PHOTON Laboratory s inverter test

PHOTON Laboratory s inverter test How PHOTON conducts its test...2 How PHOTON efficiency is calculated... 3 Overview of test results Aros... 4 Conergy... 4 Delta Energy Systems... 4 Diehl AKO... 5 Fronius... 5 Ingeteam... 6 Kaco New Energy... 6 Kostal Solar Electric... 7 Mitsubishi Electric... 7 Phoenixtec Power... 7 Power-One... 8 Refu Elektronik... 8 Riello UPS... 8 SMA Solar Technology... 9 Solon Inverters... 10 Sunways... 10 Xantrex Technology... 11 Offers for manufacturers and installers... 12 1

How PHOTON conducts its test Our goal: Helping system operators select the right inverters left: Thomas Pflaum / photon-pictures.com, cover: Udo Geisler / photon-pictures.com Random selection: Heinz Neuenstein (left), director of PHOTON Lab, and Jochen Siemer, managing editor of PHOTON Profi, our German magazine for PV installers, have picked an inverter for testing purposes at Kostal s factory. PHOTON Lab has been carrying out inverter tests successfully since 2007, informing PHOTON readers whether or not a device is up snuff. The lab assigns grades ranging from A++ to E, which correspond to an overall efficiency defined by PHOTON. Since the beginning of 2007, PHOTON Lab has employed its own inverter test methodology, and regularly published results from those tests. To make these results easier to comprehend, the editorial staff, drawing its inspiration from school report cards, launched a grading system with its own testing certificates: the grades range from A to E. The highest grade (»A«) has three different levels: an A grade, an A+ or an A++. This was deemed necessary since one of the devices tested performed so much better than its competitors that it required a more Overview of grading system subtle gradation. An E grade is assigned to an inverter with an efficiency so poor that it s essentially not worth the money paid for it put simply, so much of the produced solar electricity is either converted into heat or unabsorbed that the resulting loss in feed-in tariffs is so high that it exceeds the device s cost. In this sense, devices like these are too expensive to even give away. To assign a grade, we first need to determine the efficiency to which the grade refers. oth peak efficiency and European efficiency aren t well-suited for this purpose. This is why Grade A ++ A + A C D E PHOTON efficiency 98% 96.5 - < 98% 95 - < 96.5% 93.5 - < 95% 92 - < 93.5% 90 - < 92% < 90% 2 PHOTON Laboratory GmbH 2009

PHOTON has decided to define its own efficiency value, the explanatory value of which will far exceed conventional efficiency data. Furthermore, the goal of this grading system is to enable better comparisons of individual devices. The grades provided in the survey can essentially be associated with the devices overall utility, which is often difficult to determine from the product datasheets provided by manufacturers, as many of them omit important information on operating parameters. Currently, above all, technical data from manufacturers often omits information on the efficiency s dependence on input voltage. Also, it s uncommon to find data on an inverter s suggested MPP operating point. And information on the influence of input current limitation on the operating point, and the relation between temperature and conversion efficiency, is often nowhere to be found. To give readers of our inverter test an immediate sense of a device s value, we assign a single grade for an efficiency determined by PHOTON to each inverter that takes into account all of the above-mentioned factors. No other individual scores have an influence on the grade. The parameters reflected in the grade are reviewed on an annual basis and are be discussed with manufacturers in advance. The total grade is based on two criteria: the assessment of the efficiency determined by PHOTON and the temperature-related reduction of efficiency. The grade for this efficiency is assigned without any differentiation according to the suitability of the inverters use with a particular solar generator. The best device is the one with the highest efficiency independent of whether or not it has potential separation, is exclusively designed for use indoor or outdoors, or is a broad voltage device. In the meantime, there are now suitable transformerless inverter topologies for all known module types. Only the conversion efficiency s temperature interdependency has a relevant influence on the grade. Furthermore, we provide information on the inverter s efficiency at 25 C and the maximum temperature before any power reduction is detected. oth values are subtracted from one another. If the resulting efficiency reduction reaches or exceeds the difference from the next, lower grade (for example, there s a 1.5 gap between»«and»c«), the device receives the lower grade (i.e.»c«). The question of whether an inverter is well-suited for use with a particular module type is best answered by the manufacturer, but our tests should provide some guidance. For instance, the connections of some thin-film module types cannot be charged with negative potential against the ground. A few crystalline high-power modules require a high-impedance ground at the DC connection to avoid polarization effects. PHOTON requests approval from the manufacturers of these