Light bulbs

[Ben]

Who says incandescent bulbs don't last?

http://www.foxnews.com/scitech/2011/06/18/worlds-oldest-light-bulb-still-burning-after-110-years/?test=faces

Pretty amazing. Even if it didn't burn out from age, you'd figure a power surge or something would have taken it out by now. It seems to me that most of my incandescents used to go out right when turning them on. I can't remember one ever going out while it was on (sans power surge).

[Boomhauer]

Who says incandescent bulbs don't last?

http://www.foxnews.com/scitech/2011/06/18/worlds-oldest-light-bulb-still-burning-after-110-years/?test=faces

Pretty amazing. Even if it didn't burn out from age, you'd figure a power surge or something would have taken it out by now. It seems to me that most of my incandescents used to go out right when turning them on. I can't remember one ever going out while it was on (sans power surge).

I've got to say that my incandescent bulbs that are in fixtures that rarely get turned off last the longest.

[Perd Hapley]

I will never buy a CFL.

They are poisonous, they are carcinogenic, and they cast a sterile, hospital-white light.

They are disgusting.

You should pay attention to the "color" of CFL you're buying. They come in Warm White, Cool White, Daylight, etc.

[zahc]

It doesn't matter what color temperature you buy. The output spectrum still sucks. Some people have lower standards so they don't care, but some of us like our lives to be lit with black body radiation like God intended.

[erictank]

Who says incandescent bulbs don't last?

http://www.foxnews.com/scitech/2011/06/18/worlds-oldest-light-bulb-still-burning-after-110-years/?test=faces

Pretty amazing. Even if it didn't burn out from age, you'd figure a power surge or something would have taken it out by now. It seems to me that most of my incandescents used to go out right when turning them on. I can't remember one ever going out while it was on (sans power surge).

I'm sure they've got like 18 layers of line conditioners and a week-long APS feeding that thing...    Carbon filament, IIRC?  Reads bulb website... yup.  Neat.  On a separate ckt at only 4 watts, but but it actually DOESN'T look like they're taking extraordinary precautions with it.

It doesn't matter what color temperature you buy. The output spectrum still sucks. Some people have lower standards so they don't care, but some of us like our lives to be lit with black body radiation like God intended.

I find the soft-white CFLs to be acceptably-close, visually, to the equivalent incandescents.  And I draw, so light quality matters somewhat to me.  Not keen on the frequently-slow startup times, although those are reportedly less common now, and I've had several burn out in equal or less time than incandescents in the same outlets - not cool, considering they cost several times what the incandescents did.  Restocking regular bulbs and re-replacing them as the "green" ones fail, and hoping that LEDs come down enough in cost to be practical by the time it's illegal for incandescents to be sold...

[230RN]

Ben said,

Pretty amazing. Even if it didn't burn out from age, you'd figure a power surge or something would have taken it out by now. It seems to me that most of my incandescents used to go out right when turning them on. I can't remember one ever going out while it was on (sans power surge).

That "turn-on-POP!" phenomenon is pretty well known.  The electrical resistance of the filament is low when it's cold, so when you flick the switch, a big surge of current flows through the bulb.  When it warms up (which doesn't take long), the resistance goes up and the running current is a lot lower.  Most electrical conductors (like tungsten) show this low-temp, low-resistance phenomenon.

As the bulb ages and the filament loses tungsten through evaporation and other processes*, there is inevitably one spot that's weakest and sooner or later that on-surge is too much for that one weakest spot and that's when you get the turn-on-burn-out phenomenon.  (It also depends a little on when in the AC cycle you happen to hit the bulb with power.)  The same thing is true if a bulb is already running and a power surge hits it.  The bulb can't get hot enough soon enough in the surge to have its resistance increased quickly enough, so that inevitable "weakest spot" burns out.

You may have noticed, especially with lower-Wattage bulbs, that they actually get brighter for a while before they burn out.  Once again, that inevitable "weakest spot" in the filament is getting hotter and hotter as the bulb ages.

In many old-time tube audio amplifiers, where the tubes (valves for you Englanders) are very expensive, they had a surge protector in the filament circuits to prevent that surge from popping the tube filaments.  They usually consisted of a heating element in series with the filaments which was situated physically below a bimetallic thermal switch so that the switch was open until the heating element (and thus the tube filaments) warmed up, and then the bimetallic switch would then close, applying the full voltage to the tube (valve) filaments.

When they turn on those very high-powered TV and radio station transmitters, it's not a question of walking in the transmitter room and flipping a switch and you're "on the air."  There's a long and involved process involved, which includes bringing the power up gradually.  (Nowadays a lot of this process is handled automatically.)  These high-powered transmitting tubes have filament currents well up in the hundreds of amperes.  See, e.g.,

http://www.oldradio.com/archives/hardware/WE320A.htm

That old-timer had a filament rating of 35 Volts at 435 Amperes (15kW).  A wee bit different from flicking the wall switch and hitting your 100 Watt $0.79 ceiling light with the full line voltage.

