1

As doctors we are trying to frame some domestic-level recommendations to simulate medium intensity daylight indoors for patients in bed without proper access to actual sunlight. This is also an important need for us personally since our house lacks natural light during many parts of the day and almost everyone agrees that sunlight is the best light. I am currently using an 85w CFL and find it lacks the proper color sense of sunlight, since such bulbs with 6500k color temperature emphasise the 'blue' end of the spectrum. This creates a 'cold' ambiance which is rather disturbing. I would appreciate a more natural and balanced color scheme, but 'warm white' bulbs are far too yellow to serve that purpose, so I am considering a combination of white CFL and warm white incandescent bulbs in a 80:20 proportion of light intensity to supply the 'red' end of the spectrum that is lacking with just the CFL.

Practical considerations:

In order to provide both ends of the visual spectrum as also uv rays and infrared, I was thinking of installing one 85 watt spiral CFL side by side with one 100 watt incandescent bulb in a pair of simple open fixtures controlled by a single switch. The color temperature of the CFL is 6500k and it is supposed to output 5000 lumens of light, while the incandescent bulb is 2700k and outputs 1500lm. Voltage = 220v and current = 5amp. Are there any technical or electrical factors that counsel against this arrangement? If anybody has tried something similar, I would like your inputs. All suggested improvements on this setup are most welcome.

Background:

Since regular intermittent exposure to real or simulated sunlight is beneficial for various health purposes including sleep rhythm, mood and vitamin D3 synthesis, our aim is to design a composite artificial light that is near as possible to daylight in visual effect, as also electromagnetic content such as visible spectrum, infrared and uv rays, without possessing the extreme intensity of powerful sunlight or the technological complexity of professional arena/street/stage lighting.

Commonly available domestic lightbulbs including fluorescent tubes, incandescent lamps, CFL and LED would be used in appropriate combinations for this purpose.

Although a color temperature around 5000k is said to be in the 'natural daylight' range, such lights are not readily obtained here in India, where the commonly available options are 3000k ('warm white', an outright yellow light) 4000k (labelled 'neutral white', but really a less-yellow shade of gold) and 6500k ('cool daylight', which appears very artificial and very white). In my personal experience, each of these color temperatures has its own distinctive 'color cast' and each by itself doesn't really simulate the visual effect of a neutral sunlight which has no color cast and basically contains all color temperatures in a graded proportion - a full spectrum lighting.

Although complex machines can indeed simulate sunlight, they are beyond the scope of our project. However, each type of ordinary domestic lamp does represent some important characteristic of an ideal sunlight. For example, compact fluorescent lights produce ultraviolet rays and regular incandescent bulbs provide infrared (heat), while LED can give a strong semi-directional glare: moreover, at least 3 color temperature variants are commonly available. In fact such lamps are actually being combined in multiple ways to stimulate and support plant growth indoors in the form of full-spectrum grow lights.

That's what leads me to seek the advice of the home improvement experts here: what is the best way to combine various household bulbs in appropriate proportions of color temperature and light output to create an accurate simulation of medium intensity neutral daylight?

  • 2
    While certainly an interesting question, it has little to do with home improvement, and more to do with the physical properties of light and the implications of intensity vs. color temperature. You need a university physics lab, not some home tinkerers. :) – isherwood Feb 17 at 21:51
  • The application and benefits comes directly in home improvement @isherwood. Recommendations could be implemented by any home owner to approximate sunlight indoors. A practically viable combination of multiple spectrum lights is already being used to grow plants indoors - now added link in question by edit. – English Student Feb 17 at 21:55
  • That also is not strictly about home improvement, but horticulture. Our scope is fairly clear and narrow. diy.stackexchange.com/help/on-topic – isherwood Feb 17 at 21:56
  • Now, if you're asking specifically about how to install the hardware (and not about measurement and calculation of light output), feel free to revise to that end. You'd want to add detail about your installation conditions. – isherwood Feb 17 at 21:58
  • Thanks for pointing out the scope of this site @isherwood.I am primarily concerned with mixing lamps of various (known and rated) color temperatures and light outputs. I felt that a question about practically applied lighting is on-topic here. – English Student Feb 17 at 21:58
2

I have basic knowledge of what we did in a industrial plant to combat SAD so this answer covers what I learned when we changed the lighting to reduce the affects of SAD at our facility.

