I found this simple detector circuit:

It use the solar itself as the sensor, so no opamp needed

I haven't tried it personally yet though...

 

The solar panel can turn it on and off, but it will need a more powerful transistor to turn 40 LED's on and off.

The Op Amp was just for monitoring battery discharge voltage. May not want that.

 

Dug back to find some data on 15 W SP, not all from same day, some high clouds but enough to give some idea of battery charging capability. Top plot is peak power vs time of day, bottom current available above 14 V.
Average Ah for an 8h day is 5.48, which does not account for charging efficiency, latitude, or weather.
As for current limiting R , just a consensus from AAC? ; but would suggest ballancing the strings for = forward voltage with or with out R's

 

Thanks for the data, Bernard - that's very useful for a number of projects.

That "LED wizard" is actually overly simplistic; if you tell it that your LEDs have a Vf of 3v and you want to power them using 12v, it will generate an array of 4 by n LEDs with 1 Ohm resistors, every time, no matter how inappropriate that is. It would be rather wasteful to use resistors to drop a great deal of voltage, but you'd wind up with more uniform brightness from minimum to maximum voltage output. The best option would be a constant current supply. Burger2227's suggestion of leaving the 1 Ohm resistor out and just going with the four LEDs in series is not a good one; even though the original schematic was ill considered, leaving the resistor out made a bad situation even worse.

The link Burger2227 provided to the ten 5252F IC's with coils - I can't seem to find a manufacturer's datasheet regarding these ICs, and the directions for use provided in the E-bay listing are not exhaustive. At this point in my life, I like to eliminate as much guesswork as possible, as guessing frequently leads to errors, and errors frequently results in burned-up parts.

 

Thanks wayneh, had a look at them, but they look like mains voltage units so not ideal for me. However it spurred me on looking and found a solar charge controller that also has a timer control, so i guess it uses the solar cell as a photosensor like AfdhalAtiffTan suggested building, thanks for that AfdhalAtiffTan.

Thanks burger, had a look at op-amps, not sure if i understand them correctly yet, might just leave them at the moment.

Thanks for that bernard, great info! Looks like you are on the 33rd parallel north, and i am up at 56 degrees north in scotland, so not sure if my results would be as good! But thanks for that, gives me an idea of what to expect anyway.

Sgtwookie, thanks for going over that, now I understand why it isnt a great setup suggested, still i am not sure what to do now! I missed burgers post on the 5252f ic coils, will have a look at what they do tomorrow as was working late out of time tonight!

 

Interesting project. I was thinking about doing something of the like using glass blocks. But designed in a way that it would come on with a motion sensor in the middle and the blocks would light up gradually and sequentially depending on which way the person is coming from. Some PIR sensors you can buy have the ability to tell which direction the person is approaching from and this is what I was going to design around. Anyway, I'll read more later and hope to see how you made out.

 

Burger2227's suggestion of leaving the 1 Ohm resistor out and just going with the four LEDs in series is not a good one; even though the original schematic was ill considered, leaving the resistor out made a bad situation even worse.

The link Burger2227 provided to the ten 5252F IC's with coils - I can't seem to find a manufacturer's datasheet regarding these ICs, and the directions for use provided in the E-bay listing are not exhaustive. At this point in my life, I like to eliminate as much guesswork as possible, as guessing frequently leads to errors, and errors frequently results in burned-up parts.

What is 1 ohm going to do for the LED'S. Nothing!

That link includes pictures of schematics I found after a lot of searching. The chips came directly from China so a 1/3 of the cost was shipping from the manufacturer. The other schematics were designed by me. Unfortunately there isn't much information on the internals, but the chips are very flexable. They will last longer than the LED's!

Limits are 5 volts and 90 ma for the LED's. I have used capacitor trickle charge DC voltage directly from AC line to keep power failure light circuits charged and off without any problems. The charging voltage should be close to the battery voltage used. The LED's will turn on when Pin 1 voltage drops below .4 volts. I have not tested discharge voltage cutoff for more than one battery. One battery cutoff is .9 volts. I'd assume that would be true with 2 batteries also.

Since it is a Joule thief oscillator, it will boost AC voltage to the LED's to double the battery voltage. That's why 3 volt LED's work with one battery. If anybody would like to build an LED circuit with transistors doing the same thing, I have that information too. The chip saves a lot of components including the internal Shottkey diode from Pin 1 to 2 (+ battery).