JPH0283469A - Inspection of semiconductor device - Google Patents

Abstract

PURPOSE:To reduce measuring irregularity by respectively removing a proper number of measured values from the max. and min. values of measured values before calculating a mean value from the remainder. CONSTITUTION:When (n) measured values of the same sample according to the same inspection method are averaged, (x) measured values are removed from the max. of the measured values and (x) measured values are removed from the min. therefore and a mean value is obtained from the remainder of (n-2x) measured values. By this method, irregularity due to a measuring mistake can be reduced and an accurate mean value is obtained.

Description

[Detailed description of the invention] Industrial applications The present invention relates to a method for testing semiconductor devices.

2. Description of the Related Art Conventionally, when semiconductor testing is performed by programming, an testing method has been used in which the measured values are averaged in order to suppress variations in the measured values of the same sample and the same test.

Problems to be Solved by the Invention However, in the conventional inspection method described above, when a measurement error results in a measurement value that is significantly different from the original measurement value, that measurement value is also included in the averaging. , there was a problem that a correct average value could not be obtained.

The present invention solves the above-mentioned conventional problems.

The purpose is to suppress dispersion by using a correct average value.

Means for Solving the Problem In order to achieve this object, the inspection method of the present invention, when averaging n measured values, removes X from the maximum and X from the minimum, and calculates the remaining (n- 2x) to obtain the average value. however.

The condition is that n>2x.

Effect: By this method, a more appropriate average value can be obtained, and the dispersion of measured values can also be reduced.

EXAMPLE Hereinafter, an example of the present invention will be described with reference to the flowcharts of the drawings.

In this example, the number of measurements, that is, the number of measured values, is 10, 2 from the smallest measured value and 2 from the largest.

Each is removed and averaged.

First, as an initial setting, the number of measurements is 1, and the minimum value is MIN.
Set I to an appropriate value X1. The next smallest value MrN2 after the minimum value is set to an appropriate value x2. Set the maximum value MAXI to an appropriate value Y++ + the next largest value MAX2 after the maximum value to an appropriate value y2. Let the total TTAL of the measured value be O (.

Next, measure and let the measured value be DAT.

Next, DAT is added to TTAL, and the resulting value is set as TTAL.

Next, compare DAT and MINI, and if DAT is smaller than MINI, put DAT in MINI and proceed to the next step. If not, compare DAT and MIN2 and DAT is MI
If it is smaller than N2, put DAT in MIN2 and proceed to the next step. If not, compare DAT and MAXI and DA
If T is larger than MAXI, put DAT in MAXI and proceed to the fifth step. If not, compare DAT and MAX2, if DAT is greater than MAX2, add DA to MAX2.
Enter T and move on. Next, check whether the number of measurements (■) has reached 10 times, and if it has not yet reached 10, add 1 to (■) and repeat the same process starting from the measurement.

When ■ reaches 10, from the total TTAL of the measured value,
MINI.

MIN2. Subtract MAXl and MAX2, and divide that value by the value obtained by subtracting the number of removed measurement values, 4, from the number of measurements, 10.

As a result of this calculation, the resulting average value AVE is taken as the final measured value.

As described above, according to this embodiment, when 10 measured values are averaged, a more accurate average value can be obtained by removing the two largest and two smallest measured values and averaging them. , it is possible to further reduce measurement variation.

In this embodiment, the number n of measured values is 10 divided by the maximum and minimum measured values (the number X is 2, but these numbers may be any value).

However, the number of branches in the flowchart changes depending on the number of branches.

Effects of the Invention As described above, the semiconductor inspection method of the present invention, when averaging n measured values of the same sample and the same test, removes By obtaining the average value for 2x) pieces, measurement variation can be reduced,
This is extremely advantageous in practice.

[Brief explanation of the drawing]

The figure is a flowchart showing a control procedure of a semiconductor testing method according to an embodiment of the present invention. ■・・・Number of measurements (number of measured values), MINl・
...Minimum value, MIN2...The next smallest value after the minimum value, MAXI...Maximum value, MAX2...
...The next largest value after the maximum value, DAT...Measurement value, TTAL...Value added to the measurement value, AV
E: The average value of the total measured values, excluding the minimum value, the next smallest value after the minimum value, the maximum value, and the next largest value after the maximum value, XI... ...MINI initial setting value, x2...MIN2 initial setting value, yl
...Initial setting value of MAXl, y2...
Initial setting value of MAX2.

Claims (1)

[Claims] A method for testing semiconductor devices that suppresses variations in measured values in semiconductor testing by using a program that removes and averages an appropriate number of maximum and minimum measured values for the same sample and the same test.