In-Situ Measurement of High Frequency Emission Caused by Photo Voltaic Inverters

Cees Keyer1,2, Roelof Timens2, Frits Buesink2, Frank Leferink2,3 1 Amsterdam University of Applied Science, dept. Electrical Engineering, Amsterdam, the Netherlands

c.h.a.keyer@hva.nl 2 University of Twente, Enschede, the Netherlands.

3 Thales Nederland, Hengelo, the Netherlands.

AbstractModern photovoltaic (PV) installations are being

installed in large numbers. Manufacturers are using different harmonized standards to show compliance with the legislative requirements such as the European EMC Directive. Their selection of harmonized standards is not based on adherence to (essential requirements of) the European EMC Directive, but more on what is most convenient, i.e. lowest cost; Some manufacturers install an Ethernet interface and then claim the PV inverter is a computer and use standards applicable for IT (information technology) equipment. Furthermore no limits have been published for the DC bus, which acts, due to its large length, as an antenna. The experimentally obtained data presented in this paper show that PV inverters are causing serious interference, and thus do not fulfil (the essential) requirements of the EMC Directive.

Index: EMC Directive, Solar Panels, PhotoVoltaic inverters, interference Ham Radio.

I. INTRODUCTION The number of complaints [1] of unwanted emission,

causing interference in radio reception due to PV installations, is increasing rapidly. Alternative energy sources are heavily subsidized and growing in numbers. Grid connected inverters have to comply with international, harmonized, standards in the European community. From a first check it appeared that some suppliers use a creative way to obtain the simplest ElectroMagnetic Compatibility (EMC) requirements. These suppliers are effectively misusing product standards for other equipment than it should be used for.

In the paper by Hamza et al. [2] a grid tied PV system classified as fixed installation is investigated. The conducted EMI measurement results show significant emissions generated by the inverter module propagating on the DC side of the PV system. Large PV installations are investigated in the paper published by Areneo et al. [3]. Experimental results show that the DC cabling is forming conductive coupling paths and may also become an antenna. In the paper by Piazza [4] et al it is shown that DC cabling are perfect antennas. RF interaction and unwanted emission measurement results are shown in this paper. In papers [2,3] high power >10 kW PV installations are investigated, in [4] a low power < 3kW is investigated.

We performed in situ measurements to investigate interference caused by three different inverters for low power energy conversion, < 3kW, in residential environments, and we

compared the declarations of conformity from the three suppliers.

II. EUROPEAN EMC DIRECTIVE Preamble 2 of the European EMC Directive [3] states: Member States are responsible for ensuring that radio-communications, including radio broadcast reception and the amateur radio service operating in accordance with International Telecommunication Union (ITU) radio regulations, electrical supply networks and telecommu-nications networks, as well as equipment connected thereto, are protected against electromagnetic disturbance. Preamble 13 of this directive states: Once the reference to such a standard has been published in the Official Journal of the European Union, compliance with it should raise a presumption of conformity with the relevant essential requirements, although other means of demonstrating such conformity should be permitted.

This means that national agencies are responsible for controlling and checking that products fulfil the essential requirements (Article 5), and if they do not fulfil, take them of the market. However, the Dutch law enforcer, Agentschap Telecom, does only focus on checking if apparatus are fulfilling harmonized standards, and not considering violations of the essential requirements of the Directive. Actually, in nearly all European countries the national agencies, responsible for enforcing the law, are not able to stop violation of the essential requirements and immoral suppliers due to budget constraints;

A case where Article 5 (essential requirements) and Article 6 (presumption of conformity based on harmonized standards) of the European EMC Directive have been applied in a proper way has been published on television by the German West Deutsches Rundfunk [7].

Problems with market surveillance are known, [8, 9] but the differences between member states in the EU are huge.

