You are at The Krib ->Lighting [E-mail]

How Fluorescent Lighting Works


  1. Ballasts
  2. random notes about fluorescent lights
    by "Shimoda, Wade" <WShimoda/> (Fri, 20 Mar 1998)

Date: Wed, 9 Oct 1996 17:15:19 -0400
Subject: Magnatek ballasts, and other ballasts.

Arthur Johannson asks about Magnatek Ballasts in APD #244. I do not have my catalog at hand, so I am answering his question from memory - a risky endeavor! However, I believe he is describing a very common type of "magnetic" ballast. Magnatek is a very well-known brand of magnetic ballasts. They are well made. Ballasts seem to be poorly understood, and the differences between the various kinds are also poorly understood.

I will stick my neck out, and try to describe how ballasts work, and the differences between them. Again, this is risky, as it is a lot easier to describe these devices with drawings. Also, I am NOT an engineer. This will not be as organized nor as clear as an Engineer would say it, but I will do my best.

The "ballast" for any fluorescent bulb is mostly a current limiting device. Once the "arc" is started between the ends of the tube, an essentially unlimited direct "short circuit" is developed. The best illustration that immediately comes to mind is a bolt of lightning. Lighting does not just suddenly blast a giant spark from cloud to ground. First, in an extremely short time just before the bolt of lightning, an "ion trail" is established between the cloud and the earth. In other words, a "pathway" is established in the air, the pathway made up of "ionized" molecules of the gases that air is made of. Then, the giant "spark," the actual "lightning bolt" follows that pathway in an enormous burst of energy. The loud "Bang" is generated from the suddenly heated air as the spark jumps from ground to cloud. (The main "bolt" most often actually jumps from the ground up to the cloud, not the other way around.) The arc in a fluorescent tube is similar, in that it must be controlled to keep a sort of miniature "lightning bolt" from ruining the fluorescent bulb.

Now I will 'backtrack' a bit. If you could take the white "phosphor" coating off of the inside of an ordinary fluorescent bulb, you would see that there is a "filament," a tightly coiled wire, that runs from one of the pins in the end of the bulb to the other pit at that end of the bulb. There is a "filament" on both ends of the bulb. They are usually covered with a whitish powdery stuff. The powdery stuff is a material that "boils off" lots of electrons when heated. The filament is the thing that actually "fails" when a fluorescent bulb wears out. Either the filament burns out, or all of the powdery stuff has "boiled off"

What is in the clear space from one end of the bulb to the other? That depends upon several Engineering decisions made by the folks that designed that particular bulb. However, it is always a pretty good vacuum, with a small amount of some gas or other. If you look at a Neon sign that is not lit, it has a pretty good vacuum, with a teeeny tiny bit of Neon gas. It looks like just "clear glass" until it is lit. When an arc of electrons passes through that gas, it glows with the familiar reddish orange "NEON" color. In the fluorescent bulb, you get an arc that generates ultraviolet light. The Ultraviolet from the arc bangs right up against that white powder that coats the inside of the bulb, and the powder glows. Electricity passes through the very thin, almost a vacuum, gas in the bulb, generating ultraviolet light, which bangs into the powder, which generates visible light. Ultraviolet light has a very short wavelength, and we cannot see it. (Bees and many other insects, on the other hand, have really tiny eyes, and they see ultraviolet light just fine! In fact, flowers (that show pretty colors to our big eyes) reflect the Ultraviolet light in sunlight very strongly, and appear "bright bee white" to a honey bee. We see colors, they see "white.") ( :-) )

When a fluorescent bulb is "started," one or both of the filaments at the ends of the bulb is heated by electricity, and "boils off" a cloud of electrons, which quickly moves down the tube until this cloud of electrons fills the bulb. (actually, the electrons "ionise" the gas in the bulb, just like the ion trail that forms the path for a lightning bolt) Then the arc jumps from one end of the bulb to the other, and the bulb is "lit."

Without the ballast, the arc would be a direct "short circuit" between the ends of the bulb, and a huge blast of electricity would jump through the bulb, and it would either instantly burn up the filaments, or the bulb would literally explode. Rather like the lightning bolt I mentioned earlier. An uncontrolled arc through a fluorescent bulb is BIG trouble for the bulb! Actual bulb explosions from a "shorted ballast" are very rare, but they can happen. Most often, if the "ballast" in a fluorescent fixture fails by shorting out, you hear a loud "pop," accompanied by a very bright flash from the bulb, then nothing. No noise, no light. The bulb has burned out. (Actually, the filaments were destroyed by the suddenly uncontrolled big arc.)

