LED bulbs that flicker, and CFLs that almost never did

This is a stroboscope disc used to verifythe speed of a record player’s turntable.

You can easily find these online and printthem out.

Under fluorescent lighting, these alternatingwhite-black bars will appear stationary even though the turntable is rotating.

This happens because the A/C electricity poweringthe light is a 60 hz sine wave, and each time it crosses the zero line, the light brieflygoes dark.

Essentially, fluorescent lights actually flash120 times per second, and the spacing of these bars is calibrated so that if the turntableis going the right speed, they will move the same distance as their width with each pulseof light, which makes a blurred pattern appear that’s completely stationary.

Slight variations in speed will cause thepattern to appear to move.

You can see this as I switch the turntablebetween 33 and 45 rpm.

You may have noticed a similar effect whiledriving at night under common street lighting, particularly the orange-gold glow of highpressure sodium lamps.

These lights also pulse 120 times per second, in the US at least, which can make slow-moving patterns appear on the wheels of vehiclesdriving past you.

Sometimes the patterns move backwards whichis particularly trippy.

This stroboscopic effect is the primary reasonthat some people are sensitive to fluorescent lighting.

Though it’s not directly visible, it cangive some people headaches and cause eyestrain.

But it’s important to note the the fluorescent-nessof the light source is not what’s causing it.

What I mean by this is that it’s very verywrong to assume that all fluorescent lights produce a strobing effect like this.

In fact, nearly all CFLs used in your homedon’t.

Here’s the same disc on the same turntablewith a garden variety CFL providing illumination.

This time, the disc’s lines just blur together.

CFLs have worked like this for a looong time.

In fact, here’s an old IKEA fluorescentlamp.

It’s so old it starts like this.

(forced coughing) And yet, the lines still blur together.

You might notice a very slight pattern inthere that looks stationary, but even an incandescent light will produce such a faint pattern.

In reality, these CFLs are just as flicker-freeas the old bulbs of yore.

But in an odd twist, many newer LED bulbsare re-introducing this stroboscopic effect.

Some are far worse than others, and firstlet me say that I’m glad the CFL is being replaced.

I am in no way trying to say that LED bulbsare bad, and CFLs are somehow better.

Of course, a huge reason to be pro LED isthe lack of mercury in the bulbs.

And the list goes on–the slow warmup andpoor operation in cold weather of CFLs was annoying, and LEDs don’t suffer from theseproblems.

Poor color rendering indexes were common withcheap CFLs which caused their perceived quality of light to be not-so-great, whereas LEDsalmost always have better color rendering characteristics.

Dimmability of CFLs was generally questionableat best, and new LEDs go so far as to mimic the warming effect that incandescent bulbsnaturally produce as their filaments burn less intensely.

There’s virtually no reason to hold ontothe incandescent lamp anymore.

Even clear LED bulbs which look like theyhave filaments are cheap and widely available.

So to explain why CFLs don’t flicker andLEDs sometimes do, it’s important to look at the electronics that drive each of thesetechnologies.

Fluorescent lights, along with all other dischargelamps such as sodium vapor lamps or metal halide bulbs, have a pesky electrical characteristicknown as negative resistance.

Provide a set voltage to the lamp, and itwill consume more and more current until it, well basically explodes–or if it can manageit, exhausts its electrical supply and trips a breaker.

A ballast is therefore required to both strikethe arc and start the lamp, and most importantly to limit the current it can receive and keepthings nice and safe.

In older fluorescent fixtures, this ballastwas nothing more than a specialized inductive transformer, so-called magnetic ballasts.

These are what is responsible for the hummingor buzzing sound in older fixtures.

A magnetic ballast sends the same 60 hz electricityto the tube, but with a limit in place.

This means the light will pulse on and off120 times per second, which generally isn’t directly perceptible, but can cause eye strainin sensitive individuals.

Now, magnetic ballasts have two huge drawbacks.

One, they’re generally bulky, and two, thefact that they send straight AC current to the tube means the tube doesn’t run as brightas it could because it spends a not-insignificant period of time producing no light at all.

The pauses in light production reduce itsoverall light output considerably.

When the Compact Fluorescent Light came along, the compact nature of these compact bulbs meant less actual glass tube was availablein such a compact space.

To compact a 16 watt 2 foot linear tube intoa space as compact as an ordinary light bulb required some creative compacting action inthe form of glass bending acrobatics.


First was the curly-q nature of the tube itself.

Forming the glass in a repeating spiral patternincreases its surface area tremendously, while still confining it into a small volume.

Then there was the problem of the ballast.

Remember, magnetic ballasts are bulky andheavy.

A better solution was needed both to overcomesize constraints and to increase the light output of such a small lamp.

Enter the electronic ballast.

These guys work entirely differently frommagnetic ballasts and were, uh what’s the word, oh, compact and lightweight.

Electronic ballasts work similarly to theswitched-mode power supplies you find in virtually everything today.

