IPC-9850

Surface Mount Equipment Characterization

Assembly Equipment Sub-Committee (5-41)

Official Proposal
May 2001

Published by IPC
2215 Sanders Road
Northbrook, IL 60062-6135
847/509-9700
Fax 847/509-9798
URL: http://www.ipc.org
 IPC-9850
Official Proposal
May 2001
INDEX [This will be generated by the typesetter when the document is published; it’s included just to assist
reviewers and should not be commented against.]

1 Introduction 4.2 Reliability Parameters

1.1 Scope 4.2.1 Mean Placements Between Assists (MPBA)
1.2 Purpose 4.2.2 Mean Time To Recover from Assists (MTTRa)
1.3 Background 4.2.3 Mean Placements Between Failures (MPBF)
1.4 Implementation 4.2.4 Mean Time To Repair Failures (MTTRf)
1.4.1 Characterization Limitations 4.2.5 Total Time
1.4.2 Binding Requirements 4.2.6 Non-Scheduled Time
1.4.3 Test Components 4.2.7 Operations Time
1.4.4 Test Panels 4.2.8 Equipment Dependent Uptime
1.4.5 Measurement 4.2.9 Preventative Maintenance (PM) Time
1.4.5.1 Equipment 4.2.10 Mispick Rate (in PPM)
1.4.5.2 Reporting 4.3 Reporting Sites
1.5 Forms 4.3.1 Number of Machines
1.6 Data Methods 4.3.2 Total Number of Placements
1.7 Terms and Definitions 4.3.3 Total Pickup Attempts
1.8 Units of Measurement
5 Measurement Capability Verification–Form IPC-
2. Referenced Documents 9850-F3

3 Placement Performance Metric 5.1 Gauge Repeatability and Reproducibility Capability

3.1 Placement Performance Form IPC-9850-F1 5.2 Accuracy Capability
3.1.1 General Performance
3.1.2 Performance Validation 6 Test Vehicles
3.2 Characterization Methodology
3.2.1 Background 6.1 Placement Verification PVP Specifications
3.2.1.1 Component-to-Component Variability 6.2 PVP Carrier Specifications
3.2.1.2 Machine’s Component Accomodation 6.3 Glass Slug Specifications
3.2.1.3 Panel-to-Panel Variability 6.4 Sticky Media Application
3.3 Machine Performance Parameters 6.5 Placement Program
3.3.1 Test Conditions 6.5.1 1608C Component
3.3.1.1 Number of Heads/Spindles 6.5.2 SOIC-16 Component
3.3.1.2 Type of Heads/Spindles 6.5.3 QFP-100 and BGA-228 Slug
3.3.1.3 Type of Camera 6.5.4 QFP-208 Slug
3.3.1.4 Number of Feeders/Trays
3.3.1.5 Type of Nozzles 7 Forms and Diagrams
3.3.1.6 Number of Nozzles
3.3.1.7 Number of Panels Built Appendices
3.3.1.8 Number of Parts Per Panel
3.3.2 Time-Based Parameters A. Capability Indexes
3.3.2.1 Build Time B. Specification Limits for Cpk Values
3.3.2.2 Transfer Time (Tt ) C. Guidelines for Adhesive Used for Machine Capability
3.3.2.3 Total Tact Time Testing
3.3.2.4 Tact Time D. Suggested Methodologies for Measuring
3.3.2.5 Net Throughput Components using an Optical CMM
3.3.3 CMM Measured Parameters E. Component Locations for Placement Verification
3.3.3.1 Repeatability Panel Population
3.3.3.2 Accuracy F. Component Location for Accuracy Verification Panel
3.3.3.2.1 Spec limits for Cpk G. How to Perform a GR&R Test
3.3.3.2.2 Cpk for Termination-to-Land Coverage H. Instruction for Using the Gauge R&R Spreadsheet
I. Placement Verification Panel Carrier
4 Attribute Defect Rate and Reliability Performance J. Component Location for GR&R evaluation
Metric-- Form IPC-9850-F2 K. Instruction for Using The ‘Termination-to-Land
4.1 Attribute Defect Rate Calculations.xls’ spreadsheet

2
 IPC-9850
Official Proposal
May 2001
1 Introduction

1.1 Scope
This standard establishes the procedures to characterize machine placement capability of surface mount assembly
equipment in specification documents, as well as in documentation used to verify a specific machine’s placement
capability conformance to the specification, while maintaining a placement accuracy to placement speed
relationship.

1.2 Purpose
IPC-9850 has been developed to standardize the parameters, measurement procedures, and the methodologies used
for the specification, evaluation, and continuing verification of assembly equipment characterization parameters.
These standardized tools shall be used to develop and report the information called out in this standard.

1.3 Background
With the proliferation of Surface Mount Technology (SMT), placement equipment users have struggled with the
question of which machine will perform best in a given manufacturing environment. The advantage of the SMT
assembly process to rapidly place components in precise alignment to the land patterns on the printed wiring board
(PWB) was the initial yardstick by which machines were selected. Machines that could place components the
quickest and with the least amount of scrap were considered the best.

Initially, the most common evaluation method was placement yield. For this evaluation, a machine is made to
populate a large number of the user product where visible placement errors are counted as defects. Machines with
the least defects and the most robust operation were considered best. The high yield and reliability of modern SMT
placement systems require that very large amounts of data be collected to meaningfully assess yield and reliability.
This standard provides new tools for gauging the yield and reliability of placement equipment yet presents
performance results in the traditional metrics.

In addition to the high yield and reliability expected of modern placement equipment, the SMT assembly process
has become significantly more demanding. Components have decreased in size, component terminations are smaller,
and placement locations have moved closer together. All this while the number of components on the PWB and
product volumes have increased significantly. Placement equipment must now place components more rapidly and
with extreme precision to be financially viable. This has made requirements on placement machines more
demanding.

