Last update: December 31, 2001
Lighting is one of the most discussed and most confusing issues with planted aquariums. Everyone wants to know "What's the best light for plants?" This question seems to get the most varied answers simply because "best" can be defined in so many ways. Are you concerned about cost? Trying to optimize intensity? Looking for the best spectrum? Do you desire a "natural" appearance? Do you want to take prize winning photos? Are energy costs critical? Is the initial expense a stumbling block? Do you have a closed hood or open top? How deep is your tank? Do you want metal halide or fluorescent?
Note: In case you get bored before you get to the end of this page, I would like to recommend Aquarium Hobbyist Supply as a most excellent source of compact fluorescent lighting kits and supplies. The owner, Kim, has gone out of his way to develop a superb series of lighting kits to suit any need. If you are retrofitting a standard canopy or building a custom hood, AH Supply has what you need. Joe Bob sez, "Two thumbs up!"
Measured Bulb Data - Part 1 (Color Temperature and Relative Intensity)
Measured Bulb Data - Part 2 (Relative Intensity)
Measured Bulb Data - Part 3 (Inverse Square Law ... NOT!)
PennPlax Bulb Data
Common Bulb Data
Before we discuss our favorite lights, let's look at some terms that are useful for lighting discussions.
Watts Bulbs come in various wattage ratings. This describes how much electrical power a bulb uses and does not describe how bright it is. Different bulb technologies will produce differing amounts of light per watt. A fluorescent bulb is 4x as bright as the same wattage incandescent bulb. A metal halide bulb is 2x as bright as the same wattage fluorescent bulb. Even within a technology, light per watt will vary. All "40 watt" fluorescent bulbs draw 430 ma of current but vary greatly in how efficiently they convert the energy from the resulting arc into light.
Unfortunately, wattage is the simplest way to say how much light is needed, even though it is not an accurate measure. It is very common to see "use 2-3 watts per gallon of fluorescent light for plants". This is a very rough rule because it doesn't take into account the efficiency of the bulbs, the type of reflector, the form factor of the tank or the type of plants being grown.
Lumens This is the total amount of light that a bulb is capable of generating. This information should be readily available, either on the bulb package or from manufacturer's data sheets. It may take some digging to find it though. There are two values usually quoted: initial lumens and design lumens. When a fluorescent bulb is first turned on, it will be about 20% brighter than it will be after some amount of burn-in, typically 100 hours. Initial lumens describe how much light it produces when first turned on. Design lumens describe how much light it will produce for a much longer term. After the initial 20% drop in brightness, the light output will slowly decrease.
Lumens are a better way to specify lighting since it describes how much light is available. But it still falls short in that all the lumens might not reach the plants and it doesn't describe the energy bands that are generated. A bulb might have lots of lumens but it might be poorly focused (most of the light is lost) or it might be a spectrum that plants can't utilize very well.
Lux This the actual intensity of the light falling on a specified area and is a much better way to describe lighting requirements. Unfortunately, since it depends on the how the light gets from the bulb to the area, it can't be specified by the manufacturer - it has to be measured by the aquarist. Hobbyist luxmeters are available for under $150 and are invaluable for the serious light cognoscente.
Lux is defined as lumens per square meter. If all the light from a 3000 lumen bulb was perfectly focused on a 1 square meter area, the intensity at any spot in that area would be 3000 lux. Note that typical reflectors are far from perfect, both in shaping and in reflectivity.
CRI This is the Color Rendering Index. This describes how closely a light source renders colors compared to "standard" sunlight as observed by a "standard" human eye. Obviously, there is some interpretation in the numbers that are used here. A perfect source would score 100. Full spectrum bulbs rate in the 90s. I'm not sure if this number is useful in selecting aquarium bulbs or not.
Kelvin Temperature (K) This describes the color of the light but not the spectral components. If a light source is stated as 5000K, it is the same color as a radiating black body (a physics term) that is that temperature. Incandescent bulbs are around 2700K (reddish), "daylight" is 5500K (white), higher temperatures have more blue.
