Solar power LED

Discussion in 'The Projects Forum' started by takeitapart, Apr 25, 2011.

  1. takeitapart

    Thread Starter New Member

    Jun 20, 2010
    21
    0
    I have recently purchased some solar garden/walk lights. I would like to increase the light output (lumens). Is it reasonable to change the LED and the resistor without remaking the whole board or do I need to start from scratch? I realize the battery may not last as long per charge. Any Ideas? Thanks in advance.
     
  2. designnut

    Member

    Apr 21, 2011
    33
    1
    I think your way out is more solar lights.
     
  3. mbxs3

    Active Member

    Oct 14, 2009
    141
    3
    You could always take just one of the units, do as you described, and test it out for yourself.
     
  4. Bernard

    AAC Fanatic!

    Aug 7, 2008
    4,173
    397
    Mostly there is no current limiting resistor, but a small boost converter to up V from 1.3 to 3V; will it carry 70 mA??
     
  5. wayneh

    Expert

    Sep 9, 2010
    12,145
    3,056
    The LEDs in these things aren't always the brightest, since they may have been chosen for other reasons, such as low current draw for prolonged brightness. So you MIGHT be able to get a brighter or different colored LED to make a small improvement. You also MIGHT get away with adding a second LED in parallel with the first. This puts the circuit at risk of overload, but they're very cheap to experiment with.

    I don't think you'll get far trying to mod the internal electronics.
     
  6. russpatterson

    Member

    Feb 1, 2010
    351
    16
  7. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    Oddly enough there are LEDs in the same color, some of which may only produce 700 mcd with 20 mA whereas others can put out 7,000 mcd supplied with the same 20 mA. Obviously the less efficient ones are less expensive, that's why you'll see them in things like those walkway lights.

    While there's no easy way to measure the actual light output one could measure the current draw then look for LEDs that put out the most mcd at that same current, swap one and compare.
     
  8. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    A 700mcd LED has a wide beam angle and a 7000mcd one has a lense that focuses the beam into a bright narrow angle. You probably want a wide angle.
    Your vision's sensitivity to brightness is logarithmic so doubling the current or number of LEDs makes the brightness only a little brighter.
     
  9. RRITESH KAKKAR

    Senior Member

    Jun 29, 2010
    2,831
    89
    Very deep discussion, good for me..!!
     
  10. wayneh

    Expert

    Sep 9, 2010
    12,145
    3,056
    It's absolutely true that the OP should consider directing more of the light to the eye of the observer, if that's at all possible for the application. An ordinary LED, if it's the narrow beam type, can be painfully bright if your eye is in the beam. So a simple reflector arrangement can "double" the perceived brightness of a solar light, at the expense of darkening some other area.
     
  11. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,394
    1,606
    I just spent some time with various solar LED lights just trying to get them to work once again after leaving them out for the winter.

    First thing off is to realize these are made with the least expensive components as cheap as possible so as to still make a profit off a $3 retail light. Excluding the lamp assembly I would be hard pressed to just buy the electrical parts for $3, no less make a profit off them.

    Additionally, of the 3 controllers I have pulled apart there seems to be two different controller designs. The simpler one, from a “small” 2 1/4” light looks like this:



    [​IMG]

    It runs at about 310 KHz and drives about a 8 mA average current into the LED. The circuit makes a very clean sine wave at the junction of L2 & C1. The current is controlled by the size of L1 and the period of oscillation. Dropping L1 may let more current build up, but it may also change the frequency of oscillation, so lots of experimentation will be required to change the current.

    This is a simple circuit where one could conceivably buy all the parts from say Digi-Key.

    The next controller I have in two different versions, same topology but different values in the caps. The lamps are larger (2 1/2 and 5” tops). The 2 1/2” version seems to get away with just 0.1 uF for C1. However, it is dead so I can’t make any measurements.

    [​IMG]


    Here we have a real switching power supply with a regulated current output. However, the chip is done as a “glob top” so there are no markings to read. To use this IC you would have to pull it from an existing lamp (or get luckier in your searches then I have).

    The following analysis is filled with (educated) guesses: The solar cell is typically isolated from the battery with a schottky diode for efficiency, that seems to be built into the die. Right end of L1 is seen to switch to ground, then flyback to about 3V at a 640KHz rate. For 620 uS L1 charges C1, for the next 700 uS L1 freewheels. L1 is then charged for 860uS when U1 switches it to ground. I see no drop between right side of L1 and C1 so there may be a synchronous switch in the chip.

    C1 keeps a fairly steady 3V across it. R1 is a current sense resistor that shows a triangle wave of about 53 to 84 mV (68mV average); for a 12 ohm resistor this translates to an average current of 5.7mA.

    This second version seems much more hackable. You can vary the value of R1 to change the regulation current. You may also have to change L1 (smaller) if you want much more current.



     
    bug13, scopesys and russpatterson like this.
  12. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    All of my cheap solar garden lights have a Ni-Cad battery cell with iron contacts that rust away within one year. The LED has iron wires that also rust away. If it has an IC then its iron pins rust away. The circuit board is cheap phenolic-paper that absorbs rain and humidity like a sponge.
    The screws are iron that quickly rust away.
    The solar panel and LDR get a sunburn which causes them to fail within two years.

    I got 10 solar garden lights for free from my electrical utility company. They are 5 years old and after replacing a few rusty battery cells and one rusty LED they still work perfectly.
     
  13. wayneh

    Expert

    Sep 9, 2010
    12,145
    3,056
    Thanks for posting schematics.

    Anyone following this might also enjoy this description, or this from Maxim.

    There are a number of commercial chips to perform the boost function of the glob-top circuit (along with battery management chores), although most of the lights I've looked at seem to use the globs instead of identifiable ICs.
     
    russpatterson likes this.
  14. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,394
    1,606
    wayneh: Yeah, I've been to those sites. SOLAR GARDEN LIGHT -1 is essentially the same as the first "simple" schematic I posted, except it uses the BC547 and associated parts instead of the photocell to detect night.

    The only dedicated/single IC controller I can find is the Prema PR4403 (link).

    However, it may be tough to get (just one sole source in the US who is asking me for my complete bio to even quote a "minimum order.)

    I actually have more interest (and fun) working out circuits like this with tight constraints and small numbers of parts.
     
  15. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,394
    1,606
    Going thru the store today I had to pick up this year's model of a solar LED when I noticed the CdS cell was missing. The light is a "large" single lamp, metal pole base ("stainless steel" on the label but I believe that is just the color of the housing. $3.99 for a single lamp.

    A tear down shows the controller now has a proper minimal set of components, someone must have built the ASIC for this, and the PCB only has 3 parts on it:

    [​IMG]

    This is even better then the PR4403 circuit as it uses fewer components, in fact it has the fewest components possible.

    The device looks like a 4-lead TO-92 case and is marked with "ANA618" and "20270." Neither number leads to a data sheet on a web search, but curiously the returned lists begin with several part number data bases.

    If anyone has any info on this device (like where to get me some!) holla back.
     
  16. overlord2015

    New Member

    Sep 6, 2011
    1
    0
    I just got off a few sites i was checking out about this IC. No one seems to have a PDF file on it with the pin outs. Ha. Maybe it's one of the bootleg chips. Ha. Have a good one.
     
Loading...