## Time-Interleaved Oversampling A/D Converters: Theory and Practice (1997)

Venue: | IEEE Trans. Circuits Syst. II |

Citations: | 14 - 3 self |

### BibTeX

@ARTICLE{Khoini-poorfard97time-interleavedoversampling,

author = {Ramin Khoini-poorfard and Lysander B. Lim and Er B. Lim and David A. Johns and Senior Member},

title = {Time-Interleaved Oversampling A/D Converters: Theory and Practice},

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

year = {1997},

volume = {44},

pages = {634--645}

}

### OpenURL

### Abstract

In this paper, the design procedure and practical issues regarding the realization of time-interleaved oversampling converters are presented. Using the concept of block digital filtering, it is shown that arbitrary 16 topologies can be converted into corresponding time-interleaved structures. Practical issues such as finite opamp gain, mismatching, and dc offsets are addressed, analyzed, and practical solutions to overcome some of these problems are discussed. To verify the theoretical results, a discrete-component prototype of a second-order time-interleaved 16 analog/digital (A/D) converter has been implemented and the design details as well as experimental results are presented. Index Terms---Converters, time-interleaved, oversampling. I. INTRODUCTION O VERSAMPLING converters have become a popular technique for data conversion [1]. One reason for their popularity is their outstanding linearity which comes from the fact that they usually exploit a 1-b quantizer. Even with a tr...

### Citations

884 |
Multirate Systems and Filter banks
- Vaidyanathan
- 1992
(Show Context)
Citation Context ...ry interleaving number . A. Block Digital Filtering A block digital filter is basically a multirate system in which parallelism is exploited to reduce the speed requirement on each processing element =-=[6]-=-. In this section, the basic concept of these filters are reviewed. Consider the single-input singleoutput transfer function . An equivalentsKHOINI-POORFARD et al.: TIME-INTERLEAVED OVERSAMPLING 635 s... |

255 |
The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms
- Welch
- 1967
(Show Context)
Citation Context ...output spectrum of both structures is illustrated in Fig. 7. The spectrums were generated using the fast Fourier transform (FFT) algorithm. Specifically, the Welch method of power spectrum estimation =-=[9]-=- with 50% overlapping and Hanning windowing was used. The size of the periodogram for the conventional modulator was set at 32k while the timeinterleaved modulator at 64k. Thus, the spectrums for the ... |

74 |
Temes. Analog MOS Integrated Circuits For Signal processing. Wiley series on Filters
- Gregorian, C
- 1986
(Show Context)
Citation Context ... of the charge of the input capacitor will be dumped onto the integrating capacitor resulting in a leaky integrator. It can be shown that the transfer function for this leaky integrator is as follows =-=[10]-=-: Assuming is large and after some algebraic manipulations, (8) can be approximated by The further away is from unity, the more nonideal the integrator becomes. In the context of modulators, this noni... |

66 |
The Design of Sigma-Delta Modulation Analog-to-Digital Converters
- Boser, Wooley
- 1988
(Show Context)
Citation Context ...and applicable to arbitrary architectures. In fact, a double-loop topology was chosen in the next section to illustrate the concept. C. Simulations Consider the second-order A/D converter, taken from =-=[8], -=-as shown in Fig. 6(a). The equivalent block digital filter for assuming is Hence, the equivalent TIM structure for can be derived based on a similar procedure as depicted in Fig. 6(b)–(e). (7) Based... |

39 |
A higher order topology for interpolative modulators for oversampling A/D converters
- Chao, Nadeem, et al.
- 1990
(Show Context)
Citation Context ...for the single-input single-output transfer-function r@�A. Fig. 2. Interpolative structure. Hence, and . Therefore, B. Derivation of the Structure Consider the interpolative structure shown in Fig. =-=2 [7]-=-. Although the interpolative structure does not cover all modulator architectures (such as double-loop or error-feedback topologies), the procedure presented here is general and the above structure is... |

37 |
Temes, “Oversampling Methods for A/D and D/A Conversion,” in Oversampling Delta-Sigma Converters, Theory and Simulations
- Candy, C
- 1992
(Show Context)
Citation Context ...ails as well as experimental results are presented. Index Terms—Converters, time-interleaved, oversampling. I. INTRODUCTION OVERSAMPLING converters have become a popular technique for data conversio=-=n [1]-=-. One reason for their popularity is their outstanding linearity which comes from the fact that they usually exploit a 1-b quantizer. Even with a trilevel quantizer, linearity performance up to 20 b h... |

15 |
Oversampling parallel delta-sigma modulator A/D conversion
- Galton, Jensen
- 1996
(Show Context)
Citation Context ..., Canada M5S 1A4. Publisher Item Identifier S 1057-7130(97)03649-5. 1057–7130/97$10.00 © 1997 IEEE into several subbands which are then applied to separate A/D converters whose outputs are recombin=-=ed [4]-=-. In this paper, an alternate approach is described which is also based on multirate signal processing concepts. By using interconnected modulators working in parallel with each running at the same cl... |