Their first goal is actually to convert theincoming 60 hz AC power to DC, where it’s filtered with a capacitor.

The ballast then converts this DC into veryhigh frequency AC power, around 20 thousand hertz.

It’s this high frequency power that’ssent to the tube.

The phosphors that line the inside of theglass don’t react instantly to UV emissions from the mercury vapor.

In fact, there’s a delay between when theystop receiving energy from the excited mercury molecules and when the stop emitting visiblelight.

You can actually see this–the green phosphoris the usually the slowest, and you might have caught a slight green flash of lightwhen turning a off a light fixture with a CFL if you’ve ever moved your eyes rightat the same time.

You see this because the red and blue phosphorsstop producing light in a tiny fraction of a second, but the green phosphor hangs arounda little longer.

Anyway, the high frequency AC entering thetube of a CFL is literally too fast for any of the phosphors, and the delayed action bridgesthe gap between pulses.

The result is that the light provides nearlyconstant illumination, and the stroboscopic effect is essentially eliminated.

Which can be proven by using one of thesedo-dads.

Most newer linear fluorescent fixtures alsouse an electronic ballast.

Even the old fashioned T12 tube will see asignificant increase in light output and efficiency if high frequency A/C switching is applied.

For this reason, ceiling light fixtures usinglinear tubes are nearly always equipped with an electronic ballast these days.

Meanwhile, LED bulbs require a different kindof circuitry to make them work.

LEDs only work with direct current, so fora bulb on an AC supply, this AC needs to first be rectified into DC using a bridge rectifier.

It’s not as simple as sending DC power throughthe chips, though.

Without the proper voltage, the LEDs witheither be instantly destroyed or they won’t work at all.

See LEDs have a very narrow range of operatingvoltage, bumping it up by as little as half a volt will dramatically increase currentconsumed.

Drop it much below and it won’t light upat all.

Because of this, they also need a ballastof sorts.

Usually these are referred to as drivers.

The most important thing the driver has todo is limit the current that passes through the chips.

Without a way to limit the current, any voltageabove an LED chip’s forward voltage will cause an exponential increase in current flow, which will make the diode run extremely hot and severely shorten its life.

In many conventional LED bulbs meant to replacea 60 watt incandescent, there will be 9 or 10 chips, each rated around a watt.

These are usually arranged in a circle, andare attached to a heat sink.

The heat sink absorbs the heat they produce, and spreads it out over a wide area.

This bulb contains nine chips.

Each of these chips actually contains threediodes in one package, so there’s a total of 27 diodes arranged in series.

Most of the blue diodes used in white LEDchips–the yellow circle is a phosphor which converts some of the blue light into red andgreen, thus producing apparently white light–have a voltage drop of just over 3 volts.

The driver therefore needs to produce at least81 volts, and indeed it produces about 85.

The driver must also limit the current goingthrough this chain of diodes to ensure they don’t overheat and waste energy.

It also uses a large capacitor hidden in thebase to store and release some energy between the pulses of AC power coming from the socketthrough bridge rectifier.

This helps to eliminate the stroboscopic flicker.

This capacitor is rather large and it’sone of the biggest component of the driver.

But there’s also a way to cheat a littlebit.

LEDs can be driven off a direct voltage supplyif the voltage is equal to the voltage drop across the LED chip.

Many so-called “filament” LED lamps aredesigned with a bunch of blue diodes in series along a glass rod covered in the yellow phosphor, and the voltage drop across them adds up to just about the same as the AC line voltagepowering the lamp.

If you dim one of these, you can see the individualdiodes along the filament’s structure.

These tiny diodes will also have a voltagedrop of about 3 volts, and since 120 volts is what’s coming into the socket here inthe US, that could be divided across 40 individual diodes.

Each of these rods has 20 diodes or so ina line, and two rods are wired in series, with another series-pair being in parallel.

In European countries running on 230 volts, all four of these rods will be wired in series.

This cheat is what allows the driver to beso small that it can be crammed into just the space inside the socket.

This creates a beautiful bulb that you mightnot even know it’s an LED unless someone told you.

But there’s one huge drawback.

There’s so little space for the driver thatit doesn’t really do all that much.

In reality, nearly all it does is use a bridgerectifier to convert the AC into pulsed DC.

That’s just taking this waveform and flippingthe bottom half back up.

This means these bulbs will often exhibitstroboscopic flicker just as bad or worse as a fluorescent bulb running from a magneticballast.

In fact, that footage from earlier? It was from this bulb, just with the colortemperature messed up a bit.

And now, a note from the editor’s desk.

Oh, hello, I’m the editor, and this is mydesk.

I’d just like to clarify that I’m surethe driver is doing more than just rectifying the AC into pulsed DC.

It’s actually a complicated little thingwith a driver chip, an inductor of sorts, and other goodies.

What’s more likely the cause of the flickeris simply that the driver’s tiny little filter capacitor, a requirement with the driverconcealed in the socket, can’t store enough charge to provide completely steady DC voltagethroughout the system as the incoming AC voltage crosses the zero line.

