Maker Pro
Maker Pro

converting lumens to W/m^2

  • Thread starter Nobody Important
  • Start date
N

Nobody Important

Jan 1, 1970
0
I need to make a light that will give a particular power density at a
particular distance, say 1 mW/cm^2 at 50 cm. The LEDs I'm considering
all have luminous flux specs in lumens.

How does one convert from lumens to mW/cm^2 for point sources at 470 nm?
 
D

Don Klipstein

Jan 1, 1970
0
Important wrote:
I need to make a light that will give a particular power density at a
particular distance, say 1 mW/cm^2 at 50 cm. The LEDs I'm considering
all have luminous flux specs in lumens.

How does one convert from lumens to mW/cm^2 for point sources at 470 nm?

This needs more data. Such as radiation pattern of the LED and luminous
efficacy of the emitted light.

Now for some guesses/assumptions that I dare to make:

1. That the LEDs are Lumileds Luxeons or similar, of the "lambertian"
radiation pattern. (They also come in other radiation patterns that make
this even more difficult.)

2. I am "guesstimating" that 1 watt of "470 nm" blue LED light is
typically 70 lumens.

So, a nice feature of the lambertian radiation pattern: Candela is
lumens times pi times the cosine of the angle from the LED axis (and zero
where the cosine is negative - as in angles to the rear and diagonally to
the rear of the light source).
Straight ahead, candela is lumens times pi.

This is assuming that the "lambertian" LED has a perfectly lambertian
radiation pattern - which is a big "if".
I have some experience with Cree XLamps, and they are slightly more
directional than truly lambertian in my experience and according to the
datasheet last time I looked. I would guess for those candela on-axis
about 3.5 times the lumens. I have also seen some radiation patterns for
some lambertian Luxeons in their datasheets that are not perfectly
lambertian also.

Divide candela by the square of the distance to get lumens per square
unit. Divide candela by 2500 to get lumens per square centimeter at 50
cm.

If you decide to measure this instead of calculating this, be aware of
some quirks of some to many light meters:

1. The light meter could have a spectral response very different from
that of the human eye in one or another or a couple or a few small ranges
of wavelengths - possibly heavily coinciding with the wavelengths emitted
by the LED.

2. The light meter could have spectral response significantly high or low
at many wavelengths of lower photopic function (lumens per emitted watt
small compared to 683), since this causes only minor errors in photometric
measurement of "most white light".

Divide the lumens per area by "luminous efficacy of the emitted light"
(which I guesstimate to be 70 for typical of "470 nm blue LEDs in
general") to get watts per unit area.

===================================================

Keep in mind the significant tolerance in light output of most LEDs -
even ones known to be sorted into a specific brightness ranking!

===================================================

Doing this will get you a rough answer - maybe very rough. For now, I
would dare to expect this to get you reasonably expectedly (not quite
guaranteed even this accurately - HA!) an answer between half and double
the actual radiation intensity. If you need better than that - get a
radiometer.

- Don Klipstein ([email protected])
 
Top