## High-Frequency Distortion Analysis of Analog Integrated Circuits (1999)

Venue: | IEEE Trans. Circuits Syst. II |

Citations: | 5 - 0 self |

### BibTeX

@ARTICLE{Wambacq99high-frequencydistortion,

author = {Piet Wambacq and Georges G. E. Gielen and Peter R. Kinget and Willy Sansen},

title = {High-Frequency Distortion Analysis of Analog Integrated Circuits},

journal = {IEEE Trans. Circuits Syst. II},

year = {1999},

volume = {46},

pages = {335--345}

}

### OpenURL

### Abstract

An approach is presented for the analysis of the nonlinear behavior of analog integrated circuits. The approach is based on a variant of the Volterra series approach for frequencydomain analysis of weakly nonlinear circuits with one input port, such as amplifiers, and with more than one input port, such as analog mixers and multipliers. By coupling numerical results with symbolic results, both obtained with this method, insight into the nonlinear operation of analog integrated circuits can be gained. For accurate distortion computations, the accuracy of the transistor models is critical. A MOS transistor model is discussed that allows us to explain the measured fourth-order nonlinear behavior of a 1-GHz CMOS upconverter. Further, the method is illustrated with several examples, including the analysis of an operational amplifier up to its gain-bandwidth product. This example has also been verified experimentally. Index Terms---Analog integrated circuits, harmonic distortion, nonlinear ...

### Citations

103 |
Operation and Modeling of the MOS Transistors
- Tsividis
- 1987
(Show Context)
Citation Context ... be written as the product of three functions: (8) The function models the drain current in absence of mobility reduction and velocity saturation. An accurate expression for this function, derived in =-=[17]-=-, is given by in which (9) is the surface mobility in the absence of mobility reduction, is the gate oxide capacitance per unit area, is the flat-band voltage, is the surface inversion potential, and ... |

95 |
Direct-conversion radio transceivers for digital communications
- Abidi
- 1995
(Show Context)
Citation Context ...conductance-C filters [2], [3], the amplifiers are open-loop circuits, which can cause considerable nonlinear distortion. Further, in analog radio frequency (RF) front-ends of integrated transceivers =-=[4]-=-, [5], a knowledge of the nonlinear behavior of building blocks such as low-noise amplifiers and mixers is essential. In this paper, it is described how nonlinear effects in weakly nonlinear, continuo... |

52 |
Distortion Analysis of Analog Integrated Circuits
- Wambacq, Sansen
- 1998
(Show Context)
Citation Context ...ent from the measured ratio. III. CALCULATION OF NONLINEAR RESPONSES In weakly nonlinear analog circuits, harmonics and intermodulation products can be computed using Volterra series [7]. In [23] and =-=[24]-=- a similar method is derived that circumvents the use of Volterra series. The analysis is limited to circuits with at most two input ports or with two input frequencies applied at one input port. In t... |

40 |
BSIM: Berkeley Short-Channel IGFET Model for MOS Transistors
- Sheu
- 1987
(Show Context)
Citation Context ...is the surface inversion potential, and is the body-effect coefficient. In order to have a function that can be evaluated more efficiently, the function very often approximated as follows [16], [18], =-=[19]-=-: is (10) in which various values are used for the parameter . In several models, is a function of only, and not of the other terminal voltages. As a result, the third-order nonlinearity coefficient i... |

23 |
Sansen W.: ISAAC: A symbolic simulator for analog integrated circuits
- Gielen, Walscharts
(Show Context)
Citation Context ...s or intermodulation products are generated by the coupling of the computation method with a symbolic simulator that generates symbolic expressions for network functions of linearized analog circuits =-=[10]. -=-This is possible since the computation method described here reduces to a repeated solution of sets of linear equations. Modern symbolic analyzers [10]–[14] are able to generate approximate expressi... |

17 | Frequency-domain analysis of nonlinear circuits driven by multi-tone signals - Ushida, Chua - 1984 |

15 |
Efficient symbolic computation of approximated small-signal characteristics of analog integrated circuits
- Wambacq
- 1995
(Show Context)
Citation Context ...f the circuit parameters. Aspects of the coupling of the calculation method of harmonics, and intermodulation products with the symbolic network analysis approach for linearized circuits described in =-=[12]-=- are discussed in Section IV. The analysis technique is illustrated in Section V with two examples: the second harmonic at the output of a CMOS Miller-compensated operational amplifier, and the third ... |

