Optimal electric wave propagation parameters on a transmission line--Schottky diode system.
Today, electromagnetic susceptibility threshold for complex electronic components decreases continuously, due to two trends . First, needs in speed transmission and wide frequency bandwidth result in a large increase in operating frequencies. Second, there is a drastic decrease in device size and in bias currents. With the increase in operating frequency, the length of the electromagnetic waves emitted by devices becomes smaller, reaching the size of the devices themselves. The electromagnetic wave will then couple itself to tracks and get processed towards devices. Hence, it becomes drastic that in circuit design risks due to electromagnetic susceptibility be reduced. For this, one needs the best information possible on the conditions for optimal coupling of an electromagnetic wave to circuit board tracks. These coupling problems are addressed in this paper in a concrete way, with a long microstrip line placed in front of a Schottky diode. To exhaustively study both wave coupling and propagation on the line, a high frequency signal is injected in a near-field injection mode. The impact of the aggression on the diode is investigated as a function of its frequency, power and place of injection.
2. DEVICE UNDER TESTS AND EFFECT ON ITS STATIC CHARACTERISTIC OF AN EMI
The device under test, a HSMS 2850 Schottky diode from Agilent Technologies, is placed on a microstrip line. For purpose of wave propagation analysis the line is L = 10 cm long to highlight stationary waves in the [500 MHz-3 GHz] frequency range. A block inductance [L.sub.t] with a capacitance [C.sub.t] prevents the high frequency aggression from propagating towards the dc generator. The electromagnetic interference (EMI) is injected through a probe. Probe-line distance h is h [much less than] [lambda]/10 to work in the near field zone. Electric-field probes are made from semi-rigid coaxial cables and are characterized . Places for injection are noted [P.sub.1] to [P.sub.10] with [P.sub.1] located nearest to the diode and [P.sub.10] at the furthest. A distance of 1cm separates each possible location. Mean values for diode voltage [V.sub.d] and current [I.sub.d] are measured (Fig. 1) under EMI with frequency [f.sub.a] = 1 GHz.
[FIGURE 1 OMITTED]
A change in [I.sub.d] and [V.sub.d] values impacts the [I.sub.d]-[V.sub.d] curve. This is due to a rectification effect  because of nonlinear behaviors inherent to semiconductor devices.
In the following sections, experiments will be completed by simulations performed with the commercial software ADS of Agilent Technologies. S-parameter simulations are undertaken for a linear analysis while harmonic balance simulations [4-6] are performed for non linear analysis.
3. EXPERIMENTS AND SIMULATIONS OF TRANSMITTED POWER
3.1. Model of the System Under Test
The diode is modeled using ADS libraries . The Sot-23 package includes parasite inductances and capacitances . After adjustment, a good matching is obtained between static curves and S parameter measurements and simulations. The coupling between the electric probe and the microstrip line is modeled by a capacitance [C.sub.1] . To determine [C.sub.1], one first calculates it approximately and then adjusts it for a good correlation between measured and simulated transmission coefficients on an open-ended line. The whole model is given on Fig. 2(a) and results on Fig. 2(b) with the electric probe set at 0.5 mm over the middle of the line.
[FIGURE 2 OMITTED]
Figure 2(b) shows a good agreement between simulation and measurement results with [C.sub.1] = 37 fF. One must adjust the value of [C.sub.1] between 19 fF and 37 fF depending on the frequency of the aggression signal. Fig. 3 shows the [I.sub.d]-[V.sub.d] curves for an EMI injected in position [P.sub.10] at [f.sub.a] = 1 GHz, probe-line distance kept to h = 0.5 mm.
