Why a high voltage AC always present in SMPS AC to DC power adapters

Question

I'hv seen every 2 pin SMPS AC to DC adapters shows a little high voltage respective to the ground on the out put, on both + and GND wires(or more if more wires present).

This voltage is high enough to light small neon bulbs or a line tester, but very low current, barely could be felt. This voltage vanishes if I connect the output to ground somehow, e.g. touch it while standing on floor bare foot.

As the size of the adapter increases, the AC voltage and current increases. When laptop adapters are powered from a ungrounded AC outlet, they gives reasonable shock from the 19V DC(or whatever else) output.

Perhaps this is the reason why many 3 pronged laptop adapter carries a warning CONNECT ONLY TO GROUNDED OUTLET Connecting to a grounded outlet magically vanishes the high voltage.

I am aware about the EMI filtering capacitor connected between high voltage and low voltage ground, I tried to solve this by removing them, but this trick is not working.

1. Why and how this mysterious AC voltage is created ?
2. Is it harmful to sensitive electronics components ?
3. Is there any way to solve this without grounding the DC output side ?

Answer 1

1. The internal transformer used to step down (and isolate) the power voltage has capacitance between primary and secondary - this can be around 10pF to 1nF depending on construction and size and is enough to produce several to several hundred micro-amps of current.
2. It can be harmful to sensitive electronics - the open circuit AC voltage normally produced on an un-grounded wall-wart might be AC voltage divided by two so this means 60V AC for an 120V AC power supply.
3. Grounding is the only method I can think of but, remember that the amount of current that can be supplied is not harmful enough to cause you any damage providing you are sensible and don't push the wires into your eyeballs or connect through your body to ground via your torso.

Answer 2

Why and how this mysterious AC voltage is created ?

The transformer has stray capacitance, this creates some leakage in itself, but worse it interacts with the switching action of the converter in ways that lead to the PSU acting as a high frequency source between input and output. To prevent this high frequency interference being radiated from the PSU leads further capacitance is added between the input and output sides of the PSU.

Usually the capacitor is after the rectifier, so you will likely find that the current has a component with a frequency double the input frequency.

Is it harmful to sensitive electronics components ?

It can be, but usually the current is too low to affect most stuff.

Is there any way to solve this without grounding the DC output side ?

There are a couple of mitigations available.

If the AC supply provides a ground then the EMC suppression capacitance can be split into two capacitors in series. The mid-point of these capacitors can then be grounded. This allows leakage currents to be diverted away from the output without DC-grounding the output. This is what most laptop PSUs do.

Another is to build the PSU around a split-bobbin transformer. These can have much lower input to output capacitance, but it tends to come at the price of efficiency.