module types for the inverter under examination. As a matter of principle, one always needs to know the inverter input s potential in relation to the ground. Naturally, PHOTON Lab also measures the efficiency and the MPPT adjustment efficiency, both based on the specified P MPP power the product of which is the overall efficiency. This is then applied across all the measured input voltages to establish the average at each power level. This average is then weighted according to European and Californian efficiencies, and included in the evaluation. The overall efficiency is based on Heinrich Häberlin s definition of»total efficiency,«which is described in his book on the efficiency of PV inverters published in 2005. The PHOTON efficiency for medium and high irradiation levels is an artificial value that represents an image of the voltage and power interdependencies of an inverter s efficiency. The European and Californian weighting system reveals the dependence of the average overall efficiency from the geographic latitude at which the PV system is installed. This dependency is expressed with different weighting factors that result from the inclusion of meteorological data. This data allows one to make frequency distributions for certain solar irradiation values, which in turn provide weighting factors for particular power levels. The innovative part of the calculations used to establish PHOTON s efficiency is that it includes all measured input voltages as specified in the manufacturer s description of the device s input voltage range even if the device cannot perform as required in all parts of continued on page 12 The path to PHOTON efficiency Efficiency: η The MPP voltage range specified by the manufacturer is divided into 20 steps, and the DC power range into 24 steps. This results in 480 different solar generator characteristic curves, and therefore a measurement cycle consisting of 480 measurements. Using this series of measurements, we obtain 480 measurements for conversion efficiency h = P AC /P DC = f(p DC ) with V MPP as the parameter, via the specified power P MPP. The result is a three-dimensional color diagram. MPPT adjustment efficiency: η MPPT In the above-mentioned 480 measurements, we determine how much of the specified DC power at the MPP of the PV generator IV curve (P MPP ) is statically absorbed by the inverter, i.e. η MPPT = P DC /P MPP (P DC = actual measured power absorbed by the inverter, can be different from the P MPP ). These values, as a function of the P MPP, are presented with the predetermined DC power (P MPP ) in relation to the inverter s DC nominal input power P DCN, with V MPP as a parameter. The result is a three-dimensional color diagram. Overall efficiency: η Sum Multiplication for all 480 measurements: η Sum = η * η MPPT To correctly calculate overall efficiency, the η MPPT and η Sum have to be provided as a function of P MPP (the power in MPP offered by the solar generator simulator), with the V MPP as a parameter. The conversion efficiency η and the MPPT adjustment efficiency η MPPT have to be measured simultaneously. The result is a three-dimensional color diagram. Maximum overall efficiency: η SumMax at x%*p MPP This is the maximum value of overall efficiency at a predetermined DC input power P MPP. The result is one single value. Average overall efficiency: η AvgSum All efficiency results for all voltage levels are combined into a single power level. This value is then divided by the number of voltage levels. The result is a two-dimensional characteristic curve. The evaluation is conducted across the entire measurable V MPP range, including the ranges usually disregarded. The average is calculated including total MPP voltage range, which is provided by the manufacturer, including the current limit range. These ranges are marked in black and correspond to the value zero. The average is established for power levels of 5 to 100 percent. Maximum average overall efficiency: η AvgSumMax at x%*p MPP This is the maximum value of average overall efficiency at a predetermined DC input power P MPP. The result is one single value. PHOTON efficiency at medium irradiation levels: η Pmed European weighting η Pmed = 0.03*η AvgSum5 + 0.06*η AvgSum10 + 0.13*η AvgSum20 + 0.1*η AvgSum30 + 0.48*η AvgSum50 + 0.2*η AvgSum100 In this context, for instance, η AvgSum5 characterizes the average overall efficiency at 5 percent of the inverter s predetermined MPP nominal power P MPPnom. The result is one single value. PHOTON efficiency at high irradiation levels: η Phigh Californian weighting as fixed by the CEC: California Energy Commission η Phigh = 0.00*η AvgSum5 + 0.04*η AvgSum10 + 0.05*η AvgSum20 + 0.12*η AvgSum30 + 0.21*η AvgSum50 + 0.53*η AvgSum75 + 0.05*η AvgSum100 In this context, for instance, η AvgSum5 characterizes the average overall efficiency at 5 percent of the inverter s predetermined MPP nominal power P MPPnom. The result is one single value. PHOTON Laboratory GmbH 2009 3

