SIMULATION OF CONDUCTED EMISSIONS FROM ELECTRONIC DEVICES

Lyuben Iliev a, Nikolay Panteleev b, Nikola Mihaylov c

a Department of Electrical Power Engineering, University of Ruse

Corresponding author, E-mail address: lailiev@uni-ruse.bg

b EMC Testing Laboratory, Bulgarian Institute of Metrology, Chief of Laboratory

c Department of Electrical Power Engineering, University of Ruse

Abstract. Electronic devices can legally be sold on the market only if they correspond to the standards of electromagnetic compatibility (EMC). In order to shorten the process of development and to lower the expenses required for multiple tests, it is convenient to use simulation for EMC. The current paper shows a simulation of the conducted emissions in the power supply network from switched mode power supply for LED lighting systems. The simulation is made using the CST STUDIO software product and the results are compared with real measurements made in an EMC testing facility.

Key words: Electromagnetic Compatibility, Simulation.


1. Introduction

The usage of electronic devices worldwide is growing constantly. This leads to worsening the electromagnetic environment, which leads to a number of negative phenomena - inappropriate functioning of devices, wrong data acquisition and even malfunctions. To avoid these unwanted events, the electromagnetic radiation from the devices should be kept at minimum levels and their susceptibility to electrical noise should be higher.

The testing for EMC of electronic devices is held in laboratories, which have the required equipment, according to the current standards. When a device fails one measurement, it has to be redesigned and tested again, which is time consuming and expensive.

The current paper describes an example of simulation of electronic device for conducted emissions using a specialized software CST STUDIO SUITE. The results are compared with real measurements made in an EMC testing facility. The simulation is made for different stages of the design process: schematic of separate blocks, full schematic, printed circuit board (PCB) accounting the placement of the components.

The main target is to define the earliest stages of the design process, where potential conductive emission problems can be detected and how they correspond to real measurements.

2. Methods and instruments used

In order to avoid expensive repetitive tests in an EMC laboratory, it is convenient to use software evaluation before the production of prototypes. There are several software products which account not only the schematic of the electronic devices, but also the placement of the components and the tracks on the PCBs. This way an optimal approach can be found during the design stage.

The software product CST studio includes several modules which have the opportunity to simulate a whole device including the cabling, which is often the strongest noise source [1]. Radiated emissions can be simulated with the help of the CST Microwave Studio by defining complex 3D structures. Other modules are capable of evaluating low-frequency electromagnetic processes. The modules are integrated in a single product capable of simulating all aspects of EMC. [ 1,2 ].

The object of the simulation is a switched mode power supply with the following parameters (fig. 1).

  • input voltage - 230 V;
  • output voltage - 12V;
  • output current (max) - 2.1ņ;
  • output power (max) - 25W.

The power supply is intended to drive LED stripes and lamps. The usage of this type of power supplies is wide spread because of the more frequent usage of power saving LED lighting.

Fig. 2. Schematic of the simulated device

The schematic is classic for switched mode power supply for 230V AC to 12V DC voltage. (fig. 2). The input AC voltage is rectified and then, using a MOS transistor, it is inverted and lowered using a transformer. The switching cycles of the transistor is controlled by a special integrated circuit (CR6850T). The feedback is provided via an optron for ensuring galvanic separation of the input and output. At maximum power the duty cycle of the control signal is 50% and decreases when the power consumption is lowering.

The full schematic can be replaced with a simplified one which includes the main modules - filter, rectifier, transistor, transformer (fig. 3).

Fig. 3. Simplified schematic

The PCB of the device is imported in the CST Studio software so that the simulation can be made (fig. 4). After this stage the parameters of the components are defined. For greater precision a maximum number of parasitic parameters have to be defined - resistance, inductance and capacity of the passive components. For the active components a predefined model as set (IBIS or SPICE) [7,8] for the current simulation the models of the components used are very close to the physical parameters of the real components. The componentsí model must correspond to the real physical characteristics. The parameters of the PCB are also entered (thickness of the layers, placement of the components, physical parameters of the materials used and other).

