Defect Code Pareto Analysis (Oracle Quality Help)


	Previous	Next		Contents	Index	Navigation	Glossary	Library

Defect Code Pareto Analysis

You are the quality manager for a manufacturer of radio frequency (RF) transmitter equipment used in fiber optics communications applications. Your long-term quality goal is to achieve zero-defect manufacturing. However, you realize that small step improvements are required before your long-term goal can be reached. Therefore your immediate short-term goal is zero-defects in one particular area -- customer shipments.

To meet this goal you must implement procedures that ensure that every transmitter is tested with a power meter at an Inspection station before it leaves the factory. To ensure a "flat" response over a wide range of frequencies, two samples are taken based on a given frequency input. The resultant output measured in decibels (DB) is an indicator of how well the RF transmitter meets customer requirements. The choice of input frequency used depends on the type of transmitter you are building, i.e. on the item model. For instance, for one model, the RF transmitter model, has output of 20 dB +/- 0.5 at an input frequency of 10 megahertz (Mhz) but has an output of 19.5 dB +/- 1.0 at an input frequency of 500 Mhz.

When a transmitter fails test, a Symptom Code that clearly describes the observed results is assigned and the unit is moved to a repair operation. Based on the Symptom Code, directed analysis in the repair area leads to the assignment of a Cause Code and subsequently a Disposition Code.

You can use inquiries, reports, and Pareto Charts to analyze the most common causes of defects. Appropriate corrective actions can then be taken to ensure that your long-term goal of zero-defect manufacturing is met.

Figure 1 - 8 illustrates the quality data collection points discussed above.

Setup

Based upon the above information, you have determined that you need to create a RF Transmitter collection plan that includes the following collection elements:

Item

Serial Number

Transaction Date

Assembler ID

Symptom Code

Cause Code

Disposition Code

Date of Manufacture

Low Frequency

Low dB Reading

High Frequency

High dB Reading

Quality Control Id (Employee Id of the person performing the analysis

Collection Elements

Of the collection elements listed above, six (Assembler ID, Low Frequency, Low dB Reading, High Frequency, High dB Reading, and Quality Control Id) are user-defined. All others are predefined (Symptom Code, Cause Code, and Disposition Code) or context elements (Job, Item, Serial Number, and Transaction Date). See: Defining Collection Elements.

Specifications

Specifications ensure that products are useable and are typically based upon customer requirements. Therefore the next step is to define an item specification, 'RF Transmitter1', specifically for your Model 1 RF Transmitter. Note that you can define separate specifications for each transmitter model that you manufacture.

Add the Low DB and High DB collection elements to the RF Transmitter1 specification. If the required specification limits for these collection elements, 19.5 dB +/- 1.0 and 20 dB +/- 0.5 respectively, have been defined for the collection element they are automatically copied as you add them to the specification. If they have not been defined, you can define them in the Specification Elements window. See: Defining Specifications and Defining Specification Elements.

Finally, add the Low Frequency and High Frequency collection elements to the specification. If the required target values for these collection elements, 10 megahertz (Mhz) and 500 (Mhz) respectively, have been defined for the collection element they are automatically copied as you add them to the specification. If they have not been defined, you can define them in the Specification Elements window. If you set the QA:Default Specifications Target profile option to Yes, these target values are automatically defaulted when the RF Transmitter collection plan is invoked by a move transaction.

Collection Element and Collection Plan Element Values

You can define lists of values or SQL validation statements for the required predefined attribute collection elements (Symptom Code, Cause Code, and Disposition Code) and for your user-defined collection elements. Lists of values can be defined at the collection element level and copied to their corresponding collection plan elements. However, you can also define unique lists of values for the collection plan elements. For example, the list of values for the RF Transmitter collection plan elements Symptom Code, Cause Code, and Disposition Code -- Okay and Failed, Not Applicable and Assembler Error, and Not Applicable and Scrap respectively -- are unique to only this collection plan. See Defining Collection Element Values and See: Defining Collection Plan Element Values.

Attention: Upper and Lower Specification Limits are not required for the Low Frequency and High Frequency specification elements since frequency is an input in testing RF transmitters.

