7
\$\begingroup\$

Flickering LEDs are everywhere. I can think of a number of reasons why an LED might flicker:

  • Supplied from an AC outlet (traffic lights in ex-USSR)
  • Addressed one chunk at a time (most multi-position 7-segment displays, for example)
  • Fading (dimming)

But some occurrences have me stumped. In particular, the LED-based stop lights on some cars. I can see no reason for these to flicker: they are supplied with DC, and they are either entirely ON or entirely OFF, so no addressing issues and no dimming. And yet they flicker.

Is there some reason why someone would intentionally make LEDs flicker in such circumstances? Perhaps one can get more light per watt out of them this way or something like that?

\$\endgroup\$
0

6 Answers 6

6
\$\begingroup\$

On automobiles, the same LEDs are used for both taillights and brakelights. When braking, they're on full — no flickering — but when they're just taillights, they're dimmed, so they flicker. Very annoying.

\$\endgroup\$
2
  • \$\begingroup\$ So they dim them by driving them at a lower frequency? \$\endgroup\$
    – Tony Ennis
    Nov 18, 2012 at 15:19
  • 1
    \$\begingroup\$ Actually, a higher frequency. For braking, they're driven with DC (0 Hz), but otherwise, they're driven with a PWM signal (a few hundred Hz). \$\endgroup\$
    – Dave Tweed
    Nov 18, 2012 at 15:54
4
\$\begingroup\$

LEDS used for illumination are pulsed for two main reasons:

  • Dimming
  • Power savings.

Dimming by pulsing is ok, but the flicker and generated noise if the LEDs are remoted are not ideal. LED drivers are better because they are essentially switching regulators that vary their voltage to keep the LED current constant. This eliminates flicker, and you don't get varying currents in the line, plus dimming is still possible.

On the other hand, pulsing can save power because of the way our eyes work, being a combination of light peak detectors and rms detectors.

What this means is that even if the average power delivered to an LED is the same, a pulsed LED will appear brighter to the eye, because during the ON cycle it was brighter than average. There have been studies regarding the sweet spot for this, and IIRC it lies somewhere around 5-20 % duty cycle.

\$\endgroup\$
2
\$\begingroup\$

Did you not mention what the flicker rate was? < 10Hz ? > 30Hz? only on activation of brake lights?

There are at least three question/answers.

  • Why does it flicker? (=PWM effect)
  • Why do we perceive it off center more as flicker (human optics)
  • Why do we perceive it more as modulated streaks on center when the car is moving? (human optics)

Now some details: Let me address some general common issues;

1) Cadillac LED brake lights "were" notorious for causing eye flicker response. The cause was a PWM rate was high enough to be unnoticed when direct viewing the rear end of a stationary car, but noticeble when the car is moving while braking and more so, from your peripheral vision. Both conditions increase the human eye-brain sensitivity to flicker. This is well documented.

**added* This graph shows the eye distribution of Rods & Cones. Rods, used for night vision, have a broader Field of View (FOV) from center than cones. Rods are also more sensitive but active only at low ambient light levels. This accounts for the increased sensitivity of flicker on peripheral vision at night.

enter image description here

2) When the eyes move, the human vision is more accute to the flicker while they are in the peripheral. Once the eyes view the car directly while it is slowing down at night, they produce a stream of modulated streaks, which are stored in the retina . However the cause of both is the LED PWM intensity control used while braking, introduced by Cadillac around 2000. PWM helps to regulate the full brightness at a range of voltages such on low idle. This can be smoothened by adding a large Cap across the LEDs, but the RC time constant increases as duty cycle is reduced, so it is mroe effective when dim. given brake LED's have an ESR << 1ohm, it would take a rather large cap to filter out 5,000 uSec. Increasing the PWM> 500Hz would make it less noticeable.

"Digital projectors flash images on the screen 144 times per second, for flicker-free motion. But, as with older films, there are only 24 different pictures (frames) per second. Film projectors generally flash each image on the screen 2 times to reduce flicker, using a special shutter. Digital 3-d projectors flash the image for each eye 72 times per second upon the screen." ref