The ballast regulates, or controls, the flow of current through the arc, one way or another. The magnetic ballast literally "chokes" off the current at a set point. It does this by taking advantage of some basic physics. Remember, the electric power in your home is (depending upon what country you live in) either 50 or 60 cycle alternating current. In a "cycle," the flow of electicity goes from no flow at all, up to a set voltage, then back to zero, then goes THE OTHER WAY, from zero up to the set voltage, then back to zero. The current in that arc in the bulb does the same thing. It goes from zero to maximum, back to zero, then the other way, then back to zero. The ultraviolet light from the arc also goes from zero to maximum, back to zero, up to maximum, then back to zero. The glow from the phosphor powder coating the inside of the bulb does the same thing, except the powder never completely stops glowing.

An alternator, for example the one in your automobile, often spins magnets inside a set of coils of wire. Any time a wire moves through a magnetic field, it generates an "ElectroMotive Force," a force measured in volts. The amount of the force (voltage) depends upon how strong the magnetic field is, and how FAST the wire moves through the field. Or, you could move the magnet, and let the wire stand still. It does not matter which is moving, the magnet, or the wire.

If you pass a current of electricity through a wire, it generates a magnetic field. Since one of the most basic Laws of Physics can crudely be stated as "There ain't no such thing as a free lunch," that magnetic field is exactly opposite to that magnetic field that we were moving the wire through just a moment ago. If the ends of the wire (or the ends of a coil made of many turns of wire) are connected, an electrical current is generated when you move the coil through a magnetic field. Or, you could move the magnetic field past the coil. Is does not matter which is moving. Yes, I suppose you could move both of them at the same time if you wished. ( :-) ) And, any time an electric current moves through a wire, it generates a magnetic field.

The most difficult concept for most of us is, both things happen at the same time. You generate a current in the coil of wire by moving it through a magnetic field, and then that current in the coil of wire generates another magnetic field. That magnetic field is exactly opposite to the magnetic field you are moving the wire through to generate the current. The magnetic fields "fight", as does the current. I did not say that particularly well. Sorry!

In a magnetic ballast, the rapid changes in current direction (the "cycles per second," or "Hertz") makes equally rapid changes in the magnetic field generated by the coil of wire in the ballast. The magnetic field reverses 120 times a second. (Each "cycle" or "Hertz" goes from zero up, than back to zero, then the other way, then back to zero.) The corresponding current in the arc in the bulb is doing the same thing.

The "fighting" magnetic fields and currents in the magnetic ballast "fight" much stronger as the amount of current flowing through the ballast increases. This "fighting" impedes, or chokes, the current flow. An individual ballast is designed by the Engineers so that it will let just exactly the right amount of current through to light that particular fluorescent bulb, but not let too much current through. It gets hot from the effects of the "fight." That heat is totally wasted electricity. You pay for it, but you don't get any light from it. Magnetic Ballasts are lots simpler and cheaper to make than Electronic Ballasts. That is probably why they are used so much.

Now, lets go to the actual generation of the light. The white coating on the inside of the tube, the phosphor, glows from the ultraviolet light generated from the arc. The arc goes from zero, up to a maximum, then back to zero, then the other way, than back to zero. So does the Ultraviolet light. So does the light from the glowing phosphor! The phosphor is a bit goofy, though. It does not quite stop glowing between current pulses. It almost stops, but not quite. If you are in a dark room, you can wave your hand under a single fluorescent bulb, and you will see multiple hands, as the light turns on and off and on and off, following the current cycles. A "strobe" light turns on and off very abruptly and completely. The fluorescent bulb is kind of lazy, and does not turn either on or off all that quickly. Its brightness depends mostly upon a combination of the phosphors used, and upon the amount of current flowing through it. Too much current, and the filaments will burn up. Too little, and you don't get much light.

In a common magnetic ballast, you start the arc by flowing a current through the filaments, heating them until enough electrons have boiled off to "start" the bulb. Once the bulb is started, the filaments stop glowing, and the light comes entirely from the arc. Starting a "pre-heat" bulb can be accomplished by using the common "starter," which looks like a very small aluminum can with two pins on one end, or by using a "starter switch." If your light is started with the switch, you hold the button down for a few seconds, then release it. If you watch the bulb at the same time, you will see the ends of the bulb start to glow, first very dimly, then brighter. When you release the switch, you change the flow of the current from through the filaments to through the bulb itself. The "starter" does this for you automatically. The hand starting switch is lots cheaper to make than the circuitry and the socket for the "starter." It also takes a lot less room in the "hood" or reflector over your tank.

