More on compact fluorescent lights

July 18, 2009

I compared a new14w CFL designed to replace a 65W incandescent recessed light (Commercial electric, model  EDXR -30-14) and an new 65W incandescent recessed light (GE Reveal 65) by measuring their spectra with a NIST traceable calibrated spectroradiometer.  In each case the bulb pointed down, like a typical recessed light, with the spectroradiometer measurement point 108 cm below the bulb.  The measurement was repeated seven times for each bulb: first with the spectroradiometer directly below the bulb, then with the spectroradiometer moved about 15 cm horizontally, then 30 cm horizontally…out to about 90 cm horizontal shift. 

Note that the GE Reveal 65 had an “enhanced color spectrum that used a neodymium glass filter to reduce the amount of light in the middle part of the visible spectrum to yield more vivid reds and blues.  I would have been better off with a simpler incandescent lamp for this comparison. 

The first graph below shows the spectral irradiance for the CFL.  Note that most of the irradiance is in the visible part of the spectrum.  The seven curves correspond to the seven horizontal positions, with the highest irradiance being directly below the bulb.  The second graph is the same, but zoomed in to the visible part of the spectrum.


CFL irrad 400-1400

CFL irrad 400-750

The following two graphs show the same thing for the incandescent lamp.  Notice the dip in the middle of the visible spectrum.  This is due to the neodymium glass filter.  If that filter were not present the total irradiance of the incandescent lamp would have been higher.  I will repeat this experiment at a later date with the simpler incadescent lamp.

Incan irrad 400-1400


Incan irrad 400-750

Irradiance only tells the beginning of the story.  The human eye is more sensitive to some colors than to others.  It is more sensitive to the middle of the visible part of the spectrum than to the red or the blue.  Of course, it is totally blind to the UV and the IR.  So, the irradiance is multiplied by  a Luminosity Function  and a constant to give a measure of how bright a light is.  The following plot shows the typically used Photonic Luminosity function.

Luminosity function

The following two graphs show the products of the Photonic Luminostiy function, a constant (683 lux/W/m2), and the spectral irradiance of the CFL and the incandescent bulbs.  The total area under any curve gives the “brightness” for the lamp at a particular horizontal shift.  I have deliberately left the Y axis the same on both graphs to make them easier to compare.  It is clear that the CFL is very bright over two narrow wavelength bands centered on about 545 nm and 620 nm, while the incandescent light is spread more evenly over the visible spectrum.  This is probably why people feel that colors look less natural under a CFL.CFL photonic

Incan photonicAfter all the graphs and the math, which light is brighter?  It depends on the horizontal position, as shown in the following figure.  The incandescent is brighter directly below the lamp, but the CFL is brighter off to the sides.  This should not be too surprising, because the light from the incandescent comes from a small filament, which is more easily reflected in the same direction than the light from the extended source of the CFL.  But when integrated over all directions, the incandescent and the CFL are probably a very close match, as claimed by the CFL manufacturer.

relative brightness

It would be interesting  to repeat this experiment with bulbs that have accumulated about 1000 hours.  But that is an experiment for another day.

Warm-up time.

I also measured the irradiance of the CFL as a function of time.  This was done for the lamp after it had been off and cool for hours, and again after it had been fully warmed and then allowed to cool for three minutes.   It takes about 4.5 minutes to get to full irradiance for a cold lamp, and about 3 minutes for a warm lamp.  Of course, the warm-up time for the incandescent is essentially zero minutes.



warm-up time


There are  hundreds of different configurations of CFLs  and incadescent bulbs being used in the world.  My sample is miniscule.  However, some of my numerical results are probably fairly representative, and there are common observations reported by many users. 

As shown above, at least in my case, the 14 Watt CFL was about a bright as the  65 Watt incandescent it was designed to replace.  However, the color quality of the CFL was much poorer.  This poor color quality is a function or the flourescent nature of the lamp, and is likely common to most CFLs. 

The CFL takes a long time to warm up, compared to the instant-on of an incandescent.  The warmup time probably varies from one type of CFL to another.  I have data to indicate that the irradiance vs. time for the warmup minutes can look quite different for a new CFL vs. and an identical CFL with several thousand hours, but that data is not presented here.

As indicated in a previous post, my experience is that a CFL will save money compared to an incandescent that it is designed to replace.  But as shown here, the color quality of the light is worse and there may be an annoying wait for it to warm up.

I will continue to use CFLs where they make sense, but I am also stockpiling some incandescents for the day when they are no longer available by government mandate.  Short duration use of many CFLs reduces their lifetime, and as seen above, it may take several minutes for the CFL to get to full brightness.  So I will use incandescents in closets and storage rooms, etc., and CFLs in the main living areas.

Last comment

I have presented this information as a small part of a large issue.  My endorsement of CFLs, despite some of their drawbacks, is most definitely not support for the government mandate to force us to use CFLs.  I am stockpiling incandescents for certain situations and would suggest that others do the same.  Perhaps the price of LEDs will drop enough to make this issue irrelevant.

Ultimately, I would like to see abundant amounts of energy available to all Americans and to all the people of the world.  Then the issue of light bulb choice would simply be moot.  My fear is that we are moving in the opposite direction.


  1. What matters is the quality of light. Affects every aspect of the function of any organism.

  2. What are the Dim Bulbs up to? It’s amazing! Can people be so stupid?

  3. Well-presented, objective (albeit with a small sample size) data. Thanks for the excellent post.

  4. Tom, thanks for this data! I am interested to know if the CFLs emit more UV than incandescents. Would you be willing to repeat the measurements (or repost the data) including wavelengths between 100 and 400nm, or whatever portion of the UV spectrum is possible with your equipment?

    I agree with kokdikas. I have always been hypersensitive to flicker. Exposure to fluorescent lighting (or CRTs at < 75Hz) for a few hours produces in me at best lethargy and at worst a migraine. I will be stockpiling incandescents too and when they run out seriously contemplating candles and oil lamps at home.

  5. Hi, I would also be very interested in UV data on these bulbs. Thanks for the data.

    I’m pretty sure that I have smelled ozone coming from a couple of my bulbs when they were new (UVB and UVC will cause oxygen to combine into ozone). (I read somewhere that the glass becomes more opaque to UV with exposure). Any info on this would be much appreciated. The last thing we need is more indoor air pollution.

  6. Is the life of the CFL shorter
    or equal to an incandescents bulb
    in an older light sockets
    (i.e. sockets installed 10-15 years ago).
    I read somewhere,that this was discovered in Austraila.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: