We have known for decades that the sun radiates enough energy to meet the world’s needs for power, both now and in the future. As in any energy conversion from original source to a “usable” form, however, the various stages of adaptation introduce efficiency losses. In the case of the photovoltaic system, this has limited what could be exponential growth. In fact, the best performance available from solar cells of any kind (excluding the concentrated approach) never rose above 15% efficiency. In cascade to the solar cell itself, a classical centralized inverter would add barely more than 90% efficiency to the chain. This “poor” performance and the obvious lack of large successes has also helped to keep the installation costs very high.
Only in the last ten years have multicrystalline cells gone beyond the limit, reaching the 20% efficiency dream. We also have seen the introduction of thin film panels, with efficiencies just above 10% at a cost less than one third of multicrystalline panels. Inverter technology, with the best use of new materials, has also gone above 95% efficiency and dramatically reduced system costs.
All of these advances, championed by a population leaning more and more toward green energy sources, and sponsored by national, corporate, and local incentives, has led to worldwide yearly installations in the 10GW range, from humble beginnings of way below 1GW just a few years back.
Regardless of the changes in technology, both in panels and inverters, the basic principles always remain the same:
- Solar energy hits the solar cells (any technology) (Figure 1).
- Due to the photovoltaic effect, the cells release a DC voltage and a certain amount of current.
- Depending on how the cells are arranged, the resulting DC voltage and current will be converted into AC usable electrical energy by the inverter (performing other functions as well to maintain a safe and reliable grid connection) (Figure 2).
The Evolution of Solar Systems
There are basically three main types of solar cell installations:
- Residential (on top of single family houses or town houses).
- Commercial (larger roofs on commercial buildings).
- Ground, also called solar farms.
Typically, installed power systems are 3kW to 10kW for a residential installation, up to 20kW-100kW for a commercial installation, and practically limitless for solar farms.
Research and development investments in solar technologies have increased together with the worldwide growth in adoption of photovoltaic systems. Furthermore, solar cell technologies and inverter topologies, as well as system architectures themselves, have evolved and consist of more complex aspects than in the past.
Let’s consider the four main architectures:
- Until very recently, a typical installation of any kind of power would consist of a certain number of panels, connected through junction boxes in series and parallel, feeding one single inverter. Such an installation was called centralized. This main inverter would use communication technology in order to send data back to data servers about several aspects of the total energy conversion.
- In the last ten years, in order to improve system redundancy and efficiency, the string inverter was introduced and achieved success, by means of dividing the solar field into strings, each connected with a smaller inverter. The string inverter communicates with a concentrator that dispatches the cumulative information about the system’s energy flow data.
In both of these cases, all the processing power, or intelligence, resides in the inverter where several functions are performed including the very important Maximum Power Point Tracking (MPPT) that allows the system to adapt the input impedance to what the photovoltaic field offers, due to several environmental conditions. This can clearly be seen in a typical panel MPPT curve, as shown in Figure 3.
- The distributed MPPT approach, also known as Power Maximizer, arose from the understanding of how important it is to control the maximum power point at each photovoltaic panel. In this case, while the inverter can be of any of the previous two types (centralized or string), the MPPT is now embedded in a “smart-box,” and can track and adapt the impedance at each panel output, reaching an energy harvesting improvement of up to two digits. Each unit can now communicate with a central data concentrator or leave this function to the main inverter.
- The next architecture integration step seems obvious and is delivered in those architectures making use of microinverters attached to each panel. It completely removes the need of either centralized or string inverters, and goes straight into the grid, with a final implementation of what is called PVAC — a solar panel with an embedded microinverter. Each microinverter incorporates communication technology, collected by energy management units, and sends it back to the data servers.
Typical Microinverter Block Schematic
Because the microinverter must perform all the inverter functions while being attached directly to the panel, its block schematic could use up to three power stages to perform the DC to DC function and so deliver a maximized power at the output (adapting the impedance to the real MPPT curve), an inverter stage and a possible line matching stage. Several alternatives exist to decrease the stages to two or even one.
The reduction of stages improves the efficiency of the system dramatically (today, available microinverters range from 92% up to above 96%), potentially reducing the precision of the full control of MPPT and other important variables.
In addition to the power stages, the communication block is of extreme importance for a microinverter, for both external monitoring reasons, and also reliability and failure prevention.
Main Semiconductor Categories Used
To best assess the categories of semiconductors needed in a microinverter, we need to consider these points:
- Most microcrystalline panels reach up to 72 cells based on 6" wafers, resulting in an open circuit voltage above 50V.
- The biggest markets for photovoltaic expansion are Europe and North America, with single phase voltage mains of, respectively, 220 VAC and 120 VAC, going in poly-phases up to 408 VAC and 380 VAC.
- The environments of installations are very different and might go from completely shielded to wide open air (communication- wise).
With this in mind, the following semiconductor categories should be considered:
- Low voltage power MOSFET (from 75V up to 200V depending on topology);
- High voltage power MOSFET (from 500V up to 1200V depending on topology and installation voltage mains);
- High voltage SCRs (800V and above);
- SiC diodes (600V and 1200V) Schottky diodes (low voltage);
- Digital controllers (micro-based or DSP-based);
- Power line modems;
- RF transceivers (2.4 GHz with IEEE802.15.4 seems to be the most adopted);
- Isolated gate drivers
Best in Class Semiconductor Needs
With the need for efficiency and reliability at high temperatures and critical working environments, the semiconductors to be selected for a successful microinverter implementation must be always the best in class available in the specific design window.
In the above mentioned semiconductor categories, some suggestions come to mind:
- Low voltage power MOSFET: Trench technology with best RDS(on) and gate charge parameters give the very high frequency switching need;
- High voltage power MOSFET: “SuperJunction” technologies;
- Digtal controllers: The bare minimum for a microcontroller would be a 32-bit core with very precise timers in order to cope with the complex MPPT function and the high precision power processing topologies.
- Power line modem: Given the variability of installation sizes, a minimum of 28.8 kbps should be guaranteed to be able to reliably transmit data from a large number of microinverters in commercial and ground installations. High network reliability is mandatory for this category.
- RF communication: At this early adoption stage it is important to “future proof” installations for smart grid integration. Flexible solutions would include some sort of RF communication interface.
- ASIC technology: With the intelligence and intellectual property behind any successful microinverter company both now and in the immediate future, a good and reliable mixed signal ASIC technology partner would be necessary to keep “the secret sauce” secret and drive costs down.
The System Solution
Understandably, all these technologies must connect seamlessly in the microinverter system for performance and reliability. Obviously, much more semiconductor content per installed KW is needed in a microinverter as compared to a centralized inverter. In fact, the semiconductor content in a microinverter can account for 30% to 50% of its total cost, while in the case of a centralized inverter, the semiconductor content is below 20% of the system cost. However, the improved efficiency and higher energy gain of the microinverter, particularly within the concept of the smart grid, secures its importance.