Harmonizing Circuit Design and EMC DesignPart 19 EMC Calculation Methods and EMC Simulations (4): Trial Calculation of Conducted Immunity (CI)

Hello! I’m Inagaki, at ROHM.

This 19th article is the fourth dealing with EMC (electromagnetic compatibility) calculation methods and simulations. Here, trial calculations of conducted immunity (CI) are explained. This relates to the electromagnetic compatibility (EMC) of semiconductor integrated circuits, and so has to do with IEC 62132-4, DPI method (Direct RF Power Injection Method).

The DPI method (IEC 62132-4), together with the 150Ω method (IEC 61967-4) is an international EMC standard that is frequently used for EMC measurements of semiconductor integrated circuits. Targets for calculations may include RF signal generators, power amplifiers, bidirectional couplers, DC blocks (capacitive elements C), decoupling networks (inductive elements L), DUTs (devices under test), and EMC countermeasure circuits (here, capacitive elements C), among others. In the DPI method, 278 frequencies (between 150 kHz and 1 GHz) are stipulated as test frequencies. Hence as a method of analysis, a circuit analysis (transient analysis) is repeated 278 times, and the results are plotted on a graph with a frequency axis to obtain calculated results. Also, a method may be employed in which measured values are used to create a computer model (simulation model). It should be noted that the measurement results of the DPI method are electric power values (dBm units) for traveling-wave power, which may be observed using the bidirectional couplers mentioned above. In circuit analysis, power values cannot be directly handled, and so a little innovation is necessary.

Calculations are explained in order. In trial calculations, there are two processing stages; the first stage is IB (malfunction threshold) model extraction, and the second stage is calculated prediction, and each is automated using a (shell) script. The first stage, IB (malfunction threshold) model extraction, uses the calculation procedure described below. Regarding the IB (malfunction threshold) model, please refer to “16. EMC Calculation Methods and EMC Simulations (1): Overview of Calculation Methods”.

■First Stage: IB (Malfunction Threshold) Model Extraction

1. ① First, a calculation circuit diagram is created for the device under study. The measurement circuit is connected as-is and used as the calculation circuit. It’s easy enough if you refer to circuit diagrams in reference works and the relevant IEC standards. A bidirectional coupler can be replaced by a TL (transmission line), and for a DUT (device under test), an LCR meter is used to measure the impedance between terminals, and the electrical characteristics are used to create an LCR (passive-element) circuit.
2. ② Power values obtained in DPI method measurements are converted into voltage or current values. Here, conversions use 50Ω.
3. ③ The voltage or current values converted in ② are applied to an RF signal generator as the signal source, and circuit analysis (transient analysis) is used to calculate the IB (malfunction threshold) (LSI pin voltage or current) for one frequency. It must be remembered that DPI method measured values are not the LSI pin values when an LSI has malfunctioned, but rather are traveling-wave power values in a bidirectional coupler (this is important!).
4. ④ Step ③ is repeated for all frequencies (278 total). If (shell) scripts and macros are used so that calculations can be repeated, then there should be no problem in running even a large number of analyses with one execution. The results are saved in a file, and plotting along a frequency axis yields the following graphs.

The IA (electromagnetic interference) model used in verification of conducted and radiated emissions is, even in standalone form, versatile for producing numerical results. That is, even when calculation circuit diagrams are different, the values themselves are meaningful. On the other hand, in the case of the IB (malfunction threshold) model, it should be noted that the voltage and current values are intrinsic and of limited applicability, depending on the calculation circuit diagram and the LSI model (impedance characteristics). By using it together with a calculation circuit diagram and LSI model, the IB model can be used as a computer model (simulation model) capable of reproducing, on a computer, malfunctions at the time of measurement.

The second stage, calculated predictions, rely on the following calculation procedure.

■Second Stage: Calculated Prediction

1. ⑤ A calculated prediction circuit is created. The difference with a circuit for IB (malfunction threshold) model extraction is the addition of a malfunction evaluator (comparator). Here again, the reader may be assisted by consulting circuits in a relevant reference.
2. ⑥ Next, the signal source of an RF signal generator is set to a damped oscillation waveform. A damped oscillation waveform can be created relatively easily in SPICE. Conversely, a waveform that gradually increases in size cannot be set because it is headed toward divergence.
3. ⑦ In the circuit analysis (transient analysis), when the analysis is executed for one frequency, we find that, for example, the LSI transitions from a malfunctioning state to a non-malfunctioning state. A malfunctioning state means a case in which a large output from an RF signal generator is applied, and a higher voltage or current than the IB (malfunction threshold) is reached at the LSI pins. Conversely, a non-malfunctioning state is a state in which voltages or currents are reached that are smaller than the LSI pin IB (malfunction threshold) levels. It is then sufficient to save to a file the voltage or current of the RF signal generator at the time of switching from a malfunctioning state to a non-malfunctioning state. This is repeated for all the frequencies (278 in all).
4. ⑧ The calculated prediction values to be found are obtained by converting the saved voltage or current values to their 50Ω power value equivalents. The figure on the lower left shows calculated predictions for the same circuit as that used in IB (malfunction threshold) model extraction, without adding an EMC countermeasure circuit. The measured and calculated values match exactly! If the calculations are performed without error, this is the natural result.
5. ⑨ In addition, the figure on the lower right shows the results of calculated predictions with an EMC countermeasure circuit (here, a capacitive element C) added. We see that the added capacitive element is effective in improving LSI malfunction levels.

In actual design, EMC countermeasure circuits are studied, and these calculated predictions are executed repeatedly to determine application circuits and LSI circuits confirming to IEC standards. (An EMC countermeasure circuit is chosen such that calculated prediction values are larger than limit values for all frequencies.)

Thank you for your kind attention.