This kind of "lore" is what I was talking about in terms of the newer LED lamps.  There doesn't yet seem to be enough "lore" about them in the common pool of knowledge.

Terry, 230RN
----------
* Carbon filaments do not have the same aging process as tungsten filaments, and will last a lot longer than tungsten filaments barring shock or other catastrophe.  Google "tungsten-water cycle."

[RoadKingLarry]

Bah, all this elctrical lighting stuff is just a silly fad.
That reminds me, I need to get some more coal oil next time I make a trip to town.

[230RN]

Oh, so you've got one of these as well, eh?



Terry, 230RN

(Pic credits in properties --steam punk section)

[cordex]

Wildly off topic, but does anyone know why an incandescent bulb would glow dimly for a few minutes after the bulb burned out?  I had a lamp on my bedside table that popped loudly as it burned out.  It was late at night and no other lights were on in the room so it was immediately dark.  The whole bulb (not the filament, mind you) glowed sort of like a florescent bulb might, very dimly for a short while.

I figured the arc from the breaking filament somehow excited something in the "soft white" coating inside the bulb, but I wasn't sure.

[230RN]

For two possible reasons, probably both.

Partially because they add argon and nitrogen and some other things (including iodine sometimes)at low pressure in the bulb to inhibit some of that water cycle I mentioned above and to aid heat conduction out of the envelope.  Argon glows blue (and strongly ultraviolet) when it has current passing through it.

The blue flash is due to current flow between the broken ends of the filament energizing the gas to glow, just like a neon bulb or a photographic xenon flash lamp glows when it ionizes and current goes through the gas. 

The gas cools rather quickly and de-ionizes and the blue glow stops. In a bright room, you may not notice this glow, and it is kind of a fragile phenomenon.  After all, an incandescent lamp is not designed to give a gas disharge, like a neon or xenon lamp is.

In addition to the visible blue glow, the bright ultraviolet flash from the arc  was probaby exciting materials in the bulb's internal coating, in other words, by phosphorescence.

Phosphorescence usually has a fairly long persistence, and they  probably add regular phosphorescent materials to the bulb coating as well to minimize flicker and change the color temperature of the light output, but I do not know this for sure, so don't cite me on that.

The phosphorescent materials in the old cathode-ray TV set screens and computer monitors have a persistence as well, but it's designed to be just long enough to stay brightly glowing between scan intervals in the raster.  If you come up on a TV in a dark room and shine a flashlight on it, the phosphorescent glow will remain for a little while.

Therefore it might also be an afterglow from the coating independent of any arcs in the argon.  Easily tested by shining a light on a regular bulb in the dark and seeing if there's an afterglow from that. (Keep your eyes closed while shining the light on the bulb so you don't lose your dark adapatation.)

That's my theory, anyhow.

Terry, 230RN

[birdman]

White LED's aren't simply three color chips, they are most often (for the high intensity ones) phosphor based (same white-ish phosphors as CFL's, illuminated by a Blue/purple diode.  (to address an earlier point made).  There are RGB ones, but not as common for high intensity.

[230RN]

The persistence of the phosphor may hold down the flicker.  I was only thinking of the indicator-type  light LEDs, which have essentially no persistence.  Thanks for reminding me about that.

[birdman]

The persistence of the phosphor may hold down the flicker.  I was only thinking of the indicator-type  light LEDs, which have essentially no persistence.  Thanks for reminding me about that.

Also, most LED bulb drivers that work off AC use a high frequency switching driver circuit, running at 10's of kHz (see: http://www.edn.com/contents/images/6648790.pdf for example)... It's not a simple rectify and smooth driver, so little observable flicker.

As for LED persistence, it's nanoseconds (non phosphor) to maybe milliseconds (phosphor), and given the non-linearity of a LED, you would definitely see a flicker if the drivers weren't high speed.

[230RN]

Also, most LED bulb drivers that work off AC use a high frequency switching driver circuit, running at 10's of kHz (see: http://www.edn.com/contents/images/6648790.pdf for example)... It's not a simple rectify and smooth driver, so little observable flicker.

Of course.  I should have thought of that --a low-voltage inverter. Duh.

As for LED persistence, it's nanoseconds (non phosphor) to maybe milliseconds (phosphor), and given the non-linearity of a LED, you would definitely see a flicker if the drivers weren't high speed.

I knew that, for "straight" LEDs.  I forgot that for illumination (as opposed to indicator) purposes, they had phosphors.

Thanks for the link.

I'm going to lay off this thread for a while until I accumulate more "lore" on LEDs for household illumination. As I mentioned in a prior post, otherwise, I'm talking through my hat, but hoping that someone like you would come along with firm knowledge.

Got a message from the library that another one of those power meters is available for me.  I'm going to spring for a straight LED light-socket lamp and take it apart after I do some measurements on it.

THANKS!

Terry, 230RN