Led turning on at dark with green led as sensor ?

[Externet]

Am doing something wrong...

Just because I have hundreds of green leds that output 1.8V under sunlight; and hundreds of pnp 2N3906 transistors, I wanted to make many circuits turning on white leds at night.

Connected the green 'sensing' led between base and emitter to reverse bias the pnp transistor during daylight but it does not turn off. (anode to base).
20mA white led is connected to collector, with 51Ω in series to minus 5V.

Tried several new transistors, no way. Can you spot where am goofing ?

 

[ronsimpson]

Hopefully no (little) light is going from white LED to green LED.

First short out B to E and see if the white LED goes out. OR remove the green LED and see if the white LED goes out.

If the above works then maybe the green makes some current in the dark and you need to put a B-E resistor on the transistors. 100k??

 

[unclejed613]

hard to tell without a schematic.... there might be a better way to do what you are doing.

 

[MikeMl]

If the LED turns off the transistor, what turns it on?

 

[ronsimpson]

If the LED turns off the transistor, what turns it on?

I think he said the transistor is on all the time. Maybe he has the pin-out wrong on the transistor.

 

[schmitt trigger]

Indeed, a schematic is required.
He may be wiring things correctly, but the component values are incorrect.
The polarities may be transposed.
etc, etc.

 

[Externet]

Thanks.
The schematic is very simple :

5V(-)-----------470KΩ-----------------------------B--------------------green|>|-------------(+)
5V(-)----------51Ω-----------|<|white-----------C E------------------------------------------(+)

 

[ronsimpson]

You have a 470k resistor from B to 5v(-). That is what turns on the transistor. I don't see the green LED does anything.

 

[Externet]

Thanks.
Correct, the green led does nothing while dark; but generates 1.8 V at daylight with the intention to reverse the 470KΩ bias.

Would yo acomplish it in another way ?

 

[schmitt trigger]

You have a 470k resistor from B to 5v(-). That is what turns on the transistor. I don't see the green LED does anything.

From what I can gather from this circuit, is that he is using the LED in the photovoltaic mode, and applying an opposite bias to the one provided by the 470k resistor (which as you correctly mention, is biasing on the transistor)

I've read about green LEDs to being photovoltaic, but never have experimented with it myself...so take the next suggestion with a grain of salt:
Should the LED be reversed? Maybe the photovoltaic polarity is opposite.

 

[Externet]

Thanks.
The photovoltaic green led generates with + at its anode. Shorting base and emitter does shut the white led off, as supposed to.

 

[alec_t]

the green led does nothing while dark; but generates 1.8 V at daylight

1.8V is the typical Vf of a green diode. I've just tried a variety of LEDs. Best I could get in photovoltaic mode was ~ 0.17V. That happened to be a green one . But I think it's unlikely you'll be able to use the green LED to get your circuit working in PV mode. Might be better to reconfigure it and use photo-current as the transistor base current, e.g. put the green led and 47k (or lower value) in series between base and ground. That would mean, of course, that the white LED could only light if the green LED was illuminated. If that's the opposite of what you want then you would need a second transistor to switch the white LED.

 

[Externet]

Generation reaches over 1.8 V only at direct sunlight exposure; day room light levels generate less than half a volt as you say.
Perhaps loaded it is not enough to bring the 0.7V base treshold while fighting against the bias.

Well, the idea of putting to use hundreds of those green leds will not be simplistic for turning on a white led at dark; ...back to the drawing board

 

[schmitt trigger]

Alec;
since the OP has stated that he has hundreds of LEDs, could he put several in series to achieve the deisred back-bias voltage? Four of them would provide about 0.68V, which should be sufficent to deprive the PNP transistor's base of the turn -on bias.

 

[Externet]

Thanks.
I like that !

Will be back with results of 4 in series...

Edited... am back : 5 in series did not do it with ambient room light. Will try later that clouds move away on full sunlight when shines by my west window.

 

[alec_t]

I think part of the problem is that, even if the green LED generates a decent open-circuit voltage, connecting a load to it immediately drops the voltage. I was using a DMM (resistance ~10 meg?) to check the LED voltages.

 

[tvtech]

I think part of the problem is that, even if the green LED generates a decent open-circuit voltage, connecting a load to it immediately drops the voltage. I was using a DMM (resistance ~10 meg?) to check the LED voltages.

Hi alec

You probably have nailed the problem.

Put them in parallel rather. More current

Regards,
tvtech

 

[MikeMl]

The voltage produced by the LED in photovoltaic mode is capable of turning on a FET, due to its high gate impedance, but not likely to produce enough current turn on a transistor.

 

[ronsimpson]

I read several application notes on using LEDs to detect light. One uses a PIC to measure the voltage. My memory is bad but I remember they switched on/off the load at about 0.1 or 0.05 volts. They only put a 1uA load on the LED.

 

[audioguru]

An old dim green LED was in a green case and had a forward voltage of about 2V.
A new very bright green LED is in a clear case and has a forward voltage of about 3.5V. They are completely different.

I doubt there is much green light from the sun so the old green LED will be poor as a photovoltaic generator.
But the new green LED allows its clear case to pickup IR, all colors plus UV.

Due to the very small size of the chip in an ordinary LED then its photovoltaic output current is very low.