Spectrum This describes the wavelengths of light that make up the light source. Visible light is a continuous band of colors ranging from violet to red (wavelengths of 380 nanometers to 700 nanometers). Sunlight and incandescent light is composed of all visible wavelengths. Fluorescent and metal halide bulbs emit only a few wavelengths, depending on the phosphors or rare earths used.
Since there is so much information on the Internet about the "best" bulbs as seen by various authors, I will only describe the bulbs we like and why.
We are primarily interested in the most intensity we can get in the space we have available and are also interested in the best color rendering possible, both for general appearance and for photography. Our two 100 gallon tanks (60"x18"x22" high) use four 40 watt fluorescent bulbs; two Triton and two PennPlax Ultra Tri-Lux. Both these bulbs are "tri-phosphor" bulbs, meaning they use three types phosphors rather than the usual two to produce a fuller spectrum. The third phosphor is a rare earth that accounts for the higher price. The Ultra Tri-Lux has a strong green component which accounts for its extra intensity.
Many aquarists believe that only red and blue light is needed for photosynthesis. This is true for algaes; this "fact" comes from older studies of photosynthesis in algae. However, higher plants have evolved to use many wavelenths of light and the green wavelengths are not wasted. Also, the stronger green component also helps bring out the natural green color in plants.
With these four bulbs about 3" from the surface in our hoods, we achieve around 13,000 lux at the surface. This has proven to be suffucient light for almost any plant we've tried.
The Triton bulb has a color temperature of around 8300K and the Ultra Tri-Lux has a temperature of around 6500K. The combination seems to match well with normal "daylight" film, even though the temperature seems too high. I suspect that the water acts as enough of a filter to balance the colors fairly well. All the photos in this site were taken with only the aquarium lights and regular print film.
Our 120 gallon tank (48"x24"x24" high) uses two 175 watt 5500K metal halide bulbs. These are in a hood suspended 14" above the surface of an open top tank. Even though the two metal halide bulbs have much more intensity than four fluorescent bulbs, less of the light gets to the surface due to a poorer reflector and a greater distance from the water.
The 5500K Daylight bulbs are more expensive than standard 4300K bulbs, but the appearance of the tank is much nicer. The 4300K bulbs give a very yellow color to the plants, even though they will grow as well.
The 5500K bulbs also work very well with regular daylight film. We have the hood suspended from the ceiling with decorative swag chain and can lower the lights close to the surface for photography.
The one downside of metal halide lights is they don't bring out the color of red plants very well.
I made some intensity and color temperature measurements on a bunch of fluorescent and metal halide bulbs. My goal was to determine the relative brightness of various bulbs (to check out the claims of the vendors) and to get a feel for the color temperature of the bulbs. A secondary goal was to determine the effects of age on both color temperature and brightness.
To measure intensity, I used a lux meter offered by Marine Invertebrates, Inc. This is advertised in FAMA and costs about $100. I'm not sure about its absolute accuracy, but, being digital, its resolution is quite good. The instructions claim an accuracy of +/- 5 lux and it seems very repeatable. I'm also not sure about how color temperature affects its accuracy. Its response curve peaks at about 600 nm, so there is probably some effect.
To measure color temperature, I rented a Minolta Color Meter II from a photographic supply store in Denver. It's primary use is to determine the correction filters required to balance a specific film for a specific light source, but it also indicates color temperature in degrees Kelvin. I'm not sure of its absolute accuracy (and I did see some strange results with metal halide lights), so take the numbers with a grain of salt. This is a neat instrument that sells for about $700, so I feel that numbers aren't too far off for the fluorescent bulbs. Minolta also makes a different meter for "scientific and industrial colorimetery", but it's not available for rental locally.
BTW, I tried to call the Minolta Customer Service Department to determine the validity of my measurements, but the "representative was busy and his voice-mail was full", so I couldn't leave a message.
This posting is the raw data I collected. I have not as yet drawn any conclusions aside from obvious things like Penn-Plax Ultra TriLux bulbs are brighter than Triton bulbs and they have a color temperature closer to sunlight (whether or not that is good, I'm not saying).
To determine if the light fixture and ballast affected the temperature and intensity, 2 new 40 W Cool White bulbs (2 for $2.49) were tested in two fixtures. The readings were taken 8" from the bulbs, centered over the middle of the fixture.
