Using EIP Data Rules

Before you begin to configure and define the Quality EIPs, you should understand the data rules associated with the EIPs.

Every Sample Requires Its Own Traceability Entry

Except for defectives type characteristics (y defective out of n), every individual sample in the measurement plan requires its own traceability entry, even if all the entries are identical (the lot number is the same for all samples in the session). This means that the number of rows of traceability has to equal the maximum subgroup sample size of all the characteristics in the measurement plan that defines the session. Defectives type characteristics always count as one sample for traceability purposes regardless of the subgroup size.

In EIP terms, take the maximum of the QS_NVALUES for all the subgroups in the session, and that's the number of rows of traceability information that is required.

Examples

This table lists examples of measurement plans and the EIP requirements.

Characteristic	Data Type	Subgroup Size	Comments
Length	Variables	4
Width	Variables	5	Largest subgroup size.
Bad units	Defectives	1000	Only counts as one row.
Bad electrical	Defectives	0	Only counts as one row.
Defects	Defects	4

Note: Defectives data types always count as one row regardless of subgroup size.

In this example, specify five rows of traceability for this session dataset, because the largest subgroup size is five.

Characteristic	Data Type	Subgroup Size	Comments
Length	Variables	4
Width	Variables	5	Largest subgroup size.
Bad units	Defectives	1000	Only counts as one row.
Bad electrical	Defectives	0	Only counts as one row.
Defects	Defects	0	Variable subgroup size.

Note: Defectives data types always count as one row regardless of subgroup size.

The same rule applies in this example. The number of rows of traceability for the session dataset is still equal to the largest subgroup size. Because the defects subgroup size is variable, the dataset reflects this. Now the number of required traceability rows is five except for those datasets where the defects sample size is greater than five. As a result, the number of traceability rows equals the defects sample size.

Variable and Defects Data Types Have a Maximum Sample Size

Variable and defects have a maximum sample size of 20.

Subgroup Completion Only Applies to Variables Data Type

A subgroup completion only applies to variables type characteristics with accept incomplete option set to Average.

Order and Entry of Subgroup Values is Important

Subgroup values must start at the beginning and be contiguous (no holes in the values). Readings must start at 1 and continue until there are no others. You cannot specify the second and third values without specifying the first value.

This is a valid example:

<QS_NVALUES>3</QS_NVALUES>
<QS_VALUEREADING_1>1</QS_VALUEREADING_1>
<QS_VALUEREADING_2>2</QS_VALUEREADING_2>
<QS_VALUEREADING_3>3</QS_VALUEREADING_3>
<QS_VALUEREADING_4></QS_VALUEREADING_4>
                  .
                  .
                  .
<QS_VALUEREADING_20></QS_VALUEREADING_20>

These are two invalid examples:

<QS_NVALUES>3</QS_NVALUES>
<QS_VALUEREADING_1></QS_VALUEREADING_1>
<QS_VALUEREADING_2>2</QS_VALUEREADING_2>
<QS_VALUEREADING_3>3</QS_VALUEREADING_3>
                  .
                  .
                  .
<QS_VALUEREADING_20></QS_VALUEREADING_20>

Note: You cannot specify values 2 and 3 without specifying 1.

<QS_NVALUES>3</QS_NVALUES>
<QS_VALUEREADING_1>1</QS_VALUEREADING_1>
<QS_VALUEREADING_2></QS_VALUEREADING_2>
<QS_VALUEREADING_3>3</QS_VALUEREADING_3>
                  .
                  .
                  .
<QS_VALUEREADING_20></QS_VALUEREADING_20>

Note: You cannot skip a value. In this example you cannot skip 2.

This is a valid averaging point example, however, QS_VALUEREADING_3 is to be interpreted as 0 that yields a different average:

<QS_NVALUES>3</QS_NVALUES>
<QS_VALUEREADING_1>1</QS_VALUEREADING_1>
<QS_VALUEREADING_2>2</QS_VALUEREADING_2>
<QS_VALUEREADING_3></QS_VALUEREADING_3>
                                       .
                                       .
                                       .
<QS_VALUEREADING_20></QS_VALUEREADING_20>

Be careful when it comes to averaging values associated with the subgroup. In this example, if you were calculating the average for a point calculation, it would be(1+2+0) / 3 = 1. QS_VALUEREADING_3 is interpreted as zero in this particular case.

If you change the QS_NVALUES to 2 and the subgroup size is still 3, with the accept incomplete option on, the average would be (1+2)/2 = 1.5. Therefore, the point calculation would be (1+2+1.5)/3 = 1.5. You can calculate a totally different outcome depending on how you enter the data for the EIP.