High CRI E27 LED bulbs

10 views
Skip to first unread message

Grumps

unread,
Sep 27, 2021, 10:08:34 AMSep 27
to
I'm getting confused!
FIL wants an expensive reading lamp, but surely the only (main)
difference is the LED being used compared to cheap lights.

I can find high CRI bulbs but they have low colour temperatures.
How can a CRI bulb of >95 only have a colour temperature of 2700K? Isn't
daylight nearer to 6500K than the warm 2700K?
Or am I being confused about colour rendition and temperature?
Ta.

Sysadmin

unread,
Sep 27, 2021, 10:19:03 AMSep 27
to
I use these, but £15 for 4 on Amazon:
Sauglae LED Corn Bulbs,

B22 Bayonet Cap, 12W, 1200Lm, 6000K Daylight White, Pack of 4

They are very bright.

Roger Hayter

unread,
Sep 27, 2021, 10:19:36 AMSep 27
to
To an extent you are. The bulb you describe should have good and accurate
colour discrination, but colours should look like they would in tungsten
incandesceent light, not as they would in daylight. This might be an advantage
if the user wants restful articial light as in past days, and less good if he
wants to do accurate colour matching standardised on daylight, as might be
needed for some commercial or artistic purposes.


--
Roger Hayter

Theo

unread,
Sep 27, 2021, 12:57:45 PMSep 27
to
Roger Hayter <ro...@hayter.org> wrote:
> To an extent you are. The bulb you describe should have good and accurate
> colour discrination, but colours should look like they would in tungsten
> incandesceent light, not as they would in daylight. This might be an advantage
> if the user wants restful articial light as in past days, and less good if he
> wants to do accurate colour matching standardised on daylight, as might be
> needed for some commercial or artistic purposes.

AIUI you can have a 'white' bulb that is a black body radiator
that has a full spectrum, and a 'white' bulb that has three sharp R/G/B
spectral lines. It is the smoothness of the spectrum that establishes CRI,
while the position of the spectral lines sets the colour temperature.

I don't know how good the LED phosphors are these days, as to whether you
need a special CRI bulb or just one with a regular phosphor is good enough.

Should probably find a prism to try it out...

Theo

Brian Gaff (Sofa)

unread,
Sep 28, 2021, 4:48:45 AMSep 28
to
OK, many so called white lights only have peaks to simulate the light to
appear white to our easily foxed eyesight. But they can cause colours to
appear different and some eye tiredness effects. If you fill in the gaps in
the spectrum to create something similar to the true spectrum of sunlight,
you are going to use more power in doing so, so effectively they seem less
efficient. That was how it was explained to me. Others may know more.
Brian

--

This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
bri...@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!
"Grumps" <n...@here.com> wrote in message news:sisj8u$q14$1...@dont-email.me...

John Rumm

unread,
Sep 28, 2021, 6:23:29 AMSep 28
to
Partly perhaps... the CRI is tied to the spectrum produced by the lamp
rather than its overall colour temperature.

Most "white" LEDs are actually blue LEDs coated in a fluorescent
phosphor to convert the light from the LED into something approximating
white light.

This tends to mean you don't get a continuos spectrum, but one with
various peaks and gaps. So the more even the intensity across the
spectrum, the better the CRI.

The colour temperature will also affect the appearance of colours - but
if the LED is a good mimic of tungsten for example, then things should
look much the same as they do under incandescent light. A good daylight
lamp, should mimic daylight (if its bright enough).

(you can still get some odd effects with some fabric dyes that also
include fluresscent material, since they may look different in "real"
daylight since they emit light of their when stimulated by the UV in
sunlight).

Nice comparison slide here:

https://www.researchgate.net/figure/Emission-spectra-of-different-light-sources-a-incandescent-tungsten-light-bulb-b_fig1_312320039

A (filament lamp) shows a perfect black body emission spectra - but not
much energy at the blue end.

G, H, I show daylight at various times of day. The notches and non
linearities caused by atmospheric absorption and scattering.

E is a blue LED on its own
D is a blue LED with a "white light" phosphor coating, actually not a
bad representation, but a bit lacking in red, and too much blue, hence
the familiar slight green cast.

--
Cheers,

John.

/=================================================================\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\=================================================================/

Grumps

unread,
Sep 28, 2021, 6:30:42 PMSep 28
to
Thanks for the link.
If G is daylight, this is normally described as a colour temp of 6500K.
From the spectrum of G, how could you work out the colour temp?
And if you take a high CRI LED lamp, say 98, the spectrum also looks
similar to G, but the colour temp is given as 3000K. How does that work?
If you have a dimmable lamp, does the colour temp vary with brightness?

I am prepared for the situation where I will never understand this!

John Rumm

unread,
Sep 29, 2021, 9:35:06 AMSep 29
to
Yup - it has lots of energy well up into the blue end of the spectrum,
which pushes the average colour temperature up.

(the colour temperature being just the temperature you would need to
heat a perfect black body radiator to so as to get a similar light
spectra).

> From the spectrum of G, how could you work out the colour temp?

There is a thing called Wien's "displacement" law (an approximation of
what was later refined by Planck) that relates the colour temp to a peak
wavelength. It's given by :

lambda = b / T

where lambda is the wavelength, and b is a constant ~ 2.9 x -10^6 m.K,
and T is the colour temp in Kelvins.

So a T of 6500K gives you a peak wavelength of 2,900,000 / 6500 = 446 nm

It basically shows how the whole spectra shifts in peak energy toward
the blue at ever higher black body temperatures.

(it fits in with our normal expectation of how things heat first to a
dull read, then orange, then white, right through to the blue of a plasma)

There is a nice graph on the page her:

https://en.wikipedia.org/wiki/Wien%27s_displacement_law


> And if you take a high CRI LED lamp, say 98, the spectrum also looks
> similar to G, but the colour temp is given as 3000K. How does that work?

Usually there will be less short wavelength content, see the A filament
spectra. However you can also "fake it" to an extent by boosting the red
/ yellow content of light that does also contain more blue - as is often
the case with "while" LEDs that use a blue LED light source to excite a
phosphor.

> If you have a dimmable lamp, does the colour temp vary with brightness?

With a tungsten lamp, yes - very much so. With a normal dimmable LED,
then no - it stays the same at all brightnesses and looks a bit odd
because of it (our expectation is that higher colour temperatures are
usually experienced at much higher intensities - so daylight looks
natural when very bright, but starts to feel "cold" when experienced at
lower intensities (hence why the light looks "colder" on an overcast
day, or why a single LED daylight bulb can look very cold or blue (but
lots of them together seems quite "sunny")).

Some posher dimmable LEDs have more than one chip - one having a very
yellow filter, and they blend the output of both in varying amounts
depending on brightness. These do dim in a way similar to an
incandescent lamp, with lower levels appearing more "warm" in tone (i.e.
lower colour temperatures)


Also keep in mind that the term colour temperature works in the opposite
direction to the terms we normally use to classify light qualities. i.e.
we say "warm white" meaning a *lower* colour temp, and might describe a
very high colour temperature lamp as being very "cold"!
Reply all
Reply to author
Forward
0 new messages