An "electronic" shop light ($15.99 at Ace hardware) with a small coil/ capacitor ballast circuit produced 5000 K and 5500 Lux. This is an "energy saving" type fixture. My experience indicates this type of fixture is hard on bulbs such as Tritons and shortens their life.
As a side note, the October '92 Consumer Reports tested various types of energy saving light bulbs including compact fluorescent bulbs with electronic and magnetic ballasts. They found that a large number of the bulbs with an electronic ballast had premature failures.
A "standard" Sears shop light with a normal magnetic ballast produced 5300 K and 6650 Lux. The same fixture with aluminum foil behind the bulbs produced 5300 K and 7900 Lux. Also with this setup, the color temperature ranged from 5200 K to 5500 K when the distance from the bulb varied from 3" to 24".
I also tried to check the lights in a Coralife metal halide and fluorescent fixture to see if an "expensive" ballast and polished reflector did anything, but the bulb spacing was very different giving readings which could not be meaningfully compared to the first three.
I had hoped to determine color temperature and aging effects of metal halide bulbs, but the temperature meter responded very strangely to most of the bulbs I tested, giving readings of 9000 K to 14000 K depending on the distance from the bulb. One older, 1000 W MH bulb read 6200 K, but I don't have much faith in that number. By the way, an actinic bulb caused an "overrange" reading (above 40000K).
The local fish store allowed me to check some of the bulbs in the store. Most of the bulbs are about 1 year old. I was surprised at the temperature range of the 40 W Power-Glo bulbs; possibly caused by differing age or fixture differences or just bad quality control - I don't know. I believe the higher K temperature indicates more blue, which shows up some fish colors better.
20w Ultra TriLux 5200 K new (in a cheapie hood) Phillips UltraLume 5800 K 6 months Artic Brite 6000 K 1 year Power-Glo 7200 K " " 7850 K " " 8700 K " " 8700 K " " 9350 K " " 9800 K
A local lighting store has a light bulb comparison center with various commercial bulbs. The age of the bulbs was unknown. Unfortunately, the store was remodeling and some of the display was not working, so I didn't get a chance to try a Chroma 75.
Warm White Delux 3460 K Warm White 3630 K Regal White 3640 K Natural 3900 K Spectra 35 3930 K Spectra 30 3930 K Chroma 50 4600 K Cool White 5000 K Cool White 5150 K Daylight 7500 K
We have quite a few bulbs of various ages on hand. Each was tested in the Sears shop light. The second bulb in the fixture was covered with a piece of cardbard so it would not affect the readings. The age of the the bulb is indicated, if it was known. All readings are 8" from the bulb.
Agro Lite 3590 K Wide Spectrum 3900 K Advantage X 5300 K 5100 K in electronic fixture Artic Brite 5850 K new 5900 K in electronic fixture " " 5700 K Ultra TriLux 6750 K new " " 6350 K 6050 K in electronic fixture " " 6150 K " " 6150 K Triton 8150 K new " 8400 K new " 7550 K 300 hours " 7550 K " 7350 K 4800 hours
Combinations of bulbs in use on our various planted aquariums were tested. The average age of the bulbs is about 6 months.
1 Power-Glo, 1 Ultra TriLux 6550K 1 Triton, from 1 GE Wide Spectrum, 6000 K 1 Triton, to 6300 K 1 Ultra TriLux to 6500 K 1 Ultra TriLux, from 1 Ultra Trilux, 6150 K 1 Triton, to 6050 K 1 Ultra Lume to 5900 K 1 Artic Brite, 1 Ultra TriLux 6400 K
From the local fish store (meter was 4" from the bulbs except for the metal halide bulb):
1000 W Metal Halide, 12" from fixture 69,000 Lux 2 Power-Glo bulbs, 11,000 Lux 1 Power-Glo, 1 Ultra TriLux 8,200 Lux 1 Power-Glo, bad fixture 2,400 Lux From the local lighting store (meter was 4" from the bulbs): Regal White 3300 Lux Chroma 50 3500 Lux Warm White 4600 Lux Cool White 5000 Lux Our bulb stock (meter was 8" - twice as far - from the bulb): Wide Spectrum 2400 Lux Artic Brite 3750 Lux new 2780 Lux in electronic fixture " " 3200 Lux Advantage X 4200 Lux 3370 Lux in electronic fixture Triton 3890 Lux new " 3820 Lux new " 3730 Lux 300 hours " 3610 Lux " 3300 Lux 4800 hours Ultra TriLux 5000 Lux new " " 4720 Lux 3520 Lux in electronic fixture " " 4580 Lux " " 4480 Lux
We have had our dual 175W metal halide fixture in use for over 1 year and have tracked the intensity of 3 of the 5 bulbs in use over that period. The bulbs are on for 10 hours per day.
