When the Sun is shining and illumination is needed inside a dimly lit interior space, a popular, proven, and highly efficient solution is to utilize the energy of available sunlight in the simplest and most sustainable fashion conceivable: Opening a window!
Sometimes, however, details of access to the outdoors make this traditional solution inconvenient, impractical, or downright impossible. Then, a more topologically flexible approach may be needed, even if it’s more complex and less efficient than the window gambit. Enter the solar day lamp.
A solar day lamp is an illumination system comprising a solar photovoltaic panel mounted outside – sustainably converting sunlight into electrical power – a run of wire to conduct said power into the interior, then suitable circuitry and LEDs to re-convert the delivered power back into a useful light source.
It’s admittedly more complicated than a window, but still better than stumbling around in the dark!
For such a double-conversion scheme, converting light into electricity and then back into light, to work with a reasonable size (and cost!) solar panel and still be bright enough to be useful, puts a premium on achieving high efficiency for both conversion steps. This design idea (see the Figure 1) presents some ways to achieve these design imperatives.
 |
|
| Figure 1. | Solar day lamp with maximum power point tracking and high voltage, constant-current LED drive. |
By definition photovoltaic panels work by converting light into electrical power. It follows that the amount of power a panel can produce depends on the amount of light shining on it. Duh! What’s perhaps less obvious is that a panel’s power output also depends on the voltage to which it’s loaded, and that the voltage of maximum conversion efficiency and power output (maximum power point voltage = MPPV) varies significantly with the amount of light and (to a lesser degree) temperature.
For example, the spec’ sheet for the panel illustrated rates it for “30 Watts” and “12 Volts”. But this should never be read as saying it can source 30 W into a 12 V load, because it won’t – not even in full direct sunlight. In fact, the most it could ever deliver into 12 V is barely 20 W. To hope to get the rated 30 W, the load voltage must be allowed to rise to 156% of the nominal 12 V rating – to 18.7 V (the so-called maximum power voltage = MPV). What’s going on?
This situation is actually typical of solar panel specifications. The rated output voltage is usually deliberately underrated. This accommodates the fact that panels seldom get to bask in full perpendicular sunlight, and that a user would rather get something rather than nothing in the way of usable output (e.g., enough to charge his 12 V battery) in less than perfect conditions.
And in fact, nothing is about all this panel actually would output into an 18.7 V load if, for example, anything less than about 20% of full Sun were shining on it.
In order to extract maximum power from the panel, optimum loading must vary with incident illumination and temperature. This stratagem is typically called maximum power point tracking (MPPT) and is the purpose of U2, A1 and surrounding components.
U2a and U2b oscillate to generate a ~100 Hz “perturbation” square-wave that is summed with the duty-cycle control signal applied to U1. This results in periodic variation of the solar panel loading voltage. Panel power efficiency therefore also varies, generating a signal at synchronous rectifier U2c pin 4, where it is sampled and applied to feedback integrator A1. The resulting MPPT signal is accumulated, becoming feedback to 25 kHz voltage-multiplier oscillator U1 that increases or decreases U1’s duty cycle in the correct direction to maximize power accepted from the solar panel.
A generalized description of how “perturb-and-observe” active MPPT works is detailed in “Solar-array controller needs no multiplier to maximize power” (Ref. 1).
The power extracted from the panel must then, of course, be input to the LED array and used to generate useful light. The usual way this is usually done is to connect the LEDs in a low voltage serial/parallel matrix. This topology unfortunately incurs inherent inefficiency due to the need for current-balancing ballast resistors that compensate for unavoidable mismatch between LED forward voltages. About 10% or more of total available power is typically lost in this way.
The circuitry shown avoids this inefficiency by boosting panel voltage to a value high enough (~90 V) to accommodate a pure series connection of thirty1-W LEDs. Hence the need for ballast resistors is eliminated along with their undesirable power losses, resulting in a significant further improvement in lamp efficiency.
A complication arises, however. What if continuity of the LED series string is lost and the current delivered by D1 has nowhere to go?
If this should happen and nothing were provided to safely control the accumulation of charge on C8, the voltage there would rise dangerously (theoretically without limit) until destruction, perhaps violent, of many components including Q1, D1, and C8, became inevitable. Voltage comparator transistor Q2 is configured to prevent this catastrophe, setting U1’s RESET input low and shutting down Q1 drive should a hazardous overvoltage condition threaten to occur.
Reference
- Woodward, Stephen. “Solar-array controller needs no multiplier to maximize power”
