Does this provide galvanic isolation from an AC supply?

 

Yes so long as the capacitors don't break down.

In theory you only need one of the capacitors since there is no return path.

 

How do i find out how much voltage the caps need to withstand?

 

This is not a safe or prudent approach to obtaining power from the AC line. Any fault instantly creates a lethal shock hazard, plus the power delivered is extremely limited. A wall transformer is capable of more power in a safe manner.

The only safe way to achieve galvanic isolation in a predictable manner is through the use of a transformer.

 

Is your load constant?

The two capacitors are in series with the load across the ac supply (ignoring the diode drops)

Their actual voltage drops will depend upon their reactances and thus frequency of operation and their capacitance.

A safe option would be to rate each one for the peak value of the supply, you do not need peak to peak.

Incidentally if you drew the diagram, you should not have both positive/negative and ac symbols in your supply box. I assume you mean the supply is ac.

 

This is not a safe or prudent approach to obtaining power from the AC line.

If that is your aim I agree, even though there are commercial products which use this approach.

 

This is not a safe or prudent approach

Hypothetically then..

In theory you only need one of the capacitors since there is no return path.

If I have this right then what's needed for "won't kill you when you touch it" isolation is just one big bipolar eletrolytic cap. In this case at least 700 µF works ok along with a 2mF filter cap. I'm simulating it with a constant-power load of 1 kW. The final output is supposed to be 12v which to me suggests the use of a buck regulator. The buck regulator not requiring a transformer, just an inductor. The idea here is that the whole thing works out being much cheaper than a 60hz transformer at this power level, though i can't speak for Mhz range SMPSs. But before 1 good idea comes 100 bad ones.

 

You're going to get around 170V out of that rig. 1.4 X RMS applies.

 

AND SO ON in the graphic implied that there was a load there that wasn't shown, and in this case it was a constant power load of 1000w. It represents a buck regulator with a fixed resistive load. It pulls the voltage down from 170.

 

This doesn't really provide isolation, does it? As in if your standing on wet soil with bare feet and you decide to touch the output, you could still die. Because the cap only blocks DC. So the isolation isn't at all the same as using a transformer.

 

That's correct - there is no isolation. If a voltage surge causes the cap to short, there is also a more lethal situation.

 

If the transformer is high voltage you could still die, but yes, this is not safe. It is one of the reasons we recommend wall warts around here.

 

700uF?? that's bigger than a coke can. For true isolation, use an earth ground using a ground rod that goes at least six feet underground