Directly below the bulb at the acrylic shield: Bulb 1: 78,000 90 to 270 days 72,000 320 days 57,000 360 days Bulb 2: 126,000 10 minutes 110,000 4 hours 102,000 2 days 94,500 4 days 90,000 13 days 90,000 21 days 86,500 62 days 82,000 110 days 72,000 285 days 60,000 315 days At the water surface, 13" from the bulb: Bulb 2: 9,900 Lux 1 year old; replaced at this point Bulb 3: 19,900 Lux after 2 hours 17,600 1 day (12 hours) 17,100 2 days (22 hours) 16,900 3 days 16,400 4 days 15,700 6 days 15,700 16 days
I was very surprised at the rapid initial drop off of intensity from the MH bulbs; I had expected a linear drop for most of the lifetime. The data indicates that MH bulbs should be changed once a year if you have intensity sensitive applications.
The following intensity measurements were made on a number of bulbs that we have, both new and used. All bulbs are standard 48" 40 watt fluorescent bulbs. The measuring setup was different than the setup used above, so these values are relative only to other values in this section.
The bulbs were put in a Sears magnetic ballast shop light with black cloth covering the second bulb and the reflector so the readings are direct light only. The measurements were taken 3" from the bulb in the middle of the bulb. A piece of PVC tubing 2.5" in diameter was used to support the lux meter sensor to get fairly repeatable readings. Each bulb was allowed to warm up for 5 minutes before a reading was taken.
These numbers are for comparison only. Because of the test conditions, they cannot be used to calculate any quantitative data.
Bulb Age Lux ------ -------- --------- TriLux new 7800 " new 7600 " 1800 hrs 7100 " 1800 hrs 6800 " 3600 hrs 6800 " ??? 6800 " ??? 6700 AX50 ??? 5800 Triton new 6100 " new 5700 " ??? 5600 " 1800 hrs 5400 " 1800 hrs 5000 " 1800 hrs 4900 " 3600 hrs 5300 Daylight 1800 hrs 5400 WarmWhite ??? 5000 BioLume 1800 hrs 4600 Chroma50 1800 hrs 4100 " 1800 hrs 3800 VitaLite ??? 3900 TriLux = Penn-Plax Ultra TriLux AX50 = Phillips AX50 Advantage X Triton = Triton Daylight = GE Daylight WarmWhite = GE Warm White BioLume = Rainbow Lifeguard Biolume with internal reflector Chroma50 = GE Chroma 50 VitaLite = Vita Lite
The following data was collected to settle a discussion on rec.aquaria about the "inverse square law".
Dustin Lee Laurence (firstname.lastname@example.org) wrote: > > Grant.Gussie@phys.utas.edu.au (Grant Gussie) writes: > >The famous inverse square law applies ONLY to a point source of light that > > In fact, the 1/r intensity rule for line sources is not terribly > relevant either as far as fluorescent lit reefs is concerned. 1/r^2 > is more than accurate enough for considering metal halide lighting, > which is for these purposes a small number of point sources which can > be superimposed incoherently.
Here's what I measured last night. These were quick and dirty measurements made after consuming a lovely cocktail and as such, probably don't qualify for the Journal of the Society of Lighting Engineers. But they should be good enough to add fuel to the fire :-). I didn't get around to the MH light yet, maybe tonight.
