IPC9850_Surface Mount Equipment Characterization.pdf - 第21页
IPC-9850 Official Proposal May 2001 21 level of performance over 3 million placements. The same is also true for the mean placements between assists (MPBA) parameter (see 4.2.1), were at least 3 times as many placements …
IPC-9850
Official Proposal
May 2001
20
Table 4-3 Calculation Example
Machine
Number
1 (Note 1) 2 3
Run
Series 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Run
Series 2
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Run
Series 3
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Number
of
Attribute
Defects
0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
Note 1: The first machine built will have 20 runs of 4400 components.
4.2 Reliability Parameters
In contrast to the machine placement performance parameters that specify worst case performance, this standard
uses reliability parameters to describe typical performance. Reliability metrics are difficult for a supplier to specify
because there is no reasonable way to verify these parameters for an individual machine prior to shipment, without
rigorously exercising the machine for thousands of hours. Not only would this be costly, it would result in a new
machine becoming a used machine. Since it is not reasonable to expect all suppliers to dedicate machines for
reliability tests, in this standard the suppliers shall use reliability information collected by users of their equipment.
The problem of specifying reliability metrics is now compounded by the fact that users are collecting the data with
which reliability estimates are made by the suppliers. Each user has its own set of understanding and ability to track
reliability metrics amidst the pressures of daily production. The application of the equipment and skill of its
operators and technicians are believed to be significant factors that contribute to the final observed reliability and
maintainability. In this standard, terminology and metrics have been developed that help suppliers and users collect
and exchange reliability data.
The high yield and reliability of modern SMT placement systems requires very large amounts of data in order to
estimate these metrics. The desire for large data sets having fairly precise estimates of reliability performance must
be balanced with the desire to report reliability as early as possible. Confidence intervals for many machine
parameters can be made on the basis of what is observed in a sample because the underlying statistical distributions
associated with those parameters are fairly well understood. However, at the initiation of this standard, some
suppliers indicate that they do not have a complete enough understanding of the statistical distributions associated
with one or more of the reliability metrics in this standard.
Some types of placement machines produce failure data with χ
2
distributions. At first glance, somewhere around
half the machines would have somewhat better than the reliability levels experienced during the field studies, while
the rest of the machines would experience somewhat worse reliability levels. In fact, if the machines have constant
failure rates, 63% of the machines will have assists and failures more frequently than the observed MPBF and
MPBA. This is due to the fact that most reliability distributions (such as χ
2
) are skewed to the right, causing their
median to be smaller than their mean.
Performance similar to the typical MPBF and MPBA performance of the equipment should be enjoyed when the
recommended preventive maintenance (PM) activities are completed in a regular and timely manner. When
preventive and corrective maintenance is not performed in accordance with the maintenance manual, reliability may
be degraded.
To address confidence interval concerns, equipment suppliers shall only complete the reliability parameters with
data based on at least 3 times as many placements as the reported mean placements between failures (MPBF)
parameter (see 4.2.3). For example, it is not proper to claim MPBF of 1 million placements, without observing that
IPC-9850
Official Proposal
May 2001
21
level of performance over 3 million placements. The same is also true for the mean placements between assists
(MPBA) parameter (see 4.2.1), were at least 3 times as many placements as is reported for MPBA need to be made.
Since several machines shall be utilized for the collection of data for the reliability metric on multiple machines, the
following method shall be utilized for merging the data sets. For example, assume a supplier has data for two
machines of a particular model type. The first machine was observed for 300,000 placements and experienced 2
assists while the second machine was observed for 500,000 placements and experienced 3 assists. The calculation
for MPBA for the model type would be the total number of cycles divided by the total number of assists. In this
case, MPBA would be
000,160
3
2
000,500000,300
=
+
+
cycles.
Other metrics shall also be calculated this way when multiple machines have been observed.
An assist is defined as an unplanned interruption that occurs during an equipment cycle where all three of the
following conditions apply:
Ø The interrupted equipment cycle is resumed through external intervention (e.g., by an operator or user, either
human or host computer).
Ø There is no replacement of a machine part (defined specifically to distinguish from component parts being
placed by the machine), other than vendor specified machine consumable parts.
Ø There is no further variation from specifications of equipment operations.
This definition was clarified not to include replenishment of components since it is tied to the type of feeders and
tape capacity choices made by the user. It was agreed that replenishment is to be specified as a scheduled downtime.
A failure is defined as any unplanned interruption or variance from the specifications of equipment operation other
than assists.
Preventive Maintenance (PM) is a machine stop required by the supplier's published PM schedule.
See table for 4-4, below, for examples of the terms ‘Assist’, ‘Failure’, and ‘Preventive Maintenance” as applicable to
this standard:
IPC-9850
Official Proposal
May 2001
22
Table 4-4 Terms Definition
Feeder assist time: A machine stop required to recover/replace a damaged feeder or "un-jam" a feeder.
Excludes parts replenishment. Measured from the time the operator starts assisting
the feeder until the time that the machine starts running. Excludes operator latency.
Excludes out of spec materials --such as sticking components in tape, delaminated
cover tape, etc.
Software assist time: The machine stops due to a program sequence error, but excludes operator-
programming mistakes.
Fiducial assist time: A machine stop required to find a fiducial that is unrecognizable or outside of the
capture range. Measured from the time the machine stops until the machine starts
running. Excludes operator latency.
Transport assist time: A machine stop required to correct a board jam. Measured from the time the machine
stops until the time the machine starts running. Excludes operator latency. Excludes
assists caused by a board that has been loaded into the machine upside down or
rotated.
Software failure time: A machine stop required by a software defect. Includes the time needed to abort
current operations, reboot (if required) and to restore all data and material to a
production ready condition. Measured from the time the machine stops until the time
the machine is ready to resume normal operation. Excludes operator (and supplier)
latency.
Hardware failure time: A machine stop required by a hardware failure of a factory approved part. Includes all
unscheduled calibrations, adjustments, troubleshooting and part replacements.
Excludes replacement of consumable or wear parts identified in the machine's
manual. Measured from the time the machine stops until the time the machine is
ready to resume normal operation. Excludes operator (and supplier) latency. Excludes
preventive maintenance time. Excludes failures of unauthorized third party parts or
materials.
Preventive maintenance time: Includes all inspections, cleaning, lubrication, adjustment, calibrations and part
replacements. Includes time required to replace all consumable and wear parts
identified in the machine's manual. Measured from the time preventive maintenance
activity begins until the machine is ready to resume normal operation. Excludes
operator (and supplier) latency. Excludes repair time.