14
\$\begingroup\$

Continuous light - do LEDs flicker like fluorescent lights?

Under fluorescent light there is a flicker.

As I understand LEDs work only with DC current, thus they have a AC to DC transformer circuit pumping electricity into them. That would mean that they emit continuous light.

\$\endgroup\$
  • 3
    \$\begingroup\$ Actually, fluorescent lights flicker at twice the line frequency, or 100-120 Hz. \$\endgroup\$ – Dave Tweed Feb 26 '13 at 12:15
  • 1
    \$\begingroup\$ Florescent lights flicker at the ballast frequency, which for T12 is 100-120hz, but for T8 and T5 can be several khz. \$\endgroup\$ – Bryan Boettcher Feb 26 '13 at 15:05
  • \$\begingroup\$ I tested it with the swinging the extension cord method mentioned in Olin Lathrop. Surprisingly I found my Osram Led lamp do flicker, while the Philips fluorescent lamp doesn't flicker. I swing the lamps along a circle in the condition of radius = 0.2 meter, rps = 2. So the flickering is expected to be 2cm for 60HZ. So indeed flickering happened markedly in the led lamp. While the swing in the fluorescent lamp is imperceptible. I can't just comment because of my low reputation. \$\endgroup\$ – user26810 Jul 26 '13 at 22:50
  • \$\begingroup\$ Using solar cell and a scope, you can see the "flicker" if there is any. The output of the solar cell changes with light intensity, so probing the output of the solar cell will show if there is a steady or chopped supply. If you have a multimeter with a frequency counter, you can probe one of the LEDs and see what frequency it is switching at. If you get a zero, you have constant current and no flicker. \$\endgroup\$ – Enemy Of the State Machine Nov 17 '14 at 3:18
19
\$\begingroup\$

LEDs will emit steady light if given steady current. The question then becomes what kind of current waveform LEDs are driven with in a light assembly.

The exact answer depends on the circuit in the light. LEDs run on just a few volts. The much high line voltage has to be converted to the lower LED voltage somehow. Most likely, this will include the cheapest, dumbest, and most stripped down switching power supply possible.

I see other answers mention a transformer, but I think that is quite unlikely. There is no need for isolation in a sealed unit with no external connections other than the power line. Even if there is a transformer, it's not going to be fed directly from the line frequency. The extra savings in cost and size of a transformer that works at 10s to 100s of kHz far outweighs the cost of the components to produce that frequency.

Most likely, there is a full wave bridge to rectify the AC line voltage directly. That will then be chopped thru a inductor to drive the LEDs. Depending on how cheap the lamp is, it might chop at a fixed duty cycle, which would make the LED brightness vary with power voltage. Even a little current feedback would keep the LED current reasonably constant over most of the line cycle, perhaps dipping only briefly at the power line zero crossings. A small cap would reduce that, but caps cost money and take space, so may not be included.

LED lamps are made in large volumes, so serious manufacturers probably develop custom ICs just for this purpose. In such a case I'd expect at least some attempt at regulation, so the LED brightness will be largely constant, perhaps with short dips at the power line zero crossings.

However, all this is speculation. Why not just look? Put the lamp at the end of a extension cord and swing it around in a otherwise darkened room. Whether it flickers or not will be immediately obvious.

|improve this answer|||||
\$\endgroup\$
  • 6
    \$\begingroup\$ I'm loving the mental image of someone with a lamp on an extension cord whipping it around over their head. \$\endgroup\$ – Justin ᚅᚔᚈᚄᚒᚔ Feb 26 '13 at 16:10
  • \$\begingroup\$ Gave it a little spin... i didn't see any alterations. It seems it's continuous. I used to do this as a kid with music players to hear the sound "stretch" and "contract" as you spun it faster and faster. I'm using Ikea led lights ikea.com/gb/en/catalog/products/00218141 Are there any other tests i can undertake? \$\endgroup\$ – unom Feb 27 '13 at 21:14
  • \$\begingroup\$ The "trail" it has is continuous and uniform in color. Final vote... it continuous. I'm 99% sure of it, i'm quite sensitive of these things. For example, if i use my macbook near a window, under natural light, i can't dim the screen below 80% because i start noticing flicker. I noticed a distinct feeling of comfort while reading under this light. I love it, very similar to incandescent. \$\endgroup\$ – unom Feb 27 '13 at 21:29
  • 1
    \$\begingroup\$ I've taken a number of US LED bulbs apart. A few do actually use transformers, complete with isolated secondaries. They generally seem to use primary-side feedback, though, so they can skip the cost of the optoisolators. \$\endgroup\$ – Connor Wolf Jul 27 '13 at 9:23
  • 3
    \$\begingroup\$ @Peter - Basically, the idea is to move the lamp fast enough that the flickering becomes visible as interruptions in the light-trail the bulb leaves in your eyes. Basically, your eyes integrate light for a period of time. If the bulb is moving fast enough that it travels more then the diameter of the bulb in the period of the flicker, you will perceive an interruption in the trail of light you see. \$\endgroup\$ – Connor Wolf Jul 27 '13 at 9:25
8
\$\begingroup\$