In a "rapid start ballast," a differently designed fluorescent bulb is used. "Rapid Start" bulbs are designed to be started by Rapid Start Ballasts. In these, a small current always flows through the filaments. The inside of the bulb - the amount of vacuum, the gases, etc. - is designed to start the arc very very quickly, without much heating of the filaments. Most "Pre-Heat" bulbs will not start with magnetic rapid start ballasts. Some can be induced to start by rubbing the bulb from end to end, thus capacitatively inducing the arc. If you do get a Pre-Heat bulb to work with a rapid start ballast, the end or ends of the bulb will get really black really soon. This is because the filaments always have a current going through them, something that does not happen with the normal "pre-heat" ballast. Interestingly enough, Rapid Start Fluorescent Bulbs work just fine with either the hand start switch or the "starter." It just does not work the other way! Rapid Start ballasts are made in Magnetic, Electronic, or Hybrid kinds. They don't look any different on the outside. The differences are inside.

Why The Electronic Ballast Gives More Light For The Same Amount Of Electricity

Remember how the bulb flickers, up to maximum brightness, then almost to zero, then up, then down, over and over? Well, the Electronic Ballast does its wizardry and changes the 50 or 60 cycle alternating current into (usually) 400 cycle alternating current. This (to make a very crude analogy) keeps the bulb phosphor glowing at almost the peak brighness almost all of the time. The current cycles back and forth, going up and down 800 times a second, and the phosphor never has a chance to quit glowing like it did on ordinary alternating current. Another advantage to the Electronic Ballast is the fact that is wastes very little electricity as useless heat. The new Electronic Ballasts we are installing here in the Aquarium Center draw slightly over 60 watts of total energy. Yet, they light two, 40 watt bulbs, and light them much brighter than the Magnetic Ballasts they are replacing. The Magnetic Ballasts draw more than a hundred watts of electicicy to light two, 40 watt bulbs. They waste a lot of electricity as heat. And, the bulbs are not nearly as bright!!

During the Iowa State Fair, there is a severe voltage drop in our area. Too many other exhibitors, all drawing electricity from under-sized main wiring. Anyway, we sometimes have voltages UNDER 100 volts. Many Magnetic Ballasts just quit then. The lights just dim, then go out. They cannot be re-started until late at night, when all of the other current users close up shop and turn off their lights. The Electronic Ballasts we are now using are very cleverly designed, and give the same amount of light with line voltages from a low of only 90 volts up to over 140 volts.

Why doesn't everybody use Electronic Ballasts? Probably because Magnetic Ballasts are cheaper to make and install.

What About High Output And Very High Output Fluorescent Bulbs?

High Output bulbs and Very High Output bulbs are just bulbs designed for much higher current arcs. A 24 inch long, 40 watt High Output fluorescent bulb needs the same amount of current as a 48 inch long standard fluorescent bulb. Don't try to use a 40 watt magnetic ballast on a 24 inch, 40-watt High Output bulb. It probably will not work. Some newer magnetic ballasts will work, not as well as a proper ballast, but they will work. Sort of. At least some of our old ballasts won't work at all with such bulbs. You turn on the light, and exactly nothing happens. We found both "Hybrid" 40 watt Rapid Start Ballasts and Electronic 40 watt Rapid Start Ballasts worked just fine with 24 inch long, 40 watt High Output bulbs. (I have a friend who is a Licensed Professional Engineer, with a PhD in Electronic Engineering who told me to try the Electronic Ballasts.) If you watch Electronic Ballasts as they are starting bulbs, they seem to work really weird. They "pulse" current through the filaments to start the bulbs, then turn off the filament current when the bulbs start. You can see this happen if you only have the pins on one end of the bulb connected to the ballast. The filaments at the ends of the bulb glow bright then dim then bright and so on. (This can happen if you are using those rubber "plug in" special aquarium end caps. You don't always get the pins lined up just right.)

``Ice Cap Electronic Ballasts are very good ballasts.'' My friend tells me it is entirely possible to make an Electronic Ballast that can "sense" what kind of bulb is connected to it, and automatically regulate the current in the arc to the right level. I have never tried an Ice Cap, but I would guess they are designed to operate that way.