Collection Plan

Next, create the "RF Transmitter" collection plan and add the above listed collection elements to it. All collection plan elements, except Cause and Disposition Code, should be defined as displayed and mandatory. Cause and Disposition Code should be defined as non-mandatory so that they can be skipped when you enter your initial results. Later, after further results analysis has been done, they can be updated using the Update Quality Results window.

Collection Plan Element Actions

Next, define appropriate actions for each of your collection plan elements. For example, when a Symptom Code results value of 265 (indicating no signal output) is entered, you can send a message to the power meter operator that says "Db reading is too low. Please move this item to the repair station." See: Defining Collection Element Actions.

Transactions

Next, associate the RF Transmitter collection plan with the WIP move transaction. The move transaction should be made mandatory to ensure that quality data is collected before the move transaction is saved.

Based upon the moves shown in Figure 1 -- from Final Assembly operation to Inspection, from Inspection to Repair, from Repair back to Inspection -- define the following collection triggers and conditions:

From Operation Sequence Number <= 50 (Inspection)

To Operation Sequence Number > 50 (Inspection)

The above collection triggers and conditions accomplish the following:

Assemblies cannot be completed into inventory from Operation 40 (Final Assembly) without passing through Operation 50 (Inspection)

All moves into Operation 60 (Repair) trigger mandatory data collection

Attention: Multiple results entries for a serial number indicate that the item was repaired, perhaps more than once.

Specification Type

Last, assign a default specification type to the collection plan, in this case an item specification. Assigning a default specification type ensures that an appropriate specification can be chosen during quality data collection.

Data Collection

As transmitters are moved into the Inspection operation, the RF Transmitter collection plan is invoked. Because a specification type has been associate with the RF Transmitter collection plan, a specification -- in this case the RF Transmitter1 specification -- should be selected before data collection can begin.

The inspector is prompted for the serial number of the transmitter and for the assembler's ID (the assembler ID is noted on the routing sheet that the inspector receives along with the transmitter). Low and High Frequency default values, defaulted from the Low and High Frequency specification element target values, are automatically displayed. The inspector tests the transmitter using the suggested frequencies, then enters the corresponding Low DB and High DB results values. If both DB values are within the specification limits defined for the RF transmitter, the inspector enters the "Okay" Symptom Code and "Not Applicable" Cause and Disposition Codes. The tested transmitter can now be completed into inventory.

If however, the Low or High DB value is outside the specification limits, the message action "Item DB is outside specification. Please route item to repair station." is invoked. The inspector selects the failed Symptom Code, but because they are not mandatory, skips data entry for the Cause and Disposition Code. Once quality data collection is complete, the failed transmitter is moved to the Repair operation.

In the Repair operation, a technician determines the exact cause of the problem, fixes it, then updates the Cause and Disposition Code values using the Update Quality Results window. Once the problem is fixed the technician moves the transmitter back to the Inspection operation where the process is repeated until the transmitter passes inspection. Transmitters that pass test can be completed into inventory.

Figure 1 - 9 illustrates the process operations discussed above.

Results Analysis

You can view and report quality results in a variety of ways. You can view all results for a particular serialized item or for a particular Symptom Code = 055 (no signal output). You can also create custom reports based on your results.

You can view your results in graphical form. For example, you can create Pareto chart with using the symptom code as the X-axis element and limiting the results selected to those Cause Codes that are equal to 265. You can them create similar Pareto charts using the same Symptom Code but different Cause Codes to visualize the differences between different cause codes. Cause Codes resulting in a significant symptomatic occurrences indicates that the product or process should be redesigned.

Similarly, you can create then compare Pareto charts for one Cause Code but different Assembler IDs. You can then determine if a particular assembler is contributing to increased transmitter rejects and repairs. Once problems are identified, corrective actions -- such as training or assembly process improvements -- can be taken.

The identification and analysis of production problems are essential if the goal of zero defects is to are realized. Studies have shown that operator and other familiar with manufacturing processes are adept at devising solutions once they are convinced that problems do indeed exist. Pareto and other charts that graphically illustrate problems are vital to any quality improvement effort.