I'm building a solar powered light. The idea of this light is that its output is proportional to the brightness of the sun. At night, no light. During daylight, full light. So it's basically as simple as attaching my 12v 0.5A LED setup to a 12v 6w solar panel.

But not so fast. The 12V 6w panel won't produce enough voltage at lower sunlight levels - I want the LEDs to be brighter in the AM than a 6w panel will drive them. So I want to use a 12V 10W panel. In full sunlight, however, I fear it'll overdrive the LEDs. So I need to cap the output at 12V 0.5A. Below that, the output should just follow the output of the panel.

I assume I'll need a zener diode circuit, but am not sure how to make the circuit. Please help me design one with the correct values for my system.

EDIT: I'm hacking an existing LED strip that runs off of a 12V .5A wall wart to operate from a hobbyist-type panel. This isn't a pre-existing system.

EDIT: like this http://www.amazon.com/Solaro-Energy-SD1000-Daylight-Simulator/dp/B0053AL5LA only not $600

EDIT: so-called "12V" panels are actually >12V because the typical use is to charge a 12V battery. Both models available (the 5W and 10W) actually have nominal voltage of ~18V in full sun and OCV of ~21V in full sun. I can limit the amount of sun to prevent the voltage from peaking so high, but that kinda defeats the purpose of having an efficient system that will drive the LEDs fairly brightly at lower sunlight levels.

  • \$\begingroup\$ Please could you provide a datasheet link, or product details on the 12v .5A LED setup to be used? \$\endgroup\$ – Anindo Ghosh Apr 12 '13 at 11:59
  • \$\begingroup\$ Without a schematic or reference for your specific panel, this is just an educated guess, but the panel should already have current limiting resistors, in order for the leds to function properly at a stable current. Pictures of the panel would be best. That said, unless the panel has no regulation inside, or current limiting resistors on the leds, providing a higher current source is not a problem. Current is pulled, not pushed. \$\endgroup\$ – Passerby Apr 12 '13 at 12:08
  • 1
    \$\begingroup\$ Sound like what you really want is called a "mirror". That's a rather efficient light-in light-out device. \$\endgroup\$ – Olin Lathrop Apr 12 '13 at 12:12
  • \$\begingroup\$ @OlinLathrop only for mirrors of sufficient high quality. But sure, Op's goal might not be practical, still no reason not to make it work. \$\endgroup\$ – Passerby Apr 12 '13 at 12:17
  • \$\begingroup\$ @AnindoGhosh the lights are an LED strip normally connected to a 12V .5A wall wart. I'm replacing the wall wart with a solar panel. \$\endgroup\$ – Rip Rowan Apr 12 '13 at 12:20

A Zener would work in this situation, where the maximum excess power supplied by the panel isn't very high. 1N5349BG (12V, 5W) seems suitable. You should make sure it has adequate heatsinking; perhaps firmly screwing the ground lead to metal casework would be sufficient.

  • \$\begingroup\$ Hey, now we're talking. Do I need a resistor or do you think the excess power can just go directly to ground? \$\endgroup\$ – Rip Rowan Apr 12 '13 at 13:18
  • \$\begingroup\$ Got more specs on the panels, see third edit, above. \$\endgroup\$ – Rip Rowan Apr 12 '13 at 13:39
  • \$\begingroup\$ The Zener itself behaves as a resistor of the right value to dissipate power. It doesn't go to ground, it goes to heat in the Zener. Passerby is also correct that the efficiency of panels falls as you go beyond the maximum power point. It might be worthwhile selecting a zener with voltage slightly above 12V (13V? 1N5350BG) which will have to dissipate much less and take advantage of the MPPT properties of the panel. \$\endgroup\$ – pjc50 Apr 12 '13 at 13:45
  • \$\begingroup\$ Gotcha, makes perfect sense. I have a great heatsink - the thing can be attached to an enormous steel girder :) \$\endgroup\$ – Rip Rowan Apr 12 '13 at 13:49
  • \$\begingroup\$ I'm giving this one the answer b/c it was the first to actually spec out the correct zener to use. Thanks to all for a lot of help and good, polite discussion. \$\endgroup\$ – Rip Rowan Apr 12 '13 at 13:53

LEDs are current driven devices, not voltage driven. In other words, it is not the voltage provided to the LED that matters (except peripherally), as the current allowed through it.

Intensity of an LED is controlled by regulating the current, or by using Pulse Width Modulation of variable duty cycle (on-off time ratio of pulses) with a predefined, fixed on-state current.

The question does not specify which specific 12v .5A LED setup is to be used, so it is conceivable that the set-up may already have dimming circuitry built in - LEDs by default are not 12 Volt devices, so there must be some additional electronics in the setup.

If the LED setup does provide a dimming interface, then that should be used to alter the intensity.

  • Some LED units contain inbuilt constant-current regulators and favor PWM for control
  • Others have their own remote control protocol (and control device) - less common
  • Yet others are essentially just LED + resistor in series, so voltage, current or PWM based intensity control will work

If the LED has integrated current regulation and a PWM interface, a microcontroller circuit would be recommended, to modify PWM duty cycle as per requirement.