3.2. Highlight on Resonance Phenomenon
The influence of the frequency of the EMI on the system is modeled. Probes are sufficiently unmatched to disregard their effect. Hence, the line can be considered as a cavity loaded on one end by the choke inductor [L.sub.t] with a high impedance at operation frequency and on the other side by the diode. [L.sub.t] is equivalent to an open circuit and the diode to a variable load [[bar.Z].sub.d]. According to line theory , resonance frequencies [F.sub.r] for which the power transmitted to [[bar.Z].sub.d] is maximum can be calculated by:
[F.sub.r] = c / 2[pi]L[square root of [[epsilon].sub.eff]] [[tan.sup.-1] (j[Z.sub.c] / [[bar.Z].sub.d]) + m[pi]], (1)
with [[epsilon].sub.eff] the effective permittivity, [Z.sub.c] the characteristic impedance of the line, c the speed of light.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
Values of [F.sub.r] depend on the length L of the cavity, on [Z.sub.C] and on the value of [[bar.Z].sub.d]. If [absolute value of [[bar.Z].sub.d]] [much greater than] [Z.sub.c], values for [F.sub.r] can be calculated for which the power transmitted to [[bar.Z].sub.d] is maximum. To check these resonances, we simulate the power ratio [P.sub.Zd]/[P.sub.HF] where [P.sub.Zd] is the power transmitted to [[bar.Z].sub.d], and [P.sub.HF] the power provided by the high frequency source. The EMI is injected in [P.sub.10] with power set to 10 dBm. Results are shown on Fig. 4.
When [absolute value of [[bar.Z].sub.d]] [much less than] [Z.sub.C], two [F.sub.r] values are measured. When [absolute value of [[bar.Z].sub.d]] become closer to [Z.sub.C], the power ratio at [F.sub.r] decreases. When [absolute value of [[bar.Z].sub.d]] = [Z.sub.C], no resonance frequency can be calculated, resonance peaks disappear. These resonance frequencies impact directly the quantity of transmitted power to the disturbed circuits.
3.3. Effects of Power Resonances on the Detection Phenomena
We simulate [P.sub.t] the power transmitted to the diode in Fig. 5(a). In reverse bias, [F.sub.r] values, for which [P.sub.t] is maximum, are measured at 1.07 GHz and 2.14 GHz. These values have been previously calculated. No [F.sub.r] peaks appear in forward bias. The [I.sub.d]-[V.sub.d] curve is measured experimentally (Fig. 5(b)) for an EMI injected in the middle of the line at [P.sub.5] with [f.sub.1] = 1.07 GHz (resonance frequency) and [f.sub.2] = 1.15 GHz (non resonance frequency).
[FIGURE 5 OMITTED]
Figure 5(b) shows that the power of the EMI detected by the diode depends on the EMI frequency. For reverse bias at [f.sub.1] = 1.07 GHz, [P.sub.t] [approximately equal to] -5 dBm. The diode detects the aggression signal and a voltage change [DELTA][V.sub.d2] [approximately equal to] 200 mV is measured. For [f.sub.1] = 1.15 GHz, [P.sub.t] [approximately equal to] -40dBm. This value is too small to be detected; hence no change of [V.sub.d] is measured. In forward bias, whatever the frequency of the aggression signal, [P.sub.t] [approximately equal to] -20 dBm. The diode detects very small power values and nearly no disturbances are measured. The EMI has a maximum effect on the diode for resonance frequencies [F.sub.r] which appear when the diode is under reverse bias. It is then interesting to study the influence of the injection point at these resonance frequencies.
3.4. Influence of the Position of the Injection Point
[P.sub.t] is simulated for an EMI injected in a conducting mode. EMI is injected in [P.sub.10], [P.sub.7], [P.sub.4], [P.sub.1] with power set to 10dBm (Fig. 6). For each location, resonance phenomena occur for the same [F.sub.r] values. However, depending on the place of injection, anti-resonance frequencies may also appear. When the probe is placed over a zone with minimum voltage, the probe-line coupling can be neglected. Therefore even if the aggression signal is injected with a frequency equal to [F.sub.r], no power is transmitted to the diode, there is no disturbance.
A field mapping (Fig. 7(a)) of the line is carried out for an EMI with [f.sub.a] = [F.sub.r] = 1.07 GHz and power 10 dBm. The diode is biased at 0V.
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
The line is loaded on its left by the choke inductance, a maximum voltage occurs in [P.sub.10] and a maximum coupling is measured. On the other end, the line is loaded by the diode non-biased, hence with a high dynamic impedance. A maximum voltage then appears in [P.sub.1], a maximum coupling is measured. Maxima and minima voltages are separated by a distance equal to a quarter effective wave length [[lambda].sub.eff]/4. If a maximum voltage appears in [P.sub.10] then a minimum voltage appears in [P.sub.5], in the middle of the line, and a minimum coupling is measured. In Fig. 7(b), Figs. 7(c) and 7(d), when the probe-line coupling is maximum, the power transmitted to the diode is maximum, a large change in the [I.sub.d]-[V.sub.d] curve is measured.