Distance from center of bulb (cm) 10 20 40 80 120 ------------------------------------------------------------------------ 1. Sears shoplight; 1 Triton, 10,800 5,300 2,400 940 530 1 Coralife TriChromatic 1.5 measured at end of fixture 6,900 3,400 1,600 780 450 2. Sears shoplight, 1 Triton, 15,050 8,000 3,600 1,520 810 1 Ultra Trilux 3. Sears shoplight, 1 Triton, 13,500 7,200 3,100 1,370 780 1 Ultra Trilux 4. Triton Enhancer, 1 Triton, 13,300 8,400 4,100 1,630 830 1 Ultra Trilux Notes: 1. All fixtures use magnetic ballasts. 2. All measurements made at center of fixture (except 1.5). 3. Bulbs are various ages. 4. Intensity values are lux (lumens per square meter). 5. Sears shoplight reflector is "w" shaped; Triton Enhancer is "sort of" hemispherical. Enhancer seems to help a little at longer distances.
Distance below surface of water (cm) 0 10 20 30 40 ------------------------------------------------------------------------ 1. Bulb and fixture from (3) 16,000 8,800 7,700 6,500 5,100 Notes: 1. Measurements made in a 29T gallon tank with water only. 2. Fixture positioned at back of tank and overhung the sides by 9". 3. Measurements made below center of bulb. 4. Bulb centerline was 5 cm above water surface. ----------------------------------------------------------------------------
The intensity in water was surpising at first. From 0 to 10 cm, the intensity dropped by half. Below that, the intensity dropped much less and linearly with distance. I speculate that reflections from the sides of the glass were augmenting the intensity at lower depths and cannot be depended on in a real setup with plants and all. When I get the 120 gallon tank setup, I will repeat these measurements. It is 24" wide and 48" long, which should reduce the reflections.
The following describes the hood I made myself for our 85 gallon tank. I have also had two more professionaly made out of oak (for better appearance). The goals of the hood design were:
The hood is designed for an 85 or 100 gallon acrylic tank: 60" long by 18" wide (front to back). The basic design is a box, open at the bottom and split across the top with a piano hinge holding the two halves together. Two shop light fixtures are attached to the inside top of the hood. The height of the hood is enough to allow the bottom of the light fixtures to move over the tops of any external equipment. The back is partially open to allow it to slide backwards over the equipment and to provide ventilation for the lights. The hood fits over the rim of the tank and rests on some 1/2" sq recessed wood strips on the side of the hood. If you have a 48" by 12" tank (55 gallon), you will need to make adjustments for the length and width of the light fixtures.
front hinge back _____________ ______________ side view: | MMM -o- MMM | | /=====\ : /=====\ : | / O O \ : / O O \ : | : : wood strips-> |n___________:____________n: | | | tank | | |
To feed the fish or do minor maintenance, the front half of the hood is pivoted upwards and allowed to rest on the rear half:
............... : n| :\ / | : \ O O / | : \=====/ | :____WWW______| o______________ side view: : MMM | : /=====\ : : / O O \ : : : _________________________n: | | | tank | | |
To do cleaning and heavy maintenance, the front part is raised like the previous picture and the whole assembly is slid back 4 or 5 inches until it hits the wall behind the tank. The first hood had some push-button switches on the side to turn on the front and rear lights independently (turn off the front half so you aren't blinded when you raise the front!). On the other two, I mounted the switches inside, out of sight (you are only temporarily blinded when you raise the hood to get to the switch!).
To build the hood, I made two frames of 1/2" square pine strips (kind of a "wire frame" model of each half of the hood). I then nailed and glued 1/4" tempered (waterproof) Masonite on the front (or back), sides and top of the frames. I didn't overlap the corners of the Masonite, so there was a 1/4" "notch" all the way around the top and sides. I obtained some 1/4" square basswood from the local hobby shop and glued this in the notch. Using a "Surform" tool, I rounded the bassword to give a nice edge (It would be hard to round off the Masonite). A little Plastic wood or spackling filled in any imperfections, so the hood looks like a one piece unit. The only "precision" woodworking required is to cut the Masonite nice and square. The frame simply serves as a way to connect the Masonite pieces and act as a support rails for the hood; all the strength is in the Masonite.
