Cost of electricity for light dimmer

Thst's not because they "burn off" the excess voltage. They still work under the principle of delaying switching the current on for part of the AC cycle. The reason for the large heatsinks is they are dimming large amounts of power. In a commercial bldg, a ceiling of lights is a lot more power than the 600W of light the usual residential dimmier is rated for.

See earlier followups.

[snip]

And it should be said that incandescent lights are very inefficient to begin with. The halogen bulbs are the most efficient of the common incandescents, but they can't be dimmed without shortening their life.

Someone needs to point all these questions to the authoritive website for this, Don Klipstein's lighting website.

potentiometer, so

connection).

A TRIAC is a thyristor, as is a SCR. In Google, type in define thyristor and see what comes up.

You get a lot more efficiency from LEDs making *colored* light because they don't create white light, most of which has to be filtered out. For white light they are currently between incandescent and fluorescent lamps in efficiency, but MUCH more expensive for room lighting, as I understand it.

Yeah, but it's pretty complicated- taking into account how much usable light you actually get of what wavelength and how it's distributed around the center axis etc.. including physiological factors, of course. I think John Fields knows a fair bit about this.

Best regards, Spehro Pefhany

dimmed,

incandescent

reading

Follow the other followups. Don Klipstein's LED pages and his posts.

One figure I've seen (in a manufacturer's ad) cited an 80% reduction in electricity use in stoplight applications.

The power saving seems to be one big motivation for switching stoplights over to LEDs. Reduction in the number of person-hours to do bulb replacement seems to be the other.

Here in the Silicon Valley, almost all of the stoplights seem to have been completely converted over... all three colors are now LED-based.

'Cause if so, it seems like

Some are bluer, some are more white... depends on the manufacturer, phosphor, etc. The better ones seem to be roughly comparable to a "cool white" fluorescent tube.

I'd be tempted to see how well one might do, in terms of color temperature, with a set of "white" LEDs, plus a few ultrabright red or red/orange to warm up the spectrum.

I believe so. Most high-brightness LEDs seem to be spec'ed in terms of candellas of brightness, over a specific angular range, at a given current, with a specified forward voltage.

| You're wrong both-sorry but that's it.A dimmer is *not* a potentiometer, so | there's no energy consumed on it.Usually it has a triac (which is a simple | power electronic device, like two thyristors in anti-parallel connection).

There is SOME energy consumed there. They do get hot (some get very hot). Certainly less than a potentiometer (unless something has gone wrong), but the consumptions not zero.

| The illumination output of a standard light bulb varies approximately by the | square of the applied power. Cut the power in half and you get about one | fourth the lumens. That is only approximate and varies somewhat with the | type of filament.

Assuming a constant filament. If you get a lower wattage filament, then you save power, lower cost, and have proportionally less light. But one factor to consider is the human sensitivity and perception of light. And that's non-linear.

| And, as an additional issue, it's generally a bad idea to use a dimmer | with quartz-halogen lamps. When dimmed, the bulbs run a good deal | cooler, the halogen-sequestering-and-redeposition of the tungsten | doesn't work as well, the tungsten tends to plate out on the inside of | the tube and dim the bulb, and the bulb lifetime is greatly decreased.

And this is a frequent problem with those torchiere floor lamps which usually have a dimmer and a 300 (more than you need most of the time) watt QH bulb. So people dim them, usually. Unfortunately, getting a lower wattage bulb is more expensive, if they can even be found in the same size.

| All in all, as others have said, it makes more economic sense to use | smaller bulbs at full power rather than dimming a high-wattage bulb.

Agreed. Getting variability should be done in the form of turning more or fewer bulbs on. A lamp with 8 small 40 watt QH bulbs and a switch to control how many are on would be nice ... and expensive.

| This in particular is why you shouldn't use dimmers with lamps that are | mounted so that the filament is above the socket. When such lamps burn | out it is not uncommon for a broken piece of filament to fall and | briefly short across the two thick wires that feed the ends of the filament.