III. PV INSTALLATION AND MEASUREMENT SET UP A PV installation consists of photovoltaic cells, or solar

panels, connected to a DC bus. This DC voltage is converted to a 50 Hz sinusoidal voltage, which is fed into the grid, by means of a PV inverter. The PV inverter thus converts solar

978-1-4799-3226-9/14/$31.00 2014 IEEE

Proc. of the 2014 International Symposium on Electromagnetic Compatibility (EMC Europe 2014), Gothenburg, Sweden, September 1-4, 2014

74

power into grid power. Local Dutch rules, NEN 1010, for grid connected installations in the Netherlands allow a low power 600 VA, to be connected straight in to a socket without special measures. PV inverters which can supply more power > 600 VA should be connected on a separate feeder line with fuse. No other equipment is allowed to connect on that feeder line. Dutch standard (NEN 1010) on low voltage (230 V) grid systems requires this due to the fact that if you deliver power back into the grid the current flows through the fuse. When equipment is connected on the same fused line, for example, 16 A fuse, and the PV installation is delivering 10 A from your PV inverter, the equipment should use a fault current of 26 A, to blow the fuse out, which is undesirable.

A current probe clamp, Fischer Custom Communications F-35 and a Rigol DSA 815 spectrum analyzer, as shown in Figure 1, were used for the measurements. Although different measurement methods are described for CM mode current measurements [2, 5] on the DC bus, in this case there was no possibility to open up the DC bus for connecting a Line Impedance Stabilization network.

The measurement setup for all three inverters was the same. The DC bus is clamped with both wires to prohibit DC saturation of the clamp, and to measure the high-frequency common mode (CM) current. The measurements on the DC bus are done with the inverter switch on and switched off. This is done to check if the inverter causes the CM current, or not. The CM current loop is created by the actual installation so measurements have to be performed in-situ.

The first measurement was performed on Friday January 31st at 08:00 UTC, in Schagen, the Netherlands.

Figure 1: Measurement set up.

Figure 2: Soladin 600 emissions with 30 mtr cable

Figure 3: PV cells with Ham Radio antennas

A ham radio operator complained about a raised noise floor and interference on the reception of other hams.

The PV installation, shown in Figure 3, used approximately 30 meters of DC cable, which behaved like a perfect tuned antenna for the HF (High Frequency) band. The emissions on both the DC as the AC grid connection were measured more than a year ago to diagnose the interference problems from this installation. The results, which is a snapshot of that situation, are shown in Figure 2. We tried to find the source of the disturbance the ham operator complaint about. The cables were shortened to a few meters to reduce the efficiency of the unwanted antenna (the DC cable). This action reduced the interference drastically but not enough.

Proc. of the 2014 International Symposium on Electromagnetic Compatibility (EMC Europe 2014), Gothenburg, Sweden, September 1-4, 2014

75

Figure 4: SMA sunnyboy (red) and Soladin 600 (blue one)

IV. INTERFERENCE ON HF BAND DUE TO DC BUS CONDUCTED EMISSIONS.

Emission measurements on the DC cables were performed. The ham radio operator uses two different inverters, two from SMA (type: Sunny Boy), and one from Mastervolt, (type Soladin 600), as shown in Figure 4. Measurements have been performed just after sunrise, so the power delivered to the grid was quite low. The results of both inverters are shown in Figure 5 and Figure 6.

Up to 5 MHz the emission from the SMA inverter is 10 to 20 dB lower than that from the Soladin inverter. Above 5 MHz the emission from the SMA inverter is of the same order as that from the Soladin inverter. For some frequencies in this range, for example 10 MHz and 50 MHz, the emission from Soladin is significantly higher. When listening to the 10 MHz ham radio band, the noise divergence is quite huge.The SMA Sunny boy is approximately 10 to 20 dB quieter.

Another Soladin 600 was measured at another location. The measurement was performed on a bright sunny morning, and the power delivered to the grid was increased in comparison to the previous one. The inverter was mounted on a mobile rack and was directed to the sun as shown in Figure 7. A picture of the measurement results is shown in Figure 8, scale is in dBV. The length of the length DC cables is 2 meters. The measurements did show an increase in higher frequency emissions on the DC cable compared to the first measurement that morning.

The third inverter measured is an Omnik inverter mounted in an average family house with 12 PV panels on the roof, as shown in Figure 9. Two DC bus connections where present so 6 panels per bus.

