A new general RF breakdown simulation tool - · PDF fileA new general RF breakdown simulation tool ... distribution of the filter is then exported from HFSS and CST, ... maximum difference

A new general RF breakdown simulation tool T. Pinheiro-Ortega1, J. Armendáriz1, S. Anza1, J. Gil1, C. Vicente1, B. Gimeno1, V. Boria1

(1) Aurora Software and Testing S. L. Edificio Desarrollo Empresarial 9B

Camino de Vera s/n 46022 - Valencia, Spain

Email:[email protected]

ABSTRACT

This paper introduces SPARK3D, a general software tool for Radio Frequency (RF) breakdown analysis, which allows predicting both Corona and Multipactor breakdown onsets in a great variety of RF structures. SPARK3D is a new module of the software CAD Tool: “Full-wave Electromagnetic Simulation Tool” (FEST3D). It has as input data the electromagnetic field distribution of the device under study at a single frequency. This new FEST3D module imports electromagnetic field formats corresponding to widespread electromagnetic simulation tools, so that it extends the high power capabilities of FEST3D to all kind of passive components.

INTRODUCTION

Today, telecommunication systems demand both for a higher component integration and for an increase of services implying larger bandwidths, thus higher frequency ranges. To achieve these objectives, on the one hand, the size of microwave devices incessantly decreases whereas, at the same time, the power levels increase. Both trends lead to a higher electric field density inside the components, which in turn may produce serious problems with respect to Radio Frequency (RF) breakdown due to Multipactor and Corona discharge. As a consequence, the design of devices free of microwave breakdown becomes, in many occasions, an extremely challenging task. Therefore, it is desired to have at one's disposal a software tool capable to determine the breakdown power of such structures with a reasonable accuracy.

There are plenty of electromagnetic simulation tools which perform the analysis of RF devices. Their capabilities and strong points basically depend on the numeric algorithms on which they are based. In many occasions they can be considered as complementary working tools whether to ensure the reliability of the obtained results or to use the less time consuming one, depending on the type of device under study. On the other hand, among these software suites just a few of them include the possibility to analyze high power effects inside microwave components and the comparison of their results is not always straightforward.

In this framework, we present in this paper SPARK3D, a powerful module of FEST3D for the analysis of breakdown phenomena in RF devices unique in its area. SPARK3D is also available stand-alone. It accurately predicts Corona and Multipactor threshold breakdown power in a remarkable variety of RF components using the electromagnetic field computed with some of the most widespread electromagnetic simulation software tools. It offers a great versatility and the possibility to compare high power results using different electromagnetic kernels.

CAPABILITIES AND FEATURES

As it was mentioned above, SPARK3D permits accurately analyzing both:

Corona discharge

Multipactor

which are two of the main high power effects that could severely damage a device. Corona module is based on a numeric algorithm that uses an adapted FEM technique to solve the free electron density continuity equation. On the other hand, Multipactor module is based on a full 3D electron tracker that employs a Leap-Frog algorithm for the path integration and the Vaughan model for SEY characterization of materials. These techniques allow the analysis of Corona and Multipactor in complicated structures which involve arbitrary shapes in short computational times.

In order to carry out the high power analysis, SPARK3D needs as input data the electromagnetic field distribution of the device under study at a single frequency. It is prepared to import the electromagnetic field computed and exported with some of the most powerful and widespread electromagnetic simulation tools like:

HFSS

CST

Besides, for devices with rectangular geometry, SPARK3D is also capable to import the computed electromagnetic field in comma-separated values (CSV) file format, as long as the field is calculated in a structured mesh. So, SPARK3D offers a great versatility and allows the comparison of high power results based on the electromagnetic fields computed by different software suits. This new FEST3D breakdown simulation module has been extensively used for breakdown prediction in passive microwave components, whose results have been verified with experimental available measurements for several devices. The flexibility of SPARK3D allows not only analyzing Corona and Multipactor in passive components based on waveguide technology, but also extending high power analysis to other space hardware configurations.

