I recently switched my bedroom/office overhead lights to CFL from 40 watt GE Reveal (which I learned about from this BBS; thanks guys!!). Trying to replicate the full spectrum brightness of the Reveals, I chose with a 5500K "bright white". Seriously, I hate the yellow glow from ordinary lights, so I had to compensate.

Um, but I compensated a bit too much. 5500K CFL's, even at 9 watt / 40 watt equivalent, are DAMN white. In fact, they're so bright, they're dark. The corners of my room, ~12 feet away from the light source itself, seem horribly dark. The CFL's are exposed; there is no cover over the light (because it makes them look yellow and dim. Maybe that's the problem.

Well, looks like I'm stuck basking in STARK white light for the next seven years. How I wish for full spectrum CFL...

What kind of fitting (base) is it?

Part of the problem is the color temperature. GE Reveal bulbs are 2850K. (Normal incandescents may be as low as 2200K.) To get light like your Reveal bulbs, there are CFLs available in "bright white" (typically 3000-3500K) rather than "soft white" (trying to emulate traditional incandescents).

I also bought a bunch of 5000K CFL bulbs, put them in, and my wife hated them, saying that our house felt like the inside of a hospital at night. I've since replaced them with ~2900K halogen bulbs. This makes my wife much happier and does kinda help you wind down a bit at night.

The trick is to install Flux on your computer, which adjusts the color temperature of your screen, after dark, to match your lights. Makes quite a difference, but don't forget to disable it if/when you're retouching photos. Otherwise, you'll end up with spacey colors...

Neat!

Have you tried halogen desk lamps where needed? My computer workshop has a main ceiling light I fitted with an 18W 6500K daylight spectrum buld a few years ago, which produces a VERY white light. It's almost too bright to look directly at which in my experience is unusual for a CFL. But it does light up the benches well, and for reading lots of paperwork or working on finely detailed stuff is very good.

I added a 50W 5 foot 6500K fluorescent buld on retracting hangers over the electronics bench last year, which makes working on the really fine pitch stuff much easier, it's like having full sunlight on it all the time. But it does have a similar effect to what you're talking about, which is that the bright bits are REALLY bright, and everything else looks rather dark by comparison. Effectively I end up with a very high contrast room.

If I'm just sitting at the computer and don't need that amount of light, I use a few 20W gooseneck halogen lamps in strategic positions and turn off the main lights. Much easier on the eyes when you don't need clinical lighting. They're around the same colour temperature of those GE bulbs, last a very long time if you don't drop the lamp while it's on and are really cheap.

pca

I'm not too familiar with the Reveals other than seeing them on the shelves with their odd blue tint, but it sounds to me like full spectrum is a marketing ploy anyway since the whole visible spectrum (and a lot of invisible is already present in even the cheapest bulb). Like others have suggested, you just need to try a different color temperature of CFL that more closely matches the 'color' of the white light from your Reveals. Maybe you'd like one of these.

As for the cover, I assume you mean something to diffuse the light more? The Reveals you like have no diffusion but CFLs are diffuse by design due to their phosphor coating on the glass, but GE does supposedly make some CFLs inside a frosted incandescent shell. Maybe that would work?

Stu

If you didn't kill the packaging, you ought to be able to return them for a refund, or exchange. I did the same thing, trying to match the Reveal, but instead of living with the wrong choice for 7 years, I went back to the store, and exchanged the CFLs.

From what I recall, full spectrum is used to imply something similar to "daylight" which represents (or simulates) the colour temperature, the balance/ratios of the amount of output at each point of the spectrum and colour rendering of noon sunlight.

Great for stuff like shopping, including clothing and food, but not the most attractive for the home because of its stark/moodless appearance. They're also good for certain types of animal habitats, including specialty products with decent UVB output.

If you look at a graph of the output spectrum of various lamps you'll see dramatic differences, including the miniscule amount of output at certain wavelengths that would "logically" disqualify a particular lamp from being called "full spectrum."

If you look closely at CFL bulbs, many have a "CRI" rating. A CRI of 100 implies that you're getting the same kind of "full spectrum" that you get from the sun, which is to say that you're getting a nice smooth curve if you plot energy against wavelength. Incadescent/halogen bulbs tend to have high CRIs, while florescent bulbs tend to have lower CRIs. The latter improve their CRIs with funky coatings that absorb light and emit it at different wavelengths.

(One place that this matters, in practice, is that your clothes may look different in your closet than they look when you walk outside. Two things that matched suddenly don't match, since different textiles respond differently to different wavelengths of light. Not an issue for me, but would really piss off my wife.)

The *lowest* CRI you could get, I suppose, would be to use a video projector sending out an all-white signal, since it's really just sending out a light spectrum with three spikes for pure red, green, and blue and nothing in between. Sony makes a funky screen which takes advantage of this, reflecting the specific projector wavelengths and absorbing the rest, thus giving you deeper blacks.

Have you seen those ceiling fans?

Install a CFL on the tip of each blade and then turn on the fan.

You may need some slight adjustment to the motor to achieve sufficient speed for the desired doppler shift...

Anyone care to calculate the fan speed required to give a full spectrum from a single frequency LED.

/me goes to the patent office...

Any ceiling fan that could achieve the relativistic speeds necessary to achieve doppler shifts of the light... well you'll have some issues with tidal forces on the blades, among other technicalities.

If you can get the fan to go at relativistic speeds while carrying a current, you don't need the CFLs in the first place.