A commercial Schottky diode placed at the output of a microstrip line is studied when subject to an EMI. This aggression has a frequency value out of band of operation of the diode. The diode exhibits a change in its [I.sub.d]-[V.sub.d] curve due to rectification effects. A complete model of the diode in package, bias system, microstrip line and near-field injection for aggression signal, is established. The non linear behavior of the diode is considered. The diode detects power if the EMI has a frequency equal to resonance frequencies dependent on the load of the line; hence it is a function of diode bias. Resonance frequencies impact the quantity of transmitted power, leading to more or less disturbance on the behavior of the diode.
Location of aggression injection plays a major role in the probe-line coupling. If the probe is placed over a zone with minimum voltage, the probe-line coupling can be neglected. Even if the EMI is injected at a resonance frequency value of the system considered, no power is transmitted. Hence there will be no diode disturbance.
This work should help give parameters for estimating susceptibility to high frequency signals of a system.
Received 15 September 2011, Accepted 4 November 2011, Scheduled 7 November 2011
[1.] Ramdani, M., E. Sicard, A. Boyer, S. Ben Dhia, J. J. Whalen, T. H. Hubing, M. Coenen, and O. Wada, "The electromagnetic compatibility of integrated circuits-past, present, and future," IEEE Trans. on Electromagnetic Comp., Vol. 51, No. 1, 46-52, 2009.
[2.] Jarrix, S., T. Dubois, R. Adam, P. Nouvel, B. Azais, and D. Gasquet, " Probe characterization for electromagnetic near-field studies," IEEE Trans. on Instrumentations and Meas., Vol. 59, No. 2, 292-300, 2010.
[3.] Larson, C. E. and J. M. Roe, "A modified Ebers-Moll transistor model for RF-interference analysis," IEEE Trans. on Electromagnetic Comp., Vol. EMC-21, No. 4, 283-291, 1979.
[4.] Hattori, Y., T. Kato, H. Hayashi, H. Tadano, and H. Nagase, "Harmonic balance simulation of RF injection effects in analogs circuits," IEEE Trans. on Electromagnetic Comp., Vol. 40, No. 20, 120-126, 1998.
[5.] Maas, S. A., Non Linear Microwave and RF Circuits, 2nd edition, Artech House, 2003.
[6.] Rizzoli, V. and A. Neri, "State of the art and present trends in nonlinear microwave CAD techniques," IEEE Trans. on Microwave Theory and Tech., Vol. 36, No. 2, 343-365, 1988.
[7.] Antognetti, P. and G. Massobrio, Semiconductor Device Modeling with SPICE, 2nd Edition, McGraw-Hill, 1993.
[8.] Angelo, E., Electronic Circuits, McGraw-Hill, 1958.
[9.] Kanda, M., "Standard probes for electromagnetic field measurements," IEEE Trans. on Antennas and Propagat., Vol. 41, No. 10, 1349-1364, 1993.
[10.] Chand, C., RF and Microwave Wireless Systems, 2nd edition, Wiley, 2000, ISBN: 978-0-471-35199-3.
T. Dubois (1), J. Raoult (2), S. Jarrix (2), *, A. Blain (2), and A. Doridant (2)
(1) Poly-Grames, Montreal, Canada
(2) Institut d'Electronique du Sud, Universite Montpellier 2, Place Bataillon, Montpellier 34095, France
* Corresponding author: Sylvie Jarrix (email@example.com).
|Printer friendly Cite/link Email Feedback|
|Author:||Dubois, T.; Raoult, J.; Jarrix, S.; Blain, A.; Doridant, A.|
|Publication:||Progress In Electromagnetics Research Letters|
|Date:||Jul 1, 2011|
|Previous Article:||Super-wideband printed asymmetrical dipole antenna.|
|Next Article:||A wavelet operator on the interval in solving Maxwell's equations.|