We propose a new I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> testing signature, the graphical I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> signature. We discovered that noise, in the entire set of current measurements for a chip, is a vastly superior feature for classifying chips as good or bad, compared to present methods. The measured I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> current as a function of vectors is defined here as the signature. We examine the shape of the waveform defined by the total set of the I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> measurements, to extract the number of bands that all of the current measurements cluster into, the width and separation of the bands and current glitches or noise among all I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> measurements. We examined the I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> signatures of all SEMATECH experiment chips that were classified as good or bad by a combination of functional, delay, and scan voltage tests. A single I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> threshold, whether absolute or differential, cannot separate good/bad chips reliably. Good chip signatures contain discrete levels (or bands) of varying widths and separations. A faulty chip almost always displays noise and glitches in the band structure. The graphical I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> classifier shows very high accuracy for SEMATECH test data with a test escape rate of 5.97%, compared with 7.5% for the single threshold method, 7.6% for current differences and 7.5% for the DeltaI <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> method. The graphical I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> method had a 1.2% test overkill, compared with 2.3% for the single threshold method, 6.1% for current differences and 7.0% for DeltaI <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DDQ</sub> .