For the type of LED setup with just current limiting resistors in there, both voltage and current regulation are options, voltage regulation perhaps being more effective.

The additional stated requirement is that the LEDs must be driven brighter than default when the ambient light gets low. For this, you may need to put together a boost regulator, to raise available voltage to a higher value, and vary LED intensity through a microcontroller program as per some set table, in relation to the solar panel's power output (i.e. the ambient light). Without additional information about the solar panels and associated circuitry, it is not feasible to speculate further.

Regarding input voltage from the solar panel exceeding the rated 12 Volts of the LED strip: LED strips with integrated resistors are quite resilient to a small overvoltage, but you'll only know through destructive testing. Yes, as the question rightly mentions, a 12 Volt Zener diode connected in reverse bias as a shunt at input would be one way of limiting supply voltage to the LEDs.

  • \$\begingroup\$ "The additional stated requirement is that the LEDs must be driven brighter than default when the ambient light gets low" - isn't in the original question, it requests that the LEDs track ambient light on the solar panel. \$\endgroup\$ – pjc50 Apr 12 '13 at 13:04
  • \$\begingroup\$ @pjc50 That is my impression from this statement: "... won't produce enough voltage at lower sunlight levels - I want the LEDs to be brighter in the AM". If my interpretation is wrong, I can edit it out. \$\endgroup\$ – Anindo Ghosh Apr 12 '13 at 13:13
  • \$\begingroup\$ @AnindoGhosh you have the right idea, but honestly, attaching the LEDs directly to the solar panel provides exactly the behavior I'm looking for. No fancy microcontrollers needed. However I need to attach them to a panel that is large / efficient enough to get the lights near full brightness even at 50% sunlight. My concern is that at 100% sunlight, I'll kill the LEDs. Thus the need for voltage regulation. \$\endgroup\$ – Rip Rowan Apr 12 '13 at 13:24
  • \$\begingroup\$ @RipRowan Then with a zener shunt you're all set. :-) \$\endgroup\$ – Anindo Ghosh Apr 12 '13 at 13:36
  • \$\begingroup\$ Right! I have new info (see 3rd edit above) that might help select the right components to build the shunt. \$\endgroup\$ – Rip Rowan Apr 12 '13 at 13:44

Assuming it's a simple led strip, in a 3 leds + resistors per segment setup, then no, you have nothing to worry about, by using a bigger solar panel. A 12V 6W solar panel should allow 0.5A of current at full unobstructed sunlight. A 12V 10W would allow 0.83A of current. But current is pulled, not pushed.

The led strip like below, are set up for 12V, at a fixed current (educated guess, 20mA per segment). The resistor on each segment, will take the Input Voltage - the Led Forward voltage (x3 because there are 3 leds in series, and leds in series share the same current), and limit the current.

Assuming 3.3v Forward Voltage, that's 12V - (3.3V * 3) / 20mA or 2.1V / 20mA = 105 ohm resistor or close to it.

And since that's just 20mA per Segment, and that's 3 leds per segment, you just multiply 20mA by the number of segments your strip has, to find out how much it draws.

enter image description here

What does this mean for you? Basically, the led strips are fixed to 20mA per led, at 12V. Using a bigger Solar Panel increases it's potential current source, but the led strip will not pull more than it needs and can be sourced. At lower light levels, a bigger (or more efficient) panel will probably increase the current capacity. At high light levels, nothing will change. And that's only if the current panel cannot drive your light strip at the moment. How many segments do you have?

As for the Solar Panel Voltage output, it's a curve, and depends on the panel/cell, but in general: enter image description here

As the voltage increases past the nominal design point, in this case 12v, the current capacity starts to sharply decline. And the leds will not be able to pull more current than exists, even when the voltage is higher.

  • \$\begingroup\$ Thanks for an excellent reply. Can I ask a followup? What if the panel is rated 12V nominal but capable of almost 18V in full sunlight? That would be a problem, assuming the LEDs aren't regulated, right? I suppose I could just work with a voltmeter and place the panel such that it never gets enough light to output >12V. Or am I totally wrong? \$\endgroup\$ – Rip Rowan Apr 12 '13 at 12:55
  • \$\begingroup\$ No, you're right and that's a problem with this answer. Solar panel certainly can put out more than their nominal voltage. \$\endgroup\$ – pjc50 Apr 12 '13 at 13:01
  • \$\begingroup\$ @pjc50 but current sharply drops off in relation to the a higher than nominal voltage. \$\endgroup\$ – Passerby Apr 12 '13 at 13:06
  • \$\begingroup\$ Note - I'm hacking an existing wall-wart powered light strip to be solar powered. There isn't an existing panel, just the wall wart. I will purchase a panel to drive the LEDs. Off the shelf ~12V panels are designed to charge batteries and typically output somewhat more than 12V, up to 18V in full light. \$\endgroup\$ – Rip Rowan Apr 12 '13 at 13:15
  • \$\begingroup\$ Also note I'm new on this SE so I have no rep but this is definitely the sort of answer I need. No complex circuitry should be necessary. EDIT just saw @pjc50's answer below and I think he's on it. \$\endgroup\$ – Rip Rowan Apr 12 '13 at 13:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.