CHAPTER 4 Substrate Noise Coupling Effect in
4.2 Measurement results
The micrograph of the testkey is shown in Fig. 4.2. The row 1 and row 4 are the test patterns. The measurement setup is presented as Fig. 4.3 and the Agilent 8364B network analyzer and a pair of GSG probes with pitch size 100 um are used during measurement. Therefore, The S21 measurement results of the two-port network as the quantity of substrate noise in substrate, the substrate noise as shown in Fig. 4.4 is reducing when the distance(d) becomes longer. And the relationship between quantity of noise coupling and distance seems to be linear. Moreover, the measurement data show the quantity of substrate noise coupling is almost independent of frequency.
Therefore, the measurement results imply that substrate noise coupling effect is almost resistive in the lossy substrate.
Fig. 4.2 Micrograph of the testkey.
N.well with P+ contact
P.well with N+ contact
(a)
(b)
Fig. 4.3 Measurement setup. (a) Connect to network analyzer. (b) Test pattern (P-well with N+ contact and d=125 um) with probing.
Agilent 8364B
P+ P+
N-well P-substrate
G
G G
G
S S
RF probe with pitch =100um
-60 -55 -50 -45 -40 -35 -30
0 5 10 15 20
Frequency(GHz)
S21(dB)
50um 75um 100um 125um
(a)
-50 -45 -40 -35 -30 -25
0 5 10 15 20
Frequency(GHz)
S21(dB)
50um 75um 100um
(b)
Fig. 4.4 Measurement results. (a) P-well with N+ contacts. (b) N-well with P+ contacts.
CHAPTER 5
Conclusion and Future Works
5.1 Conclusion
In the thesis, the coupling effect of two adjacent inductors is presented first.
Moreover, a wide band model of single spiral inductor is also built up. The inductor model can fit the effective inductor and Q factor from 0.2 GHz to 30 GHz in correlation with the measurement. In the part of inductor coupled pairs, the measurement data indicate the coupling effect between two inductors is less than -20dB. Moreover, we also realize the coupling effect of inductor coupled pairs is slightly dependent on the separate distance. As a result, to increase the distance and suppress the noise coupling effect between two inductors is not an efficient way.
From the measurement results of type A and type B, the quantity of S41 are not equal. The results give a hint that the coupling effect is related to the layout style of inductors. And the proper layout style can reduce the coupling effect of adjacent inductors. After the measurement results of inductor coupled pairs are obtained, an equivalent circuit model is established to illustrate the coupling effect. The proposed model can precisely predict the coupling behavior from 1 GHz to 15 GHz.
The substrate noise coupling phenomenon in TSMC 0.18um CMOS process is discussed, too. In this part, the substrate coupling effect of two P+ (or N+) contacts in N-well (P-well) is analyzed. According to the measurement result, we find the substrate coupling effect is not strongly dependent on the frequency.
for RF circuit designers. The designers can refer to the information and decrease the parasitic effect of inductors. Therefore, a more efficient RF circuit design procedure can be realized.
5.2 Future works
As the model established in Section 3.3.4, the model of inductor coupling can explore the coupling behavior from 1 GHz to 15GHz. The frequency range is not wide enough for RF circuit designers. As a result, how to increase the frequency range is the most important thing to do.
Moreover, the inductor coupled pairs used in RF circuit designs is not only the type we propose. There are many different layout styles of inductor coupled pairs.
Hence, to find their coupling mechanisms and sum up those results is the significant task, too
Finally, we investigate the substrate noise coupling effect in TSMC 0.18um CMOS process. The situation we investigate is passive like. The coupling effect of active device, such as MOSFET and BJT, is not discussed in the thesis. Therefore, as the operation frequency goes to high (up to 60 GHz), the study about the substrate noise coupling effect of active device is an urgent issue for RF circuit designs.