The Omnik inverter was mounted in the attic. The house is not occupied, just used for storing equipment and furniture. So any background interference from other grid connected equipment is minimized on the local grid.

Figure 5: Low sun power measurement Soladin.

Figure 6: Soladin (blue) SMA (red) spectrum measurements

Figure 7: Movable PV array, Soladin mounted on the back

Proc. of the 2014 International Symposium on Electromagnetic Compatibility (EMC Europe 2014), Gothenburg, Sweden, September 1-4, 2014

76

Figure 8: Measurement with more sun illumination.

Figure 9: 12 panels on a roof with the Omnik inverter.

On the bottom, Figure 10, of this inverter are two black/red cables for the DC bus, one 230 Vac cable (greyish) and an antenna for WIFI. The length of the DC cables is unknown. The spectrum measured is shown in Figure 11, and is considerably, approximately 10-20 dB, lower than the Soladin 600.

Figure 10: Omnik inverter.

Figure 11: Spectrum of the Omnik inverter.

V. COMPARISON OF THE DECLARATION OF CONFORMITY

The declarations of conformity [10, 11, 12] lists the standards used for testing. The emission standards are listed in Table 1.

TABLE I: LIST OF EMISSION STANDARDS. Soladin 600 Omnik 3K SMA Sunny Boy IEC 55022 IEC 61000-6-3 IEC 61000-6-3 IEC 61000-3-2 IEC 61000-6-4 IEC 61000-6-4 IEC 61000-3-12 IEC 55022 IEC 61000-3-11

Both the SMA Sunny Boy and Omnik inverters are tested against the generic emission standards, IEC 61000-6-3 and IEC61000-6-4. Hence the DC bus is tested as well. The Soladin 600, however, is tested against the IEC 61000-3-2 and IEC 55022 standards. The IEC 61000-3-2 is for harmonic distortion on the grid, for apparatus with a rated current up to 16A. IEC 61000-3-12 is for harmonic distortion on the grid for apparatus with a rated current between 16 and 75 A. IEC 61000-3-11 is for voltage fluctuations flicker and voltage changes. Comparison of the Declarations of conformity [10,11,12] shows that the Soladin 600, and SMA Sunny Boy are tested against EN55022 (CISPR 22). This is the product standard for information technology (IT) equipment the SMA inverter contains a wifi adapter hence CISPR 22. In IEC 61000-6-3 limits are described for a DC power bus. In the definitions of this standard it is stated: 3.5 power port port at which a conductor or cable carrying the primary electrical power needed for the operation (functioning) of an apparatus or associated apparatus is connected to the apparatus. Hereinafter in the standard is stated: 3.8 d.c. power network local electricity supply network in the infrastructure of a certain site or building intended for flexible use by one or more different types of equipment and guaranteeing continuous power supply independently from the conditions of the public mains network[13]. Hence a DC bus of a PV installation is not a DC bus within the scope of this standard.

VI. MIS-USE OF ARTICLE 6 All apparatus being put on the European market shall fulfil the requirements of the EMC Directive. A common approach is to follow Article 6 stating that presumption of conformity is raised when harmonized standards have been used. These

Proc. of the 2014 International Symposium on Electromagnetic Compatibility (EMC Europe 2014), Gothenburg, Sweden, September 1-4, 2014

77

harmonized standards are used for testing the equipment. If the equipment passes the test the manufacturer has a presumption of complying to the EMC Directive as stated in Article 6 of the directive 2004/108/EC [6]. The manufacturer of the apparatus can select the applicable harmonized standards. Due to the fact an Ethernet port has been integrated in this inverter the manufacturer of the Soladin inverter states that it is a computer with some special features. No limits on conducted emission from the DC bus are defined for IT equipment. If the reader looks in the other 2 Declarations of Conformity there is no reference towards any standard regarding the DC bus. The SMA and Omnik use the generic standards from the 61000-6 series to test their equipment so the DC side is tested as well. The Soladin 600 is only tested on the AC side (230 V) due to the fact that standard IEC 55022 does not mention a DC bus. This shows a large caveat in the standards which can be used for PV inverters, because the measurement results shows that the ham radio operators are seriously interfered. This can be a nice court case where the suppliers of (all?) PV inverters are summoned because of violating the essential requirements of the EMC Directive. But who will start this court case?