In our office areas we used 50% 6500k lighting and 50% 4500k lamps were matched with each ballast or 2 lamps at 6500k on 1 side and 2 at 4500k. Matching lamp type extends the life of the pair of lamps and reduced ballast problems.

In walkways we used 6500k lighting 100%

The reason for the difference was when walking the 6500k was closer to true daylight and the affect was similar to exercising out side even though just walking. The reason for 50/50 split was that 6500k was hard on our eyes when reading so reducing the level and mixing with 4500 ended up being the best over all.

We also had clean room areas that they wanted to get similar results however 6500k has two much blue light and was a problem for photo lithography (the process of making different layers on wafers). Some rooms only have yellow light no blue at all. The 6000k light was the max we could use without affecting the processes and still get some advantage from the light.

The last thing we did was offer some very concentrated UV lights in the bathroom lounges. Designed for short exposure on the neck and shoulders.

The results after several years was an increase in productivity and a reduction in accidents and sick time taken. I was a manager and I saw all the reports I think it was 19% improvement with no other changes attributed except lighting.

We did not work with 2500k-4000k at the plant but occupational Health and Safety recommended that in our homes we use 3500k lamps in areas like bed rooms because the bright white / blue lights affect our Circadian Rhythms and low levels of yellow / orange lights are more natural and if dimmed tell the body it’s time for sleep, where 6500k says its day time get to work.

We were told that the blue / white lights on the back of the neck , shoulders and head was the most effective location for intense light therapy. This site is in Corvallis Oregon and at one time the site had ~10k personal working on the site. Prior to the rollout of 6500k, 5000k, 5500k & 6000k lighting levels were also tested in different areas of the plant.

I think your plan to have the light to the rear as you mentioned is a good plan from what I had learned, I don’t know if the 3500k is really needed I remember that low levels did not affect the body in a negative way where sleeping was concerned but 6500k may prevent a person from sleeping up to an hour after exposure depending on the intensity.

I would get some additional references material on SAD as there is a ton of info on the web today. The lighting work I am describing was done 25years ago and they still use the same color frequency except in areas close to photo they went to all 4500k as geometry of the components got smaller the light had a larger affect. Hope this helps.

I just remembered, we had to remove the plastic bezels in the lamps and installed reflective grids , the plastic was absorbing the blue light and we had to have some kind of covers so the reflective metal grids increased the light as measured at desk top level. I do not remember how many foot candles / lumens the target was.

| improve this answer | |
  • Thanks a lot for a very detailed and useful answer @Ed Beal. Does mixing an even number of 6500k and 4500k (as you did in that facility) readjust the overall color temperature to 5500k in effect? In fact that's the color temperature of 'neutral' daylight but we only get 3000, 4000 and 6500k here. – English Student Feb 19 at 6:49
  • We were not concerned with the + and minus wavelengths but it was discussed. (our occupational health and safety group found in the office areas the 6500k was two bright and creating eye strain), mixing the lights softened the effect and we still had enough of the blue to reduce SAD. – Ed Beal Feb 19 at 14:35
1

I am currently using an 85w CFL and find it lacks the proper color sense of sunlight, since such bulbs with 6500k color temperature emphasise the 'blue' end of the spectrum. This creates a 'cold' ambiance which is rather disturbing.

That's not because of the color temperature. That's because of the light quality aka CRI or Color Rendering Index. From a classic CFL, CRI is terrible - especially when it's cheaply built Indian/Chinese stuff. This is why the light is "rather disturbing" - it is freakishly unnatural, probably a CRI of 40 or 50. What you want is an 85-95 CRI (100 is perfect).

The 6500K is exactly what you say you want. It is the color spectrum of a cloudy day. (A sunny day is closer to 5500K).

Try it with a 90+ CRI source, and I expect you will be much happier.

But regardless, 6500K is pretty blue. It's also the color of a cloudy day, so make sure the science is really saying that's the right spectrum. A sunny day is warmer - 5500K.

