I would not bother with LEDs in that small a space. The inexpensive ones are generally crap and the better ones are designed for larger areas. If you are not sure you are going to expand into a larger area a single 26 Watt CFL grow light can be had for under $15 and certainly should be good enough. Both your lettuce and your kale will do fine at 6500K.

Surrounding the plants with reflective material is helpful but aluminum isn't the best choice, any shiny white plastic is fine.. or even gloss white paint on board.

How many lumens should I be looking for?

A good 26(ish) Watt CFL should give 1800lm, the super high output grow light versions can give 2200 - 2500lm.

I've been really pleased with these.

http://www.amazon.com/gp/product/B00AKKU7SA?keywords=sunblaster%20grow%20light&qid=1444090554&ref_=sr_1_2&sr=8-2

T5ho puts out more lumens per watt than led.

This myth that led's are the most efficient is just that. A myth.

That's not entirely true pip313.. the lumen is a unit of measurement that is entirely based around human perception, it is used to allow consumers to cross compare different bulbs used for home lighting.

There is a chart on the Wikipedia page here

https://en.wikipedia.org/wiki/Lumen_%28unit%29

from the chart a 100Watt incandescent provides at a minimum 1,600 lumens. To get the same number of lumens you would typically expect to use 24 -28 Watts of CFL and 18-22 Watts of LED. Not a big advantage for LEDs and certainly there are some specialized Fluorescent bulbs that produce more lumens but there are also specialized LEDs to do the same.

The biggest difference from the standpoint of growing plants is that the LED lights for growing are tailored to the wavelengths that the plant find most useful. In human perception yellow and green score highly helping provide a high lumen count, plants make little use of these wavelengths, rather they use Red and Blue light both of which score rather low on the scale of human perception.

Ok I have one more question. I know that the bulb needs to be positioned close to the plant, but I also know that CFL bulbs give off heat, and heat isn't good when you want to grow lettuce and kale. So how do I balance bulb distance with heat so that it doesn't hurt my lettuce?

Also, are reflectors required? If they are, I'm afraid of them holding in more heat.

Because of the heat issue, I'm more inclined to use an LED bulb...even if it costs up to 3x more than a CFL of the same specs.

You used CFL to compare to led. T5 ho is the florescent grow lamp that should be compared. Remember I specifically said t5 ho.

Your right that lumen is what the eye sees. Your wrong in the ideas that led spectrum is better. Led has one single peak fluorescent has multiple peaks, while not ideal it is closer to what natural light is.

plants don't want red, plants don't want blue, plants want both with higher levels of one or the other during different seasons. This is why some led fixtures have both. But even then it's only 2 single peaks of spectrum. Plants want wide spectrum.

honestly the best bulb is cmh. Ceramic metal halide. But they push out crazy uv.

As far as heat xx watts is xx watts is xx watts. Hid cfl t5 led they all put out heat. Another myth that needs to be busted.

In order to not put out heat a bulb would need to be 683 lumens per watt which is impossible right now. At that rate 100% of electricity will become photons. No light does it, so all lights waste energy as heat.

just some physics for you.

s_c_f ... in all honesty in a 16 inch by 10 inch area I would be torn as to what to use. From what I've seen the inexpensive LEDs (the type that thread into a standard socket) are not very good at all, and the specialized fixtures at 75 or 90 "Watts" (made of 25 or 30 x 3 "Watt" chips) are overkill for your small area yet not really big enough if you expand.

I can't recommend strongly enough that you ignore the thread-in type LEDs and anything where the individual emitters (chips) are less than 3Watts. (Yes I know they are not really 3 Watts but it is a descriptive term for the class.)

I use reflectors with my CFL bulbs since I figure that if I'm paying for the light I should use it... but is also allows me to keep the lights higher above the plants and makes heat less an issue. For me they work fine 12 to 16 inches above the plants. I keep the LED fixtures much higher.

pip313 the Wikipedia chart did not list T5 bulbs so I used CFLs which emit a similar amount of light per watt, but the point is still the same.. a bulb can pump out tonnes of yellow and green light and get a high lumen rating because yellow is what our eyes want, but yellow and green are pretty close to useless for plants. A GOOD LED can be better because it can be customized to produce light that plants can use and not waste energy on the spectra that plants don't need. Decent LED fixtures will have 7 to 12 "bands" covering the Blue and Red ends of the spectrum from UV through IR but virtually ignoring the greens and yellows.

Ironically because our eyes don't react as strongly for Blue and Red they would score low on the lumen scale.

Your not comparing apples to apples. You keep mentioning the different spectrums. Which is fine. But you then compare blue to yellow and such. You need to compare 5000k to 5000k, 6500k to 6500k and so on.

40000 lumens of 6500k light comes from one of my t5 fixtures at 430 watts. What's the lumens at 6500k and wattage of your led so we can truly compare them.

or if your spectrum isn't 6500k post what it is too so we can find a fixture to compare them.

