Debunking the Myth: Do LED Lights Flicker? Get the Definitive Answer Here!

[etotheipi]

I was in disagreement with my dad about this and hoped that someone could provide a definitive answer.

My take was that a LED driven by a rectified DC power supply will not flicker (provided the rectification is sufficiently smoothed out). Whilst he texted me this, taken from a manufacturers website

Many LED bulbs produce flicker - a rapid switching between on and off states. Because the flickering happens at a very fast rate - typically 120 times per second or faster - it is not immediately observable to humans and appears to us as a light bulb of steady and constant brightness.

I thought this was strange. A LED being driven by an alternating current of 60Hz will flicker on and off 60 times per second, but surely that is not how most LEDs are designed?

Thanks!

 

[Keasy]

The light output of a LED follows the current through the device. If you use a constant current to drive it, it will not flicker. This is normally from a battery or well regulated rectified AC source. If you use something like an unregulated AC supply it will flicker on and off at the frequency of the AC supply. For the common 60 hz power line supply, the flicker is normally too fast for us to notice.

 

[etotheipi]

If you use something like an unregulated AC supply it will flicker on and off at the frequency of the AC supply. For the common 60 hz power line supply, the flicker is normally too fast for us to notice.

Awesome, thanks. And I suppose also for simple full wave AC rectification with no smoothing it is also possible to get flickering at 120Hz. As opposed to the 60Hz you'd get if you just connected the LED to the 60Hz AC source with no rectification.

 

[anorlunda]

In theory, one could drive a LED with half wave rectification. But I haven't seen it done.

There are also some LED lights deliberately made to flicker for decorative purposes.

 

[hutchphd]

If you take a string of "Xmas" lights and spin them around your head it is easy to see the half wave duty cycle... nicely dashed colored lines. They will usefully follow the current up to ~50 MHz...useful for phase sensitive detection of fast temporal events.

 

[anorlunda]

If you take a string of "Xmas" lights and spin them around your head it is easy to see the half wave duty cycle... nicely dashed colored lines. They will usefully follow the current up to ~50 MHz...useful for phase sensitive detection of fast temporal events.

I foresee all PF members doing that experiment out in the snow at night next December. LOL

 

[Merlin3189]

I don't know about shoulds and oughts, but LEDs do flicker. The two I notice a lot (more in winter when it's dark) are traffic lights and car rear lights. I notice some flicker in household LED lamps, but I'd say there is some smoothing, as it is much lower contrast than the road lights.

Since the flicker rate in household leds will be at least 50 Hz (probably 100 Hz), I'd have thought flicker would be acceptable and people might not bother with smoothing. I wonder if the smoothing I observe is simply a byproduct of using a capacitor to control current?

For the car lights I wondered whether they are driven with raw AC from the alternator or whether it is inverted from the battery DC. I suppose it must be the latter, else you wouldn't have lights if the engine were off.

 

[anorlunda]

If you look at the list of related threads at the bottom of this page, you'll see that flicker has been discussed before. If I remember correctly, one conclusion was that individual ability to detect flicker varies widely. Some see it, some don't. Maybe drinking carrot juice might change your flicker frequency threshold. That makes is more of a physiology question, than a physics question.

 

[AZFIREBALL]

You can tell if they are flickering if you place a fan close by. The light reflecting on the rotating blades will make the blades look like they are moving slower, standing still or even reversing in rotation if the flicker rate is just right.

 

[OmCheeto]

I was in disagreement with my dad...

I did a test a few years back and discovered that all my household lamps flickered: incandescent, fluorescent, & LED.
The only thing was, I couldn't sense it. (Except in one "dollar store" case.)
It was only detectable with an oscilloscope.
So, I would say that you are both correct.

 

[AZFIREBALL]

Incandescent lights do not flicker (turn off - on), they tend to vary in intensity as the filament cools off between cycles; but may never go out completely, unless the cycle rate is very slow.

 

[Tom.G]

Just to add to the inanity here:
Back when there was 25Hz AC power distribution the, 50Hz flicker of incadescent lamps was quite noticable to most people, especially in their peripheral vision.

And at least some LED automotive brake lights are pulsed, as can be seen if you rapidly scan your vision past them at night.

 

[DaveE]

Whether lights flicker or not depends entirely on the power supply they are connected too.