\$\endgroup\$
3
  • \$\begingroup\$ Flicker of a glowing moving object will cause it to appear as multiple disjoint streaks rather than one. This is the principle upon which spinning "persistence of vision" displays operate. \$\endgroup\$
    – supercat
    Nov 18, 2012 at 15:42
  • \$\begingroup\$ The frequency is higher than 42 Hz, which is the frequency at which I personally stop noticing flicker when looking straight-on. But I can easily see the flicker by just moving my eyes around. Last time I saw them was during daytime, and yes they were activated upon braking. \$\endgroup\$ Nov 18, 2012 at 17:38
  • \$\begingroup\$ If an object is moving, flicker may be visible at speeds up to (rate of motion) divided by (size of object). If a small light which is spinning at e.g. 60.00Hz is modulated at precisely 10,000.0Hz 10% duty cycle, it would appear as 200 little lights each flickering at 20Hz, whereas if it was on solidly it would simply appear as a continuous circle of light. To avoid flicker, the light would have to be modulated sufficiently fast that "little lights" that appeared could not be visually separated. \$\endgroup\$
    – supercat
    Nov 18, 2012 at 19:38
2
\$\begingroup\$

The real problem lies in the misapplication of the figures for flicker response in the human eye to completely different subject matters. In a darkened room, viewing a brightly lit screen with a full range of colors, the figures of merit were arrived at long ago, in the mid-20th century. a refresh 70 hertz or more does impart a flicker-free image for viewing by observers equipped with human eyes. Old style CRT's could get away with 60 hertz mainly because the phosphors used continued to glow for some time after they were excited by the scanning beam. However, the typical automotive tail-light with a mono-chromatic source, such as red LEDs, on a dark night, with a duty cycle in the flicker being less than 50%, the percievable flicker rate is much higher, on the order of 120 hertz or more, depending somewhat on how quickly the given LED comes on and shuts off when power is applied. Yes, I can see the flicker in the newer LED car tail-lights also, some are bad enough to give me motion sickness in some situations!

\$\endgroup\$
1
\$\begingroup\$

Automotive Incandescent Bulbs for Combination Parking and Turn Signal or Brake Light have two filaments in them, one is a dim filament and the other a bright filament. There are two circuits to the bulb, one to light the dim filament and the other to light the bright element or Flash it.

Most of the LED bulbs designed to retrofit for these type of bulbs have two sets of LED's, dim LED's for the dim circuit and bright LED's for the bright circuit. At the same time, I've only have experience with a very limited amount of Automotive Retro-Fit LED bulbs, so take this with a grain of salt.

I doubt the flicker from the bulbs would be from a schema of dimming the bulb for the parking lamp, since its clearly simpler to take advantage of the two separate and discrete circuits. I suppose you could make an LED bulb that dims or brightens triggered by the two circuits.

I suspect, since Automotive retrofit LED's are in the Wild West environment right now, technically none of them are legal and many people fail state safety inspections if they have them installed, there are a lot of manufacturers producing extremely bright LED bulbs, to the point where heat becomes a concern and thus use switching regulators within the bulb that causes flicker.

I've used the Sylvania retro-fit LED bulbs, and like you'd expect from a large reputable company, their bulbs are well designed and manufactured and have no flicker or other issues. But Sylvania doesn't make a Fog Lamp bulb for my vehicle, I got some cheap LED's bulbs off the internet, that design mimics some of the really bright LED's for headlamps and fog lamps. These noticeably flicker when looking directly at the bulb, they create noticeable noise/EMI within a few feet (cell phone picks up noise) or for a faint distant AM radio station noise is picked up from the LED's. Not noticeable from the light reflected from the road surface, so for the exception of trying to listen to a distant weak AM radio station the flicker really doesn't interfere with operating the vehicle. Since these Fog Lamps LED Bulbs are significantly brighter than Parking Lamp or interior illumination bulbs, in such a tiny package, its reasonable to speculate there may be some driving/regulation going on to handle the heat that is causing the flicker and noise.

\$\endgroup\$
0
\$\begingroup\$

Enough LEDs will suck down some power. For this reason, it's common to drive them individually, but fast enough that the eye can't see the flicker. Drive them too slowly and they flicker. If you really want 50 LEDs on at once, it will cost you 50x the power of just driving them one at a time.

\$\endgroup\$
2
  • \$\begingroup\$ If you're driving them one at a time, your maximum possible duty cycle is 2% per LED. That'll cause them to be quite a bit darker. Can't you achieve the same by using a bigger resistor for each LED? (which you're probably going to have there anyway to balance the currents and keep brightness close to each other) \$\endgroup\$ Nov 18, 2012 at 17:35
  • \$\begingroup\$ They could be driven by multiple circuits - there are innumerable ways to do it. What I posted is true but may not apply as directly to the use of LEDs in the automotive industry. For specifics, I'd look at other answers, to be honest. \$\endgroup\$
    – Tony Ennis
    Nov 18, 2012 at 17:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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