There is a lot of miss-information around LEDs and I think our
discussion is basically hinged around one of the more common issues.
There are white LED chips designed and used mostly for lighting, these
chips are rated by colour temperature and lumens just like florescent bulbs.
These lights are really a hybrid technology, basically a deep blue LED
is filtered through a yellow phosphor, some of the blue light is
converted to yellow light by the
phosphor, the remaining blue light, when mixed with the yellow light,
results in a white light. Absolutely everything you say regarding these types of lights is true. Those things are fine for home use,
which is what they were designed for, but they are quite inefficient for
growing since the useful blue light is converted to almost useless
yellow light.

That there are LED grow lights that not only use but ADVERTISE that they contain white chips is just one more unfortunate reason why there is a lot of distrust about the products. (It appears that they are often used so the manufacturer can refer their lights as "Full Spectrum" in promotional material)

You can't really even use the terms spectrum or colour temperature when discussing good LED fixtures. A spectrum in the traditional sense is a continuous wave form, the spectrum for visible light starts around 400nM and ends around 700nM, every type of grow light EXCEPT LED uses a single emitter and produces light as a wave. In an LED grow light there are a large number of chips, and each individual LED chip produces light in a very narrow band typically only 10 to 20nM wide. In, for example, the Red "spectrum" that runs from about 620nM to 680nM a good LED Grow light would typically use 3 preferably 4 different types of chip, to cover that range. My last set of home builds had 11 bands and the next set will probably have 13.

No spectrum is not a continuous wave form and blue light is not filtered through yellow filters. Show me the filter if that's true.

You sound like you know what's up, but you use the terms wrong and give misinformation.

Keep in mind before you buy a led fixture that just like any other bulb they loose efficiency with time. But you can't change them easily. So your more expensive fixture would need to be tossed and replaced instead of the bulbs.

All light regardless of source is in wavelengths, the spaces between the peaks determines wavelength.

also the idea that all "emitters" as you say put out a single wavelengths is false too. Sodium from hps lights are the closest to what your saying and no sodium does not put out a single wavelength. It just has one that peaks much higher. Do you think plants know or care what the source is?

ive had high quality led's, I bought into it. Then checked the spectrum of light being output and the levels. T5 ho spanks led. Maybe in 5 years when today's lab led's make it to market it will change. Not today.

i stand by all my statements including the fact that the best light is cmh, ceramic metal halide. If you don't mind the insane uv output.

you say

No spectrum is not a continuous wave form ..

From the Collins English dictionary

Spectrum:

1. ...the distribution of colours produced when white light is dispersed by a prism or diffraction
grating. There is a continuous change in wavelength from red, the longest wavelength,
to violet, the shortest.

you say
and blue light is not filtered through yellow filters.

sorry maybe "filtered" isn't the best choice of words but a phosphor isn't a filter.

From Merriam Webster

Phosphor:

2. a luminescent substance that emits light when excited by radiation (as
electrons) and is used especially in fluorescent lamps and cathode-ray
tubes

in the case of LEDs the Phosphor is excited by blue light.

From LEDs Magizine article titled "How to make a white LED."

2 Phosphor downconversion

The light emitted from an LED chip passes through a phosphor
material which "downconverts" a portion of the light. The phosphor
absorbs the light and re-emits it with a lower energy, which means the
light has a longer wavelength.

Light from the phosphor mixes with unconverted light from the LED chip
to form white (if the phosphor has been correctly chosen).

The most common example combines a blue LED chip and a phosphor emitting in the yellow region of the spectrum...

Article here...

Edit Added..

you say

all "emitters" as you say put out a single wavelengths is false too

I didn't say they produce single wavelengths, I said "..LED chip produces light in a very narrow band typically only 10 to 20nM wide"

From OSRAM "one of the two leading light manufacturers in the world" s website

OSRAM

And the following is LED's special feature: Each LED light color is
limited to a very narrow range of wavelength (keyword: dominant
wavelength) which accordingly only represents a specific light color.

End Edit

It is too bad that you feel that you were not well served by your LED fixtures there has been quite a bit of independent research that supports the case that properly constructed LED lighting is more efficient.

No.

the distribution of colours produced when white light is dispersed by a prism or diffraction grating. There is a continuous change in wavelength from red, the longest wavelength, to violet, the shortest.

You must have missed the part where they specifically mentioned white light. And even then not all white light has a continuous change. Some has gaps. Led anyone? They must have been thinking sunlight.

a luminescent substance that emits light when excited by radiation (as electrons) and is used especially in fluorescent lamps and cathode-ray tubes

What at produced the blue light? You misinterpreted that too.

yes you can absorb and reemit light. What's your point? You specifically said filter. I said no not filter. To prove me wrong you proved yourself wrong.

