Data processing: structural design – modeling – simulation – and em – Modeling by mathematical expression
Reexamination Certificate
1998-11-16
2001-07-31
Teska, Kevin J. (Department: 2763)
Data processing: structural design, modeling, simulation, and em
Modeling by mathematical expression
C716S030000, C700S051000, C700S121000
Reexamination Certificate
active
06269326
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to the testing of electronic components.
When they come off the production line, semiconductor electronic components are tested by a testing machine when they are still joined together on a wafer and/or in the state of unit components, packaged or not packaged, possibly in the form of strips.
The testing machine comprises electrodes laid out respectively so as to come into contact with the conductive zones (or regions) of the components. Signals are applied to these electrodes and picked up by them by means of an interface configured by testing means of the machine.
The testing machine comprises a set of electronic modules and software programs whose general function is to compare each measured value with a reference value or interval (or field of validity), in order to make a general decision on whether the component is sound or defective. Finer decisions may be taken, for example when similar components are manufactured together but guaranteed with different tolerance brackets.
More specifically, the software programs may be broken down into a library of tools, a definition of constants, variables and data tables (multiplets) especially for the reference intervals, a sequence of elementary tests for the acquisition of measurements and their comparison with the reference intervals and a final decision mechanism.
The elementary tests of the sequence are stored in a memory of the testing means and subdivided into steps managed by a sequencer. Thus, each test generally comprises the following steps:
a) configuring the interface with a view to the performance of a chosen elementary test (this step consists of the selective designation of the electrodes to which the chosen electrical signals must be applied),
b) after the application of the electrical signal (or signals) chosen for the elementary test, waiting for a first nominal fixed stabilizing period (or at least the longest of the first stabilizing periods) corresponding to the test,
c) selectively picking up, on the designated electrodes, the electrical measurable quantity or quantities designated by the elementary test and representing the responses of the component to the chosen electrical signals,
d) if necessary, waiting for a second fixed measurement period (or at least the longest of the second measurement periods) corresponding to the elementary test in order that the measurement of the designated electrical quantity may be stabilized in the measurement machine,
e) reading the measurement or measurements of the electrical quantity or quantities designated by the elementary test, and
f) comparing the measured electrical quantity with the domain of validity corresponding to the elementary test.
2. Description of the Related Art
Since the number of test operations to be performed is generally very great, the overall wait that results from the adding together of the first and possibly the second waiting phases (or periods) is great (each waiting phase may last at least some milliseconds). This is a major problem in an industry that is constantly seeking higher productivity.
The wait is further aggravated by the fact that the duration of each elementary waiting phase of a testing operation is fixed rather on the higher side, for quality is still more important than productivity. This precaution makes it possible to cover the entire range of variations between the batches.
OBJECTS AND SUMMARY OF THE INVENTION
It is therefore an aim of the invention to enable the reduction of the durations of the waiting phases of electronic component testing machines of the type described in the introduction.
To this end, the invention proposes a method for the testing of electronic components in which:
electrodes of these components are subjected, at an initial date, to electrical potentials,
a measurement is made, at a measuring date, of the values of the potentials set up at the terminals of these components, and
these measured values are compared with typical values in order to accept or reject the components as a function of this comparison,
characterized in that, for the comparison:
a nominal statistical image is determined at a nominal date corresponding to a nominal waiting duration after this initial date, for an acceptable batch of components and for a given test,
an intermediate statistical image is determined for this acceptable batch at least one intermediate date prior to or after the end of the nominal waiting duration,
the earliest possible intermediate date during the measurement duration is chosen by the comparison, using a criterion, of the nominal and intermediate statistical images of the measured values, and
the earliest possible intermediate date is chosen as the date of measurement of the test.
In the invention, the waiting duration is the nominal duration at the end of which the potential is normally stabilized at the terminals of the components. This duration is indicated in practice in the technical specifications of the component. An acceptable component, also known as a component to be accepted, is a component which is furthermore known to be acceptable. A component of this kind may be identified individually. More generally, a component to be accepted is virtual: its characteristics may be the result of statistical measurements performed on a set of components deemed to be individually acceptable. Or again, the characteristics of the component to be accepted are the result of statistical measurements performed on a set of components after the elimination of aberrant measurements.
In the invention, it has thus been realized that the best method for looking for the earliest date consists in carrying out a statistical study on samples for a given test, at a given measurement date, and in searching for a compatible statistical distribution with respect to another measurement date.
Thus, in a preferred embodiment, for a given population, namely for given samples for a given test, at a given measurement date, a computation is made firstly of an average of the measurements, referenced M and, secondly, of a standard deviation of the measurements referenced S (for sigma). It can already be seen that it would have been possible to choose other statistical elements such as the median, the mean of the 1st, 2nd, 3rd or 4th quartile, the differences between these mean values or the like. For the time being, it will be noted that these useful elements will be the mean and the standard deviation.
Then, for the same samples and for the test, but for another earlier (or later) measurement date known as an intermediate date, measurements are made and the same elements, namely a mean M′ and a standard deviation S′, are computed.
Furthermore, a criterion of appreciation referenced CP is defined. In one example, this criterion of appreciation CP is equal to the ratio between a difference of limits and S. In this example, the difference of limits is given by manufacturers tolerance To and CP=To/S. The difference of limits for a measurement in which a result of 5 volts is expected is equal for example to ±0.5 volts. The difference of limits is therefore To=1 volt in this case. If, instead of measuring values achieved at the test points (for example 5 volts), a difference is measured between a value achieved and a value to be achieved, the result can be expressed with greater precision. Above all, in this case, it becomes similar for all the different tests so that one and the same selection method may be used.
A computation is made in the invention for another measurement date, known as an intermediate date, of the result of the same criterion, for example CP′=To/S′. The manufacturer's tolerance remains the same. The standard deviation has changed. According to the invention, it has been planned that, for a given structure of samples, the standard deviation S′ and hence the criterion CP′ depend on the date on which the measurement is made. It has been discovered in the invention that if we g
Frejd Russell W.
Nilles & Nilles SC
Softlink
Teska Kevin J.
LandOfFree
Method for testing electronic components does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for testing electronic components, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for testing electronic components will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2499275