I would appreciate a more natural and balanced color scheme, but 'warm white' bulbs are far too yellow to serve that purpose, so I am considering a combination of white CFL and warm white incandescent bulbs in a 80:20 proportion of light intensity to supply the 'red' end of the spectrum that is lacking with just the CFL.

What's natural is lamps of appropriate color temperature and 100 CRI.

You should not be trying to mix and match, unless you are doing some very precise science with color spectrum, which would mean having the necessary color spectrometers to do that. Choose the color temperature you want. Then, buy quality bulbs for that color temperature that actually deliver a high CRI. You've been working with obsolete or low quality bulbs, so of course you've been getting a bad result.

our aim is to design a composite artificial light that is near as possible to daylight in visual effect,

Then simply have the color temperature of daylight, and the highest CRI obtainable.

Commonly available domestic lightbulbs including fluorescent tubes, incandescent lamps, CFL and LED would be used in appropriate combinations for this purpose.

CFL is a dead technology that was never very good, and has now been left behind. It's high time you dropped it altogether. I also don't see much use in using incandescents in this day and age. That leaves LED.

options are 3000k ('warm white', an outright yellow light) 4000k (labelled 'neutral white', but really a less-yellow shade of gold) and 6500k ('cool daylight', which appears very artificial and very white). In my personal experience, each of these color temperatures has its own distinctive 'color cast' and each by itself doesn't really simulate the visual effect of a neutral sunlight which has no color cast

That's not accurate. Every type of lighting has a color cast. Natural sunlight is 5500K which is fairly blue on the spectrum. Not as blue as 6500K "cloudy day", but still, pretty darn blue, especially compared to typical home lighting of 2700-4000K. There is no "colorless" light, that simply does not exist. Everything is relative. And the eye automatically adjusts. You can't see the color temperature of light unless you have something to compare it to.

Again, what you may be mistaking as "color temperature" is actually poor color rendering (CRI) from the relatively poor bulbs you've been working with.

while LED can give a strong semi-directional glare

That's only true for cheap or old LEDs. This has been cleaned up and LEDs can now be obtained with good CRI. Excellent in fact. However the best CRI is now linear tube fluorescents, which have evolved greatly in the last 15 years.

what is the best way to combine various household bulbs in appropriate proportions of color temperature and light output to create an accurate simulation of medium intensity neutral daylight?

The short answer is "Don't mix; it will only muddy up the color spectrum. Buy the one color temp you want at the best CRI obtainable".

The long answer is, "Get a spectrum analyzer and combine various light sources to produce a relatively flat spectrum". However, don't expect human eyes to see it. CRI is measured against what our eyes can see. This is not necessarily the same thing as what causes SAD.

| improve this answer | |
  • Thanks for a very informative answer @Harper - Reinstate Monica. May I summarise your advice that high CRI lamps of LED type with a color temperature of 5000 - 5500k would be the best approximation to 'neutral' daylight? I shall try to find if such lamps are available on a color-adjustable format, but as for fixed-color-temperature lights, we only get 3000, 4000 and 6500k here. – English Student Feb 19 at 6:55
  • @ Harper, to combat SAD higher numbers are needed small higher power SAD lamps are UV and that is what the 6500k + lighting targets. It’s not so much visual perception. – Ed Beal Feb 19 at 14:41
  • @EdBeal that would indicate the problem is more complex than OP is figuring for, then. Color temperature and CRI only apply to visible light, so if non-visible portions of the spectrum are essential, then OP is wasting time thinking about color temp. OP must do the more complicated spectral analysis I spoke of. – Harper - Reinstate Monica Feb 19 at 14:48
  • No read what I wrote. HP spent hundreds of thousands as this was one of their largest plants. Prior to the lighting upgrades SAD was a big issue especially for night shift. 1-2 years after the lighting upgrade we saw almost a 20% decrease in “sick call in’s” I did not write the report but was involved in some of the planning and testing. A large portion of the employees had no idea what was going on as not to skew the data set. The one exception was the high intensity SAD lamps in the lounges. – Ed Beal Feb 19 at 14:59

Not the answer you're looking for? Browse other questions tagged or ask your own question.