Cyril, there's almost no good reason to avoid the AMS1117. It operates with the given overhead and provides many important features. A zener design could be modified (not the one you show, exactly) to use an NPN BJT and an added diode, with the zener. And it may satisfy -- maybe. But why? Other discrete techniques may even be "reasonable" over some temp range. But it is just more work without the advantages, unless you are making a lot of these and the difference between a handfull of 0.1 cent parts vs the AMS1117 matters. (I bought my lifetime supply of BJTs at 90 per penny. But thousands.)
↧