TECHNICAL ARTICLE
Originally Printed in the November 2002 Issue of
Circuits Assembly
Magazine
REAL–TIME DATA COLLECTION AND ANALYSIS
Enterprise productivity software can be used to address numerous data
collection and analysis problems faced by electronics manufacturers.
The
first article of this series discussed the use of the web-centric
manufacturing software platform to deliver paperless documentation to
the factory floor through browsers. It explored how process and product
data was prepared in the factory office and made available in a
controlled fashion throughout the factory. This second article extends
the discussion of the web-centric platform into manufacturing execution
system (MES) capabilities. Now, rather than only disseminating
information out to the factory floor, the system is also used to direct
product flow and collect the information emerging from the process as it
executes.
The synergy of process and product preparation in the factory office
tools, and the dissemination of this information to the factory has been
discussed. The prepared process and product data is also critically
important for tracking and quality data collection systems. The benefit
of shared and accessible information is key to the web-centric
manufacturing system. While paperless documentation focused on pushing
information to the process, work in process tracking (WIP) and quality
collection pull data from the process back into the web-centric system.
This data is leveraged to analyze workflow, archive product and process
history, automate rework routing, generate alarms, expedite repair, and
assess quality.
This article elaborates on the data collection and analysis problems
faced by electronics manufacturers and how enterprise productivity
software addresses these problems. It explores the use of the system by
personnel on the factory floor as well as management and engineering
personnel, and the benefits for each.
THE PRODUCT INFORMATION CHALLENGE
Many manufacturers are turning their attention to production information
in their pursuit of greater efficiency. They seek visibility to the
actual performance of their process, while eliminating the overhead
required to gather and compile product flow and quality information.
Still other manufacturers are driven to improve their factory
information management by customer requirement.
Many manufacturers are grappling with traceability requirements for
mission-critical devices. Their customers require efficient recall of
every process a single serialized product experienced, the data for each
component added to the product, the operators who worked upon it, rework
history, etc. Often these customers also expect the data to be
immediately accessible, sometimes over the web.
Even manufacturers not required to maintain traceability data are
interested in learning from product movement. This information is very
useful for identifying and resolving bottlenecks in real-time or
historically. Labor tracking and the ability to refine future quoting
and scheduling practices depend upon such information. The ability to
electronically access a single product’s location in the factory can
eliminate wasted time and effort.
Perhaps the greatest challenge faced by electronics manufacturers is the
effective collection and analysis of quality data and the tracking of
rework. Many generic software systems exist for managing quality data.
However, electronics manufacturers typically require solutions extending
beyond defect collection through on-screen forms. They require graphical
means of collecting defects, simply because product density in
electronics is so much greater than most other types of discretely
manufactured products. Traditional solutions involve paper or on-line
forms into which operators enter defects manually. This information is
later gathered, compiled, and analyzed. Unfortunately, such systems
preclude real-time analysis and alarms, and often the analysis comes so
long after the actual process that it renders the information relatively
useless. They also demand significant labor overhead. Another role in
the factory that is made more difficult by information access challenges
is that of the repair or rework technician. Systems collecting defect
data by form entry make the subsequent localization of defects at the
rework bench time consuming and difficult, as there is no graphical
record of location.
THE WEB–CENTRIC SOLUTION
Manufacturing execution systems (MES) have been in use for many years to
fulfill these requirements. Manufacturers often employ collections of
traditional client-server applications to help satisfy their
requirements, or use an extension of an MRP system. Unfortunately,
several factors conspire to limit the value of these traditional
solutions for the manufacturer. First, the IT architecture of such
systems makes them difficult to maintain and deploy, and often require
customization. In addition, traditional MES solutions are “generic”, in
that they lack inherent knowledge of the product and process, and are
therefore unable to associate collected information to rich product data
and imagery. Web-centric systems offer functionality beyond these
collections of systems to better fulfill the requirements of electronics
manufacturing.
Consider the challenges cited in the previous section. Issues such as
defect data collection on graphical product displays and detailed
process and product traceability archival rely upon a great deal of
product and process information. A web-centric system is capable of
delivering such functionality simply because it has inherent access to
the information required.
To understand why the synergistic relationship of the various sections
of the web-centric system is so critical, and why it so radically
exceeds the functionality of disconnected solutions, begin with a simple
examination of a PCB rework station. At this station, the operator scans
the PCB to recall inspection information. At this moment, a traditional
quality system might display a form of text-based defects and require
the operator to manually search for the defects on the physical product.
Such software is not able to open accompanying data such as the
appropriate process instructions or specification drawings because it
has no connection to revision data. It is not be able to present
drill-down data to AVL and AML data because the relationship of this
information to this particular revision product is not available to it.