Overview of inverters tested by PHOTON Laboratory until July 2009 Aros Conergy Delta Energy Systems Tested: Aros Sirio 4000 Tested: Conergy IP 5000 vision Tested: ed: der Delta Energy SI 3300 A 95.1 % at medium irradiation 12/2008 94 % at medium irradiation 7/2007 93.9 % at medium irradiation 5/2008 The Aros Sirio 4000 s operating range: 250 to 450 V; DC nominal power: 4 kw. The Conergy IP 5000 vision s operating range: 300 to 700 V; DC nominal power: 5 kw. The Delta Energy SI 3300 s operating range: 150 to 435 V; DC nominal power: 3.6 kw. The Aros Sirio 4000 is a transformerless inverter produced by Aros Srl, located in Cormano, Italy. This device is quite similar to a 4 kw device from Taiwanese OEM manufacturer Phoenixtec Power Co. Ltd., but the software is different. In general, the device, with a nominal power of 4 kw, makes a clean impression, which is only disturbed by the multiple circuit boards and power cables with additional ferrite cores. A limited decline in conversion efficiency at higher MPP voltages and a decrease of a few percentage points at higher powers gives the device a rather high PHOTON efficiency for medium irradiation of 95.1 percent. When designing a PV system s MPP, only the middle third of the MPP voltage range should be selected. All in all, the device is a small and light inverter with a very good PHOTON efficiency. If judged solely on its PHOTON efficiency, this device is one of the best inverters to walk through our laboratory doors. The maximum efficiency was measured at 96.1 percent. This result is slightly below the figure of 96.7 percent specified by Conergy; we also detected a 1-percent discrepancy in regard to European efficiency between the laboratory measurements (95.1 percent) and the data sheet (96 percent). ut what s more alarming is the product s maximum MPP voltage of 679 V. Conergy has reduced its original MPP voltage range from 220 to 750 V to 300 to 700 V, although this is not enough. Moreover, current limiting in the lower voltage range restricts the device s ability to fully process power, which is of less consequence on account of the narrower voltage range. The IPG 5000 vision, which does not have a grid separation transformer, is designed for system powers of 5 kw. The SI 3300 inverter is produced by Delta Energy Systems GmbH. This device s performance is refreshingly superior to much of the competition at both medium and high irradiation levels. Visually, it also makes a strong, positive impression. We were able to verify all of the manufacturer s specifications and we couldn t find a single point of complaint with this device. The PHOTON efficiency for medium irradiation is 93.9 percent, which is fairly high. The PV system s MPP can easily be placed in the medium MPP voltage range. The inverter doesn t have an overload reserve, and must therefore be precisely adjusted to the PV generator s demand, so that solar power peaks aren t wasted. Still, all in all, it s a solid, good device. Model: Aros Sirio 4000 PHOTON efficiency at medium irradiation: 95.1% PHOTON efficiency at high irradiation: 95.7% Detailed test report in PI 12/2008, p. 106 Model: Conergy s IP 5000 vision PHOTON efficiency at medium irradiation: 94.0% PHOTON efficiency at high irradiation: 94.7% Detailed test report in PI 7/2007, p. 96 Model: Delta Energy s SI 3300 PHOTON efficiency at medium irradiation: 93.9% PHOTON efficiency at high irradiation: 94.7% Detailed test report in PI 5/2008, p. 110 4 PHOTON Laboratory GmbH 2009

Diehl AKO Diehl AKO Fronius photon-pictures.com (6) Tested: Diehl AKO Platinum 4600S Tested: Diehl AKO Platinum 6300 TL Tested: Fronius IG 30 C A + D 92.9 % at medium irradiation 4/2008 96.8 % at medium irradiation 2/2009 91.4 % at medium irradiation 1/2007 The Diehl AKO Platinum 4600S operating range: 320 to 628 V; DC nominal power: 4.1 kw. The Diehl AKO Platinum 6300 TL s operating range: 350 to 710 V; DC nominal power: 5.7 kw. The Fronius IG 30 s operating range: 150 to 397 V; DC nominal power: 2.7 kw. The Platinum 4600S manufactured by Diehl AKO Stiftung & Co. KG ranks in the middle of the devices we ve tested thus far. Its PHOTON efficiency for medium and high irradiation levels is satisfactory. When used with thin-film modules, there are limitations on the MPP range due to higher open-circuit voltage of these modules. When setting the MPP of a PV system, it s best to select the lower MPP voltage. At nearly 120 percent, the inverter has a large overload range, together with a large thermal-time constant, and therefore allows for simpler PV system layouts in terms of power peaks. This inverter is one of the heavier models as it uses a 50 Hz toroidal core transformer. All in all, a reasonable device. The Platinum 6300 TL, a device with a 5,500 W AC nominal power, a 5,700 W DC nominal power and a 6,300 W maximum DC power, offers clean construction and operation, and an informative display. The conversion efficiency s dependence on voltage and power is very low. The overall efficiency is highly consistent across the specified voltage range. The result is a very high PHOTON efficiency. The device s large overload range of 112 percent also simplifies system layout and the cooling system s thermaltime constant does the rest: the inverter s temperature range is very wide. Its A+ PHOTON efficiency, along with its other advantages, makes the Platinum 6300 TL a leading device among the candidates we ve tested thus far. Conceived for systems with powers between 2.5 to 3.6 kw, this product belongs to a family of inverters equipped with transformers. Yet, it s strange that the IG 30 actually has