After experiencing a notable shocking type sensation after coming into skin contact with my laptop and phone (both metal and plugged into their chargers), I decided to probe the chassis of both devices to discover a 50Hz 120V transient...

To investigate this further, I directly recorded the voltage from the grounds of the charger connectors to find the same transient. Swapping between two different laptop and two different phone chargers, all permutations yield the a large AC signal from their grounds. That said, swapping between a 10W to a 5W usb charger did drop the difference to 60V between laptop and usb charger grounds.

I’ve even checked the sockets to ensure that its not a wiring problem, and sure enough the earth, live and neutral are all where they should be and connected properly. N.B. I’m in the UK, so 240V outlets.

This is quite baffling to me since not only are the voltages produced by the charging circuits direct current, but neither voltage exceeds 19.5V! Any explanations/diagrams to explain this would be fantastic!



The DC has a common mode AC from the SMPS capacitance across the insulation transformer. The leakage should not exceed ~ 100uA shorted to earth gnd. This depends on Vac(f)/Zc(f)=Ic(f) and the leakage capacitance might only be 50 pF but at 50kHz or much higher.

The "f" can be from the SMPS rise time, rep. frequency, PWM duty cycle and modulated by line f, 2f etc. so in general, we can only say it is not from the DC.;}

Electric sensations are not determined by the leakage voltage, rather induced current where I=V/Z as a function of some frequencies.

These symptoms are common to all types of laptops and may be caused by the laptop earth gnd EMI filter noise reduction filter along with galvanic capacitance coupling of the SMPS RF transformer in either or both devices.

My cure when using the laptop charger is to connect the laptop case or 0Vdc signal to earth ground with a bigger RF Cap such as 10nF.

A direct connect of Earth gnd to 0V dc also works but; may induce hum on external microphones , exposes you to ground-fault currents, since it is no longer "floating".. i.e. if you touch a live low impedance line to frame fault situation, it could be fatal. A laptop VGA cable to an earth grounded Monitor also connects this 0Vdc to 0Vac earth ground on 3 prong plugged monitors.

This 100 to 200 uA leakage current can be painful to a knee or wrist but is not fatal. But then the EMI filter is supposed to suppress reduces accumulation of functional issues on other electronic devices, somewhere nearby. the RF transformer leakage is unavoidable unless by fluke the leakage of 2 devices matched amplitude and phase to null the interface current of line frequency modulated RF current.

If you wanted to specify accurate specs, device models and accurate current waveform measurements, then a more precise answer is possible.

Final remarks

Unfortunately, this situation is not 100% unavoidable between two electronic devices each with (RF) SMPS AC-DC converters. The galvanic isolation and grid isolation is guaranteed to some HIPOT test leakage current max and this protects you from grid voltages and stray lightning noise while EMI to grid in both directions and isolates you from live ground faults to some extent and all the while Low Pass Filters transients in both directions to the grid (both X (Diff) and Y mode -CM ).
But it can create nuisance leakage noise current between earth gnd or other mobile charged units. So you can avoid this situation, or shunt the current to a ground common to both devices with a small RF cap. (ceramic or plastic)

  • 1
    \$\begingroup\$ Sunnyskyguy How should I alter my answer, to align with your (excellent) text? \$\endgroup\$ Sep 7 '19 at 15:18
  • \$\begingroup\$ @analogsystemsrf replace 50/60Hz line transformers with equiv AC-DC SMPS circuits to show CM leakage with Buck regulated transformers. \$\endgroup\$ Sep 7 '19 at 16:12

something like this will cause the "shock"


simulate this circuit – Schematic created using CircuitLab

How much current can flow?

1nanoFarad at 1Hertz is 1/(2 * PI * 1Hertz * 1nanoFarad) = 0.16/1nano = 160Million ohms

Now lets use that 1Hertz value, at 60 and 60,000 and 60,000,000 Hertz.

1nanoFarad at 60Hertz is 160M/60 = 2.66 Million ohms. Round down to 2Million ohms; there is no sense in pretending to have numeric precision here.

1nanoFarad at 60,000Hz (approx. the chopping rate of a switching regulator) is 2,000 ohms.

1nanoFarad at 60,000,000 Hz (approximately the parasitic ringing frequency of a switching regulator) is 2 ohms. TWO.

Now lets compute some maximum-possible currents, using voltage of 200 volts.

200 volts at 60Hertz is 200v/2Million ohm = 2/20,000 = 1/10,000 = 0.0001 amp = 100 microAmps. This is 10 to 100 X smaller than electrocution numbers.

200 volts at 60,000 Hertz is 200v/2,000 === 0.1 amps; this will kill you, thus this current must be greatly reduced, and EMI filters are used inside those switching regulators.

200 volts at 60,000,000 Hertz is 200/2 === 100 amps; this means whatever ringing exists inside those switching regulators will attempt to use the human body as a return path; EMI filters are used inside those switching regulators; PCB layout and metallic shields are part of the design task.

  • \$\begingroup\$ This has the right concept but wrong values. usually it converts AC-DC-AC-DC so transformer is RF (50k to 2M) \$\endgroup\$ Sep 7 '19 at 16:26
  • \$\begingroup\$ Yet only laptop charger is 3 pin, while mobile is lower power 2 pin AC plug, so only laptop has line Y caps for CM Pi line filter to add to leakage impedance and voltage. \$\endgroup\$ Sep 7 '19 at 16:33
  • \$\begingroup\$ You should edit your answer to reflect the actual circuit with a 3 pin AC to DC SMPS and a 2 pin AC to DC SMPS to show both leakage paths, of Ycaps and high f transformers to show the current path. \$\endgroup\$ Sep 12 '19 at 13:57

I would recommend trying it in another house or building, you may have a failing mains ground in your house if it only applies there.

The EMI capacitors on devices leak a little current between the phases to ground, if its not going to the mains ground due to a break or similar, it can be enough amplitude to make a noticeable zap,

The other way it can happen is more rare, but the transformer in the device can act as a capacitive divider giving half mains modulating your output DC voltage, but this is generally a very low curent, thus hard to notice when the EMI filter is working.


Not the answer you're looking for? Browse other questions tagged or ask your own question.