Advice for small AC-DC power supplies?

So, I have been looking into trying to build a few things that I want to be powered from line (120V where I live) voltages, but also use microcontrollers. At the moment, I am using a transformer I rescued from a clock along with a 7805 as a makeshift power supply for my projects. However, I need something much much smaller (square inch maybe?).

I really want a 5V (even 3.3V would be fine) power supply that sources 100mA just to power a microcontroller and a few small external components that is tiny. However, many small designs I have seen appear unsafe or inefficient and I was wondering what people normally use for this kind of thing. Certainly there has to be an accepted form for tiny bootstrap voltage supplies from line voltages?

I tried looking into 0.5VA transformers, but digikey doesn't appear to have any in their Power Transformer category of that size which are reasonably priced. I've also thought of going the cap plus zener route, but seems inefficient and doesn't strike me as particularly safe.

use flyback smps. link switch or tiny switch from power integration will be sufficient for your project if you need only one smps then search ebay

www.powerint.com

@Los Frijoles: You mentioned safety several times. All transformerless power supplies are potentially unsafe because there is no isolation between the main voltage and your circuit. That being said, there are places for transformerless power supplies - things that are self-contained and have no electrical connection to the outside world (other than that one power line connection). But developing that type of circuit usually involves making some kind of connection to a voltmeter or scope. That is potentially very dangerous. Have you thought through how you intend to avoid being electrocuted during debug as well as during deployed use of this circuit?

If you insist on using a transformerless power supply, all methods are equally unsafe. You might as well use the cap and zener idea. At least it has the advantage of not dissipating 20 times as much heat in the power supply as in your application circuit, which would be the case for any linear resistive power supply. However you should study how these supplies work very carefully before using one. The size of the cap is related to the maximum current you need. So it can't be too small. But it also can't be too large because then it will cause too much power to be dissipated in the zener. Already the zener has to be capable of dissipating as much power as all the rest of you circuit combined (in case the load is disconnected from the zener or the application circuit turns off). 5 volts @ 100 ma. is 0.5 watts, so you need more than a half-watt zener. And that assumes you sized your capacitor perfectly. If you made your capacitor a bit bigger, then the zener needs to be proportionally bigger. So you see that the design of a transformerless power supply is a non-trivial matter, both from a safety perspective and a functionality perspective.