Solar Inverter Efficiency

I. Introduction/Abstract

The market for solar inverters (photovoltaic inverters) is growing due to the demand for renewable energy. These inverters need to be extremely efficient and reliable. This article examines the power circuits used in these inverters and recommends the best choices for switching and rectifying devices.

The general structure of a photovoltaic inverter is shown in Figure 1. There are three different inverters to choose from. Sunlight shines on solar modules connected in series. Each module contains a group of solar cells connected in series. The direct current (DC) voltage generated by the solar module is in the order of hundreds of volts. The specific value depends on the lighting conditions of the module array, the temperature of the battery, and the number of modules in series.

The primary function of this type of inverter is to convert the input DC voltage to a stable value. This function is achieved by a boost converter and requires a boost switch and a boost diode.

In the first configuration, the boost stage is followed by an isolated full-bridge converter. The full-bridge transformer provides isolation. The second full-bridge converter on the output is used to convert the DC of the first-stage full-bridge converter into an alternating current (AC) voltage. Its output is filtered before being connected to the AC grid network via an additional double-contact relay switch, with the goal of providing safety isolation in the event of a fault and isolation from the supply grid at night.

The second structure is a non-isolated solution, in which the AC voltage is directly generated from the DC voltage output by the boost stage.

The third structure uses an innovative topology of power switches and power diodes to integrate the functions of the boost and AC generation parts into a dedicated topology.

Figure 1: Schematic diagram of solar inverter system

Although solar panels have very low conversion efficiencies, it is important to get the inverter efficiency as close to 100% as possible.

A 3kW series connected module is expected to generate 2550 kWh per year. If the inverter efficiency is increased from 95% to 96%, an additional 25kWh can be generated per year. The cost of generating this 25kWh with an additional solar module is equivalent to adding an inverter. Since the efficiency increase from 95% to 96% does not double the cost of the inverter, investing in a more efficient inverter is a natural choice. For emerging designs, the most cost-effective way to increase the inverter efficiency is a key design criterion.

As for the inverter reliability and cost are two other design criteria. Higher efficiency can reduce temperature fluctuations over the load cycle, thus improving reliability, so these criteria are actually related. The use of modules will also improve reliability.

All topologies shown in Figure 1 require fast switching power switches. The boost stage and full-bridge converter stage require fast switching diodes. In addition, switches optimized for low frequency (100Hz) switching are also useful for these topologies. For any particular silicon technology, switches optimized for fast switching have higher conduction losses than switches optimized for low frequency switching applications.

II. Switches and diodes for the boost stage

The boost stage is typically designed as a continuous current mode converter. Depending on the number of solar modules in the array used in the inverter, either a 600V or 1200V device is used.

The two choices for power switches are MOSFETs and IGBTs. Generally speaking, MOSFETs can operate at higher switching frequencies than IGBTs. In addition, the effect of the body diode must always be considered: in the case of a boost stage, this is not a problem because the body diode does not conduct in normal operating mode. The conduction losses of a MOSFET can be calculated from the on-resistance RDS(ON), which is proportional to the active die area for a given MOSFET family. When the rated voltage changes from 600V to 1200V, the conduction losses of the MOSFET increase significantly, so even if the rated RDS(ON) is comparable, a 1200V MOSFET is either not available or too expensive.

For boost switches rated at 600V, superjunction MOSFETs can be used. This technology offers the best conduction losses for high-frequency switching applications. MOSFETs with RDS(ON) values ​​below 100 milliohms in TO-220 packages and RDS(ON) values ​​below 50 milliohms in TO-247 packages are available on the market.

For solar inverters that require 1200V power switches, IGBTs are the appropriate choice. More advanced IGBT technologies, such as NPT Trench and NPT Field Stop, are optimized for reduced conduction losses, but at the expense of higher switching losses, which makes them less suitable for boost applications at high frequencies.

Fairchild Semiconductor has developed a device based on the old NPT planar technology that can improve the efficiency of the boost circuit at high switching frequencies. The FGL40N120AND has an EOFF of 43uJ/A, compared with 80uJ/A for devices using more advanced technology, but it is very difficult to achieve this performance. The disadvantage of the FGL40N120AND device is that the saturation voltage drop VCE(SAT) (3.0V relative to 2.1V at 125oC) is high, but its advantage of low switching losses at high boost switching frequencies is enough to make up for it. The device also integrates an anti-parallel diode. In normal boost operation, this diode will not conduct. However, during startup or transient conditions, the boost circuit may be driven into the operating mode, at which time the anti-parallel diode will conduct. Since the IGBT itself does not have an inherent body diode, this co-packaged diode is required to ensure reliable operation.

For the boost diode, a fast recovery diode like Stealth? or carbon-silicon diode is needed. Carbon-silicon diode has very low forward voltage and loss. However, they are very expensive at present.

When selecting a boost diode, the effect of the reverse recovery current (or junction capacitance of a carbon-silicon diode) on the boost switch must be considered, as this can lead to additional losses. Here, the newly introduced Stealth II diode FFP08S60S can provide higher performance. When VDD=390V, ID="8A", di/dt=200A/us, and the case temperature is 100oC, the calculated switching loss is lower than the parameter of FFP08S60S, 205mJ. With the ISL9R860P2 Stealth diode, this value reaches 225mJ. Therefore, this also improves the efficiency of the inverter at high switching frequencies.

III. Switches and diodes for bridge and special stage

After filtering, the output bridge generates a 50Hz sinusoidal voltage and current signal. A common implementation is to use a standard full-bridge structure (Figure 2). In the figure, if the upper left and lower right switches are turned on, a positive voltage is loaded between the left and right terminals; if the upper right and lower left switches are turned on, a negative voltage is loaded between the left and right terminals.

For this application, only one switch is on at a time. One switch can be switched to the high PWM frequency and the other to the low 50Hz frequency. Since the bootstrap circuit relies on the switching of the low-side device, the low-side device is switched to the high PWM frequency and the high-side device is switched to the low 50Hz frequency.

Figure 2: MOSFET full bridge

This application uses a 600V power

The 600V super junction MOSFET is very suitable for this high-speed switching device. Since these switching devices will bear the full reverse recovery current of other devices when the switch is turned on, fast recovery super junction devices such as the 600V FCH47N60F are ideal. Its RDS(ON) is 73 milliohms, and its conduction loss is very low compared to other similar fast recovery devices. When this device is switched at 50Hz, there is no need to use the fast recovery feature. These devices have excellent dv/dt and di/dt characteristics, which can improve system reliability compared to standard super junction MOSFETs.

Another option worth exploring is the FGH30N60LSD device. It is a 30A/600V IGBT with a saturation voltage VCE(SAT) of only 1.1V. Its turn-off losses EOFF are very high at 10mJ, so it is only suitable for low-frequency switching. A 50mOhm MOSFET has an on-resistance RDS(ON) of 100mOhm at operating temperature. Therefore, at 11A, it has the same VDS as the VCE(SAT) of the IGBT. Since this IGBT is based on an older breakdown technology, the VCE(SAT) does not vary much with temperature. Therefore, this IGBT can reduce the overall losses in the output bridge, thereby improving the overall efficiency of the inverter.

The FGH30N60LSD IGBT is also useful for switching from one power conversion technology to another dedicated topology every half cycle. Here the IGBT is used as a topology switch. Conventional and fast recovery superjunction devices are used for faster switching.

For 1200V dedicated topologies and full-bridge structures, the FGL40N120AND mentioned above is a very suitable switch for new high-frequency solar inverters. When dedicated technologies require diodes, Stealth II, Hyperfast? II diodes and carbon silicon diodes are good solutions.