A Single-Probe Method for Measuring High-Frequency Adjacent Electromagnetic Fields

Satoshi Kazama

A new measurement technique resolves electromagnetic-field characteristics down to the lead-frame level of an IC package.

Identifying the source of electromagnetic radiation and mitigating electromagnetic interference (EMI) in electronic equipment is crucial, particularly as operating frequencies increase and noise margins shrink. The commonly used method for measuring electric and magnetic fields is limited both in the frequencies at which the measurements can be made and in its resolution. This technique requires both a magnetic probe, such as a shielded loop, and an electronic probe to measure the distribution of current from the adjacent magnetic field and voltage from the adjacent electric field, respectively. The shielded loops used to measure current distribution typically require a bulky shielding structure to prevent electric coupling. In addition to size issues, shielded loops have proven, in certain cases, to be ineffective for measuring high-frequency radiation.

This article introduces a novel method to measure both electric- and magnetic-field distributions simultaneously using a single probe. The structure of the probe is simple, and the size of the probe can be easily reduced. The resulting technique allows for the characterization of electric and magnetic fields at the lead-frame level of integrated circuits (ICs) and at frequencies up to 3 GHz.

Basics of the Measurement Technique

The measurement probe, a simple metal loop, is terminated with the same impedance at each end. When the probe is in the presence of electric and magnetic fields, current will be induced in the sensor. The current that results, from the magnetic coupling of current and the electric coupling of voltage, flows in the probe. Figure 1 shows the probe coupling with voltage V and current I, both of which are parallel with the probe. The voltage and current are given by

V = A sin(wt + q_V) (1)

and

I = B sin(wt + q_C), (2)

where A and B are the respective magnitudes of voltage and current, and q_V and q_C are the respective phase angles.

Figure 1. Probe couplings.

The direction of the current resulting from the magnetic coupling is different at both ends of the probe, but the direction of the current resulting from the electric coupling is the same at both ends, thereby providing different outputs at the two ends of the probe. The current and voltage are estimated from this difference.

The outputs at the two ends of the probe are given by

O₁ = a₁V + b₁I (3)

and

O₂ = a₁V – b₁I, (4)

where a₁ and b₁ are coupling coefficients that vary with the position of the ground and the probe.

From these relationships, the current and voltage are derived as

(5)

and

(6)

respectively.

Figure 2 illustrates the new method for estimating the current and voltage distributions. It shows current I with current angle f and voltage V at a point with coordinates (x, y) on the measurement plane. The probe outputs for the four rectangular directions (O_xf, O_xr, O_yf, O_yr) are given by

O_xf = a₂A_{(x, y)}sin(wt + q_{V(x, y)}) – cos(f)b₂B_{(x, y)}sin(wt + q_{C(x, y)}), (7)

O_xr = a₂A_{(x, y)}sin(wt + q_{V(x, y)}) + cos(f)b₂B_{(x, y)}sin(wt + q_{C(x, y)}), (8)

O_yf = a₂A_{(x, y)}sin(wt + q_{V(x, y)}) – sin(f)b₂B_{(x, y)}sin(wt + q_{C(x, y)}), (9)

and

O_yr = a₂A_{(x, y)}sin(wt + q_{V(x, y)}) + sin(f)b₂B_{(x, y)}sin(wt + q_{C(x, y)}), (10)

where a₂ and b₂ are coupling coefficients; A_{(x, y)} and B_{(x, y)} are the magnitudes of the voltage and current, respectively; and q_{V(x, y)} and q_{C(x, y)} are the respective phase angles.

Figure 2. Estimation of current and voltage distributions.

Equations 7–10 can be used to derive the x-direction voltage (Equation 11), y-direction voltage (Equation 12), x-direction current (Equation 13), and y-direction current (Equation 14):

(11)

(12)

(13)

and

(14)

For these relationships to be correct, the x- and y-direction voltages must be the same.

Component-Level Field Distributions

Figure 3 shows the structure of the probe used for the measurement system. It is composed of a bundled pair of semirigid cables with a characteristic impedance of 50 W. The sensing component consists of a metal wire connected to the two center conductors of the cables. The sensing wire is about 1.0 x 0.5 mm (0.039 x 0.019 in.). The current and voltage of the measurement plane couple with the metal wire. Both cables are connected to the inputs of the measuring instrument. This simple structure makes it easy to reduce the size of the probe.

Figure 3. Structure of the probe.