Masonite top 1/4" basswood ___________________________ goes here ------> _|__________________________| O <- 48" Piano | | | | || | hinge Masonite -> | | |__________________| || | front | | | | | | | | | | | | | | o---- 1/2" pine | | | Masonite | | frame | | | side | | | | | | | | | | | | | | |__________________| | frame rests on | | | | | top of tank front and sides extend | |___|__________________|___| _______________ below frame to hide | | | | top edge of tank |_|__________________________| | tank
I painted the hood white and sealed the inside with Marine varnish. The smooth, seamless white hood gives the tank a very modern look. I guess you could get veneer and cover the outside to make it look like wood. I have been very happy with the hood; it makes access to the tank very easy and it looks good. There has been no warpage of any kind (over two years).
Lumens (40 watt) FL-40 "Aquarilux" 900 FS-40 "Sea-Lux" 1570 NL-40 "Natura-Lux" 2450 FT-40 "Tri-Lux" 2480 FUT-40 "Ultra Tri-Lux" 3350 Wavelengths cl wht FL-40 FS-40 NL-40 FT-40 FUT-40 --------------------------------------------------------------------- u-v than 380 nm | 1.7% | 1.7% | 1.5% | 1.8% | 0.5% | 0.6% | violet 380-430 nm | 7.6 | 8.6 | 11.2 | 9.8 | 3.1 | 4.1 | blue 430-490 nm | 20.8 | 24.6 | 20.6 | 21.4 | 38.4 | 29.8 | green 490-560 nm | 24.6 | 13.2 | 19.1 | 25.5 | 21.2 | 34.1 | yellow 560-590 nm | 18.3 | 4.1 | 8.1 | 13.2 | 3.6 | 4.5 | orange 590-630 nm | 17.7 | 7.9 | 12.3 | 12.6 | 23.5 | 20.1 | red 630-700 nm | 8.5 | 36.7 | 22.7 | 11.7 | 7.3 | 5.0 | i-r 700-800 nm | 0.8 | 3.2 | 4.5 | 4.0 | 2.4 | 1.8 | Wavelengths cl wht FL-40 FS-40 NL-40 FT-40 FUT-40 --------------------------------------------------------------------- u-v than 380 nm | 52 | 15 | 24 | 44 | 12 | 20 | violet 380-430 nm | 232 | 77 | 176 | 240 | 77 | 137 | blue 430-490 nm | 634 | 221 | 323 | 524 | 952 | 998 | green 490-560 nm | 750 | 119 | 300 | 625 | 526 | 1142 | yellow 560-590 nm | 558 | 37 | 127 | 323 | 89 | 151 | orange 590-630 nm | 540 | 71 | 193 | 309 | 583 | 673 | red 630-700 nm | 259 | 330 | 356 | 287 | 181 | 168 | i-r 700-800 nm | 24 | 29 | 71 | 98 | 60 | 60 |
Here are some manufacturer's specifications for various commonly available 40 watt fluorescent bulbs (sorted by intensity):
Manuf Brand Name Mfg Number Temp CRI Lumens ------ ------------- ----------- ---- --- ----- GE StayBright F40SXL/SP41 4100K 70 3400 PennPlax Ultra TriLux FUT-48 6500K ? 3350 Philips Advantage F40AX50 5000K 80 3300 GE Warm White F40WW 3000K 52 3150 GE Shoplight F40SHOP 4150K 62 3100 GE Cool White F40CW 4150K 62 3050 Philips Ultralume F40/50U 5000K 85 2950 GE Daylight F40D 6250K 75 2550 PennPlax Tri-Lux FT-48 ? ? 2480 Philips Daylight F40D 6500K 79 2250 GE Dlx Cool Wht F40CWX 4175K 89 2250 GE Chroma 50 F40/C50 5000K 90 2250 GE Soft White F40SW 3025K 52 2200 GE Dlx Warm Wht F40WWX 3025K 77 2200 GE Chroma 75 F40/C75 7500K 92 1950 Philips Corortone F40C50 5000K 92 1915 GE Plant & Aquar F40PL/AQ 3050K 90 1900 Philips Colortone F40C75 7500K 95 1720 PennPlax Aquarilux FL-48 ? ? 900 GE "Gro & Sho" F40PL 6750K ? 800