I was experimenting with LEDs, when I encountered the following strange behavior:
5 LEDs on a strip simultaneously emit light when I connect eny output wire from a mean well +36V 700mA constant current power source.
the strip is about 40cm and has a clip to connect to its neighbor strip but the contacts on the clip are marked (+, -)
Connecting only red wire to + the strip lights
Connecting only red wire to - the strip lights
Connecting only black wire to + the strip lights
Connecting only black wire to - the strip lights
Trying all the above using another 18-32v 0.6A driver same result except they are now blinking at around 2 Hz
Try it on a wooden table:NOTHING, it works only when the strip is on the ceramic floor
Try it with a paper sheet between the floor and the strip IT WORKS
Try it with the whole sketch book: IT WORKS but very dimm
OK probably some mysterious EM field is causing this but maybe the strip has some hidden capacitors or capacitance behavior :
Try it with another strip: IT WORKS
Try it with a 9v battery doesn't work(for the batery i used only 4 and less LEDs and the light is very bright when connected properly ) Try it with normal long legs red, yellow and green LEDs near the floor: doesn't work
the strip has glass buttons shaped defusers over each of the 4 leged LEDs and they don't light with one wire without removing it or maybe not visible . It also have other tow secondary side terminals that produce dimmer light for the same source than the main marked terminals when properly connected but nothing happens when a single wire is connected to either one of them.

What's going on?

Edit: Measuring the main AC voltage with the multimeter with only the red probe to the line(hot) while the black probe with its wire laying on the ground and not connected to anything gives 38 volts. Extending the black probe by conecting it to a other twi probes it reads 51 volts AC.

This is a first draft of answering my own question and I'm not an electrical engineer nor an electrician so it can make no sense :If anyone can improve it and link it to the LEDs and the drivers and explain how they emit even by connecting the black wire alone then I will accept that answer.
my theory that doesn't explain the whole story is: This is happening in the first floor in concrete building where the house is grounded or the neutral is connected to the reinforcement steel cages between the floors and when the LED strip circuit or the COM probe from the multimeter is on the floor it interacts with the steel reinforcement cage by inductance similar to what happens in a transformer, if that is at all what is happening in a transformer.
This came to my mind because some years ago after discovering that a similar house was not grounded an electrician suggested to me to do the grounding to the reinforcement cages.

  • \$\begingroup\$ Have a look at this: tingling-sensation-by-touching-stuff I have a LED light in my house which at night glows but not when I reverse the plug in the socket (Europe with phase & ground plugs) \$\endgroup\$
    – Oldfart
    Feb 29, 2020 at 8:39
  • \$\begingroup\$ @Oldfart That's cool man you have a night sky stars simulation in your room. Can you tell me the lamp exact brand and model. \$\endgroup\$
    – Amr Berag
    Mar 1, 2020 at 20:12

1 Answer 1


There is often much capacitive coupling of the net power in a building or room. On the floor and walls might be even an invisible very thin layer of condensed water and dust that results in a huge capacitor to the rebars resp. grounding.

The LED strip is lighting since the phase/hot pole is also somehow coupled via the power supply into both DC wires. Or it is coupled in the body of the person that holds the strip and then into the DC-wires.

So it does not matter which power supply's pole is connected to which LED strip's pole since the AC is anyway rectified by the LEDs.

But this current is very tiny, i.e. the strip is only very dim. Modern LEDs are very sensitive, can light up with those tiny currents. Since multimeters have huge inner resistance, it is possible to measure the (fluctuating) voltage as well.

  • \$\begingroup\$ But LED is a diode where current flows only in one direction \$\endgroup\$
    – Amr Berag
    Mar 1, 2020 at 20:04
  • \$\begingroup\$ That is why the polarity doesn't matter with AC coupling. But the flickering should be visible since the LEDs are only emitting light during less then half of each 50 or 60 Hz period. Flickering is better visible by moving a hand with straight spaced fingers up and down in front of the LEDs. \$\endgroup\$
    – xeeka
    Mar 1, 2020 at 22:30
  • \$\begingroup\$ the current was so tiny the DMM wasn't able to detect it \$\endgroup\$
    – Amr Berag
    Mar 3, 2020 at 11:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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