Fig. 4. Overview of the PCB imported in CST PCB Studio

A model of the PCB has to be defined according to the input parameters, which will be used for the simulation. The software supports a number of methods for modelling. For the particular case, the Transmission Line Method (TLM) was chosen, where the electrical connections in the PCB are calculated as long transmission lines, with their coresponing parameters (resistance, conductance, capacity and inductance). This method is relatively fast, but it does not take into account the 3-dimensional structure of the components. For frequencies up to 30 Mhz this method is relevant, but for higher frequencies it can not be used.

The simulation of a PCB, accounting all of the components, is a complex and time consuming process. In some cases replacing signals might be used in order to simulate separate blocks of the PCB where it is possible. The simplified schematic is connected to the PCB and the required signals are applied to it. The physical positions of the PCB tracks are taken into account. The measured signal is passed through a low-pass filter consisting of a 250 nF capacitor and 1 k? resistor. This filter has a cut off frequency of 636 Hz and is designed according to the specification of the Line Impedance Stabilization Network (LISN) used in the laboratory. [4]. The switching frequency of the MOS transistor is 70 kHz and it is controlled by a PWM signal with 50% duty cycle. A power load of 20W is simulated, the same is used in the laboratory measurements. The results of the simulation are shown in fig. 5.

Fig. 5. Simulation results for frequencies up to 1 Mhz

3. Laboratory measurements

The conductive disturbances are measured according to the standard EN 55015 for lighting systems [5]. The measurement is made using a peak detector with 200 Hz step for the 9 kHz - 150 kHz frequency band and 9 kHz for the 150 kHz - 1 MHz band. The following devices are used for the measurement: spectrum analyzer Schaffner SCR3502 [6], LISN NNB 52 [6] and attenuator 10 dB. The results of the measurement are shown on fig. 6.

Fig. 6. Results of the measurements in the frequency band - 9kHz-1MHz

The measurement shows clear peaks at the switching frequency (70 kHz) and the corresponding harmonics. The graphic on fig. 6 is very similar to the one obtained from the simulation. The first harmonic has an equal magnitude in the two graphics, but differences can be seen at the higher harmonics. In the laboratory measurements, the magnitude of these harmonics is higher than the simulated results. The difference is about 15 dB and is kept constant in the higher harmonics. The reason is that some parasitic parameters are not taken into account in the simulation, like the capacities of the transformer windings, the inductance of the capacitors and other. Also, the 3-dimensional structure of the components was not taken into account and some resonances were not registered. Besides the small differences between the simulation and practical results, they can be used as guidelines for the design process because of the following reasons:

∑ from practical point of view, differences in order of Ī10 dB, when alternative methods are used, are acceptable and can be used during the design stage

  • the error is systematic

4. Conclusion

The simulation of the switched mode power supply showed that the level of the conducted emissions is high and some additional changes are needed to lower the disturbances. Better filtering of the power line and removing some resonances will improve the emissions. For a future simulation, a 3-dimensional model will be created, which will provide better results, accounting all the disturbance sources. This process is slow and more complicated, but it is needed for maximum accuracy.

References

1. www.cst.com

2. CST Studio Suite Getting Started, Computer Sumilation Technology AG, 2008

3. Christopoulos C., The Transmission-Line Modeling Method in Electromagnetics, Morgan and Claypool Publishers, 2005, pp. 100

4. Bronaugh E, Mains simulation network (LISN or AMN) Uncertainty, pp. 12

5. BDS EN 55015 :2006 Limits and methods of measurement of radio disturbance characteristics of electrical lighting and similar equipment (CISPR 15:2005)

6. www.teseq.com

7. Quarles “., SPICE3 Version 3f3 Userís Manual, Department of Electrical Engineering and Computer Sciences, University of California, 1993, pp. 146

8. Baker. B, The IBIS model: A conduit into signal integrity analysis, Texas Instruments Incorporated, pp. 17