SPARK3D has a friendly user interface and its usage is very simple. Besides, the results of the analysis are given both in graphic and tabular form to make their interpretation easier. First, the user imports the electromagnetic field computed at a certain frequency choosing between the different formats compatible with SPARK3D, that is, HFSS, CST and CSV. Once the computed electromagnetic field is loaded, it can be visualized in order to detect the potential areas of the structure where the breakdown onset is more likely to occur, that is, where the electric field is maximum. In this way, the user can choose the regions of the device where the breakdown prediction will be carried out.

Secondly, the configuration of either Corona discharge or Multipactor analysis must be set. For Corona discharge, a pressure sweep is defined for analysis and it is possible to choose between air and nitrogen as filling gases. The results of the analysis involve for each region under study: the threshold breakdown power for each pressure, the representation of the Paschen curve corresponding to the pressure sweep selected and the minimum breakdown power in the whole pressure sweep. With this information it is easy to recognize which is the most critical region for Corona discharge and the minimum breakdown power supported by the device.

In the case of Multipactor, the user can choose between different predefined materials whose SEY is characterized by the Vaughan model. The user may define new material properties either specifying the parameters involved in the Vaughan model of the SEY or importing a text format file that contains for each impact energy the corresponding SEY value. In Multipactor configuration it is also possible to take into account an external magnetic field defined by the user. The results of the analysis include for each analyzed region: the threshold breakdown power, the Multipactor order and the electron evolution for each analyzed power in the process of searching the threshold breakdown power.

EXAMPLES

In this section, the capabilities of SPARK3D are shown through the high power analysis of a lowpass filter, whose design is represented in Fig. 1. As electromagnetic kernel, we use both HFSS and CST suits. The obtained results with the module SPARK3D are validated with the ones computed directly with FEST3D.

As preliminary work, we compare the S-parameter response of the filter obtained with the three aforementioned electromagnetic software tools, which, as can be seen in Fig. 2, matches reasonably well. The electromagnetic field distribution of the filter is then exported from HFSS and CST, as it will be used as input data of SPARK3D.

In order to carry out the high power analysis of the filter with SPARK3D, first of all its input parameters must be set. Fig. 3 shows a screenshot of the input parameters window of SPARK3D. Depending on the electromagnetic kernel previously used to compute the electromagnetic field distribution inside the filter, the user selects between the different formats supported by SPARK3D, which includes CST, HFSS and CSV. Then the user must import from a

Fig. 1: Screenshot of FEST3D corresponding to the analyzed lowpass filter including the 3D view of the component and the circuit in the canvas.

Fig. 2: S-parameter response of the lowpass filter analyzed with HFSS, CST and FEST3D.

specific file the electromagnetic field computed for a single frequency, which in our example is 9 GHz. As can be seen in the inset of Fig. 3, it is possible to visualize the imported field distribution. In this way, it is easy to recognize which are the potential breakdown critical areas, that is, where the electric field is maximum. Finally, the mode of high power analysis must be selected: either Corona or Multipactor.

Corona analysis mode

Once the electromagnetic field is imported, the parameters that define the specific configuration for Corona analysis must be set. Fig. 4 shows a screenshot of the Corona configuration window. The main configuration parameters that must be set include:

the pressure sweep,

the type of gas, which can be chosen between air or nitrogen,

the temperature of the device,

the initial power where the analysis starts,

and the precision, which defines the stopping criterion for the process of searching the threshold breakdown power.

Fig.3: Screenshot of input parameters window of SPARK3D module. The inset shows the result of the visualization of the imported file, which in this case corresponds to the electric field distribution of the lowpass filter.

Fig. 4: Screenshot of Corona configuration window of SPARK3D. The inset shows the electric field of the lowpass filter together with the region where the analysis will be carried out, which is shown in green.

The Corona analysis can be performed both in the whole device and on user defined regions that include potential critical areas for breakdown. Computing Corona onset on specific regions is faster than taking into account the whole circuit and besides, a denser mesh can be considered, which leads to more accurate result. The inset of Fig. 4 shows in green color the visualization of the region under study defined in our example.

