Yesterday I had a BBQ with some friends. The sun had already set and the only light source left (besides some ambient light from the world around) was a low energy light bulb.

After a while I started to see lighting changes in the faces of my friends and the number plates of their cars. It felt like someone toggled the light very fast. When looking at the wall or the light directly I didn't notice any flickering.

In my country the power grid is running at 50 Hz. Is it possible that I actually saw the flickering caused by the alterations in the power grid or am I just going insane?

  • 2
    @Criggie No intoxication whatsoever - funcact: my friends asked me the same question when I told them :D – Timo Aug 19 at 21:38
  • 3
    You didn't mention if this was an incandescent bulb. If so, it seems the filament's brightness wouldn't change that much during a 50 hz cycle, but it's possible. – rcgldr Aug 20 at 2:41
  • 1
    @rcgldr It was some sort of energy saving bulb (Not sure which type, but not LED). Not a incandescent bulb. – Timo Aug 20 at 10:42
  • 1
    @user21820: This is quite noticable (at least for me) with LED car tail lights, especially at night when I'm a bit tired. Sometimes traffic signals & advertising signs, too. Part of it, I think, is that they often use pulse width modulation to control the light intensity... Also note that a bad or failing LED or CFL bulb can flicker. – jamesqf Aug 20 at 15:53
  • 1
    @jamesqf: Yes indeed. Especially those that deliberately pulse on and off, but what I found interesting was that some fluorescent tubes that give 'warm light' are not monochromatic, and I discovered it exactly by flicking my line of sight across them, the first time accidentally. – user21820 Aug 20 at 16:00
up vote 33 down vote accepted

Short answer
Yes, the flickering of a light bulb may be noticeable, and yes, that's directly related to the mains frequency. However, since the flickering of a bulb is about two times higher than the temporal limits of our visual system, it is unlikely to be perceivable.

Background
The temporal resolution of the visual system can be quantified in a number of ways. As you are referring to a relatively simple flickering stimulus, the critical flicker fusion frequency is probably the most relevant. At a certain critical frequency, a flickering stimulus will appear as a continuous stimulus. This critical flicker fusion frequency limit is around 50 Hz, but variable between 5 - 50 Hz, dependent on the lighting conditions (Kalloniatis & Luu), see Fig. 1 below.

For example, the turn signal of a car is obviously flickering (flickering in the 1 Hz region). But an object displayed on a standard flat-screen computer seems steady. A monitor's refresh rate is typically 60 Hz, which is indeed above the critical flicker fusion frequency (Holcomb, 2009).

However, the good old CRT screens can sometimes seem to be flickering. The mains, as you indicate, is indeed 50 Hz (Europe, Australia) or 60 Hz (US), and indeed the flickering is at this frequency. Similarly, well functioning fluorescent tubes seem to flicker on occasion (when they are reaching their end they start to flicker too, but that's because of a failure of the device rather than the mains frequency peaking through). Due to a similar effect, light bulbs may seem to flicker too. However, because of the sine wave characteristic of the mains alternative current, featuring two peaks per wavelength (a negative and positive peak, the flickering of a light bulb is actually two times the mains frequency, or 100 - 120 Hz. This is quite far above the critical flicker fusion limit and hence will likely not be noticeable.

It's interesting to see that you mention that it was around sunset. Scotopic vision (night vision) is mediated mainly by the rod photoreceptors. The rod visual system mediates gray scale vision at low-lighting. While spatial resolution is poor, it's very well adapted to process fast-moving stimuli. Hence, the flicker fusion frequency under scotopic viewing conditions may indeed be higher; i.e., flickering of light bulbs may not be perceived during the day (Federov & Mkrticheva, 1938). Nice add-on there.

To add to this as alluded to in the comments, whether the flicker of mains-grid powered appliances are actually visible depends on a lot of factors other than flicker frequency. CRT screens, for instance, may have improved phosphors that have delayed response times, 'smearing' out the flickering into invisibility. In other words, it's not a simple matter of 'ON' and 'OFF'. Likewise, light bulbs heat up and hence the temperature difference might not be noticeable to us, as the temporal flickering depends on heating and cooling of the wire.

FlickerFusion
Fig. 1. Flicker fusion as a function of stimulus intensity. Note that the shape of this graph means that photopic vision is less sensitive to temporal changes; the intensity scale relates to the stimulus intensity, as alluded to in the other answer. Scotopic vision to promote the temporal resolution of vision in the sense mentioned in this answer alludes to the ambient lighting conditions conditions. source: Kalloniatis & Luu (2007)

References
- Federov & Mkrticheva, Nature; 142: 750–1
- Holcomb, Trends Cog Sci 2009; 13(5): 216-21
- Kalloniatis & Luu, WebVision, chapter "Temporal Resolution" 2007

