Related
Recommended

Solar Garden Light

jc2048

A couple of days ago, I was in a store here in the UK (in the garden department looking at seeds and bulbs) when some 'solar lights' caught my eye. The very cheap

ones didn't look up to much, but the £2 ones had a decent solar cell on them so I bought one. I was curious as to how something with a CE sticker saying it had

a 1.2V battery inside was able to light a white LED that would need a forward voltage of around 3V to illuminate.

 

Here are some photographs as I disassembled it.

 

 

 

There's not a lot to it. The battery is a 1.2V Ni-MH cell. What at first glance might appear a transistor is actually an integrated circuit (three legs good,

four legs better!) and there's a single passive component that you might, at first glance, mistake for a resistor.

 

The cell voltage was very low and it would only power the LED for a few seconds, so I put it on a window sill in the sun for a day. By the evening, the cell was

charged and the terminal voltage (off load) was up to 1.32V.

 

As you might expect, to get the 1.3V up to the 3V needed to power the LED requires a switching circuit and the device that looks like a resistor is actually a choke (coil).

 

Here's a schematic

 

 

Here are some waveforms. The yellow trace is the connection between the coil and the LED. The blue trace is the current through the LED. A switch to ground in

the IC waits for the coil to ramp up to about 18mA, then lets it go, allowing it to run into the LED. The average LED current looks to be about 4mA overall.

 

 

If I calculate (roughly) the coil value from the waveforms

 

L = V x dT / dI = 1.32V x 2.4uS / 18mA = 176uH.

 

That agrees with the coil markings: brown-grey-brown would be 180.

 

I haven't looked at the solar cell and the charging side of things as of yet.

Anonymous

  • Very interesting.

    I keep my eye out for bargin LED Lighting like these but with 2 batteries.

    I have a project in mind, and I'm now wondering if the boost converter can be made to do 3v3 to directly power the device.

     

    Some experimenting is needed .....

     

     

    Mark

  • in reply to fluxcap

    I've just done a very quick test of the solar cell. With it pointing directly at the sun, it manages 2.338V into a 33 ohm load, which is 70.85mA (166mW). If I point it north, away from the sun (bright, with cloud to reflect the sunlight), I get 1.996V, which is 60mA (120mW).

  • in reply to jw0752

    Yes, it's the same IC. I can make out the YX8018, though the lettering is very faint.

     

    Interesting development path - once the quantities get large enough it becomes worth doing a custom chip.

  • in reply to jc2048

    I still have a lot to learn so just in these comments I've bunches about something I thought was so simple. This is a reason I love engineering, it's more than meets the eye!!

  • in reply to jc2048

    Hi John,

    I just finished finding my solar light circuits of which I have two styles. The older one is perhaps 15 years old. Here are pictures of it.

     

     

     

     

    The little 5 legged ic has the number 5438 on it. Besides the IC there is a resistor, diode, inductor, and a capacitor. In this case they are bothering to smooth out the pulse train that is operating at 54 kHz.

     

    Here are the pictures of the newer unit which is perhaps only 3 years old.

     

     

     

     

    I suspect that this one is more like yours. The numbers on the IC are YX8018 and it operates at 227 kHz. They must have moved all the other components onto the chip as the only externals are the inductor 470 uH and the LED just as you described your unit.  Even at this level of simplicity one has to marvel that they can manufacture this unit, package it, ship it to the USA, distribute it to the stores and then sell it for $1.

     

    John