I have recently purchased an IKEA LED bulb with G9 fittings, to replace a halogen one. This one.

The bulb is installed in a lamp with 3 other halogen G9s in a 3 + 1 configuration with a 3-state switch. "up" lights up all four bulbs, "down" lights up only the socket where the LED is.

I have found a strange behaviour. When the lamp's switch is in the off position, the LED bulb still emits a very dim light, whereas the halogen ones do not.

What is happenning here? Is this normal?

From my limited knowledge I would guess the lamp's circuit has some residual current which is not enough to light up the halogen bulbs, but makes the LED emit the dim light observed.

  • 3
    \$\begingroup\$ Seems legit. You can get visible light from microamp-range currents. \$\endgroup\$
    – KyranF
    Commented Nov 16, 2014 at 22:15
  • 3
    \$\begingroup\$ Is there any sort of fancy light switch? Dimmer? Night light? Wifi controlled switch? Switches that need power but dont use the neutral wire often exhibit this behaviour. \$\endgroup\$
    – Grant
    Commented Nov 17, 2014 at 4:09
  • 3
    \$\begingroup\$ When you say it continues to glow, does it glow continuously, or just for seconds/minutes? If the latter, what you are seeing is the phosphors continuing to glow. Most "white" LEDs are actually blue/violet/UV LEDs exciting phosphors to produce white light. When the LEDs turn off, the phosphors continue to glow off a little while. No additional energy is being consumed. \$\endgroup\$
    – DoxyLover
    Commented Nov 17, 2014 at 7:58
  • 3
    \$\begingroup\$ It glows continuously, and doesn't fade with time. Unplugging the lamp from the wall socket or the bulb directly makes it stop emitting light. \$\endgroup\$ Commented Nov 17, 2014 at 9:14
  • 3
    \$\begingroup\$ I once had a similar thing happening for a lamp, and it turned out that turning the wall plug by 180° (thus switching phase and neutral) "fixed" it. \$\endgroup\$
    – PlasmaHH
    Commented Nov 17, 2014 at 10:17

5 Answers 5


There are two possibilities the switch is inserted:

  1. Switch switching the voltage line.
  2. Switch switching the GND line.


simulate this circuit – Schematic created using CircuitLab

The capacitors shown in the circuit are the capacities the (more or less) long lines form to GND.

If the 2nd circuit is what you have the lamp always is connected to alternating voltage. In that case there is a possibility of some very low alternating current flowing via C10 to GND even if the switch is open.

You can find out if your configuration is the 2nd circuit by testing with a one-contact neon test light. It is the case if the test light lights up when touching one of the two connections in your lamp socket even if the switch is turned off.

  • 3
    \$\begingroup\$ I took @PlasmaHH's suggestion from a comment above and flicked the wall plug 180 degrees. It is no longer happening, so it seems that this lamp is wired like circuit #2. \$\endgroup\$ Commented Nov 17, 2014 at 19:27
  • 1
    \$\begingroup\$ -1 Please do not use those single contact testers. They are incredibly dangerous at large voltages: youtu.be/AGXQNLq19FQ Please use something like a multimeter instead... \$\endgroup\$
    – George
    Commented Nov 18, 2014 at 8:28
  • 5
    \$\begingroup\$ @George: 230V is not an incredibly large voltage and single contact testers are made exactly for the purpose of testing that voltage. So there is no reason to warn about using them (assuming the person using them knows what he is doing; otherwise it'd be also dangerous to use a hammer) \$\endgroup\$
    – Curd
    Commented Nov 18, 2014 at 16:40
  • 3
    \$\begingroup\$ A multimeter is regrettably no safer when dealing with high voltages - just avoid mains and greater until you have specialist knowledge and training. Even 480v arc-flash is no laughing matter. \$\endgroup\$
    – rdtsc
    Commented Jun 22, 2016 at 17:30
  • \$\begingroup\$ @George I suppose Darwin awards await anyone who goes for 1kV lines with a test light. \$\endgroup\$ Commented Aug 19, 2019 at 14:22

I would guess that your off-low-high switch has an indicator light so you can find it in the dark, so the LED lamp ends up in series with the indicator when the switch is 'off'.

Edit: For the lamp to emit visible light, current must be leaking across the switch.

Possible reasons are:

  • Indicator light (ruled out by OP)

  • Capacitor or RC snubber across switch to avoid EMI when switch is flicked (not seen often in North America, but might be a possibility in Europe).

  • Dimmer circuit that has an internal snubber and is not switched off entirely

  • damage (arcing and tracking or moisture) to the switch. A mechanical switch by itself should not leak enough to light an LED.

It will not draw more current than it would with the halogen lamp, and all but the last cause are nothing to be concerned about.

  • 1
    \$\begingroup\$ That is not the case. There is no indicator light. \$\endgroup\$ Commented Nov 16, 2014 at 22:57

I took a hint from a previous comment and flicked the plug 180 degrees on the wall outlet. The bulb doesn't emit light when in the "off" position anymore.

It seems that the lamp's circuit was interrupting the GND line, like illustrated in @Curd's answer.


Added for safety value - even though a very old question:

Andre said:

I took a hint from a previous comment and flicked the plug 180 degrees on the wall outlet. The bulb doesn't emit light when in the "off" position anymore.

It seems that the lamp's circuit was interrupting the GND line, like illustrated in @Curd's answer.


If reversing the polarity of the circuit fixed the problem then you probably have Phase / Live connected to the circuit at all times and it is improperly wired and a "death trap".

If only lighting equipment without grounded metal accessible to users is used then you may not have anyone killed.

But if you use equipment on the circuit where neutral and ground are connected (as happens even though it shouldn't) then such equipment will be "safe enough" on a properly connected circuit BUT lethally dangerous on this one.

Getting it fixed while everyone is still alive is liable to be "a good idea".


Phosphor emits light at total darkness, but it is dangerous because of radiation, radio active radiation, so this new LED bulbs may contain a small amount of chemicals like phosphor or something like to improve the light output of the bulbs, that's why they gloing even after shut them down for quiet some time.

  • 6
    \$\begingroup\$ You have to distinguish between the chemical element Phosphorus and phosphors. And the phosphors used in LEDs don't contain radioactive elements, because their purpose is not to glow on their own, but to convert UV light into white visible light. The kind of phosphor you think that contains radiactive material uses that radiation to glow in the dark for several years (until the radiactive material is "used up"). \$\endgroup\$
    – 0x6d64
    Commented Jul 3, 2015 at 10:27
  • 2
    \$\begingroup\$ Oh dear, no. Don't listen to this answer. \$\endgroup\$ Commented Oct 10, 2016 at 8:14
  • 3
    \$\begingroup\$ Some people who have up voted here may not have understood that this answer seems to be confusing radium dials with white LED phosphors. The wording should be corrected to be less confusing so the misunderstanding is more obvious and the answer can then be voted down on its merits. \$\endgroup\$
    – KalleMP
    Commented Feb 21, 2017 at 10:20

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