The system voltage thus dips slightly betweeneach incoming pulse.

This is also probably the cause of the slightflicker produced by the CFL, but the immensely larger filter capacitor is able to providemuch more stable DC voltage to the rest of the ballast.

In regards to the number of diodes along theglass, I’m sure that’s geared towards line voltage as it is common for Europeanbulbs to have all the rods wired in series, but the driver is probably still providinga different voltage for them.

I’m thinking it just makes the design ofthe driver a whole lot simpler and cheaper if it’s got to produce roughly the samevoltage as it receives.

If we have a qualified electrical engineerin the comments, please do tell us if I’ve got this all wrong.

I’m not even going to get into how thesebulbs work with dimmers because there’s enough in there for a whole other video.

So then, here’s my point.

If you are an individual with photosensitiveepilepsy who has legitimately been affected by fluorescent lighting in the past, thistype of LED bulb probably isn’t for you.

But if you’ve casually avoided compact fluorescentlights believing them to cause eye strain and you’ve been around these lights andhaven’t noticed a problem, perhaps it wasn’t the, as I said, fluorescent-ness of the lightthat caused your headaches.

As I’ve demonstrated, most CFLs producelight just as well–meaning consistently and without flicker–as an incandescent bulb.

But some newer LED lamps actually producereally strong strobing light.

If these don’t affect you, that’s great! But it also means that perhaps you shouldn’thave been so averse to using the CFL.

One easy way to tell if a bulb has high flickeris by bringing a smartphone camera right up to the bulb.

With bright light the camera has to increaseits shutter speed a lot, which when combined with the way it captures the light via a rollingshutter, will make alternating bright dark bands appear all over the image.

If bands are barely visible, then the flickeris very minor.

Me again.

I discovered while shooting the B-roll forthis video that the old IKEA bulb exhibits less flicker than an incandescent.

You can even see that going back to the stroboscopedisc footage.

These pictures shot with my phone confirmit.

While we’re looking at pictures, light bulbmanufacturers have figured out how to produce flexible filaments, and this one on displayin a retailer is shockingly bad! However, one cool thing about the flexible“filament” is that you can see the printed circuit in the dark portion provided by theabsurd flicker of the bulb, and you can see here that the diodes are wired as two serieschains, with each trace skipping every other diode.

This means there are two parallel circuitsin each piece of filament spaghetti.

Now, I’ve long maintained a personal theorythat the folks most opposed to the compact fluorescent were really more averse to theblueish light of daylight color temperature bulbs.

In fact, I hate those things.

I have a whole drawer full of them becausethe previous owner of my place loved them, and I just can’t stand the coldness of theirlight.

I won’t go so far as to say they give mea headache, but I dread being around them.

Because a well-made warm-white balanced CFLis often indistinguishable from an incandescent, particularly if the bulb is hidden behinda shade, these people might have never noticed that they were under fluorescent lightingunless it was a cool white or daylight color temperature, where it couldn’t possiblybe an incandescent.

But that’s just conjecture.

In reality, the CFL is on its way out, andI’m happy to hear it.

So many great designs of LED bulbs are onthe market today, not even mentioning smart bulbs or color-changing bulbs that are onlypossible with LEDs inside.

But the CFL was a great innovation that helpedus start saving energy at home years before LEDs came down in cost.

And if people just took the effort to recyclethem, the mercury wouldn’t have been much of a concern.

But I’ll admit, a 100% recycling rate isa pipedream.

Best avoid the problem all together.

Thanks for watching.

I hope you learned something interesting today! I’m closing this video out with a thankyou and announcements.

To my subscribers, wow, I’m so thrilledthis channel has passed 35 thousand! It still doesn’t seem real.

Having a successful YouTube channel has alwaysbeen a dream of mine, and it’s slowly becoming reality.

But as you know, making videosis really hard.

I’m doing my best to keep videos like thisheaded your way, but I work full time and it’s hard to do two things at once.

Which is why starting at the end of November, I’m gonna stop doing two things at once.

I’m gonna concentrate on videos.

Hopefully I’ll be making weekly videos bythe start of next year, as I’ll have 4 days a week to do this, and not just 2 if I’mlucky.

There’s a lot of stuff up in this nogginand eventually it will make its way out and to your eyeballs and ears.

If all goes to plan, my next video will beon Philo Farnsworth and the invention of electronic television.

I’m overwhelmingly flattered that some peoplehave asked if I have a patreon page.

Well, I wanted wait and see if these typesof videos could earn me a following.

Apparently they have and now, I do have apatreon.

In fact, it’s right over there.

I’m really new to this whole thing and don’treally know what I’m doing, but if you’d like to become a patron you will immediatelybe rewarded with thanks and good vibes.

My biggest struggle right now is finding timeto do more management stuff, like make playlists and set up a Patreon.

But if it works, I’ll be spending all ofmy time making videos for you.

Thanks for watching.


Leave a Reply