three efficiency peaks that reach 94 percent at a single voltage point, namely 280 V. Since no PV system can consistently deliver this type of voltage, the device quickly drops to lower power levels. This also explains the device s D grade: the PHOTON efficiency for medium irradiation is about 91.37 percent. Nevertheless, this device is definitely beyond reproach when it comes to MPP voltage range. The inverter operates reliably between 150 and 400 V. It s only when the device is connected with thinfilm modules that it fails to completely exploit the upper voltage range. Fronius has no plans to change its device s voltage range.»the IG 30 was developed for use in an extremely broad voltage range,«the company writes. Model: Diehl AKO s Platinum 4600 S PHOTON efficiency at medium irradiation: 92.9% PHOTON efficiency at high irradiation: 93.3% Detailed test report in PI 4/2008, p. 96 Model: Diehl AKO s Platinum 6300 TL PHOTON efficiency at medium irradiation: 96.8% + PHOTON efficiency at high irradiation: 96.9% + Detailed test report in PI 2/2009, p. 108 Model: Fronius IG 30 PHOTON efficiency at medium irradiation: 91.4% Grade: D PHOTON efficiency at high irradiation: 92.2% Detailed test report in PI 1/2007, p. 96 PHOTON Laboratory GmbH 2009 5

Overview of inverters tested by PHOTON Laboratory until July 2009 (continued) Fronius Ingeteam Kaco New Energy Tested: Fronius IG Plus 50 Tested: Ingeteam Ingecon Sun 3.3 TL Tested: Kaco Powador 2500xi C C 94.5 % at medium irradiation 8/2008 93.4 % at medium irradiation 8/2007 92.5 % at medium irradiation 12/2007 The Fronius IG 50 s operating range: 230 to 500 V; DC nominal power: 4.2 kw. The Ingeteam Ingecon Sun 3.3 TL s operating range: 159 to 414 V; DC nominal power: 3.5 kw. The Kaco Powador 2500xi s operating range: 350 to 597 V; DC nominal power: 2.7 kw. The IG Plus 50 is indisputably a very good inverter with its 94.5-percent PHOTON efficiency. The inverter s peak efficiency value only drops slightly across all specified voltages and powers. Overall, the IG Plus 50 makes a consistent impression. Its modular construction and flexibility make it ideal for most imaginable system configurations today that s a big bonus. The upper two-thirds of the specified MPP voltage range can be selected when setting a PV system s MPP, since this is the range in which the inverter s efficiency is highest. Moreover, operators need not worry about the inverter heating up the device s temperature range is very broad. Potential buyers don t have to worry about their system yields with this device, as long as they properly set its parameters. Model: Fronius IG Plus 50 PHOTON efficiency at medium irradiation: 94.5% PHOTON efficiency at high irradiation: 94.8% Detailed test report in PI 8/2008, p. 194 Spanish company Ingeteams product, the Ingecon Sun 3.3 TL, received a C grade for PHOTON efficiency for medium irradiation levels, and a grade for higher levels. It s designed for a 3.3 kw system, and is not equipped with a grid separation transformer. Our test measured the device s maximum efficiency at nearly 96.5 percent, and its European efficiency at exactly 95.1 percent. This inverter s weakness is that it reaches its highest efficiency of 96.3 percent at an MPP voltage of 380 V, a value that already hovers at the level where this device could destroy itself. Originally, Ingeteam said it designed the Ingecon Sun 3.3 TL for an MPP voltage range of 125 to 450 V. In the meantime, Ingeteam has altered its voltage range specifications; in consequence the PHOTON grade improved the manufacturer now lists the inverter s range as 160 to 415 V. Model: Ingeteam s Ingecon Sun 3.3 TL PHOTON efficiency at medium irradiation: 93.4% PHOTON efficiency at high irradiation: 94.3% Detailed test report in PI 8/2007, p. 128 The Powador 2500xi works best at lower MPP voltages. At the minimum value of 350 V, it almost achieves the manufacturer s maximum efficiency specification of 95.6 percent, and at the same voltage it even exceeds the specified European efficiency of 94.3 percent. oth represent good values. Across the entire voltage range, which spans a broad section between 350 and 600 V, the inverter s efficiency ranges between 93 and 92 percent, which gives the device a C grade for both the middle and high irradiation levels. At no point is there danger that the inverter s MPP voltage will come critically close to the maximum MPP voltage. Kaco chose not to alter the device s MPP voltage range. With an overload tolerance of 114 percent the device is sturdy and can absorb power peaks. As a member of the Powador family, the device is a transformerless inverter that covers input powers of 2.7 to 5.7 kw. Model: Kaco s Powador 2500xi PHOTON efficiency at medium irradiation: 92.5% PHOTON efficiency at high irradiation: 93.4% Detailed test report in PI 12/2007, p. 122 6 PHOTON Laboratory GmbH 2009