Figure 4 shows the setup of the measurement system for testing a digital IC device. A three-dimensional actuator controlled by a personal computer drives the scanning movement of the probe. A fixed probe, which measures the reference signal, and the two cables of the scanning probe are connected to a three-input phase-difference measurement system. The vector outputs from the probe are measured using the three-input phase-difference measurement system, which measures both the magnitude and phase difference of the high-frequency signals between the three channels.

Figure 4. Outline of the measurement system.

Figure 5 shows an outline of the three-input phase-difference measurement system. The measured input signal is down converted to an intermediate frequency and digitized by an analog-to-digital convertor. The signal magnitude, frequency, and phase are derived by fast Fourier transform. All inputs are synchronized by using the same local oscillator and clock signal, preserving the phase difference between the channels. The upper frequency limit of the system is 3 GHz.

Figure 5. The three-input phase-difference measurement system.

The probe is shifted between two rectangular orientations while it continuously scans above the test device. The vector outputs at each of these probe positions are measured and then used to calculate the electric- and magnetic-field distributions.

Current and Voltage Distributions in a Digital IC

Figure 6 shows how the current and voltage distributions of a 300-mil small-outline-package complementary metal-oxide semiconductor digital IC (74HC04) were measured using this method. The IC was mounted on a printed circuit board (PCB), and six inverters were connected to the IC and driven by a 40-MHz signal wave. The source voltage was 5 V. The probe scanned in 0.5-mm steps on a plane approximately 1 mm above the IC package.

Figure 6. Measured IC.

Figures 7a to 7d show the estimated x-direction current and voltage distributions in the IC at the third harmonic (120 MHz) of the signal wave. Because the coupling coefficients are unknown, the magnitude scales are relative. Using this technique, estimates of y-direction distribution produced a complementary set of figures (not shown). The same technique was used to produce a family of graphics estimating the second-harmonic distribution (also not shown).

Figure 7a. Current distribution (relative magnitude) in a digital IC (120 MHz, x direction).

Figure 7b. Current distribution (relative phase) in a digital IC.

Figure 7c. Voltage distribution (relative magnitude) in a digital IC.

Figure 7d. Voltage distribution (relative phase) in a digital IC.

To verify these estimated results, the voltage distributions (magnitude and phase) in the x and y directions were compared. Both the magnitudes and phase angles coincided exactly at the second and third harmonics, indicating that this measurement system produces reliable results.

The current magnitude (I_m) distributions were calculated by using

(15)

where I_x and I_y are the current magnitudes in the x and y directions, respectively.

As shown in Figure 8, the third-harmonic current flowed mainly in the input-output lead frame of the inverter. A similar pattern was seen at other odd-numbered harmonics. As shown in Figure 9, the second-harmonic current flowed mainly in the source and ground lead frame.

Figure 8. Current distribution at third harmonic (120 MHz).

Figure 9. Current distribution at second harmonic (80 MHz).

A similar pattern was seen at other even-numbered harmonics. The waveform in the input-output lead frame is therefore trapezoidal and consists mainly of odd harmonics, whereas the current in the source and ground lead frame contains even harmonics.

Probe Output Characteristics

The output characteristics of the probe were measured using a microstrip line. The probe was set parallel to the microstrip line, as shown in Figure 1. (O₁ – O₂ is referred to as the current output of the probe, and O₁ + O₂ is referred to as the voltage output.) The microstrip line was formed on a 1.6-mm-thick PCB with a 50-W characteristic impedance. One end of the line was connected to a signal generator with a 10-dBm output, and the other end was connected to an impedance of 50 W. The current in the line was about 14.2 mA, and the voltage was approximately 0.71 V.

Figure 10 shows the current output distribution above the line at 1 GHz. The probe was set parallel to the microstrip line, and scanning was done in 0.2-mm steps along the upper part of a cross section of the line.

Figure 10. Current output distribution by probe position.

Figure 11 shows the voltage output distribution above the line at 1 GHz. (The hatched squares in Figures 10 and 11 show the width of the line.) These distributions show the relationship between the probe output and the position of the probe. Figure 12 shows the relationship between the current and voltage outputs of the probe and the distance from the center of the microstrip line. The signal frequency was 1 GHz, the current in the line was about 14.2 mA, and the voltage in the line was about 0.71 V. See Figure 13 for a diagram of the measured microstrip line.

Figure 11. Voltage output distribution by probe position.

Figure 12. Voltage and current output by probe distance.

