[Laser] Beacon to test a photodetector sensitivity limit

[Art]

I couldn't agree more Terry!

I never tried using a 1N914 diode as a photon source though. I have 
tried only old technology LED's.

But, I always STRONGLY suspected that IR or other wavelength photons 
were being produced by the LED's at low current levels. I did not try 
an IR absorbing filter to confirm this, mainly because I didn't have 
one at the time.

I started out working with red LED's at 10-100 microamps and found 
they totally swamp any sensitive light receiver-too many photons and 
had very very poor linearity! It was VERY BAD. Varying the number of 
photons out by varying the curret to the LED was very un-linear at 
these lower current levels.

 From this point, I decided to use a highly sensitive PM tube for 
setting the output of the LED to a relative level-then switching to 
the DUT to take measurements. Even this didn't work! What I found was 
that the relative output of the photomultiplier didn't correlate to 
the readings taken by the DUT!

I concluded that the LED must be producing near IR-which the 
photodiode was sensitive to and the PM tube wasn't....hence the non 
correlation between the PMT and DUT relative outputs.

The ONLY way I could make this work was to run the LED between 5 and 
30 ma and to use a pinhole in combination with a series of absorbers 
between the pinhole and the PMT/DUT. Using this method, I got 
excellent correlation between the PMT and the DUT outputs. But, the 
LED still wasn't very linear. I was happy enough though because I 
could get a weak signal to evaluate the various photodiode amps I was 
working with.


>>With our  last photodetectors, we noticed it is possible to hear the signal
>>to 1 meter  without lens while it is absolutely not possible to see the LED
>>with  eyes.

Yves, please BEWARE of evaluating a photodiode without a lens! 
Without a lens, that photodiode will pick up photons from almost any 
angle. You need a lens or some method of limiting the angle of 
admittance of the photons entering your detector. Without it, you 
will have scattered photons and very much distorted sensitivity 
data-the detector will seem much more sensitive than it really is.

To test this, go into your dark room and take a reading of the 
ambient light level with a lens in front of the PD, without a lens 
over the PD and with the PD in total darkness (covered with a light 
shield). You will find the ambient light level to be several 10's or 
hundreds of times higher without the lens-because the few photons 
that find your way onto the dark room are bouncing around off of 
every single wall surface in the room and there is allot of 
reflecting area! Even with the lens in your dark room, you will 
probably detect a small amount of light compared to the readings 
obtained with the covered photodiode.

The only way I found around this problem was to mount the photodiode 
in a black box surrounded by flat black 'flocking' paper and to have 
a pinhole several inches in front of the photodiode-the only light 
allowed to reach the PD was through the pin hole. Even then, the 
electronics had to be totally shielded from light because solder and 
copper etches reflect stray light that enters the pinhole from 
extreme angles. For high sensitivity detectors, the detector 
shielding from RF and EMF also blocks light, so it's not a problem. 
But, do be aware you have to wrap the metal shielding in flocking 
paper to prevent it from reflecting stray photons.

It's a very long story, and I won't bore you with ALL the details. 
The short version is that I eventually migrated to my landlords 
basement in the apartment building to do optical testing. It was in 
the lower basement, 20 feet below ground level and there were no 
windows or way for light to enter the room although it was a large 
open area with lots of junk in the room. I went to this area 
expecting it to be pitch black so I wouldn't have to worry about 
stray light. But, using a photodiode without a lens, it was painfully 
obvious that light was getting in somehow (the ambient light level 
was 5 to 15 millivolts while the dark level (with a cover over the 
PD) was well under 1 millivolt. I started using my receiver as a 
probe to find out where the light was coming from. Basically, light 
was coming from everywhere, even off my hand placed a few inches in 
front of the photodiode! There was absolutely no visible light 
anywhere in the room that I could detect with my eyes. Poking around 
in pitch black was not easy, but I eventually found a peak-coming 
from the furnace room door direction. In this room, there was an LED 
on the furnace. I assumed this to be the source of my light even 
though it was in the next room with the door closed. Even with this 
LED completely covered, and the door to the furnace room closed and 
taped over, I still got residual light levels in the basement area 
until I put a pinhole over the PD to minimize the admittance angle 
that of the PD's view.

With the lesson learned here, I eventually was able to build a test 
setup that allowed me to work in my kitchen (at night with the lights 
turned off). It was much more convenient this way!

I ended up with a medium sized cardboard box lined with flocking 
paper with a shoe box lined with flocking paper inside of it. All the 
optics were inside the shoe box. My light source was an LED covered 
in rubber cement and painted black except for a small pinhole to let 
some of the LED's light out.

Between the light source and the PD was anywhere from 1 to 5 pieces 
of arc welding light absorbing glass with a pin hole on the PD side. 
In order for it to work, it was necessary to wrap the arc welding 
absorbers in flocking paper in order to keep the reflections from the 
glass surface from contaminating the readings (by reflecting the 
little bit of stray light inside the enclosure).

Eventually, I just left the LED running at 5 ma and varied the number 
of absorbers between the LED and the PD to get my weak signals. It 
was all 'relative', but it was cheap, and it worked well.

Regards,

Art