9 |
A 16-mW, 120-dB Linear SwitchedCapacitor Delta-Sigma Modulator with Dynamic Biasing
- Kasha, Lee, et al.
- 1999
(Show Context)
Citation Context ...ularity is their outstanding linearity which comes from the fact that they usually exploit a 1-b quantizer. Even with a trilevel quantizer, linearity performance up to 20 b has recently been reported =-=[2]-=-. However, there is a price to be paid to achieve this degree of linearity. Due to the nature of oversampling, these converters are much slower than their Nyquist-rate counterparts. For instance, the ... |

8 |
Effective dithering of sigma-delta modulators
- Norsworthy
- 1992
(Show Context)
Citation Context ...of . Limit cycles are a disturbing phenomena in conventional converters and are traditionally removed by introducing dithering at the input of the 1-b-quantizer causing an increase in the noise floor =-=[15]-=-. In TIM structures, dithering not only alleviates the problem of in-band limit cycles but also destroys them elsewhere. By dithering, tones around disappear. Consequently, in-band harmonics due to al... |

7 |
High speed A/D conversion incorporating a QMF bank
- Petraglia, Mitra
- 1992
(Show Context)
Citation Context ...y, several authors have utilized the concept of multirate signal processing. In one approach, quadrature-matchedfiltering is used with a bank of independent A/D converters to reduce in-band harmonics =-=[3]-=-. In another approach, a Hadamard transform is used to decompose the input spectrum Manuscript received May 16, 1994; revised February 7, 1997. This paper was recommended by Associate Editor V. Gopina... |

7 |
Time-interleaved oversampling convertors
- Khoini-Poorfard, Johns
- 1993
(Show Context)
Citation Context ...ction V. II. TIME-INTERLEAVED MODULATOR (TIM) In this section, the basic concept of block digital filtering is introduced and is used to develop a technique for realizing time-interleaving converters =-=[5]-=-. A step by step procedure is provided showing how to derive the equivalent timeinterleaved version of an arbitrary converter with an arbitrary interleaving number . A. Block Digital Filtering A block... |

5 |
Mismatch effects in time-interleaved oversampling converters
- Khoini-Poorfard, Johns
- 1994
(Show Context)
Citation Context ...s a maximum dynamic range as was discussed and the optimal in this case is around 0.96. Above this value, the dynamic range quickly deteriorates as integrator outputs saturate. C. Mismatch Effects In =-=[13]-=-, the effect of coefficient mismatch on the performance of a general block digital filter is studied where it is shown that due to mismatches, the overall structure becomes time-varying and hence alia... |

5 |
How to achieve optimum performance from delta-sigma A/D and D/A converters
- Harris
- 1993
(Show Context)
Citation Context ...attenuated by a factor (related to mismatch ratio) and then folded back into the band of interest. Furthermore, modulators tend to generate limit cycles at frequencies of where is the input frequency =-=[14]-=-. Due to mismatch, these limit cycles will also fold back into the baseband and cause tones at as if they were harmonics of the input frequency . Note that these tones are originating from the time-va... |

4 |
On the Use of Chaos to Reduce Idle-Channel Tones in Delta-Sigma Modulators
- Schreier
- 1994
(Show Context)
Citation Context ...ould be noted that leaky integrators pose another problem as well. Specifically, they also cause the zeros of the NTF to move inside the unit circle and hence dead-band and limit-cycles are generated =-=[11], -=-[12]. Since a TIM structure has the benefit of having the NTF zeros closer to 1 for a given leaky integrator, it is also less prone to limit cycles. In summary, TIM’s are less sensitive to leaky int... |

2 | degrees in electrical engineering from the University of Tehran in 1987 and 1989, respectively, and the Ph.D. degree in electrical engineering from the University of Toronto - Sc - 1995 |

1 |
The effect of integrator leak in 1-6 modulation
- Feely, Chua
- 1991
(Show Context)
Citation Context ...e noted that leaky integrators pose another problem as well. Specifically, they also cause the zeros of the NTF to move inside the unit circle and hence dead-band and limit-cycles are generated [11], =-=[12]. -=-Since a TIM structure has the benefit of having the NTF zeros closer to 1 for a given leaky integrator, it is also less prone to limit cycles. In summary, TIM’s are less sensitive to leaky integrato... |

1 |
Analysis methods and time-interleaving architectures for oversampling modulators
- Khoini-Poorfard
- 1995
(Show Context)
Citation Context ...ing significantly improves the dynamic range since the in-band noise is dominated by aliasing. The -factor technique is also applicable to the interpolative as well as the double-loop topologies with =-=[16]-=-. However, the practical realization of time-interleaved modulators with , would be a challenging task. In fact, more SNR have to be traded away to keep the loops operating properly for .sKHOINI-POORF... |

1 | was a recipient of the University of Toronto Open Doctoral Fellowship - Poorfard - 1992 |

1 | M’88–SM’96) received the B.A.Sc., M.A.Sc., and Ph.D. degrees from the - Johns - 1980 |

1 | to 1981, he was with Mitel - From - 1980 |

1 | S’90–M’94) received the B.A.Sc. (Hons.) and M.A.Sc. degrees in electrical engineering from the - Lim - 1994 |

1 | to 1994, he worked on BiCMOS integrated circuits design for analog applications - From - 1992 |

1 | he has been with Lucent Technologies (formerly AT&T Bell Laboratories - Since - 1994 |