Kostal Solar Electric Mitsubishi Electric Phoenixtec Power photon-pictures.com (6) Tested: Kostal Piko 10.1 Tested: Mitsubishi PV-PNS06ATL-GER Tested: Phoenixtec PVG 2800 94 % at medium irradiation 7/2009 93.9 % at medium irradiation 6/2008 94.4 % at medium irradiation 5/2008 The Kostal Piko Sun 10.1 s operating range: 400 to 850 V; DC nominal power: 9.65 kw. The Mitsubishi PV-PNS06ATL-GER s operating range: 260 to 650 V; DC nominal power: 4.8 kw. The Phoenixtec PVG 2800 s operating range: 250 to 450 V; DC nominal power: 2.9 kw. The Piko 10.1 was the first device from Kostal Solar Electric to be tested by PHO- TON Lab. The multi-tracker device has three MPP trackers that can control the MPP of connected module strings independently of each another. Therefore, the MPP trackers can be operated asymmetrically. The PHOTON efficiency was recorded when the three MPP trackers were under symmetrical loads. At 94 percent for medium irradiation, the device received a grade. The only negative aspect is the MPP voltage range: at 400 to 850 V, the range is wide, but the distance from the maximum DC voltage is not sufficient. Therefore, the device cannot be used fully with crystalline and thin-film modules. The PV-PNS06ATL-GER is a PV inverter produced by Japanese company Mitsubishi Electric Corp. The device, with its unpronounceable name, functions efficiently and reliably. Above all, the good adjustment efficiency of this transformerless and relatively light inverter should be underscored. The device fails to exploit only a fraction of the power from the generator. Thus, the values for conversion and overall efficiency differ only slightly from each another. The maximum conversion efficiency is 96.13 percent and the PHOTON efficiency is 93.9 percent. When adjusting the MPP of a PV system, it s best to use the middle of the MPP voltage range. The PV-PNS- 06ATL-GER is a good device and lands in the medium range of those inverters already tested by PHOTON Lab. An old acquaintance resurfaced in PHO- TON Lab after a long absence: the PVG 2800 from manufacturer Phoenixtec Power Co. Ltd. The device has a slightly improved conversion efficiency of maximum 96.7 percent. At 94.4 percent, the PHOTON efficiency for medium irradiation is good. The PHOTON efficiency for high irradiation is very good at 95.1 percent, as a result of its very consistent overall efficiency curve at higher powers and a corresponding higher weighting. The around 2-percentage point difference between the conversion efficiency and the PHOTON efficiency emphasizes the dependency on voltage and power. Therefore, when designing a PV system, one should select the middle of the MPP voltage range. Model: Kostal s Piko 10.1 PHOTON efficiency at medium irradiation: 94% PHOTON efficiency at high irradiation: 94.4% Detailed test report in PI 7/2009, p. 122 Model: Mitsubishi s PV-PNS06ATL-GER PHOTON efficiency at medium irradiation: 93.9% PHOTON efficiency at high irradiation: 94.6 % Detailed test report in PI 6/2008, p. 148 Model: Phoenixtec s PVG 2800 (updated model) PHOTON efficiency at medium irradiation: 94.4% PHOTON efficiency at high irradiation: 95.1 % Detailed test report in PI 2/2008, p. 108 and PI 5/2008, p. 118 PHOTON Laboratory GmbH 2009 7

Overview of inverters tested by PHOTON Laboratory until July 2009 (continued) Power-One Refu Elektronik Riello UPS Tested: Power-One Aurora PVI-6000-OUTD-S Tested: Refusol 11 K Tested: Riello HP 4065REL A A + D 95.4 % at medium irradiation 3/2009 96.9 % at medium irradiation 9/2008 91.7 % at medium irradiation 9/2007 The Power-One Aurora PVI-6000-OUTD-S operating range: 180 to 530 V; DC nominal power: 6.2 kw. The Refusol 11 K s operating range: 380 to 800 V; DC nominal power: 11 kw. The Riello HP 4065REL s operating range: 255 to 435 V; DC nominal power: 4 kw. The transformerless Aurora PVI-6000- OUTD-S inverter, produced by Power- One SpA, was the first multi-tracker inverter PHOTON Lab tested: the two independent DC inputs, each with an MPP tracker for locating the optimal working point, allow for the connection of two different solar generators. When using the Aurora device, these generators can be different in size, since the inverter allows for asynchronous loads. Thus, the PVI-6000-OUTD-S achieved a 95.4-percent PHOTON efficiency for medium irradiation. The device uses convection cooling, which means users have many options when it comes to selecting a location for the Aurora PVI-6000-OUTD- S. Above all, the multi-tracker concept, together with its very good efficiency, makes this device a profoundly versatile unit. A top efficiency, compact construction and an unusually stable mode of operation: with its Refusol 11 K, Refu Elektronik has produced an excellent device. And one thing shouldn t be forgotten: this is the first inverter Refu Elektronik has developed. We can only guess the kind of potential this company has after the successful debut of this device on the PV market. The inverter s maximum conversion efficiency is a sleek 98 percent. Only one other unit has performed this well in the lab: SMA s SMC 8000 TL showed similarly good performance. The Refusol 11 K obtains a PHOTON efficiency of 96.9 percent for medium irradiation and an A+ grade. This inverter is top class and our staff would like to see more devices like this. The editorial team labeled this device as wasteful, on account of the fact that it can only process 100 percent of produced DC current once input voltage reaches 225 V. The Helios Power HP 4065 was originally designed for a voltage range of between 150 and 450 V. The manufacturer, Italian company Riello UPS, has now adjusted this range to 250 to 450 V. Still, this range isn t optimal, which the white hatching in the diagram