Path: utzoo!utgpu!water!watmath!watcgl!awpaeth
From: awpaeth@watcgl.waterloo.edu (Alan W. Paeth)
Newsgroups: sci.electronics
Subject: Re: Seeing IR
Keywords: visible, IR, LED, laser
Message-ID: <4227@watcgl.waterloo.edu>
Date: 2 May 88 14:45:35 GMT
References: <2871@jumbo.dec.com>
Reply-To: awpaeth@watcgl.waterloo.edu (Alan W. Paeth)
Distribution: na
Organization: U. of Waterloo, Ontario
Lines: 40

In article <2871@jumbo.dec.com> murray@jumbo.dec.com (Hal Murray) writes:
>Since the cluster of messages a few weeks ago, I've been thinking,
>browsing, and asking my friends, trying to understand what it takes to
>"see" an IR source. I know it happens ...
>...Does anybody have a reference to good info about the spectral response of
>the eye? The best chart I found didn't label the vertical axis. It showed



The photopic response of cones is good to nearly .9-1.0nm in the red, but
decreases exponentially beyond about 7000A, much like the roll-off of a simple
(one pole) filter. So with enough power, near I.R. is quite possible. 
You'd see this color as "red" -- the rods are out of the running; they become
less sensitive to cones at about .68uM.

I've stared straight into the beams of consumer (TV/VCR remote control) IR
LEDs in a darkroom with my eye fully adapted and can just make out the light
at the threshold, but not for all IR LEDS (I think that the straight GaAs
versions are too long in wavelength, but some come much closer to visible).

I saw some empirical charts on this stuff while at Xerox Electro-Optical
(they have application in IR sniperscope illuminators and such -- the military
is interested in detection of sources). We all felt that the data points were
*presumably* from Rhesus monkeys and not prison convicts, because to validate
the exponential model the beam powers stated were *VERY* high out near 1nm.

Most tables (like CIE color charts) deal with con sensitivity as linear values,
allowing color characterizations using linear models. I think I have a chart...
Let's see, here is a table in "Modern Optical Engineering by Smith (McGraw
Hill). Calling peak cone relative sensitivity "1" at .55u, we have 10^-3 at
.72u, 10^-5 at .8, 10^-8 at .9, 10^-10 at 1.0, and 3*10^-12 at 1.1um (end
of chart), all relative values.

The light source in a typical CD is is a laser diode -- typically 0.4 mW at 
30 nm wavelength. That makes it a relatively low power Class II laser,
but then, when do you know to blink?

   /Alan Paeth
   Computer Graphics Laboratory
   University of Waterloo