Peter

It's not as simple as that. When I tried this, long before reaching relativistic speeds the tips of the blades went supersonic, causing a cavitation effect that created an amazing amount of noise. Then, before I could turn off the fan, the suction or propeller effect ripped the whole fan assembly right out of the ceiling (it was set to Winter rotation) and when the fan hit the floor the housing shattered, releasing the fan blades (which were still spinnning madly by inertia, despite the wires to the fan motor having been severed in the fall), one of which narrowly missed hitting my head before embedding itself in the wall. Another smashed through the living room window, while a third destroyed an armoire with SWMBO's best crystal in it.

All in all, an expensive experiment. As much as I hated to give up the idea, I ended up having to install a conventional light fixture and call it good.

tanstaafl.

According to Wikipedia (the internet's best source for plausible information):

So, assuming that the full spectrum of light is 670nm wide, and you have a light source emitting a frequency in the dead center of that range (565nm), (that I've done my math right and at 1c a 565nm wave is 530.6THz and 750nm is 399.73THz,) and this doppler shift calculator is accurate, the fan blade's tip would have to have a maximum relative speed of about 98.15e+6 m/s, or a little less than one-third c.

Not sure it's as simple as that -- that page says "non-relativistic" at the top, and a real doppler fan at 0.3c would be experiencing time dilation at the tips relative to the observer, so the emitted frequency in the observer's frame would be changed.

Another consideration, along the lines of tanstaafl's, would be that the energy taken to spin up a fan to that rate is of order 10^16 J (probably more as that's another non-relativistic calculation), or the difference in energy use by a CFL compared to an incandescent over 6.3 million years, so you'd have to leave it turned on a long time before it paid for itself.

And if tanstaafl had suffered catastrophic fan dissection at 0.3c, it would have essentially dumped all that energy into explosive destruction -- an amount equivalent to about 2 megatons of TNT.

Peter

Yeah, I noticed that, but then had spent enough time goofing off that I didn't want to figure out what z was in the relativistic calculator. But now you've shamed me into it. Turns out:

z = (f_emitted / f_observed) - 1

So if we're still wanting to get from 530.6THz (emitted) to 399.73THz (observed), we want to get a z of about 0.33. (This is very close to the value in c of the speed I got in my last calculation. Hm.)

And the calculator shows us about 82.7e+6 m/s (0.276c).

I love this place.

I've been using it for a while and it works as advertised.

Good show on their part for having the color temperature ramp slowly so you don't get whiplash.

Bad show on their part for thinking that moving many hundreds of degrees Kelvin in about 20 seconds is anything approaching slow enough. It should approximate the amount of time of twilight, or maybe double that, starting when the sun starts to go below the horizon. (Seems like there ought to be a technical word for that, but I can't find it.)

I used flux for a couple of days, but the fact that my cursor stayed bright white was a bit distracting.

You can solve that problem by buying a Mac. This is probably where that Mac tax that everyone talks about goes.

And gain the problem that it pops out of flux-mode when switching to and from the screensaver.

I imagine that the Windows cursor problem is a video driver issue.

Isn't it a hardware aka sprite based cursor?

End of Evening Nautical Twilight (EENT for short). Well, aviators call it EENT, there are actually three types of twilight: Civil, Nautical, and Astronomical.

You can get more information from the Navy on the different twilights.

Yes, but none of those name the time Bitt is talking about, which is the time the lower edge of the sun touches the horizon. The earliest-in-the-evening of those is "sunset", which is the time the upper edge of the sun touches the horizon, somewhat after the Bitt moment.

Peter

Regardless, there is a (vague) point at which the light provided by the sun starts to change in color temperature (which I assume to be at or near the time that the sun appears to start descending below the horizon), and another (vague) point at which the direct light of the sun becomes insignificant as compared to other sources of light. Flux's time of change should be comparable in orders of magnitude to the amount of time between those two instances.

The length of official twilight can be approximated at half an hour (though it varies greatly based on time of year and latitude). I'd be perfectly happy if Flux approximated even that. But it's literally like 20 seconds.

Actually I suppose what you're really talking about here is sunset redness, which depends on the length of the in-air path of the sunlight, so it actually starts increasing very slowly as soon as the sun leaves the zenith, and might already be changing significantly before the sun's disk hits the horizon.

Peter

It has a 60 minute option for transition.

I, too, find the cursor a bit off-putting, though I do like it in concept.

-jk

The Right Answer, of course, is for flux to work out the Earth's orbital parameters, and thus the sun angles, using the JPL ephemeris. And you'll be pleased to hear that the JPL page describing various revisions of this calculation refers to them as "ephemerides".

Peter

The Mac version doesn't seem to.

Very odd. I installed flux on my Mac and at first didn't like the effect (the screen went yellow) so I hit the 'Disable for an hour' and didn't think anything of it.

So, a couple of days later I remember I installed it, the screen looked 'normal' so I assumed it was still disabled, but no, it was actually running and upon hitting the disable button again the screen went 'blue', so I re-enabled it. I actually really like the effect. It does still look a little yellow to my eyes, but in a nice 'warm' way.

Yeah, I think it goes a little too far into yellow, but it's close. It would be nice if it were more configurable.

I find I use Flux with one notch less color shift than the lights I'm using. So, with real-world incandescents, I set Flux to halogen, and so forth. It's not perfect, but it's certainly a big improvement over leaving the monitor in daylight mode.

Flux was just updated with an option to have the color temp change over an hour. Yay! Now I just have to wait for nightfall.

Just giving flux a try now on my laptop. Will see how it goes over the next few days...