I had one bulb do that in a non-brief way, once. The filament shorted in such a way that it had about 75% if it's length still working. It got brighter (in theory, from 60 watts to 80 watts if that estimate of filament length was correct. The nice thing was the light now looked a lot more "white". Unfortunately it gave out about a week later.

Someone once suggested getting 100 volt (because that is their nominal voltage) light bulbs intended for the Japanese domestic market. That would get a brighter whiter light. Just be sure to get a lot of them as you will be changing bulbs more frequently. A 60 watt 100 volt bulb run on 120 volts will be dissipating about 86 watts (NOT figuring in the change of resistance due to the change of heat). Which could be used in place of a 100 watt bulb. Power usage goes up. Light output goes up even more. Bulb replacement goes up, too.

Other people prefer bulbs rated for 130 volts so they don't have to change them so often. In hard to reach places that's definitely a plus.

| A few bulbs benefit from "soft starting", many and probably most do not. | Most (but not all) bulbs have zero or negligible fatigue damage to the | filament from a "cold start" despite a cold start jolting the filament to | cause a "ping" sound that is audible at close range.

I've found that this can be a benefit for the small candelabra style bulbs often used in small home chandeliers. Their filaments are constructed in a different way with typically 4 vertical segments. I've had these blow very often with full starting, and noticeably less often starting dimmed. Regular bulbs haven't noticed a change.

| There is a usual prelude-to-failure uneven evaporation of the filament. | That process causes a "thin spot" that is subject to a temperature | overshoot during a cold start. In most lightbulbs, such a "deadly thin | spot" is a deterioration mode of the filament that accelerates at a rate | worse than exponentially (during operation) once it becomes significant. | This means that for most lightbulbs, when they become unable to survive a | cold start their operating hours are numbered. And for most (but not all) | lightbulbs, cold starts do zero to usually-negligible damage until the | filament has aged enough for a cold start to be fatal.

[...]

| True, but they usually don't blow out immediately unless the overload is | very severe. Mild to moderate overload merely shortens their life.

I have seen at times some strange bulb behaviour. And this without any dimmer involved. The bulbs sometimes start to hum. Or maybe they will modulate the light output around 10 Hz. In these cases I have learned they are "dead already". If others are around I'll just say "that bulb just burned out" to which they reply "but it's still giving out light". Then my reply is "when have you ever seen a light bulb just stop giving out light while its on" (it happens, but it's rare enough most people don't remember ever seeing it). I insist that it is already dead and add that it won't come back on when the light is turned off then back on. Often times someone is willing to actually try the switch and every time I've been proved right. Then they start wondering about me.

| Current surge from burnout is often worse than the current surge of a | cold start. Often when the filament breaks, an arc forms across the gap. | The arc can be encouraged by the voltage gradient across the filament to | expand and go across the ends of the filament, in which case the filament | is no longer limiting current through the arc. This is what causes the | "bright blue flash" that sometimes occurs during a burnout, especially a | burnout during a cold start (when the filament resistance is less and | allows more current to flow through the arc which makes the arc hotter and | more conductive). | Most lightbulbs have fusible links in one of their internal lead-in | wires so that a "burnout arc" does not pop a breaker or blow a fuse. | However, this may be inadequate for protection of dimmers.

I've had a few cases where the surge that resulted for a burn out arc has caused other bulbs in the same fixture to burn out. In one case, all

5 bulbs in a fixture blew at the same time. In that case, all 5 bulbs also showed multiple filament breakage, and arc damage in the base, such as melting through the base. In another case, the bulb became welded to the base. In another case, there was damage in the wiring itself. Two case tripped breakers (multiple bulb burnouts in these, too).

| I believe probably true. It gets more uncertain when you have a 600 | watt or 540 watt load on a 600 watt dimmer, and it gets worse when you put | more than one dimmer in the same box since each one adds heat to the | others (despite the loss in each dimmer being only a few watts).

