using a LED as a light sensor

I once built a high input impedance amplifier and used two

1N914/1N4148 (or similar diode with transparent case) to limit the input voltage range between Gnd and Vcc. I noticed sudden voltage fluctuations when working with the circuit and suspect some oscillation problems.

It took a while, before I realised that the fluctuation occurred, when my finger was close to one of the diode, shading the light from the lamp. The illuminated diode generated a larger current, which caused the fluctuation.

Putting black tape on the diodes solved the problem. The moral of the story is to use non-transparent cases in circuits with high impedance levels.

Paul

[snip]

Heat shrink tubing works nicely for such situations.

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
I love to cook with wine.      Sometimes I even put it in the food.

I have written long threads about this in other groups before.

My simple led communications circuit connects a single led to a pair of transistors, one npn, one pnp. To get the most out of the led in receiving mode it is connected directly* to the base of a grounded-emitter npn. This way the led is loaded up which is needed for higher speed coms and the npn is quite happy as it is after all a current amplifier.

For transmitting I run a pnp from the supply down to the led via a current limit resistor. Because this would be clamped by the receiving npn *I use a 10K in the base of the npn which has negligible effect in receive mode. Of course this circuit is short-range up to a few meters and can be affected by direct sunlight but is well suited for indoor use such as a hand-held instrument to machine interface.

The single led coms circuit (I hate trying to do ASCII circuits)

+V +V | | E 10K PNP B ----1K---< TxD | C | | |-----> RXD 100R | | C |--------------10K----- B NPN | E | |

• | LED |

- | | | GND GND

BTW, for daylight and fluro sensing I use clear green leds.

*Peter*

The early GE plastic transistors used a beige translucent epoxy. Just when you had your diffamp all tuned up, microvolts of offset, the boss would step up behind you to check your progress, block the light, and mess it all up.

Fluorescent lights would cause "unexplainable" hum, too.

John

[Snipped]

Download AACircuit from

This is a CAD to draw ASCII circuits.

Regards Anton Erasmus

A very nice package. I have v 1.28.6 here, from about a year ago. Has it changed to any great extent?

--
 Some informative links:
   news:news.announce.newusers
   http://www.geocities.com/nnqweb/
   http://www.catb.org/~esr/faqs/smart-questions.html
   http://www.caliburn.nl/topposting.html
   http://www.netmeister.org/news/learn2quote.html

The latest one still seem to be V1.28.6 beta.

Regards Anton Erasmus

I finally dug out (and through) my notebooks for this data from 2001:

3/14/2001 Short-circuit photocurrent of some random, junk-box devices, 1.5m from a 50w halogen flood lamp:

LEDs green, bright 3mm, Hosfelt 25-302

Confirms about what I measured in my junkbox. Led's are poor light receptors.

greg

True, but not impossibly so it would seem. I was actually quite impressed by those numbers from the "ultrabright" offerings above. Of course these are the ones with a lens that will focus the incoming light from a much larger area than that of the led chip.

--

John Devereux

I didn't measure current, but I was working with a Phillips Lumiled, and needed to get closer to the emitter. We started milling down the top piece, plastic but not really a beam forming lens. Something funny looked in the center. As the center region was cleared, I discovered the flexible inner Sylgard like material under the plastic. It was interesting and screws me up.

greg

Would you please add whether the epoxy on the various LEDs was clear, frosted or dyed a color?

Sure.

Please note that these were *crude* measurements taken five years ago, just to see the range of outputs that might be expected from good & bad photogenerating LEDs.

I wasn't trying for ultimate rigor. The 1.5m lamp spacing, for example, is an estimate. The X-Y-Z location of the device tested in the (uneven) illumination field is critical, and the angles even more so. I chose the combination producing max. photocurrent in each case.

That said, ...

First, let's measure the spacing to the lamp: 1.54m, so 1.5m was not a bad guess.

Next, check our calibration with an identifiable type measured 5 years ago:

red, 2mm, diffused, Dialco 555-2003 (unit #1) 0.13uA red, 2mm, diffused, Dialco 555-2003 (unit #2) 0.11uA

Okay, so we're in the ballpark.

Now that I'm calibrated, I can use my recollection for device-identification wherever possible, and have the old data for sanity-checking / confirmation of ID...

Here's an updated table, with as much info on each LED as I can ascertain, plus a few new entries...

'------------------------------------------- Short-circuit photocurrent of some random, junk-box devices, 1.54m from a 590l, 50w halogen lamp:

LEDs green, 3mm, clear light green tint, bright

Yes, one would think a larger lens, e.g. a 10mm LED, would gather 4x as much light as a 5mm unit. Lower output from 10mm LEDs, then, indicates a less efficient photogenerating LED die.

Another explanation might be variations in focus, but virtually all of the 10mm units are *very* tightly focused, e.g.

Ooops -- both red LEDs above ^^^ were clear-cased.

James Arthur

"BrunoG" wrote in message news:445d31a5$0$19698$ snipped-for-privacy@news.wanadoo.fr...

Very cool.

Unfortunate about the color of the casing filtering incoming light.

Could this detect the signal coming from an IR remote (well enough to decode), I wonder?

Rufus

"Rufus V. Smith" a écrit dans le message de news:

44903395$0$28047$ snipped-for-privacy@news-taz.teranews.com...

Hi, thanks !

It's worth a try, but I don't think so, because the response time of the diode & PIC DAC would be too long.

Bruno

OK, not quite 40kHz territory, but this looks fairly quick:

from here:

Deep.

>

LEDs are decent candidates for RF VFOs (variable frequency oscillators) in place of specialized varaiable capacitance diodes. What we've found is that the LED-based VFO must be kept in a light-shielded enclosure due to the effect described. Not usually a problem, since the VFO should be in an RF-shielded enclosure anyway.