Re: "Colorimeter"

On 22/05/2025 18:34, Don Y wrote:

On 5/18/2025 4:22 AM, Theo wrote:

Don Y <blockedofcourse@foo.invalid> wrote:

Not quite, but, close enough...
>
How can I determine the spectrum of incident light on a sensor,
in general?  Then, how many corners can I cut to sacrifice resolution
and accuracy?

>
How broad and how much resolution?

 I've not been read into those details.  It's unclear if they are *known*
or being withheld (as they know I'm not interested in any more work; why
disclose information that you don't have to?)
 

 There are sensors, eg:
https://ams-osram.com/products/sensor-solutions/ambient-light-color-spectral-proximity-sensors

 Thanks, I will pass that along.
 

versions of which can be found in cheap dev boards:
https://shop.pimoroni.com/products/as7343-breakout?variant=41694602526803
>
I'm sure I remember reading recently of a consumer grade multispectral
camera part with a moderate resolution (something like 8x8 or 32x32) but I
can't find a reference to it now.  But it seems there's a phone launching
with such a camera soon (according to rumours):
https://www.gizmochina.com/2025/05/13/huawei-nova-14-series-to-launch-in-may-with-harmonyos-5-and-an-ultra-model-specs-here/

 That would target a different application (IMO).  As presented to me, they're
just looking for characterizing the light falling on a *spot*.  If
different (e.g., bandwidth sensitive) detectors were employed AT that
spot, I would assume they would have to be treated as a single point
(despite there being some obvious separations involved in their manufacture)
 I.e., almost like a photographer's "light meter" but with the interest
being on the spectral content and not the overall intensity.
 [Whether this is true or not, it has influenced how *I* have thought
about the problem -- in terms of function, size, portability, power
requirements, etc.  Assumptions are always the bane of a good design...  :< ]

If the light levels are very high then you can get LEDs with emission profiles of about 50nm width. They work as sensors in the opposite direction with some leakage for higher energy photons. Choose them wisely and calibrate against a reference white and you might have something that is both cheap accurate and durable.
--
Martin Brown