Doctor of Philosophy (PhD)


Electrical and Computer Engineering

Document Type



Power supply quiescent current (IDDQ) testing has been very effective in VLSI circuits designed in CMOS processes detecting physical defects such as open and shorts and bridging defects. However, in sub-micron VLSI circuits, IDDQ is masked by the increased subthreshold (leakage) current of MOSFETs affecting the efficiency of I¬DDQ testing. In this work, an attempt has been made to perform robust IDDQ testing in presence of increased leakage current by suitably modifying some of the test methods normally used in industry. Digital CMOS integrated circuits have been tested successfully using IDDQ and IDDQ methods for physical defects. However, testing of analog circuits is still a problem due to variation in design from one specific application to other. The increased leakage current further complicates not only the design but also testing. Mixed-signal integrated circuits such as the data converters are even more difficult to test because both analog and digital functions are built on the same substrate. We have re-examined both IDDQ and IDDQ methods of testing digital CMOS VLSI circuits and added features to minimize the influence of leakage current. We have designed built-in current sensors (BICS) for on-chip testing of analog and mixed-signal integrated circuits. We have also combined quiescent current testing with oscillation and transient current test techniques to map large number of manufacturing defects on a chip. In testing, we have used a simple method of injecting faults simulating manufacturing defects invented in our VLSI research group. We present design and testing of analog and mixed-signal integrated circuits with on-chip BICS such as an operational amplifier, 12-bit charge scaling architecture based digital-to-analog converter (DAC), 12-bit recycling architecture based analog-to-digital converter (ADC) and operational amplifier with floating gate inputs. The designed circuits are fabricated in 0.5 μm and 1.5 μm n-well CMOS processes and tested. Experimentally observed results of the fabricated devices are compared with simulations from SPICE using MOS level 3 and BSIM3.1 model parameters for 1.5 μm and 0.5 μm n-well CMOS technologies, respectively. We have also explored the possibility of using noise in VLSI circuits for testing defects and present the method we have developed.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Ashok Srivastava