demonstrates. ut because of the current limitation, and the device s good efficiency, its grade is still better than one would expect. At high irradiation levels the device s PHO- TON efficiency is almost 94 percent, giving it a grade; at medium irradiation levels the efficiency is 91.7. The transformerless device achieved very good results in maximum efficiency and European efficiency measurements. Model: Power-One s Aurora PVI-6000-OUTD-S PHOTON efficiency at medium irradiation: 95.4% PHOTON efficiency at high irradiation: 95.9% Detailed test report in PI 3/2009, p. 122 Model: Refusol s 11 K PHOTON efficiency at medium irradiation: 96.9% + PHOTON efficiency at high irradiation: 97.2% + Detailed test report in PI 9/2008, p. 88 Model: Riello s HP 4065REL PHOTON efficiency at medium irradiation: 91.7% Grade: D PHOTON efficiency at high irradiation: 93.9 % Detailed test report in PI 9/2007, p. 150 8 PHOTON Laboratory GmbH 2009

SMA Solar Technology SMA Solar Technology SMA Solar Technology photon-pictures.com (6) Tested: SMA S 2100TL Tested: SMA S 3800 Tested: SMA S 5000TL-20 C A 93.7 % at medium irradiation 6/2009 93.2 % at medium irradiation 2/2007 95.7 % at medium irradiation 5/2009 The SMA S 2100TL s operating range: 200 to 480 V; DC nominal power: 2 kw. The SMA S 3800 s operating range: 208 to 395 V; DC nominal power: 4 kw. The SMA S 5000TL-20 s operating range: 175 to 440 V; DC nominal power: 4.8 kw. The smallest transformerless inverter in the»sunny oy«family was released in 2002, and is based on the concept of SMA s first transformerless device, which was released in 1999. In the test, the conversion efficiency reached its highest level of 96 percent in the upper half of the MPP voltage range. The conversion efficiency s dependence on voltage is relatively low, while the dependence on power is rather high. The situation is similar for the overall efficiency (the product of conversion efficiency and MPPT adaptation efficiency), which reaches a maximum of 95.9 percent. A weak point is the display of inverter output power: it is very imprecise, particularly at lower powers. All in all, the inverter achieved a good grade, and takes its place in the middle field of all the inverters tested thus far. Model: SMA s S 2100TL PHOTON effi ciency at medium irradiation: 93.7% PHOTON efficiency at high irradiation: 94.6% Detailed test report in PI 6/2009, p. 144 Listing the maximum MPP voltage as 500 V with a maximum input voltage of 500 V is most certainly not common practice and the documentation should be revised,«wrote PHOTON in February 2007. Et voilà manufacturer SMA listened. The most recent data sheet lists a voltage range of 200 to 400 V. Originally, the inverter was designed for input voltages of up to 500 V, but could not be operated above 450 V without the causing harm to the device. The new values ensure that the inverter can be operated safely with all crystalline modules; only combination with thin-film modules can cause problems for the inverter in the upper voltage range. In our test, the device reached a European efficiency of 94.7 percent. The PHOTON efficiency is slightly lower at 93.2 percent in medium irradiation and 93.6 percent at high irradiation. The device is suited for input powers of up to around 4 kw. Model: SMA s S 3800 PHOTON efficiency at medium irradiation: 93.2% PHOTON efficiency at high irradiation: 93.6% Detailed test report in PI 2/2007, p. 104 The SMA S 5000TL-20 was the second multi-tracker device to make it to the inverter test bench. oth of the SMA device s DC inputs have their own MPP tracker searching for the best operating point, and both inputs can accept different loads. The symmetrical operating mode was used for grading purposes. The S 5000TL-20, a very well-organized and compact device, reaches a maximum conversion efficiency of 97 percent. This result is barely dependent on voltage, but just slightly more dependent on power. All in all, the device adjusts very well to the MPP across the entire operating range and therefore has a high overall efficiency. The result was a PHOTON efficiency of 95.7 percent for medium irradiation. Model: SMA s S 5000TL-20 PHOTON efficiency at medium irradiation: 95.7% PHOTON efficiency at high irradiation: 96% Detailed test report in PI 5/2009, p. 114 PHOTON März Laboratory 2009 GmbH 2009 9

Overview of inverters tested by PHOTON Laboratory until July 2009 (continued) SMA Solar Technology Solon Inverters Sunways Tested: SMA SMC 8000TL Tested: Solon Satis 40/750 IT Tested: Sunways AT 4500 A + C 96.9 % at medium irradiation 10/2007 92.3 % at medium irradiation 11/2008 94.6 % at medium irradiation 7/2008 The SMA SMC 8000TL s operating range: 335 to 487 V; DC nominal power: 8.3 kw. The Solon Satis 40/750 IT s operating range: 375 to 575 V; DC nominal power: 4 kw. The Sunways AT 4500 s operating range: 250 to 600 V; DC nominal power: 4.7 kw. With a PHOTON efficiency of almost 97 percent for medium and high solar irradiation levels, the SMC 8000 TL sets itself so far apart from most of the competition that it requires an additional plus signs. The inverter operates soundly and exploits its entire specified MPP voltage range of 335 to 500 V. This makes it easy to dimension the device for PV systems. Its display is precise and the temperature range