NSF Research Program


NSF ITR: Built-In Test of High Speed/RF Mixed Signal Electronics:


WORK SUMMARY:

The major activities of this project focus on developing and testing
analog and microwave IC test circuits that implement an 'alternative
test'-based RF BIST methodology. The overall project is part of a
larger effort multi-university NSF ITR program including researchers
from Georgia Tech, University of Texas at Austin and Auburn University
committed to developing low cost 'alternate test' techniques to
drastically lower the cost of production IC test.

The key goal for this task is to develop 'physical feature extractors'
in coordination with Prof. Abjihit Chatterjee of Georgia Tech. By
'feature extractors' the research requires the development of GHz RF
detectors that sample RF/microwave signals and report the output at
low frequency analog or DC to an ATE system. This can drastically
lower the ATE system requirements, increase test speed and lower
manufacturing part cost. These extractors must be selected and
optimized use in a 'alternative test' methodology. Example feature
extractors include RMS detectors and peak detectors.

Other key tasks are to develop methods of signal sampling and
generation for 'alternative test.' Thus, RF/microwave generators, RF
couplers, dividers, switches, baluns, attenuators are also very
desirable in the 'alternative test' application. IC fabrication for
this project has been performed through the MOSIS-based IBM MEP
donation program.

This project is in its first year (Nov. 2003 funding) and the focus
has been on developing new detectors, signal generators and sampling
circuits for 'alternative test' applications. As part of the effort,
the participating researchers sat down at a collaborative meeting at
National Semiconductor Corporation and redefined the major goals for
this year the effort given the revised level of NSF funding.

RESEARCH ACTIVITIES:

REVISED YEAR ONE GOALS FOR UNIVERSITY OF FLORIDA:

1) Design 1-2 GHz RMS and peak detectors.

Note: the preliminary evaluation from Georgia Tech researchers was
that RMS detectors were more useful than peak detectors for an
'alternate test' response.

2) Develop other high performance detectors for feature analysis.

3) Look at systematic methods of 'non-invasive' insertion of sensors
into high-frequency signal paths.
 
4) Examine the detectors in the presence of process variations.

5) Look at doing external calibration of detectors.

6) Plan to have IC designs for several MOSIS fabrication runs.


This material (in the paragraph and in the web page above) is based upon work supported by  the National Science Foundation under Grant No. 0325340. Any opinions, findings, conclusions or recommendations expressed in the martial are those of the author (William R. Eisenstadt) and do not necessarily reflect the view of the National Science Foundation.

Back to Top