Fig. 5 shows the results window of SPARK3D Corona mode. For each analyzed region, the Paschen curve, that is, the threshold breakdown power versus pressure, is represented and its points are given in tabular form. Besides, the minimum breakdown power in whole pressure sweep is given. With all this information, it is easy to determine the critical pressure and minimum power supported by the device, which in our example is 245 Watts at 12 mBar considering air as filling gas.

In Fig. 6 we present a comparison between the results obtained with FEST3D and SPARK3D module using the electromagnetic field computed both with CST and HFSS. As can be seen the agreement is quite good and the maximum difference between the obtained results is 0.07 dB, which is found at critical pressure, 12 mBar.

Multipactor analysis mode

The configuration parameter window for Multipactor mode is presented in Fig. 7. The main configuration parameters that must be set include:

the SEY of the material's device, which can be chosen between different predefined materials whose SEY is characterized by the Vaughan model. It is also possible for the user to define new material properties either specifying the parameters involved in the Vaughan model of the SEY or, using the Custom SEY option, importing a text format file that contains for each impact energy the corresponding SEY value;

the initial number of electrons,

Fig. 5: Screenshot of Corona results window of SPARK3D.

Fig.6: Comparison of the Paschen curve obtained with FEST3D and SPARK3D, using the electromagnetic field computed with HFSS and CST.

the initial power where the analysis starts,

the maximum power, which is the stopping criterion in case Multipactor onset does not occur,

and the precision, which is the stopping criterion in the process of searching the threshold breakdown power.

Other option that can be considered in Multipactor mode is the presence of an external magnetic field defined by the user, whether it is uniform or not. The latter case is only available for CSV format file and the external magnetic field is defined as an extra column of the input file.

It is also possible to perform Multipactor analysis both in the whole device and on user defined regions that include potential critical areas for breakdown. Computing Multipactor onset on specific regions is faster than taking into account the whole circuit and besides, a denser mesh can be considered, which leads to more accurate result. The inset of Fig. 7 shows in green color the visualization of the region defined in our example.

Fig. 8 shows the results window of SPARK3D Multipactor mode. For each region under study, the electron evolution is represented in a graph for every analyzed power. Besides, a table summarizes the results for the analyzed power sweep, showing the Multipactor order for those ones where breakdown occurs. Finally, the threshold breakdown power is given. With all this information, it is easy to determine the minimum power supported by the device. Table 1 shows a comparison between the results obtained with SPARK3D using the electromagnetic field computed with CST and HFSS and the results from FEST3D. The maximum difference is 0.4 dB.

Fig. 7: Screenshot of Multipactor configuration window of SPARK3D. The inset shows the electric field of the lowpass filter together with the region where the analysis will be carried out, which is shown in green.

Fig. 8: Screenshot of Multipactor results window of SPARK3D.

Table 1. Multipactor breakdown power calculated with FEST3D and SPARK3D with both CST and HFSS.

FEST3D SPARK3D + CST SPARK3D + HFSS

8400 Watts 7900 Watts 8600 Watts

CONCLUSIONS AND FUTURE DEVELOPMENTS

In this paper, we have presented SPARK3D, the new high power software module of FEST3D. It is the first software of its class capable to analyze Corona and Multipactor effects by importing the electromagnetic field distribution of the device under study from several widespread electromagnetic software suits. It offers a great versatility and extends the high power analysis to a remarkable variety of space hardware configurations. SPARK3D software tool is also available stand-alone and is in continuous development. Within the next close or mid-term future, it is expected to include new features as the analysis of Multipactor under multicarrier excitation, Corona discharge prediction at ambient pressure and new compatibilities with other electromagnetic software suits.

ACKNOWLEDGEMENTS

The authors would like to thank the Ministerio de Ciencia e Innovación of Spain, which has supported this work under Grant TEC2007-67630-C01 and the ”Programa Torres Quevedo” (PTQ08-03-08254).