Just put a 7 watt light bulb in a small metal case and leave it one for a while. Without good thermal dissipation, heat rise can be significant. Many a fire has happened from decorative items with even small bulbs used inside (such a Christmas light items).

| Say for instance you had four 25 watt bulbs and you needed only 25 watts | of light. Well, you switch on only one. and you get 25 watts of light | and pay for 25 watts of power. But if you dim your 100W lamp to get 25 | watts of light, then you might be paying for 50 watts of power. You get | less light per watt when it is dimmed.

The 100 watt bulb in this condition has a lower color temperature. This results in far more infrared output relative to light output compared to normal operation. It's that "orange look".

| LEDs for light seem to come up a lot - I very seldom see an incandescent | red stop light, for example. But are they really more efficient, like | lumens per watt, or however that's measured? 'Cause if so, it seems like | a no-brainer, although the one "white" LED flashlight I've seen was very | blue, and would have been icky for general room lights. Maybe a reading | lamp...

"white" LEDs are more like fluorescent lights. They are a blue LED with some kind of phosphors for other colors.

I've seen a light with multiple LEDs in various colors, with separate dimmers for each color. Choose your own color. It was expensive.

connection).

Relatively small. I would expect it would be something like a normal NP junction where the voltage drop is about 0.6 volts. Assume a 50 watt load using about a half ampere current. = about 0.3 watts. Other components in the circuit may have more disappation than that.

There was some correspondence in New Scientists about the impact of mains overvoltages on the life of incandescent lamps. Something along the lines of a 10% overvoltage halves the life.

It's just possible that the lamp manufacturers really have chosen an operating point that represents the best choice for the majority of people.

I use flourescent substitutes in some places, but they're not exactly a perfect solution either. The ones I use take a significant time (minutes, I think) to reach their peak brightness, an effect that seems to get worse as they age. However I've yet to have to replace one.

I don't know that you can use a dimmer with them, though. I rather suspect they wouldn't like that.

Sylvia.

| Relatively small. I would expect it would be something like a normal NP | junction where the voltage drop is about 0.6 volts. Assume a 50 watt load | using about a half ampere current. = about 0.3 watts. Other components in | the circuit may have more disappation than that.

Put a small wattage light in the same kind of enclosure. Experiment until you find the wattage that gives the same temperature rise over a long settling time. Then you'll have an estimate of the waste by the device.

| I use flourescent substitutes in some places, but they're not exactly a | perfect solution either. The ones I use take a significant time | (minutes, I think) to reach their peak brightness, an effect that seems | to get worse as they age. However I've yet to have to replace one. | | I don't know that you can use a dimmer with them, though. I rather | suspect they wouldn't like that.

There are dimmable ballasts. That obviously combines the role of limiting the current and adjusting the current. I bet they are all solid-state.

Here is a nice collection of articles about dimming fluoros:

LEDs are now getting more efficient than incandescent lamps, but there are extra factors in traffic lights, flashlights, and some other applications.

Traffic lights: The incandescents in those are not regular ones, but ones designed to last (typically) 8,000 hours. These are something like

30% less efficient than the usual household incandescents. But the real kicker is that a red filter or a green filter blocks about 2/3, maybe about 70% of the light. You probably get only 3-4 lumens out per watt in, as opposed to around 11-17 lumens per watt with the usual "standard" household incandescents. (Slightly less for 230V versions in part because their filaments are thinner and have to be run slightly cooler for the same life expectancy). But they make red LEDs and trafic signal green LEDs with about the same overall luminous efficacy as white LEDs - 20 lumens per watt, even more. Lumileds makes some red ones that get 40 and traffic signal green ones that get 30 or more. Nichia makes traffic signal green LEDs that get 25 lumens/watt. Similarly efficient ones are available from ETG.

Flashlights: The main problem with incandescent is that their efficiency decreases greatly as the battery weakens. On the other hand, white LEDs usually have slightly increased efficiency when moderately underpowered. Their efficiency is usually maximized with current somewhere around 20% of the "typical" current in the datasheet, and as current decreases below that they lose efficiency but not nearly as rapidly as incandescents do.

I mention more in

BTW - Many LEDs actually have some color change as current is varied.

- Don lipstein ( snipped-for-privacy@misty.com)