Led turning on at dark with green led as sensor ?

[tvtech]

Hi alec

You probably have nailed the problem.

Put them in parallel rather. More current

Regards,
tvtech

I hate guessing. That's what I did here.

Till next time

tvtech

 

[Externet]

One green led at full sunlight produced 1.8V unloaded.
Five green leds in series at full sunlight produced 1.8 V unloaded.
Do not understand. Behavior is not like photovoltaic.

Five green leds in parallel to base/emitter, produce 1.8V at full sunlight; does not overcome the 470KΩ bias; still on.

Will be unable to find the data sheet for these; are brutally bright. Removed from these signs :

**broken link removed**

Now trying with mosfet and later with jfet. Will be back at some point. Meanwhile, suggestions are welcome.

 

[Externet]

OK. Done. Works.

2N7000 N-channel enhancement mosfet, as I have 1200 of these

(+5V)---------51Ω----------|>|white-----------D S---------------------------------(-)
(+5V)---------20MΩ------------------------------G-----------5green|<|------------(-)

 

[audioguru]

Please draw a normal schematic with MS Paint or something.
Your horrible ASCII scribbles are awful.

The base-emitter voltage of a turned on 2N3906 is 0.7V. If your LEDs produce 1.8V unloaded then maybe thousands of them will produce enough current to cancel the 9.2uA from the 470k base to ground resistor.

Guess what? An LED makes a very weak photovoltaic device. Use a photocell or photo-resistor circuit instead.

 

[Jaguarjoe]

A few years ago there was a thread in the Picaxe forum about a dude who wired up many LED and 8 pin uC's. His program would utilize the LED as a sensor until it "saw the light" then it would blink the LED. With many of them spread about in a darkened room, they would "talk" to each other in a random but recognizable pattern.
He used a red LED directly into the chip's ADC pin.
The Picaxe ADC works OK, however, it's not stellar. If a bare LED can provide a useful signal to that uC, it should work with almost anything.