"These lights are really a hybrid technology, basically a deep blue LED is filtered through a yellow phosphor, some of the blue light is converted to yellow light by the phosphor, the remaining blue light, when mixed with the yellow light, results in a white light. Absolutely everything you say regarding these types of lights is true. Those things are fine for home use, which is what they were designed for, but they are quite inefficient for growing since the useful blue light is converted to almost useless yellow light."

That is a misleading statement because one could infer that "yellow light" is a very narrow band near 590 nm, when, in fact, a typical household LED bulb emits a very broad band that is almost all green and red light. If a household LED bulb emitted only blue and yellow light, a red rose would appear black which is not the case. This is a power spectral density chart for a typical household LED bulb:

http://www.gl-leds.com/led-bulbs/img/b60-led-bulb-spectrum.jpg

Compare with the spectra of typical fluorescent bulbs:

http://ledmuseum.candlepower.us/led/spectra7.htm

The fluorescent bulb spectra have a peak at 611 nm with practically no output above 630 nm whereas the LED bulbs have a substantial amount of light above 630 nm. Consider the fact that, at equal lumens, 611 nm light provides only 1/9 as much energy as light at 660 nm. Thus, at equal lumens and color temperatures, the household LED bulb provides both more red light and PAR energy than a fluorescent bulb.

Elsewhere in this forum I have mentioned that a 8.5 watt, 800 lumen, 3000 K household LED bulb for $2.50 is a good deal. Too late, that sale has expired.

sigh ... pip believe what you want, twist the language all you want, this was a discussion in a couple of paragraphs not a scientific monograph.

post a link to some data or not

Sorry but your the one who twisted the language. I'm trying to let people know the truth.

In my opinion, the spectrum of the light (what type of grow lamp you are using) is much more important than the color temperature. Lower color temperatures are really not that beneficial for plants unless the spectrum contains plenty of longer red wavelengths that the plant can use. At least with higher color temperatures, the chlorophyll can convert some of that blue light into energy, albeit less efficiently.

And if you are trying to grow red lettuce, the leaves need some UV in the spectrum to develop a red blush. With metal halide, higher color temperatures tend to have more UV.

Lettuce grows very good under LED, btw. Do not overdo the wattage, you do not want those leaves drying out.

Some cheap CFL grow lamps do not use a very good phosphor composition (meaning they put out a poor spectrum) and are scarcely better than regular ordinary CFLs.

I don't want to jump into your debate too much, but I do want to note that LEDs are now more efficient in terms of lumen/watt than fluorescent, metal halide and HPS.

They are not more economical, especially at higher wattages, due to the current cost of state-of-the-art LED chips and the attention that must be put into thermal design to cool them - heatsinks and fans are not necessarily expensive parts, but expensive to design for mass production.

Cree CXA3070 or even better CXB3070 (not as available yet) can do 150 lumens/watt at 60 watts with a single emitter ("led"). The emitters are expensive today, but get cheaper and better very quickly. Currently these emitters are around $30 - $40 each in single quantities, and you have to be a bit handy to mount them to heatsinks and properly drive them, but anyway, LED is definitely the most efficient (in terms of lumens/w) way to light plants finally.

Oh yeah, those chips mentioned above are full spectrum white that come in a variety of color temperatures and have as good or better spectral distribution than HPS or Metal Halide - funny looking blue/red lights are not required anymore, though you can of course find plenty for sale.

A hps also does 150 lumens per watt.

led has potential, but the real good ones are in labs, not mass production.

Those Cree models are not actually "emitters", they are arrays of many individual emitters (each with it's own color temp), all mounted together, which combine their individual outputs to create a field/beam of color-corrected white light. This is the major problem with measuring LED--It takes multiple individual emitters in combination (COB/array/etc.) to provide any real spectral range in the output. Notice I didn't say "spectrum", because that would be a measurement of said combined output spikes, measured at some given distance from the collimator or lens. Individual LED emitters will only create spikes of their given color, so manufacturers combine the multiple spikes into a pretty little graph they can print on the back of the box. So it's still exact, but not really the same as measuring the spectrum of total light given off by a single CFL bulb, which acts like a single wide spectrum emitter. Rant over.

Yes, at higher (600+) wattages. HPS is more efficient at higher wattages, generally LED is more efficient at lower wattages. Not many hobbiests need or want 1000W light in their home. LED will do better at lower wattages every year, and HPS is not going to improve. At 150W, HPS is closer to 100 lumen/watt. With 2 LED emitters you can do close to 50% better than a single HPS bulb at 150W, with a nicer spectrum as well. Maybe not yet available in mass produced commercially available fixtures, but all the necessary components to build them exist right now in single unit quantities (you dont need to order 100+ parts). Not just in labs. You can bet there's many places tooling up to produce fixtures as well. If you can solder and understand what volts, amps and watts are you can easily build one yourself today.

I just want to provide an alternate perspective.

I also think funny colored led grow lights will disappear, but probably not for a while because its easy to market them as something that plants crave. Though it doesn't take much physics knowledge to notice that sunlight isn't colored like a nightclub and plants like it better than anything else.