comfortably wide. Our test confirmed the device s maximum efficiency at nearly 98 percent; test results showed that the European efficiency was just shy of the impressive benchmark of 97.7 percent. The inverter is a transformerless device designed for systems with more than 20 kw. According to Solon, the device that we tested was a prototype that is only being sold in small batches. The conversion efficiency is at its highest 95.8 percent at lower MPP voltages. The PHOTON efficiency for medium irradiation is rather low just 92.3 percent. Thus, when setting a PV system s MPP, only the lower portion of the specified MPP voltage should be selected. The inverter doesn t have an overload range, but rather displays a slight tendency to limit power even at nominal power. Th is sensitivity shown at lower powers doesn t have a large effect on annual yield, but the fact that it shuts on and off could unnecessarily decrease the device s lifespan. Potential buyers should keep that in mind before purchasing this unit. Model: Solon s Satis 40/750 IT The AT 4500 is a transformerless device from Sunways Photovoltaic Technology AG. The PHOTON efficiency calculated across all efficiencies at different voltages and powers is a very high 94.6 percent. When adjusting a PV system s MPP, the middle voltage range should be selected. And another tip: since the inverter doesn t have overload capabilities, the PV generator s power should be adapted as precisely as possible to the inverter s power. The only weakness worth mentioning is the AT 4500 s display and the values it shows. Although very informative and clear, it s too inaccurate for a precise comparison of system yield values. Model: SMA s SMC 8000TL PHOTON effi ciency at medium irradiation: 96.9% + PHOTON effi ciency at high irradiation: 97% + Detailed test report in PI 10/2007, p. 106 PHOTON efficiency at medium irradiation: 92.3% PHOTON efficiency at high irradiation: 93.5% Detailed test report in PI 11/2008, p. 134 Model: Sunways AT 4500 PHOTON efficiency at medium irradiation: 94.6% PHOTON efficiency at high irradiation: 94.8% Detailed test report in PI 7/2008, p. 88 10 PHOTON Laboratory GmbH 2009

Sunways Sunways Xantrex Technology photon-pictures.com (5) Tested: Sunways NT 2600 (upper range) Tested: Sunways NT 2600 (lower range) Tested: Xantrex GT5.0SP C 92.3 % at medium irradiation 11/2007 93.8 % at medium irradiation 11/2007 94.1 % at medium irradiation 1/2009 The Sunways NT 2600 s (upper range) operating range: 476 to 749 V; DC nominal power: 2.6 kw. The Sunways NT 2600 s (lower range) operating range: 350 to 623 V; DC nominal power: 2.6 kw. The Xantrex GT5.0SP s operating range: 240 to 550 V; DC nominal power: 5.3 kw. With the NT 2600, German company Sunways AG has created an inverter that achieves an A/ grade at high irradiation levels, and a /C grade for medium irradiation levels. Since the device is equipped with a sliding switch, which allows users to adjust the voltage range for»voltages lower than or equal to 630 V«or»higher than 630 V,«the inverter was tested twice once for the upper voltage range between 476 and 750 V, and once for 350 to 620 V. The results show that the device has a better efficiency for the lower voltage range, and laboratory measurements confirmed the manufacturer s maximum efficiency at 97 percent and European efficiency at 96.5 percent. Hence, the entire range can be exploited with no limitations. In the upper voltage range, the inverter s highest efficiency is at the lower end of the scale. At the upper end of the scale, the hatching shows the critical point at which the maximum MPP voltage is no longer adequately distanced from the maximum input voltage. The GT5.0SP inverter is produced by Xantrex Technology Inc. With a PHO- TON efficiency of 94.1 percent, the GT5.0SP achieves a grade. A distinguishing feature of this device, which doesn t use a fan, is the cooling element located on its front. Its weaknesses emerge at higher voltages and lower powers. All in all, the GT5.0SP shows very consistent operation and despite its rather average maximum conversion efficiency and broad voltage range deserves its grade PHOTON efficiency. The device s consistency facilitates the layout of PV systems, since almost the entire MPP voltage range of the inverter can be selected. The solar generator s power, on the other hand, should be precisely adjusted to accommodate the inverter, since the device doesn t have any overload capabilities. Model: Sunways NT 2600 (upper range) PHOTON efficiency at medium irradiation: 92.3% PHOTON efficiency at high irradiation: 93.9% Detailed test report in PI 11/2007, p. 126 Model: Sunways NT 2600 (lower range) PHOTON efficiency at medium irradiation: 93.8% PHOTON efficiency at high irradiation: 95.1% Detailed test report in PI 11/2007, p. 126 Model: Xantrex s GT5.0SP PHOTON efficiency at medium irradiation: 94.1% PHOTON efficiency at high irradiation: 94.7% Detailed test report in PI 1/2009, p. 142 PHOTON Laboratory GmbH 2009 11

