Powering LEDs from solar cells

These days, there are fast-paced technological developments in both light-emitting diodes (LEDs) and photovoltaic solar cells. So it is not surprising to see applications that combine the two technologies. Consider street lighting as one example of where LEDs are taking hold. The city of Seattle started swapping out high-pressure sodium lamps for LEDs in 2007 and expects to complete the task for the 41,000 HPS lamps that light its residential neighborhoods by 2014. The LEDs are 50% more energy efficient and last 10 years longer than the sodium lamps they replace. Seattle figures it will eventually save $2.4 million per year in light operations and maintenance.

Or consider the Solar Gen2, a solar-powered street lamp developed by Philips Lighting. Its LEDs are bright enough to allow for a lamp post spacing of up to 50 m, which complies with EU road lighting standards. Philips says by charging street lighting during the day, Solar Gen2 can supplement the capacity of the conventional electricity grid. A microcontroller in each pole tracks the amount of sunlight received and predicts the likely amount of sunlight over the next few days based on geographical location and time of year. It dims the street light if there seems to be a danger the battery will approach discharge. China’s Guiyang area recently installed 100 of the solar street lights.

Other kinds of outdoor lighting are putting LEDs together with solar cells. For example, it is possible to find LED rope lights for lining walkways, windows, trees, gazebos, decks, and so forth, that use a polycrystalline solar panel to eliminate the need for wires and outdoor outlets. Such products typically provide 8 to 12 hours of illumination, depending on the amount of sunlight, and turn on/off automatically via a built-in light sensor.

Even solar-powered LED flood lamps, spot lights, lanterns and table lamps are available for prices below $100.

Still, the challenge for solar-powered LED lighting is getting cost down and doing so with products that get to market quickly. These are both challenging areas. For example, DOE’s Municipal Solid-State Street Lighting Consortium recently worked with the City of Sacramento on a pilot project involving LED luminaire replacements for high-pressure sodium luminares. None of the LED products evaluated could match the economics of the existing 100-W HPS luminaires, though the energy used by three of the LED systems ranged from 63 to 90% of the baseline HPS units.
The power-conversion strategy is one means of reaching cost and speed-to-market goals for LED/PV products. So it is helpful to review the power-conversion components typically involved, develop a system, and provide a high-level method to analyze behavior.

Basics

The core components in all of these systems are the solar cells, storage element (battery or super capacitor), and LEDs. The behavior of each element must be compatible with that of the others. In this case, that means the output voltage/current behavior of the solar cell must align with the battery-charging profile. In other words, the solar cell must put out enough energy at the right level to charge the storage element. And the battery-discharge profile must match the LED drive requirements. In other words, the battery must put out enough energy at the right level to run the LED for the required time. Without intervening electronics, this is not posible.

A review of the solar cell V-I curve, charging and discharging qualities of nickel-metal-hydride batteries, and LED drive requirements reveals mismatches among output voltage/current and drive voltage/current. For example, the maximum solar-cell voltage (per cell) is around 1 V, while a NiMH battery operates in a range of 0.9 to 1.4 V. LEDs require a constant current source so their light output doesn’t vary, although their forward voltage is typically above 3 V. Further, the NiMH battery must be charged in a specific way, with a charge current that depends on charge voltage, to maximize its useful life.

It is possible to develop a system that interfaces all of these components directly, but doing so involves significant trade-offs in system efficiency and robustness.

Now consider adding a simple power electronics interface in the form of a battery charger and LED driver. This gives a much higher degree of flexibility and lets the designer optimize the overall system performance. A standalone battery-charger IC can manage the NiMH charging profile, and an LED driver IC can convert the battery voltage into a constant-current source. There is no real need to add a controller.

However, there are at least two down sides to this discreet IC configuration. First, it limits the permissible range of operating conditions. ICs for managing batteries and driving LEDs often have a fairly narrow operating range, which limits the designer’s ability to make changes down the road or in response to customer requests. For instance, use of a new solar-cell configuration may necessitate use of a different battery-charging IC. If the energy-storage technology or configuration changes, it’s likely both the battery-charging IC and the LED-driver IC will have to go. And if there is a change in the type of LEDs or their series/parallel configuration, the LED driver will need to be reconfigured.

The inclusion of a microcontroller helps handle these difficulties by providing some flexibility. Reprogramming can take the place of significant hardware changes.

The second downside to the use of discrete ICs for these functions is that it makes system optimization difficult. A generic battery-charging IC, for example, probably won’t include a Maximum Peak Power Tracking (MPPT) algorithm to maximize the output of the solar cell. Thus a circuit based on discrete chips may have difficulty keeping up with new technology developments.

A computer inside

There are three advantages to an LED/solar cell circuit that incorporates a microcontroller. First, it allows optimization of the system through programming. For example, the overall charging efficiency depends on the efficiency of the solar cell. Incorporating an MPPT profile into the power-conversion algorithm should make the solar/electricity power conversion more efficient, possibly allowing a smaller solar array to meet charge objectives. Array size impacts the product’s form factor and so may provide opportunities to enhance the product’s visual appeal.

Similarly, light quality can be a critical characteristic in some applications, as when the LED provides light for reading. LED drive current affects light quality and the ability to dim the LED, so drive current waveforms typically have tight tolerances. The ability to make changes by reprogramming eases optimization of everything from the operating efficiency to the system’s overall lifetime.

Second, a controller-based architecture is entirely scalable and can work across a broad power range. A compact, portable lantern used for reading may have a single solar cell, off-the-shelf rechargeable NiMH batteries, and a few LEDs using 20 to 75 mA of drive current. Simply replacing the power components, including readily available power MOSFETs and transformers, potentially scales the design to fulfill the needs of security lighting. The same architecture can handle more solar cells, a bigger NiMH battery pack or a custom storage element, and high-brightness LEDs requiring over 350 mA of drive current. Similarly, use of a controller can make it easier to incorporate evolving solar cell or LED technology with different handling requirements. And as with other kinds of applications, a controller can also communicate the system’s health and predict when it will need maintenance.