I have designed power supplies for laser diodes that absolutely didn't flicker because our customers (high tech types) would pay a lot for low noise. OTOH, the lights you buy at the hardware store are intended to be inexpensive, and since your eyes really don't respond much above 50 - 100 Hz or so, they can get away with simpler designs.

It's not really about the LED or light bulb, it's about cost vs. customer requirements.

 

[pinball1970]

If you take a string of "Xmas" lights and spin them around your head it is easy to see the half wave duty cycle... nicely dashed colored lines. They will usefully follow the current up to ~50 MHz...useful for phase sensitive detection of fast temporal events.

Do they have to be a particular colour or just 'white' light?

 

[hutchphd]

They need to be the "cheap" strings (if they have a microprocessor they will also have a DC power supply) and I think the white ones work as well as the pure colors. Usually half the string fires on the + cycle and half on the - cycle.

 

[Dullard]

The apparent intensity of an LED is a product of the instantaneous intensity and the duration. It was not uncommon (20 years ago, anyway) to 'dim' LEDs by controlling the duty cycle. if this is done at a sufficently fast rate, the 'flicker' is not apparent. The threshold for 'sufficiently fast' definitely varies person-person; I've never seen anyone detect a 100 Hz flicker.

 

[hutchphd]

I believe the accepted "flicker fusion" max value for humans is 75 Hz. Obviously (and to the consternation of LCD TV designers) if the objects are in motion we can detect higher rates of variation.

 

[Windadct]

A number of things should be clarified to state specifically what is happening : LED = a Component, and I believe "LED Lights" here is referring to some type of "appliance" using LEDs.

A basic LED on a breadboard fed by DC will not flicker, if the DC has some ripple, then of course a variation MAY be noticeable, but not an on/off phenomena.

Many consumer products are use a LED Driver circuit, which is high frequency (PWM) control of the current in the LED to provide maximum perceived brightness. Especially noticeable on some automobile tail and brake lights. So they definitely flicker.

 

[NTL2009]

Incandescent lights do not flicker (turn off - on), they tend to vary in intensity as the filament cools off between cycles; but may never go out completely, unless the cycle rate is very slow.

Actually, I was surprised to discover that incandescent lights *do* flicker. It's minor, ~ 10% IIRC, not noticeable to the naked eye, but the filament does cool some between peaks of ~ 10 ms. Which makes back-of-the envelope sense. Turn a filament bulb off - it drops to near zero light in way less than one second, so I'd guesstimate a significant reduction in 0.1 second? So 10% in 10 ms makes some sense - though that's not full off, the peak is reducing, then increasing in that time, but close enough.

Ahhh, youtube to the rescue:

Also, some cheap Christmas lights are half-wave rectified, with half the string on one phase, the other half on the other. Makes sense with US 110 V mains. Take a 50 light string for example, 25 white LEDs @ ~3V each would reach full turn on at about 75 V, then some current limiting resistor takes effect up to the ~ 155V peak. So the duty cycle for an individual LED is only ~ 25%, which results in pretty extreme flicker. I find it very obvious just moving the string around in a dimly lit room.

And as mentioned, many LED drivers use a switching circuit that operates at a high frequency so they can use small inductors. Off-hand, I'm not sure if they commonly smooth the rectified AC or not. If they don't, the high frequency would still be modulated by the 50/60 Hz mains (100/120 if full wave rectified) and you'd still have that flicker come through.

 

[HackySac]

I did this very thing a few years ago myself and post a few videos about it because after doing a few searchs for flickering LED lighting I couldn't find anyone else who had done such a thing until I found this page

 

[Merlin3189]

Nice. But I wonder how you know it is the Leds flickering rather than camera strobe?
(I don't actually know whether digital cameras have a strobe effect like the old movie cameras did.)

 

[jrmichler]

I found that a couple of my LED flashlights flicker at roughly 100 to 200 Hz with very roughly 50% duty cycle when running at night in heavy snowfall. The snowflakes showed as dashed lines, from which I could make rough estimates of flashlight frequency and duty cycle.

 

[anorlunda]

Try taking daylight photos of LED signs of the kind used for businesses. It may look very different from what your eyes perceive.

 

[sophiecentaur]

LEDs do flicker. The two I notice a lot (more in winter when it's dark) are traffic lights and car rear lights

I find that very surprising. They are fed from 12V and is it possible that there is a DC - DC converter in them, operating at low frequency?

I wonder if what you are seeing could be due to your vision and vibrations in your car. Plus you could be particularly sensitive to it.