11 |
Analysis of nonlinear systems with multiple inputs
- Bussgang, Ehrman, et al.
- 1974
(Show Context)
Citation Context ... only 3 dB different from the measured ratio. III. CALCULATION OF NONLINEAR RESPONSES In weakly nonlinear analog circuits, harmonics and intermodulation products can be computed using Volterra series =-=[7]-=-. In [23] and [24] a similar method is derived that circumvents the use of Volterra series. The analysis is limited to circuits with at most two input ports or with two input frequencies applied at on... |

11 | DC small signal symbolic analysis of large analog integrated circuits - Hsu, Sechen - 1994 |

11 |
Sechen C.: Fully symbolic analysis of large analog integrated circuits
- Hsu
- 1994
(Show Context)
Citation Context ...k functions of linearized analog circuits [10]. This is possible since the computation method described here reduces to a repeated solution of sets of linear equations. Modern symbolic analyzers [10]�=-=��[14]-=- are able to generate approximate expressions, where the approximation is based upon the numerical values of the circuit parameters. Aspects of the coupling of the calculation method of harmonics, and... |

10 |
Design considerations for high-frequency continuous-time filters and implementation of an anti-aliasing fiiter for digital video
- Gopinathan, Tan, et al.
- 1990
(Show Context)
Citation Context ...TRODUCTION IN the analog design community, many circuits are emerging in which nonlinear effects play an important role. For example, in switched-current filters [1] and in transconductance-C filters =-=[2]-=-, [3], the amplifiers are open-loop circuits, which can cause considerable nonlinear distortion. Further, in analog radio frequency (RF) front-ends of integrated transceivers [4], [5], a knowledge of ... |

8 |
Hagreraats “A 2.5-GHz BiCMOS transceiver for wireless LAN’s
- Meyer, Mack, et al.
- 1997
(Show Context)
Citation Context ...ctance-C filters [2], [3], the amplifiers are open-loop circuits, which can cause considerable nonlinear distortion. Further, in analog radio frequency (RF) front-ends of integrated transceivers [4], =-=[5]-=-, a knowledge of the nonlinear behavior of building blocks such as low-noise amplifiers and mixers is essential. In this paper, it is described how nonlinear effects in weakly nonlinear, continuous-ti... |

5 |
Switched-current circuit design issues
- Fiez, Liang, et al.
- 1991
(Show Context)
Citation Context ...istortion, nonlinear distortion. I. INTRODUCTION IN the analog design community, many circuits are emerging in which nonlinear effects play an important role. For example, in switched-current filters =-=[1]-=- and in transconductance-C filters [2], [3], the amplifiers are open-loop circuits, which can cause considerable nonlinear distortion. Further, in analog radio frequency (RF) front-ends of integrated ... |

5 | Interactive ac modeling and characterization of analog circuits via symbolic analysis - Fernández, Rodríguez-Vázquez, et al. - 1991 |

5 |
Small-Signal MOSFET Models for Analog Circuit Design
- Liu, Nagel
- 1982
(Show Context)
Citation Context ...tage, is the surface inversion potential, and is the body-effect coefficient. In order to have a function that can be evaluated more efficiently, the function very often approximated as follows [16], =-=[18]-=-, [19]: is (10) in which various values are used for the parameter . In several models, is a function of only, and not of the other terminal voltages. As a result, the third-order nonlinearity coeffic... |

4 |
A parametric short-channel MOS transistor model for subthreshold and strong inversion current
- Grotjohn, Hoefflinger
- 1984
(Show Context)
Citation Context ... used, this is modeled with a function of the form (14) Here is the critical electric field [17] that is approximately equal to , in which is the saturation velocity and is the effective mobility. In =-=[21]-=- a more accurate model for the function is used: (15) The problem of using more accurate expressions for the functions , , and than the expressions that are used in widely used MOS models, is that the... |

4 |
Symbolic Simulation of Harmonic Distortion in Analog Integrated Circuits with Weak Nonlinearities
- Wambacq, Gielen, et al.
- 1990
(Show Context)
Citation Context ...dB different from the measured ratio. III. CALCULATION OF NONLINEAR RESPONSES In weakly nonlinear analog circuits, harmonics and intermodulation products can be computed using Volterra series [7]. In =-=[23]-=- and [24] a similar method is derived that circumvents the use of Volterra series. The analysis is limited to circuits with at most two input ports or with two input frequencies applied at one input p... |