  • 1
    Great answer, thanks! I think, besides your point, if the sun weren't already set, it would have inferred with the light from the light bulb anyway, making it impossible to notice. Glad I'm not going crazy :D – Timo Aug 19 at 18:47
  • 4
    To add a data point, with CRTs I had to go up to around 80Hz or so - 85 being one "standardised" (VESA) refresh rate. Anything less and I had a sense of the flickering "out of the corner of my eye". – Will Crawford Aug 20 at 0:49
  • 1
    LCDs generally are better, even at 60Hz, but I had the privilege of using three Hanns·G monitors in $JOB[-2]. They would operate at 75Hz, and got rid of the semi-permanent mild headache I'd had previously. – Will Crawford Aug 20 at 0:56
  • 1
    @WillCrawford - Part of the flicker effect from a CRT depends on the persistence of the phosphors. The greater the persistence, the slower the rate at which the phosphors light decays, Greater persistence is good for flicker but, can cause blurring of virtual objects moving across the screen. Some fixed field entry monitors, where scrolling wasn't used, like the green monochrome IBM 3740 terminals, had a persistence around 1 second, which allowed for very crisp fonts and no apparent flicker. For many standard CRT monitors, the flicker goes away at around 75 hz. – rcgldr Aug 20 at 2:38
  • 1
    @WillCrawford - continuing, CRT televisions with interlace, only update "pixels" about 30 times a second, but flicker is usually not an issue. The persistence of the phosphors used on CRT televisions is designed to avoid flicker at these lower frequencies. – rcgldr Aug 20 at 2:40

A lamp flickers at 2x the mains frequency, i.e. 100 or 120 Hz, and that is typically not noticeable to human eyes. It is visible to chicken and insects.

That being said, a low quality lamp or a lamp at end-of-life may also flicker at 50 or 60 Hz, and you will notice. It depends on the brightness, so an area illuminated by the lamp may not seem to flicker.

A simple way to suppress the flicker of a 60 Hz CRT is to put on sunglasses. The chemistry in your eyes is slower at low brightness, this makes the flicker less visible. The invention of 100 Hz CRT TV (I was involved) was necessary for allowing higher brightness.

  • +1 for your A lamp flickers at 2x the mains frequency, i.e. 100 or 120 Hz. Can you add sources? I adapted my answer to allude to this effect too. Thanks. – AliceD Aug 20 at 7:12
  • 2
    A 50 Hz true-sine wave will pass through zero amplitude twice per cycle: once going toward its positive extreme (climbing through zero), and once going toward its negative extreme (falling through zero). (I'm ignoring bias voltage here.) The entire cycle is from e.g. zero with slope X all the way to zero with slope X, but two zero-passes happen, at slope X and slope -X. So the instantaneous voltage will be zero 100 times per second for a 50 Hz AC waveform (120 times/second for a 60 Hz AC waveform). I don't know the extent to which this actually impacts the intensity of the lamp's light output. – Michael Kjörling Aug 20 at 9:37
  • 1
    While you are correct about the sine wave and the math, I dealt with this same issue with a set of LED Christmas lights. The LED lights are wired in a simple series, and they only get forward voltage 60 times a second (60Hz in my location). Since LEDs only work with voltage flowing one way, the flicker is very noticeable in my peripheral vision, but not straight on. If the OP was using cheap LED lights, it's possible this same thing was happening. – JPhi1618 Aug 20 at 15:58
  • @JPhi1618: LED drivers can use a diode bridge and a smoothing capacitor. Which means only cheaper LED lamps ought to have significant variation in luminous flux over the mains cycle. – RedGrittyBrick Aug 20 at 16:03
  • @RedGrittyBrick, oh, yea, of course they can and most LED bulbs I have seen look great with no flicker, but on the extreme cheap end, like the string of lights I have, there is nothing done to smooth out the light or reduce flicker. Sorry if it seemed like I was talking about all LED bulbs. – JPhi1618 Aug 20 at 16:05

Let's say there is a point source of light (it could be a lamp or a highly reflective object) that undergoes large, rapid intensity changes, say 50-100 times a second.

If you quickly move your eyes while it's in your field of vision, it will trace out a path across your retina. Some sections of this path will receive little light, while others will receive a great amount. What you see will look like a dashed line. (The flicker fusion frequency is irrelevant because it refers to fixed points in your field of vision. In this case we are dealing with many spatially distant photoreceptors.)

Let's say it takes 0.15 seconds to "flick" your eyes from right to left. This means a light flickering at 100Hz would be broken up into 30 "off" sections and 30 "on" sections during this time. So in fact you would be able to detect frequencies much higher than 100Hz. (This could make an interesting Arduino experiment.) I have noticed the effect when my laptop varies the brightness of its charging light using PWM. As it "darkens", the dashes in the dashed line get shorter, and vice versa.

But let's put this in context. The conditions you describe mean that:

  • There is a big contrast between reflective objects and the background
  • There is a big contrast on some objects when the lamp is at its brightest and at its darkest (assuming a LED lamp without smoothing capacitors, the objects could receive virtually no light from the lamp for certain periods)
  • The objects are not "point sources"

This means that whenever you move your eyes around, a quick succession of bright "ghost" objects will additively combine in your field of vision. The effect will probably look like a sped-up version of strobe lighting. The same will happen with moving objects (e.g. someone waving a hand). But if you fix your gaze on a stationary object you will probably not see flickering.

Your Answer

 
discard

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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