37. The Importance Of Accurately Quantifying Defect Losses
There are many essential factors in the effective running of a high quality and economically effective web coating process. One of these factors is to maintain a quantitative detailed accurate database of defect losses in the process. This analysis should provide an ongoing:
• Accurate summary of product yield losses
• What caused the specific losses, streaks, bubbles scratches coverage etc?
• Summary of losses by product and defect type type.
• Specific losses for coating line down, product change, start-up losses
• The ability to calculate dollar losses for each defect
• Easy access to the data for all personnel
Utilization of this database will provide:
• Understanding of defect types in process and where they occur
• The ability to compare product performance and identify better performing products
• The ability to dentify process elements that need upgrading
• With accurate yield loss number the economic Justification for programs and new hardware can be calculated
• A priority list to work on
• An aid in understand mechanisms and
A barrier to developing this database is that often, the yield data collected can be rudimentary and does not accurately reflect the total losses and is not very helpful in identifying the potential causes for defects. The losses are characterized by where they are detected, i.e. in slitting or sheeting process and when the customer uses the product and return the product because it is defective. The classification also can be very general categories, which are not very helpful in finding the sources of the defect, and quantifying losses.
The following is an example of a basic yield loss summary, which provides insufficient information:
• Finishing yield loss 9-62 % of square feet processed
o Pattern length loss 2-14%
o Pattern width loss 1.5-35 %
o Quality yield loss 0-22 %
• Customer returns 10-15 %
For example, what is needed is the exact defects, which resulted in the loss. In regard to pattern width loss there are several potential losses and each may require different corrective action:
• Poor coating weight profile in either machine or transverse direction
• A streak or scratches which prevents using optimum slitting pattern
• Substrate defect such as gauge bands
• Poor edge coverage profile
• Quality Defects
• Missing coating
The database must be maintained in a computer system with intranet access for all company personnel. With advent of computer process control systems, on-line thickness gauges and on-line inspection systems; a significant amount of the data required can be added to the database without requiring personnel costs to add data.
The following is some of the typical data that is needed for an effective database:
• Identification information
• Roll number
• Raw material batches
• slit location
• Defect type
• Amount lost in sq.ft. for each defect type
• Segregated according to product
• Type classification
• Downtime losses,
• Time lost
• Causes such as
• Mechanical failure
• Product change
• Start-up times
• Getting on target for coverage and drying conditions
• Best estimate of defect origination, not where detecte
• Bubble detected in sheeting is coating defect
• Data for each product
The sources of data for the database are:
• Coater logs & inspection data
• Losses from slitters and sheeting
• Customer complaints
• Quality control testing
• On-line inspection data
• On-line thickness gauges
Another requirement of the database design is the ability to interact with other quality control and process databases so that losses can be correlated with process variables to determine defect causes.
One of the difficulties with this database is that there are no standard names and occasionally jargon names are used. This can result in different names for the same defect. One way to minimize this effect is to classify defect by attributes as well as the name. Figure 1 is an example of a classification scheme. An advantage of this approach is that the loss from general defect classes can be identifies and the cost of losses identified, such as streaks, chatter, base defects, etc. In addition, coater losses for operational variables can be determined.
Figure 1
DEFECT CLASSIFICATION
1. Linear Continuous
a. Machine direction
i. STREAKS
ii. RIBBING
b. Transverse direction
i. CHATTER
c. Diagonal
i. DIAGONAL CHATTER
2. Linear Intermittent
a. Machine direction
i. SCRATCHES
b. Transverse direction
c. Diagonal
3. Discrete point defect
a. Spot
i. BUBBLE SPOTS
ii. GELS
b. Contamination coating
i. REPELLENT SPOTS
ii. DIRT
c. Contamination substrate
i. REPELLENT SPOTS
d. Other types
i. REPEAT SPOTS
4. Pattern & diffuse defects
a. MOTTLE
b. DRYER BANDS
c. ORANGE PEEL
5. Process specific defects
a. Raw Materials
i. REPELLENTS
ii. BENARD CELLS
b. Coating
i. AIR ENTRAINMENT
c. Drying
i. MOTTLE
ii. DRYER BANDS
d. Web Transport & winding
i. WRINKLES
ii. SCRATCHES
iii. REPEAT SPOTS
e. Metallizing process
i. BLOOMING
6. Substrate defects
a. Plastic
i. GAUGE BANDS
ii. POLYMER SPOTS
iii. COATING NON-UNIFORMITY
b. Paper
i. WRINKLES
c. MEtalized
T
Recent Comments