Conversely, the web-centric system displays the proper revision of the
product, highlights the defects, presents interactive AVL/BOM drill down
on each part, and allows access to the entirety of controlled
documentation pertinent to that product version. Defects are reworked
and graphically “closed” to aid efficient troubleshooting.
Now the product exits rework and re-enters the flow for re-inspection.
Traditional quality systems might present the inspector with table
entries for collection. Since the system contains only the basic
routing, but not the actual processes within each point, it cannot limit
access to defect attributes pertinent only to that point in the flow.
Without rich knowledge of the process engineering decisions made during
new product introduction, it cannot even filter an image to display only
components populated up to that station in the routing. In fact, most
systems cannot display an interactive product image at all.
In summary, disconnecting bill of materials, process, and revision
management systems from tracking and quality systems limits information
depth; whether it is information presented to the factory for use by
operators, or information derived from the process for use by
management. Web-centric solutions incorporating process and BOM data
preparation solutions solve these problems.
THE FACTORY FLOOR EXPERIENCE
A web-centric tracking and quality system interacts with factory
personnel through its web-browser interface, and sometimes through
automated data collectors which are also web-based. These browsers
present the full depth of product and process knowledge the factory
engineers wished to make available at each routing station. Briefly
exploring the process followed when an operator scans a product with a
barcode reader, or enters a batch movement into the browser illustrates
the scope and capability of such a system.
At the moment of product scanning, the software resolves the barcode
against job, revision, schedule, and floor location to access the
appropriate document set for that location, opening and displaying it to
the operator. Simultaneously, it logs a great deal about the scanning
event into the database including product, operator, time, etc. If the
system is running in batch mode and the operator finds a problem, it
allows that individual product to be serialized for subsequent tracking
through rework. If configured with inboard and outboard station
scanners, the system can literally resolve product location down to an
individual machine or operator station by accessing its serial number
lookup.
If the nature of the product involves variable parameters during
production such as trimming settings or other data unique to each
product built, the tracking system allows the collection of this
information at each station for instant recall in the future.
When the product moves to a station in the flow requiring inspection,
the behavior of the system changes. If the station is a manual one, the
user environment displays the product and offers defect attributes
pertinent only to that point in the flow. The operator collects defects
by clicking on the board image, then submitting them. The system then
makes pre-determined logical evaluations against trends, thresholds,
etc. and indicates where the product should be routed. If the station is
configured as automated, such as from AOI, the information is collected
automatically.
If the product arrives at a rework terminal, it is then scanned to
display all appropriate documentation, part data, and graphical maps of
the defects located for resolution. Depending on configuration,
manufacturers can move the product through multiple “states” of rework,
allowing closed-loop rework validation rather than only ‘defective’ and
‘fixed’. When finished, the system again makes logical decisions based
on its setup to re-route the product back to the primary flow. Systems
allow unlimited nesting of logic decisions for such points.
ANALYSIS FOR THE FACTORY OFFICE
Having explored the source of tracking, quality, and rework data from
the factory floor, it’s important to consider the analysis aspect of the
solution, typically accessed by engineering and management. The first
benefit of this incoming data involves the analysis of production
performance and flow.
Tracking product movement enables “snapshot” views of product locations,
as well as analysis reports of bottlenecks. It also enables serial
number lookup for product history, contents, and present location. This
information is often used to build exact retrofits for product failed in
the field that had unique properties collected during its initial
manufacturing. The virtually unlimited analysis possibilities enable
production planners and foremen to improve scheduling, quoting, and
management practices. Production management has access to the
information they need to improve the way the factory operates.
Quality personnel configure such systems to remove the great majority of
laborious paperwork from their duties, freeing their time to focus on
analysis and on providing feedback for the actual improvement of
process. Unlimited numbers and types of customized analysis reports,
along with real-time drill down tools for instant diagnostics gives
quality engineers the information they need to offer recommendations to
process engineering or management. Sophisticated logic-building tools
enable the construction of completely automated routings, rework loops,
and alarms. Real-time monitoring and reaction to process issues
identifies and copes with problems before they escalate and become
serious. In summary, quality engineers are liberated from clerical work
and are empowered to analyze the real quality information being derived
from the factory.
CONCLUSION
Web-centric manufacturing execution systems harvest information needed
to improve efficiency and meet the demands of customers. They deliver
these benefits without the overhead of manual quality data collection
and product genealogy logging, within an easily deployed and maintained
architecture. The synergy of prepared process and product data with
manufacturing execution systems offers new levels of capability for
electronics assemblers. The next article in this series extends the data
preparation, paperless, tracking, and quality functionality of
enterprise productivity software to real-time production monitoring and
remote web-based visibility of process.
Author Information:
Jason Spera, Chief Executive Officer
Aegis Industrial Software Corporation
220 Gibraltar Road, Suite 100
Horsham, PA 19044