 

[dlgoff]

Here's a "Why LED's flicker" I found by Google searching:

https://www.nbpower.com/blog/en/posts/2018/june/why-led-s-flicker-and-how-to-stop-it/

 

[anorlunda]

Here's a "Why LED's flicker" I found by Google searching:

https://www.nbpower.com/blog/en/posts/2018/june/why-led-s-flicker-and-how-to-stop-it/

None of those reasons apply to a battery operated flashlight.

 

[dlgoff]

None of those reasons apply to a battery operated flashlight.

Yes, I know. Just adding the link for why AC events may cause flicker.

 

[hutchphd]

If you run an LED from a 12V source any linear current regulator will dissipate at least 8V and hence operate at <33% efficiency. A switching regulator will give you 90% or so. I don't know what cars use for incidental LED lights.

 

[anorlunda]

The OP was asking about LED light bulbs indoors. I was thinking about outdoor advertizing signs. Each LED in a sign can be very bright and use a lot of power. Those signs are scanned in a raster pattern like old cathode ray tubes. Only a small number of them are lit simultaneously, in the extreme only one at a time. Persistence in vision makes us see the whole message. The sign below says "1/4 scanning" which I interpret to mean a maximum 1/4 of the pixels can be lit at one time.

 

[NTL2009]

If you run an LED from a 12V source any linear current regulator will dissipate at least 8V and hence operate at <33% efficiency. A switching regulator will give you 90% or so. I don't know what cars use for incidental LED lights.

I was curious also, found an example of using a switching supply so that the same LEDs (string of 3) could easily be used as tail lights (low setting) and as brake lights (high power setting), similar to the dual filament bulbs.

Here's a source from Texas Instruments with one IC controlling tail, brake, turn and reverse LEDs:

https://www.ti.com/lit/ug/tidubs0a/tidubs0a.pdf?ts=1641344881749&ref_url=https%3A%2F%2Fwww.google.com%2F

edit: they use a switching freq of 2MHz to keep out of the US AM band.

 

[Rive]

The OP was asking about LED light bulbs indoors.

I suspect the underdesigned input side caps. These are some 10-47uF sized high voltage electrolytic caps, and while the LED driver will try its best to maintain a constant current through the LED string with high frequency switching, some of the 100 or 120Hz (graetz-doubled line frequency) may still slip through.

 

[HackySac]

Nice. But I wonder how you know it is the Leds flickering rather than camera strobe?
(I don't actually know whether digital cameras have a strobe effect like the old movie cameras did.)

I'm aware of that , this is why its difficult to grasp the whole effect from a screen. when you see it with the naked eye it makes stuff look like a hologram but on a screen it just looks like frame rate effects here is a more recent video of mine ...

 

[HackySac]

Here's a "Why LED's flicker" I found by Google searching:

https://www.nbpower.com/blog/en/posts/2018/june/why-led-s-flicker-and-how-to-stop-it/

I've done many tests using LED strip lights running off different levels of AC voltage using a Auto-Transformer and controlling how much AC volts go into the strip. It was kinda surprising how much AC voltage they could handle before burning out at around 60-70 volts AC. Even though they could run on 30-40 volts AC but at this voltage the heat build up that gets put out by the strip increases dramatically to the point where the PCP pipe I had wrapped the LEDs around began getting soft and to in a sense melt the PCP pipe from the excess heat. The LED lights though to my surprise could handle the heat very well and only burnt out when AC voltage was apply at too high a level and cause the diode to pop and come on only very dimmly after but still not completely burnt out. This tolerance is only done with LED strip lights that have semi-conductor chips in them providing resistance for the DC current. These resistors is what allow for the AC voltage to power the strip without burning out the LED completely. if AC voltage is apply to LEDs that don't have resistors in its circuit will burn then out at 3-5 volts. I know from tests I've done. But strip lights with resistor chips in them can run on lower voltages of AC and when doing so give off a very holographic ghost like blinking effect behind any object that moves like a ghost blinking tail following behind the object moving on what ever these LED lights shine on. This effect is really cool to see with the naked eye and I think could have potential for stage performers or break dancing anything like that. I have yet to test different frequencies with the LED strip lights to see if I can shorten or extend the tail of blinking flashes that follows behind the objects moving when using these LED strip lights but have not gotten to that point yet. Finding time to work on this stuff can be difficult sometimes...