become a Bipolar transistor.
I think it can be done. What do you think using pure CMOS to realize voltage regulator design.
The BJTs are used only in the front end of the error amp
and many, many commercial CMOS LDOs do just fine with
MOS front ends. Might want a little attention to just what
the output does, when VIN is up and VREF is not - startup
had best not put out full VIN voltage, etc.
BJTs are used for the input differential amp because they have low offset voltages.
You could use all CMOS for the complete design but it may require a special process to get the NMOS differential stage offsets as low as desired.
BJTs are used for the input differential amp because they have low offset voltages.
You could use all CMOS for the complete design but it may require a special process to get the NMOS differential stage offsets as low as desired.
An analog or mixed-signal BiCMOS technology. Common
in power management applications. Not nearly as cheap
as CMOS.
Until you know what your output accuracy requirements
are, and the matching attributes of your technology,
you don't know whether BJT matching suffices or is
overkill, or whether large MOS devices do well enough.
If you're really fussy then you will have to think about
trimming anyhow (and then care about BJT and MOS
device stability).