this range, in which case the efficiency is then listed as 0 percent. This reflects the conditions of a real PV system: after all, if an inverter had to face these conditions, it would cease operating properly. The graphical representation shows these areas. For instance, the color diagram included in our inverter tests shows the inverter s efficiency, the MPPT adjustment efficiency, and the overall efficiency. The diagram is colored black if the maximum MPP voltage isn t adequately distanced from the inverter s maximum DC voltage, and if it doesn t have an active overload limit according to the manufacturer, which means no measurements can be conducted in this range, since the MPP tracker won t operate properly. The diagram also reflects the DC current limitation range. These black areas, which reflect a value of zero, are used to calculate an average in the new PHOTON grading system, and therefore have a strong influence on the grade. The resulting effect is desired, and a consequence of considerations about the inverter s actual, useable MPP range: an inverter will only get a good grade in the test if it actually can operate without limitations in the voltage range specified by the manufacturer. Finally, manufacturers who change their product data to reflect a more sensible MPP range will receive an improvement in their grade. The color diagram also includes white, hatched areas. These represent areas in the MPP voltage range that are considered critical when designing a PV system. They are located in the upper end of the MPP range. There are two types of hatching marks. The diagonal upward lines represent an MPP range in which the V MPPmax is generally absent for PV systems with crystalline modules. The hatching marks in the other direction (i.e. sloping diagonally downward) also represent the MPP range in which the V MPPmax is generally absent for PV systems with thin-film modules. The exact definition of these limits can be established when designing a system with actual modules. Hatching can also be seen in the lower portion of the MPP range. This denotes the area in which the activation of the DC current limitation prevents the inverter from feeding 100 percent of available DC power P MPP into the grid. A PV system s V MPP shouldn t be located in this range either, since that would result in a yield loss. The result of all of this is an efficiency number that is generally lower than the European efficiency, since this is usually measured at the»best«voltage levels, and does not take into account mismatching and unreliable operating ranges. That means that PHOTON s efficiency can make an inverter look like it will fair worse than its true performance in a real PV system, since it takes into consideration the entire input voltage range specified by the manufacturer regardless of whether that range will actually be exploited by a particular PV system. Hence PHOTON s efficiency tells us something about the least you can expect from an inverter and gives information for all system configurations that operate within the input voltage range specified by the manufacturer. Text Heinz Neuenstein, Ines Rutschmann Our offer for manufacturers and installers Admitting new devices It all began when system operators and installers began to ask us which inverters functioned best. Above all, the goal of the inverter test is to meet the interests of our readers. For this reason,a portion of the devices we test are purchased on the open market anonymously. We gladly accept inverter test requests from our readers you can simply send an informal email to Martina Siebmanns, the CEO of PHOTON Laboratory GmbH. On the other hand, we would also be delighted if manufacturers provided us with inverters naturally, preferably their latest products. Naturally, we conduct the measurements free of charge for manufacturers. For the test, we require a serially produced device from ongoing production that hasn t been modified for the test. We can gladly provide details on theprecise test conditions upon request. Results from the measurements are sent to the manufacturer prior to publication. This also applies to modules purchased anonymously. Due to increased demand on the part of manufacturers, there is already a short waiting list. We therefore suggest that you contact us as soon as possible if you are interested. At the same time, we are currently expanding PHOTON s test laboratory, so that we will be able to test several devices per month in the future. Exchange of experience and further development Once a year, PHOTON invites inverter manufacturers to an event in Aachen at which they can exchange experiences, in order to adapt test conditions to current developments. The most recent of these meetings took place in November 2008. Eighteen representatives participated in the meeting. Among other questions, the attendees discussed whether the MPP adjustment efficiency should be measured statically or dynamically, as well as how the test laboratory could appraise inverter noise emission going forward. A set of minutes is produced in conjunction with these meetings and is made available to meeting participants, as well as our lab, for use in developing test criteria. Generally, the invitation to this event is sent by PHOTON Lab to all known inverter manufacturers 4 weeks in advance. If requested, simultaneous English translation will be provided. The program always includes a visit to the lab. PHOTON Laboratory GmbH CEO Martina Siebmanns Jülicher Strasse 376 phone +49/241/4003-107, fax -307 martina.siebmanns@photon.de 12 PHOTON Laboratory GmbH 2009