VII. CONCLUSION DC bus emissions of PV installations are not described in international harmonized standards. So the manufacturer is free to do what is best for his wallet. Even the general standards do not apply if the PV inverter is equipped with an Ethernet port, as then the manufacturer can state it is a computer with special functions. Manufacturers search for creative ways to get their product on the market with disregarding the rights of the radio spectrum users. The currents can cause high radiated interference levels. Actually, the suppliers are violating the essential requirements of the European EMC Directive. Due to budget constraints of the national authorities the PV inverter suppliers can continue their illegal activities.

ACKNOWLEDGMENT We would like to thank the ham radio operator PA5KK for leading us in to these measurements.

Figure 12: The "shack"of PA5KK

REFERENCES

[1] Dutch Ham Radio society VERON, EMC-EMF Committee, 1st author is the Secratery.

[2] Hamza, D.; Jain, P., "Conducted EMI in grid-tied PV system," Telecommunications Energy Conference (INTELEC), 32nd International , vol., no., pp.1,7, 6-10 June 2010

[3] Araneo, R.; Lammens, S.; Grossi, M.; Bertone, S., "EMC Issues in High-Power Grid-Connected Photovoltaic Plants," Electromagnetic Compatibility, IEEE Transactions on , vol.51, no.3, pp.639,648, Aug. 2009

[4] Di Piazza, M.C.; Serporta, C.; Tine, G.; Vitale, G., "Electromagnetic compatibility characterisation of the DC side in a low power photovoltaic plant," Industrial Technology, 2004. IEEE ICIT '04. 2004 IEEE International Conference on , vol.2, no., pp.672,677 Vol. 2, 8-10 Dec. 2004

[5] H. Haberlin, New DC-LISN for EMC Measurements on the DC side of PV systems: Realisation and first measurements at inverters 17th European Photovoltaic Energy Conference, Munich Germany, 2001

[6] European EMC Directive 2004/108/EC. http://eurlex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32004L0108:en:NOT , checked February 22nd 2014

[7] http://www.youtube.com/watch?v=-BZSXKxyJn0, accessed 22nd Feb 2014

[8] Rajamaki, J., "Correlations between EMI statistics and EMC market surveillance in Finland," Electromagnetic Compatibility, 2004. EMC 2004. 2004 InternationalSymposium on , vol.2, no., pp.649,654 vol.2, 9-13 Aug. 2004

[9] Rajamaki, J., "Finnish safety and EMC market surveillance statistics," Electromagnetic Compatibility, 2002. EMC 2002. IEEE International Symposium on , vol.2, no., pp.686,691 vol.2, 19-23 Aug. 2002

[10] Mastervolt Soladin 600 user manual, found by google. Accessed 22nd feb 2014.

[11] SMA Sunny Boy sales leaflet. found by Google. Accessed 22nd Feb. 2014

[12] Omnik user manual, found by Google. Accessed 22nd Feb. 2014 [13] IEC 61000-6-3, Electromagnetic compatibility (EMC) - Part 6-3:Generic

Standards - Emission standard for residential, commercial and light-industrial environments (IEC 61000-6-3:2006,IDT)

Proc. of the 2014 International Symposium on Electromagnetic Compatibility (EMC Europe 2014), Gothenburg, Sweden, September 1-4, 2014

78

Welcome PageHub PageSession ListTable of Contents Entry of this ManuscriptBrief Author IndexABCDEFGHIJKLMNOPQRSTUVWXYZ

Detailed Author IndexABCDEFGHIJKLMNOPQRSTUVWXYZ

----------Next ManuscriptPreceding Manuscript----------Previous View----------Search----------Also by Roelof B. TimensAlso by Frits J.K. BuesinkAlso by Frank B.J. Leferink----------