Ordinary Electronic Ballasts are a lot cheaper!!!! We are paying from #39.50 up to about $44.00 each for the Electronic Ballasts we are using. (That is wholesale from Grainger.) There are Electronic Ballasts made for all sorts of fluorescent bulbs. You can get them to light only one bulb. Or, light two bulbs at once. Or, you can get them to light two or even three bulbs, each bulb lighting all by itself. The advantage of that is, just because one bulb "wears out," the others will keep on working all by themselves. You pretty much get what you want, and are willing to pay for.

Do You Really Get All That Much More Light From High Output Bulbs?

Sort of. There is no such thing as a free lunch!!! The bulbs are much brighter, but they also draw much more electic current. The lighting engineers say you want "the most Lumens per Watt," I would say "Most Bang for the Buck." The bottom line seems to be you get pretty much the same Lumens per Watt from any Electronic Ballast with any combination of bulbs. It takes more bulbs with "standard" 40 watt bulbs, fewer bulbs with High Output and Very High Output bulbs. The amount of electricity used per actual Lumen of light output is just about the same, no matter which bulbs you use. We found here that High Output Bulbs "leak" more Ultraviolet light, which actually burned some corals before we figured out what was happening. All Electronic Ballasts are brighter than Magnetic Ballasts. As far as I know, there are no exceptions to this rule.

Should you buy Ice Caps? Darned if I know! They sure seem expensive! My Engineer friend thinks they are pretty "pricey." They work very well indeed, and have some good safety features built in. They turn off automatically if you have a dangerous current leakage. So does any properly installed Ground Fault Interrupter. Tunze makes excellent stuff, so does Dupla. Expensive stuff! Is it worth the money? Again, I simply do not know.

I hope the above very long post will be helpful.

random notes about fluorescent lights

by "Shimoda, Wade" <WShimoda/>
Date: Fri, 20 Mar 1998

I recently took a course on efficient lighting taught by someone from the Lighting Design Lab in Seattle and I thought I'd share some of the points that might interest you. Although the class covered various technologies, I'll just cover fluorescent lighting since that appears to be what most people are interested in.


In the 1970's, the first electronic ballasts appeared.
In the 1980's, compact fluorescents and T8s appeared.

Lamp life and lumen depreciation:

Lamp life is the average or median operational life. e.g., in a sample of 100 lamps, by the time their total operating hours equals their operational life, 50 will be burned out. Lumen depreciation is the fractional loss of lumens at rated operating conditions that progressively occurs during lamp operation. At 100% of rated lamp life - Halogen incandescents have about 95% of their initial lumens T8 fluorescents (1", or 8/8ths of an inch diameter) - 90% (GEs new Starcoat lamps (F32 T8) are supposed to maintain 95% of their initial lumens) Incandescents - 80% T12 fluorescents (1 1/2", or 12/8ths of an inch diameter) - 75%

Ballast types (these comments apply only to elec. ballasts for fluorescent lamps):

Magnetic ballasts operate with an output (the elec. going to the lamps) of about 60 Hz. Electronic ballasts operate at over 25 kHz. People may notice the flicker of magnetic ballasts, but not of the electronic ballasts. Electronic ballasts usually run cooler than magnetic ballasts.

Ballast factor - the lumen output of the lamp and ballast combination compared to the rated output of the lamp on an ANSI reference ballast. Wattage consumed by the combination is approx. proportional to the ballast factor.
---Magnetic ballasts - .94 to .95
---Electronic ballast - .65 to 1.28
Note: lamp life may be reduced with ballast factors greater than 1.18.

Rapid start or instant start ballasts.

Instant start ballasts put more stress on the lamp when it's started, therefore slightly reducing lamp life, but they also use less electricity. However, at about 12 or more hours of operation per start, lamps on instant start ballasts have about the same lamp life as those on rapid start ballasts.

Energy consumption examples (unfortunately I don't know if these are for instant or rapid start, but I'm guessing rapid start): A .94 ballast factor mag. ballast running 2, 40W T12 cool white fluorescent lamps will put out about 3050 rated lumens (2867 at .94 ballast factor) per lamp and use 96 watts. A .88 ballast factor elec. ballast running 2, 32W T8 lamps (unknown lamp type) will put out about 2900 rated lumens (2552 at .88 ballast factor) per lamp and use 62 watts. This is about 11% lower lumens and about 35% less electricity. (A .87 ballast factor mag. ballast running 34W T12 energy saver cool white lamps would put out 2750 rated lumens per lamp at 82 watts).

I hope you find this helpful or at least interesting.

Up to Lighting <- The Krib This page was last updated 29 October 1998