Figure 13. Measured microstrip line.

The current and voltage distributions in the microstrip line shown in Figure 14 were derived from the relationships illustrated by Figures 12 and 13. In Figure 12, output voltage is about 0.011 V at a distance of 1 mm, and the output current is about 8 mA at a distance of 1 mm. Therefore, the voltage par output in Figure 13 is

and the current par output is

The voltage and current are proportional to their respective outputs. As shown in Figure 14, these distributions show characteristics typical of standing waves in a transmission line, indicating that the current and voltage distributions can be measured simultaneously by this method.

Figure 14. Current and voltage distribution on a microstrip line.

The output characteristics of the probe varied with the thickness of the PCB. They are sufficiently different from those shown in Figure 12; therefore, these relationships must be measured for each case. The output characteristics also varied with the frequency, as shown in Figure 15 for a 50-W microstrip line on 1.6-mm-thick PCB when the probe was set 1.3 mm above the center of the line.

Figure 15. Voltage and current output by frequency (T = 1.6 mm).

EMC Countermeasure Application

To demonstrate the effectiveness of the high-frequency adjacent-electromagnetic-distribution measurement system in the application of EMC countermeasures, an experiment was performed using a test board, the upper half of which was connected to supply voltage and the lower half to ground. A logic IC was connected to a 50-MHz input signal.

The purpose of the experiment was not only to demonstrate the performance of the new measurement technique but also to show that the careful application of the appropriate device in the circuit will dramatically reduce the effects of electromagnetic fields. The basic circuit consisted of a logic IC, bypass capacitors, power and ground connections, and a selected countermeasure inductor, a Taiyo Yuden (Tokyo, Japan) model FBMH. High-impedance devices like the FBMH are appropriate for applications in which a large withstand current (≤ 2 A) is needed to mitigate radiated and conducted noise, particularly at frequencies ≥ 1 GHz. This type of inductor should also have low dc resistance (R_dc) to reduce heat generation and power loss in low-voltage/high-current applications, including cpu V_CC lines. (The 150-W FBMH has a maximum R_dc value of 0.050 W, whereas the 220-W part has a rating of 0.070 W.)

Using the adjacent-electromagnetic-distribution system, the electric and magnetic fields were measured across the test platform both before and after the placement of the inductor. Figures 16 and 17 depict the resulting reduction in the magnetic and electric fields, respectively.

Figures 16 and 17. The levels of radiation from the electromagnetic field of an IC are compared before and after placing a high-impedance inductor in the test circuit. The "after" pictures clearly show a significant reduction in radiation after installation of the countermeasure.

Conclusion

Because the new measurement technique can resolve electric- and magnetic-field characteristics down to the lead-frame level of an IC package, the application of appropriate countermeasures at the optimum location in the circuit can be quickly and easily accomplished, thereby speeding the new-product development process. As operating frequencies continue to increase and logic voltage levels continue to decrease, the importance of effective EMC countermeasures has never been greater. This technique not only speeds resolution of EMC concerns but also has the potential to reduce costs by optimizing the use of required components.

References

1. H Wakuba et al., "Estimation of the RF Current at IC Power Terminal Using Magnetic Probe with Multilayer Structure," in Technical Report of IECE, Electromagnetic Compatibility Japan 98-6: 39–43.

2. A Namba, O Wada, and R Koga, "Measurement of Near-Field Emission from Printed Circuit Board Using Miniature E-Field Probe," in Technical Report of IECE, Electromagnetic Compatibility Japan 98-46: 41–46.

3. Y Kami and T Tobana, "Measurement of Magnetic Nearfield on Printed Circuit Boards by Using a Magnetic Loop Antenna," in Proceedings of International Zurich Symposium on EMC (Zurich, Switzerland: EMC Zurich, 1997): 591–596.

4. S Yabukami, M Yamaguchi, and KI Arai, "HF-UHF Band Electromagnetic Measurements Using Multi-Layer Printed Wiring Board," in Technical Report of IECE, Electromagnetic Compatibility Japan 97-37: 21–26.

5. T Kurouchi, A Ohneda, and A Hasumi, "Research of Noise Measurement Technology in Minute Area," in Technical Report of IECE, Electromagnetic Compatibility Japan 94-12: 33–37.

Satoshi Kazama is a member of the product evaluation team at the Taiyo Yuden EMC Center (Tokyo) and is a doctoral candidate in electrical engineering at Tohoku University (Sendai, Japan).