Your assumption is correct that when a LED is fed a stable DC current it will not flicker. However LEDs that run from an AC source such as mains will normally use a rectifier to convert the AC to DC, the following diagram from Wikipedia illustrates a full-wave rectifier:

Full-wave rectifier

A capacitor is typically used to smooth the output for applications where stability is important, but if the value is too small relative to the current drawn, or not used at all, a flicker is certainly possible.

I wouldn't be suprised if a lot of "drop in" LED replacement lamps and LED supplies omit the capacitor to reduce size and cost. Other than flicker there's no drawback I can think of so it's a reasonable design compromise.

|improve this answer|||||
\$\endgroup\$
  • 3
    \$\begingroup\$ I really don't think there will be a transformer in a LED lamp. And even if there is, it won't be running at the power line frequency. \$\endgroup\$ – Olin Lathrop Feb 26 '13 at 14:42
  • \$\begingroup\$ @OlinLathrop, I agree with your answer / comment this was more of a conceptual explanation of why they can still flicker after the AC to DC conversion. I've personally purchased a device consisting of a LED array that is just a lot of LEDs in series connected directly to the mains with a wirewound resistor and diode. \$\endgroup\$ – PeterJ Feb 26 '13 at 15:02
  • 2
    \$\begingroup\$ @OlinLathrop I've been at the LED lighting trade show a couple of weeks ago. Oddly enough, there are lots of off-line LED lamps with transformer isolation. They use flyback topology with frequency between/around 65kHz and 150kHz. The lamps without a transformer (off-line buck) and case isolation exist too. [Come to think of it, this deserves a thread on EE.SE .] \$\endgroup\$ – Nick Alexeev Feb 26 '13 at 19:42
  • \$\begingroup\$ I would think including a transformer will also take care of reducing the line voltage instead of droping the 55V rectified across the LED chain (although maybe they have enough LEDs in the package to justify the 30V RMS) \$\endgroup\$ – crasic Jul 26 '13 at 23:10
6
\$\begingroup\$

LED's are still diodes and can be used to rectify AC to DC. Some of the least expensive LED lights are simply comprised of strings of LED's wired such that the sum of forward voltages comes close to the peak line voltage with a small inline resistance for current limiting. THis is simply placed across the plug that goes into the mains. A second string of lights can be placed in the opposite direction for the opposite phase. SO the LED will light up for only 1/2 of the 50/60 Hz cycle.

OF course many variants of this wiring, like using strings of these LED's in a bridge configuration.

This is how LED christmas tree lights are wired, it essentially requires ZERO external components (except the resistor) and results in a flickery string of lights.

If you were to take those same strings of lights and apply a higher voltage DC to them you'd get steady light on only some of the lights.

This technique will be used in several situations where the manufacturer does not want to produce a control circuit.

Of course there are many different variations, the key point is that you don't need any extra circuits if wired correctly and you can accept flicker.

|improve this answer|||||
\$\endgroup\$
  • 3
    \$\begingroup\$ It's good to note that this can produce fairly hilarious results. My first LED setup was just like this, then an LED blew. That raised the voltage on every LED, and they began to blow until I had ~50 LEDs all dead. \$\endgroup\$ – Thebluefish Feb 19 '14 at 12:57
2
\$\begingroup\$

I own quite a few LED lamps (A form factor). What I do to check for flicker is to use a spinner-top with a pattern of closely spaced lines, one inner one outer. When spun under an incandescent lamp or (strangely enough) compact florescent lamps the lines will blur. If there is any flicker, then you get the "wagon wheel effect" where the inner and outer lines seem to travel at different speeds and in different directions. Cree is the worst offender and the non-dimmable 10.5w 800lm Philips is bad as well. Samsung, Sylvania, and the dimmable Philips A-line are all smooth. I should also note that I'm referring to the mostly non-perceptible flicker and NOT the flicker that occurs with dimmer switches.

|improve this answer|||||
\$\endgroup\$

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