Abstract
Sampling is one of the main processes in an analog-to-digital converter. The sampling theory is examined and the crucial elements are extensively discussed. The relation with other techniques such as modulation and sampling of noise is described. The second section discusses the design of time-discrete filters. These filters form an important building block in several conversion structures especially in sigma-delta conversion. Down-sample filters transform the bit-stream format into a more usable pulse code format. The essential properties of finite impulse response and infinite impulse response filters are reviewed.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
“No value is defined” does not imply that the value is zero! There is simply no value.
- 2.
Strange in the sense that many normal mathematical operations cannot be performed, e.g., δ 2(t) does not exist.
- 3.
For clarity this chapter uses for time-domain signals normal print, while their spectral equivalents are in bold face. The suffix s refers to sampled sequences.
- 4.
The Fourier transform definition and its inverse require that a factor 1∕2π is added somewhere. Physicists love symmetry and add \(1/\sqrt{2\pi }\) in front of the transform and its inverse. Engineers mostly shift everything to the inverse transform, see Eq. 2.15. More attention is needed when a single-sided engineering type Fourier transform is used instead of a mathematically more accurate double sided.
- 5.
Other originators for this criterion are named as E. Whittaker and V.A. Kotelnikov. Landau proved in 1967 for non-baseband and non-uniform sampling that the average sample rate must be twice the occupied bandwidth.
- 6.
Precise mathematicians will now argue that a signal with frequency f = f s, ny ∕2 cannot be reconstructed, so they should read here: \(f = f_{s,ny}/2 - \Delta f\), where \(\Delta f\) goes to zero.
- 7.
The only storage in the early days of CDs were video recorders. The 44.1 ks/s sample rate was chosen such that the audio signal exactly fits to a video recorder format (25 fields of 588 lines with 3 samples per line) of 44.1 ks/s.
- 8.
Keep track of: dsp.rice.edu/cs for an overview of all compressive sampling developments.
- 9.
In this book the term down-sampling is used for sampling an analog signal with the purpose to perform a frequency shift of the band of interest. Subsampling in Sect. 2.3.3 removes samples in a predetermined manner from an existing sample stream, but does not change the signal band.
- 10.
Subsampling by a rational factor (a division of two integers) or an irrational factor requires to calculate the signal at each new sample moment by interpolation of the existing samples. This technique is often applied in image processing and is used to combine data from sources with asynchronous clocks. Some fast-running hardware is needed to carry out the interpolation.
- 11.
Formally the result of a Fourier transform reflects the intensity of a process or signal at a frequency. Therefore the result has the dimension “events per Hz” or “Volt per Hertz.”
- 12.
Thermal noise in electronics is modeled as a noise source connected to a real impedance. For circuit calculations this works, but impedances are not noisy, electrons with random energy are.
- 13.
As thermal noise is an atomic phenomenon, its frequency span ends where sub-atomic mechanisms, as described by quantum physics, start. A rule of thumb limits the standard noise spectrum at 1 THz.
- 14.
Very annoying “T” for absolute temperature as well as for fixed time periods.
- 15.
A peak–peak value is often used for jitter, but peak–peak values for stochastic processes have no significance if the process and the corresponding number of observations are not identified.
- 16.
For some reason deciBel is spelled with single “l” although it was named after A.G. Bell. Similarly the letter “a” was lost in “Volta.”
- 17.
Here a strictly formal derivation requires some 10 pages, please check out specialized literature, e.g., [31].
- 18.
In financing a monthly moving average is a very simple FIR filter with 12 taps and simply “1” as multiplication factor.
- 19.
The human ear is sensitive to delay variations down to the microsecond range.
- 20.
The definition for Euler’s relation is: e j π + 1 = 0. According to Feynman this is the most beautiful mathematical formula as it is relates the most important mathematical constants to each other.
- 21.
An equivalent analog filter would require 10–12 poles.
- 22.
In the period 1970–1980 the charge-coupled device was seen as a promising candidate for storage, image sensing, and signal processing. Analog charge packets are in this multi-gate structure shifted, split and joint along the surface of the semiconductor. Elegant, but not robust enough to survive the digital era.
Bibliography
Rabiner LR, Gold B (1975) Theory and application of digital signal processing. Prentice-Hall Inc., Englewood Cliffs. ISBN:0-139-141014
van den Enden AWM, Verhoeckx NAM (1989) Discrete time signal processing, an introduction. Prentice Hall, Englewood Cliffs. ISBN:0-132-167557
Nyquist H (1928) Certain topics in telegraph transmission theory. Trans AIEE 47:617–644; reprinted in Proceedings of the IEEE 90:280–305, 2002
Shannon CE (1948) A mathematical theory of communication. Bell Syst Tech J 27:379–423, 623–656
Shannon CE (1949) Communication in the presence of noise. Proc IRE 37:10–21; reprinted in Proceedings of the IEEE 86:447–457, 1998
Unser M (2000) Sampling-50 years after Shannon. Proc IEEE 88:569–587
Candes E, Romberg J, Tao T (2006) Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information. IEEE Trans Inform Theory 52:489–509
Hajimiri A, Lee T A general theory of phase noise in electrical oscillators. IEEE J Solid-State Circuits 33:179–194 (1998)
Demir A (2006) Computing timing jitter from phase noise spectra for oscillators and phase-locked loops with white and 1∕f noise. IEEE Trans Circuits Syst I 53:1869–1884
Dalt ND, Harteneck M, Sandner C, Wiesbauer A (2002) On the jit-ter requirements of the sampling clock for analog-to-digital converters. IEEE Trans Circuits Syst I 49:1354–1360
Shinagawa M, Akazawa Y, Wakimoto T (1990) Jitter analysis of high-speed sampling systems. IEEE J Solid-State Circuits 25:220–224
Zogakis TN, Cioffi JM (1996) The effect of timing jitter on the performance of a discrete multi-tone system. IEEE Trans Commun 44:799–808
Clara M, Da Dalt N (2011) Jitter noise of sampled multitone signals. IEEE Trans Circuits Syst II 58:652–656
Drakhlis B (2001) Calculate oscillator jitter by using phase-noise analysis. Microwav RF 40:82–90
Nathawad LY, Wooley BA, Miller DAB (2003) A 40-GHz-bandwidth, 4-Bit, time-interleaved A/D converter using photoconductive sampling. IEEE J Solid-State Circuits 38:2021–2030
McClellan JH, Parks TW, Rabiner LR (1973) A computer program for designing optimum FIR linear phase digital filters. IEEE Trans Audio Electroacust 21:506–526 3. Sample-and-Hold
Ali AMA, Dillon C, Sneed R, Morgan AS, Bardsley S, Kornblum J, Wu L (2006) A 14-bit 125MS/s IF/RF sampling pipelined ADC with 100 dB SFDR and 50 fs jitter. IEEE J Solid-State Circuits 41:1846–1855
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Pelgrom, M. (2017). Sampling. In: Analog-to-Digital Conversion. Springer, Cham. https://doi.org/10.1007/978-3-319-44971-5_2
Download citation
DOI: https://doi.org/10.1007/978-3-319-44971-5_2
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-44970-8
Online ISBN: 978-3-319-44971-5
eBook Packages: EngineeringEngineering (R0)