Lights, Action. . .

Like most gardeners I start seeds indoors.  For years I’ve used a basic setup in the sunroom.  The seed flats were set on a folding table.  A shoplight with two 5100K fluorescent bulbs was hung from the ceiling so it was just above the seedlings.  A timer kept the lights on 16 hours a day.   

The seedlings were always spindly with this setup.  I put L-brackets on some white shelving so the shelves could be stood on either side of the light unit in order to reflect the light back onto the plants (they’re below the table in the picture).  That helped some.  I put a small fan at one end of the table to blow air over the seedlings for an hour a day.  Still the plants were spindly. 

Last spring the seedlings looked worse than ever.  The brassicaes especially were not growing.  Part of the problem may have been with the coir-based seed starting media that I tried that season, but I think most of the problem was that the fluorescent bulbs, which had never provided adequate light, were now in their third season of operation.

I knew the lighting had to be improved, and that’s when I started looking into what kind of light plants need and how to meet those needs.  My thinking about driving photosynthesis was basically that you try to duplicate what the sun does.  I had never really looked into how artificial light meets the photosynthetic requirements of the plant.  I found out that I was wrong about a lot of things.

Most home gardeners use fluorescent lights to start seeds.  They are much more efficient than incandescent lights, which radiate only a small amount of their energy at wavelengths that plants can actually use.  Incandescent lights, like the sun, are “black-body radiators” meaning that they give off light as a result of being heated to a high temperature.  The filament of an incandescent bulb is heated to about 1200 Kelvin, or K, which is its color temperature.  At that temperature most of it’s light is produced in the infrared region.

To match the sun’s light spectrum that filament would need to reach a temperature of 6000K.  Then it’s peak output will be at a wavelength of about 550 nanometers (nm), or yellow-green.  This is the spectral power distribution curve (SPD) of sunlight.

The most economical way for a home gardener to duplicate the sun’s spectrum is with a fluorescent light.  Fluorescent lights are not black-body radiators, they work by electrical excitation of a small amount of mercury vapor inside the tube.  The mercury emits UV light which causes the phosphor coating on the inside walls of the tube to fluoresce.  The selection of phosphors determine the wavelengths that the tube fluoresces at.

Most of you probably know that plant photosynthesis requires light in the red and blue parts of the visible spectrum. (I knew that but never thought about it much).  Why higher plants evolved a photosynthetic process that ignores the peak output of the solar spectrum is open for speculation, but that is what they do.  This schematic, from the website Hyperphysics, shows where chlorophyll absorbs light.  Light from about 510 nm to 600 nm is not absorbed, the reasons plants look green.

Carotene functions as an auxiliary pigment that can capture some of the light in the yellow region.  Apparently carotene works by capturing this energy and passing it on to the chlorophyll centers.  This extends the plant’s absorption spectrum somewhat, but yellow-green light is still mostly useless to the plant.

When choosing lights one factor that should be ignored, but often is not, is the human factor.  The figures above shows what the plant sees.  What do we see?  It just happens that human sight is most sensitive about where sunlight has peak strength – right around 560 nm.  Here’s a chart that measures the sensitivity of the eye, or spectral luminous efficiency at different wavelengths.  What the plant “sees” and what we see are two very different things.

In practical terms of choosing lights, what does that mean?  Basically this:  lumens don’t mean diddly (parroting what a former coach of the Colts once said after a game).   Many flourescent lights have their output rated in lumens.  Lumens are calculated by weighting the light output by the spectral efficiency at a given wavelength (note that it’s peak is 1 at 560 nm).  That means light at a wavelength of 550 nm is weighted by a factor of close to 1, while light at 650 nm is weighted by a factor of about 0.1 when calculating total lumens.  Big difference!  When it comes to plant growth, lumens are not only meaningless, they are misleading.

Here’s a spectral power distribution curve for a Sylvania warm white fluorescent bulb with a correlated color temperature (CCT) of 3000 K.  (This is copied from a technical bulletin, Spectral Power Distributions of Sylvania Fluorescent Bulbs). Fluorescents don’t have true color temperatures since they are not black-body radiators, so they are assigned a CCT that describes what color temperature they appear to be to the human eye. 

 

There is some output above 600 nm that is useful for plants, but the peak output is about 570 nm which the plant can’t use.  There’s not much going on in the blue region except for a strong sharp peak at about 430 nm (blue).  That’s a mercury excitation peak and it’s in every fluorescent bulb.  It also happens to be right in the blue sweet spot for photosynthesis.

A 6500 K daylight bulb is probably the highest color temperature bulb you can find at the hardware store.  The SPD curve for this bulb shows a big output in the blue range, and of course that mercury spike at 430 nm.  A 6500K bulb should provide the most useful light of any of the commonly available bulbs.  I know that some growers use a combination of warm white and day light bulbs to provide both the blues and reds for plants.  The bulbs at in-between CCT’s, 4100K and 5000K , like the warm white bulbs, emit very little light in the blue region and some red light. 

There’s one more option if you can find it, and that’s a bulb designed specifically for plants.  You have to be careful here as some lights are designed to make plants look good, not to provide useful light for photosynthesis.  This is true for many aquarium lights.  I did a number of searches on the Internet and could not find any spectral power distribution curves for any bulbs except Sylvania Gro-Lux bulbs.  And if a manufacturer is not willing to publish that data I can only assume they don’t want you to know.   Here’s the spectrum for a Gro-Lux bulb:

From the data at least it looks like these bulbs are the way to go.  But where to find them?  I searched around on the Net and came up empty.  Then last week I was in Menards, a midwest lumber store, and in the bulb section there they were, Gro-Lux T8 bulbs, $14.98 for a pack of two.  That’s a little pricey but since a fluorescent bulb will last several years they seemed worth the expense.  I bought a pack. 

Even if the bulbs are better suited for plants than the 5000K bulbs I was using it seems that more light intensity can help even more.  I did not want to spend the money on a 4-lamp fixture when I had a perfectly good shoplight already.  This light, which I bought from Lowe’s several years ago, has a wide reflector that should hold two more bulbs.  There is plenty of room inside for another ballast.  I bought the cheapest T8 shoplight I could find at $9.99.  Here’s the shoplight (cover removed) I use alongside the light that will be cannabalized for parts.

 

I removed the reflector and cover on the new light with the removal of a few screws to expose the ballast.  The ballast was detached from the power cord and the power cord removed.  The ballast, wiring and bulb holders were removed from the light as one unit. 

 

The new ballast was attached to the light that I use.  The power cord was fed through the end of the unit.  This picture shows the new ballast installed.  It has not been attached to the power cord yet. 

 

Then I attached the new bulb holders.  I cut notches in the reflector that matched the grooves in the bulb holders and pulled back the sheet metal.  The bulb holders were then slid into the cutouts.  The ridge at the end of the reflector actually holds the bulb holders in place very well, although this is surely not up to any electrical code.

 

This is the light unit with the hookups almost complete.  At the fixtures’s end on the right side of the picture I had to put in a longer jumper wire from one bulb holder to the other.  At this point the fixture has a power cord at each end, one to power the inside bulbs and another to power the outside bulbs.  I later hooked both ballasts to one power cord. 

 

The cover was reattached.  I had to cut out a bit of sheet metal so the wiring could exit to the new bulbs. 

 

I put in the bulbs with the Gro-Lux bulbs in the center.  A 4-bulb fixture makes a lot more light!  Those Gro-Lux bulbs put out some weird looking light, but it’s not about what looks good to me.  Once I get things set up it’s time to start some seeds.