4 |
Application of Volterra series to intermodulation distortion analysis of transistor feedback amplifiers
- Narayanan
- 1970
(Show Context)
Citation Context ...he output of the amplifier, which is the drain of , is given by (29) In order to know the second harmonic in closed loop, the second harmonic in open loop needs to be divided by , being the loop gain =-=[27]-=-. The second harmonic distortion is obtained by dividing the second harmonic by the fundamental response. Doing so, the poles in (28) and (29) are cancelled, and is computed with the above method as s... |

3 |
A 1-GHz CMOS up-conversion mixer
- Kinget, Steyaert
- 1997
(Show Context)
Citation Context ...ch can be determined. This procedure is illustrated in the following example. B. Example: A CMOS Upconverter Fig. 3 depicts a 1-GHz CMOS upconverter that has been designed in a 0.7- m CMOS technology =-=[22]-=-. The operation of the circuit is based on the mixing operation of transistor which is biased in the triode region with V. In this way, the drain and the source play an identical role. The local oscil... |

3 |
Formula approximation for flat and hierarchical symbolic analysis
- Fernández, Rodríguez-Vázquez, et al.
- 1993
(Show Context)
Citation Context ...errors that must be used for the approximation of every subexpression are computed in advance from the user-specified error for the total expression. This is performed with the algorithm described in =-=[25]-=- that has been developed for the approximation of nested (hierarchical) expressions. The final simplified expression can be expanded on user demand. It must be remarked that the intermediate result ob... |

1 |
A 3.0 V 40 Mb/s hard disk drive read channel IC
- Veirman, Ueda, et al.
- 1995
(Show Context)
Citation Context ...CTION IN the analog design community, many circuits are emerging in which nonlinear effects play an important role. For example, in switched-current filters [1] and in transconductance-C filters [2], =-=[3]-=-, the amplifiers are open-loop circuits, which can cause considerable nonlinear distortion. Further, in analog radio frequency (RF) front-ends of integrated transceivers [4], [5], a knowledge of the n... |

1 |
Distortion analysis of bipolar transistor circuits
- Kuo
- 1973
(Show Context)
Citation Context ...ar behavior of a 1-GHz CMOS upconverter. The method that is used here to compute harmonics and intermodulation products closely resembles the Volterra series approaches that have been used previously =-=[6]��-=-�[9] to compute nonlinear responses numerically. Although the same results are obtained with the Volterra series approach, the method used here is simpler in the sense that it avoids redundant computa... |

1 |
SPICE3 distortion analysis
- Roychowdhury
- 1989
(Show Context)
Citation Context ...ehavior of a 1-GHz CMOS upconverter. The method that is used here to compute harmonics and intermodulation products closely resembles the Volterra series approaches that have been used previously [6]�=-=��[9]-=- to compute nonlinear responses numerically. Although the same results are obtained with the Volterra series approach, the method used here is simpler in the sense that it avoids redundant computation... |

1 |
High-accuracy MOS models for computer-aided design
- White, Wiele, et al.
- 1980
(Show Context)
Citation Context ...on nonlinearity coefficients. The function models the mobility reduction. It has the form The factor (11) is the mobility reduction factor [17]. In many models, the parameter is often as simple as In =-=[20] a more accurate m-=-odel is used: (12) (13) Fig. 2. Relative error in percentage between the nonlinearity coefficients of an n-MOS transistor in the triode region @‡ aQPH "m, v aIR "m, �hƒ aHXRS V, �qƒ ... |

1 |
A physics-based currentdependent base resistance model for advanced bipolar transistors
- Yuan, Liou, et al.
- 1988
(Show Context)
Citation Context ...ter area and at moderate bias currents. At high base currents effects such as current crowding, base pushout and/or base conductivity modulation make the base resistance dependent on the base current =-=[26]-=-, such that its nonlinear behavior needs to be taken into account. The most important transistor parameters are: mA, , , ps, V. Fig. 12 shows the third harmonic distortion as a function of frequency t... |

1 |
Early effect modeling in SPICE
- Andrew, Nagel
- 1996
(Show Context)
Citation Context ... on the base–emitter voltage and the Early effect. The different nonlinearities are modeled with the Gummel–Poon model, with the extension that the nonlinearity of the Early effect is modeled as w=-=ell [28]-=-. However, this nonlinearity plays a negligible role in this example. Further, the base resistance has been considered as a linear element for simplicity. This is a reasonable assumption for transisto... |