## Analog Testability Analysis and Fault Diagnosis Using Behavioral Modeling (1994)

Venue: | In Custom Integrated Circuits Conference |

Citations: | 9 - 1 self |

### BibTeX

@INPROCEEDINGS{Liu94analogtestability,

author = {Edward Liu and William Kao and Eric Felt and Alberto Sangiovanni-Vincentelli},

title = {Analog Testability Analysis and Fault Diagnosis Using Behavioral Modeling},

booktitle = {In Custom Integrated Circuits Conference},

year = {1994},

pages = {413--416},

publisher = {IEEE}

}

### OpenURL

### Abstract

This paper presents an efficient strategy for testability analysis and fault diagnosis of analog circuits using behavioral models. A key contribution is a new algorithm for determining analog testability. Experimentally, we determined the testability and faults of a fabricated 10 bit digital-to-analog converter modeled using the analog hardware description language, Cadence-AHDL. Also, we applied the testability analysis at the circuit level using SPICE sensitivity analysis. 1 Introduction The goal for testability and fault diagnosis is to identify fault location and severity from measurements. By identifying the faults, we can identify design weak spots and problems in processing, thus pinpointing the critical areas for improvement in the next design iteration. In this context, the testability of a circuit corresponds to the extent faults can be identified from measurements. Fault diagnosis corresponds to the algorithm for computing the severity of faults from measurements. In gene...

### Citations

48 |
Fault diagnosis of analog circuits
- Bandler, Salama
- 1985
(Show Context)
Citation Context ...(3) # (3) where n = m = 3. In this case, U T U is singular, thus, not invertible. 2.1 Ambiguity groups Components whose sensitivity vectors are dependent are said to fall into the same ambiguity group=-=[3, 4, 5]-=-. There can be more than one ambiguity group in a circuit. Each component can either be independent or belong to only one group for the obvious reason that if the component belongs to two groups, the ... |

40 | A top-down, constraint-driven design methodology for analog integrated circuits
- Chang, Sangiovanni-Vincentelli, et al.
- 1992
(Show Context)
Citation Context ...cuits. For instance, in top-down design, designers can explore system behavior quickly without investing time in actual implementation of the components; thus, allowing rapid design space explorations=-=[6]-=-. In bottom-up verifications, designers can verify system implementations quickly because behavioral models are usually simple and quick to evaluate[7]. In fault diagnosis, designers are interested in... |

39 |
Time Series, Data Analysis and Theory, Expanded Edition”, Holden-Day
- Brillinger
- 1981
(Show Context)
Citation Context ...ted in Section 5, followed by a conclusion in Section 6. 2 Testability analysis For the linear error model[1], errors can be diagnosed by solving the system of equations in (1) using linear regression=-=[2]-=- when certain conditions are met. Using linear regression, we compute the optimal least square estimate of component errors,sc = (U T U) \Gamma1 U T s; (2) provided (U T U) \Gamma1 exists. To determin... |

15 |
Saeks: “Fault Diagnosis for Linear Systems Via Multifrequency Measurements
- Sen, R
- 1979
(Show Context)
Citation Context ...(3) # (3) where n = m = 3. In this case, U T U is singular, thus, not invertible. 2.1 Ambiguity groups Components whose sensitivity vectors are dependent are said to fall into the same ambiguity group=-=[3, 4, 5]-=-. There can be more than one ambiguity group in a circuit. Each component can either be independent or belong to only one group for the obvious reason that if the component belongs to two groups, the ... |

6 |
Ambiguity groups and testability
- Stenbakken, Souders, et al.
- 1989
(Show Context)
Citation Context ...(3) # (3) where n = m = 3. In this case, U T U is singular, thus, not invertible. 2.1 Ambiguity groups Components whose sensitivity vectors are dependent are said to fall into the same ambiguity group=-=[3, 4, 5]-=-. There can be more than one ambiguity group in a circuit. Each component can either be independent or belong to only one group for the obvious reason that if the component belongs to two groups, the ... |

4 |
Modeling and test point selection for data converter testing
- Stenbakken, Souders
- 1985
(Show Context)
Citation Context ... (faults) cannot be easily determined for analog circuits. In the case where the component error is small as in mismatches or small variations in component parameters, a linear error model can be used=-=[1]-=-. This model relates the output errors as a linear function of the component errors, s = Uc; (1) where s 2 R n is the output errors, c 2 R m is the component errors, U 2 R nxm is the sensitivity matri... |

2 | Analog system verification in the presence of parasitics using behavioral simulation
- Liu, Chang, et al.
- 1993
(Show Context)
Citation Context ...s; thus, allowing rapid design space explorations[6]. In bottom-up verifications, designers can verify system implementations quickly because behavioral models are usually simple and quick to evaluate=-=[7]-=-. In fault diagnosis, designers are interested in finding components errors in systems such as current mirror mismatches in DAC or comparator offsets in ADC. Computing the sensitivities in such large ... |