REFERENCES
[1] D. K. Su, M. J. Loinaz, S. Masui, and B. A. Wooley, “Experimental Result and Modeling Techniques for Substrate Noise in Mixed.signal Integrated Circuits,”
IEEE J. Solid-State Circuits, vol. 28, pp. 420-430, Apr. 1993.
[2] T. Gabara, “Reduced Ground Bounce and Improved Latch up Suppression through Substrate Condition,” IEEE J. Solid-State Circuits, vol. 23, pp. 1224-1232, Oct.
1988.
[3] R. B. Merrill, W. M. Young, and K. Brehmer, “Effect of Substrate Material in Mixed Analog/Digital Integrated Circuits,” in Proc. IEEE Int. Electron Devices Meeting, pp. 433-436, San Francisco, CA, Dec. 1994.
[4] J. W. Lin, “An Optimum Design of the Micromachined RF Inductor” Master’s thesis, Dept. of Electronics Engineering and Inst. of Electronics, NCTU, Taiwan, 2004.
[5] Troels Emil Kolding, “On-Wafer Calibration Techniques for giga-hertz CMOS Measurement,” in Proc. ICMTS, pp.105.110, March 1999.
[6] D. M. Polar, Microwave Engineering, Third edition, John Wiley & Sons, Inc., 2005.
[7] C. J. Chao, S. C. Wong, C. H. Kao, M. J. Chen, L. Y. Leu, and K. Y. Chiu,
“Characterization and Modeling of On-Chip Spiral Inductors for Si-RFICs,” IEEE Trans. on semiconductor manufacturing, pp. 19-29, Feb. 2002.
[8] T. S. Horng, J. K. Jau, C. H. Huang, and T. Y. Han, “Synthesis of a Super Broadband Model for On-Chip Spiral Inductors,” IEEE RFIC Symposium, pp.453-456, June 2004.
[9] Joonho Gil, and Hyungcheol Shin, “A Simple Wide-Band On-Chip Inductor
vol. 51, pp. 2023-2028, Step. 2003.
[10] C. J. Chao, S. C. Wong, C. J. Hsu, M. J. Chen, and L. Y. Leu, “Characterization and Modeling of On-Chip Inductor Substrate Coupling Effect,” IEEE RFIC Symposium, pp. 311-314, June 2002.
[11] A. O. Adan, M. Fukumi, K. Higashi, T. Suyma, M. Miyamoto, and M. Hayshi,
“Electromagnetic Coupling Effects in RFCMOS Circuits,” IEEE RFIC Symposium, pp. 293-296, June 2002.
[12] M. Werthen, I. Wolff, R. Keller,and W. Bischof, “Investigation of MMIC Inductor Coupling Effect,” IEEE MTT-S International, vol.3, pp. 1793.1796, June 1997.
Appendix I
Four-Port Transformation
Because considering the model for the pad parasitic as Zi and yp, the equivalent circuit of our DUT (with pad parasitic) is shown in Fig. 1. Therefore, we derive the transmission matrix of the pad parasitic and use the de-embedding procedure to obtain the correct measurement result of our four-port network.
Fig. 1 The four-port network with pad parasitics.
Pad parasitics Zi
yp
Zi
yp
yp
yp
Zi
Zi
Port 1
Port 2
Port 3
Port 4 DUT
A. Transmission matrix of Z circuit
The Z circuit is presented in Fig. 2. From the definition of transmission matrix, we can derive that:
[ ]
T ZcircuitB. Transmission matrix of Y circuit
The concept to obtain Y circuit as depicted in Fig. 3 is the same as Part A and we can derive:
[ ]
T YcircuitFig. 3 Y circuits.
C. Transform between four-port S-parameters and transmission matrix (1) S-parameters to transmission matrix
y1
y2
V1
V2
I1
I2
V3
V4
I3
I4
[ ]
)
Fig. 4 Four-port transmission matrix.
Transmission
(2) Transmission matrix to S-parameters