Historically, placement equipment vendors have selected their own parameters and methodologies to present the
specification of their machines’ throughput and placement capabilities. The many representations of this information
have made the comparison between similar types of placement machines very difficult. To obtain comparable data,
users have been forced to conduct on-location evaluations of various machines under the same conditions. This type
of methodology is very time consuming for users and very capital intensive for suppliers.

This standard simplifies the evaluation process by standardizing the performance parameters that describe the
placement machines’ capabilities. It also couples placement throughput and placement quality so speed and accuracy
parameters are dependent on each other. This standard also specifies the methodologies by which the capability
parameters are measured. This reduces potential user-vendor friction created when the user believes the equipment
is not functioning properly. The methodologies specified herein are consistent and verifiable, thus providing
common-ground-methodologies between users and vendors.

These methodologies were achieved by separating machine performance from the rest of the SMT process variables,
which include paste printing, component quality, packaging type and PWB quality. The speed and quality evaluation
methods of this standard specify that measurements will be made by placement of standardized components into
sticky media on clear glass panels. Experience shows that surface mount equipment must perform well on sticky
media before it can perform well in production. Furthermore, improved process capability on sticky tape usually
translates into enhanced process capability in production. Although this method does not provide information that
can be utilized to perfectly predict production quality, this methodology was selected in order to remove as much of
the variation as possible between facilities, products, process, and operators.

While the ultimate goal is to evaluate a machine’s capability to place components in paste on actual PWBs, it is not
currently possible to make such measurements at the required precision and speed. It is anticipated that future in-line
3
 IPC-9850
Official Proposal
May 2001
inspection systems will improve in their ability to measure component location and orientation. In the future it may
become possible to use in-line post-placement (pre-reflow) automatic optical inspection (AOI) systems to measure
the placement machine capabilities.

Due to the convergence of high-speed and fine pitch machines, this standard makes no attempt to separate the two
types of machines. The user is empowered to decide if a particular model is the best solution for the application
based upon the data reported by the supplier.

1.4 Implementation

1.4.1 Characterization Limitations This standard is comprised of a set of parameters that are the lowest common
denominator for surface mount placement equipment. It is be recognized that additional metrics may be of value in
some instances. The collection of parameters selected for this standard comprise the best subset for use as a core set
of requirements to be included with an equipment supplier’s general specification. This core set may change in
future revisions as technology dictates.

In addition, since there are many possible combinations of hardware and software features that are unique to
individual machine types, this standard cannot address every one of them. Such features and options affect the
overall capabilities of specific equipment model and it is left to the user to understand their implications.
Additionally, it is incumbent upon the user to understand the restrictions and leeway provided for each parameter in
this standard so the proper performance conclusions are reached.

1.4.2 Binding Requirements The body of this document is the standard. The word shall is used throughout this
document whenever a requirement is intended to express a provision that is binding. Material in the appendices is
provided only for information and reference.

1.4.3 Test Components Five component types --QFP-100, QFP-208, BGA-228, 1608C capacitor, and SOIC-16 --
were selected to represent the range of component types placed by surface mount equipment. See Table 3-1 and
section 6 for more precise documentation of these component types.

During the verification of a placement machine, one of the goals is to assess the error induced by the placement
machine. To isolate the contribution of the surface mount equipment to the placement error, which is what this
standard intends to hold suppliers accountable for, it is desirable to reduce other effects that may contribute to the
placement error evaluation process. Using nearly perfect components can best reduce the effect of components on
placement error evaluation. Such components minimize the error associated with SMT component-to-component
variation. For instance, 1608C chip capacitors were chosen as test components because the sides of the capacitors
are very precisely and squarely fabricated. Chip resistors were not selected because they are fabricated in a way that
the top edges of the component seen by the vision CMM may not be in good registration with the side or bottom
features of the component used by the placement machines. The electrodes of ceramic chip capacitors are somewhat
problematic because of their bulbous shape.

SOIC-16 integrated circuits were selected as a standard component because of their relatively low cost and sturdy
construction. SOIC-16 components are believed to be representative of a broad class of coarse pitch leaded
components. Almost all SMT placement machines are capable of processing SOIC-16 devices, so this component
makes it possible to compare the performance of the various machine models. Additionally, 1608C Chip capacitors
and SOIC-16s are economically available in tape packaging, which is how most users feed them because it enables
most machines to maximize the placement speed. Both 1608C and SOIC-16 component types must meet the Joint
Electron Device Engineering Council standard (JEDEC Solid State Technology Association, also known as
JEDEC).

Glass slugs are used to present placement machine vision systems with perfect component images that are free of
bent leads and other part imperfections. The glass slugs also allow the inclusion of fiducial markings that the
coordinate measurement machines (CMMs) use to speed up measurements. The slug fiducial markings are measured
by the CMM, instead of the actual slug component features, to represent the location of the slug component feature.
The slug fiducial markings can only be used to represent the position of the slug component feature if the slug has
been United States National Institute of Standards (NIST) certified to that effect. The slug fiducial markings shall
not be processed by the placement machines during the test and are not known to interfere with the normal operation
of the placement machine's vision systems. The clearly defined slug fiducial markings have the benefit of
4
 IPC-9850
Official Proposal
May 2001
minimizing the measurement-induced portion of the total error observed by the measurement machine, and are
particularly important because the specification limits for fine pitch components (QFP and BGA) are so tight.

For this standard, two QFP and one BGA slugs represent component types. Specifically, the QFP-100, QFP-208,
and BGA-228 component types permit the comparison of an extended range of chip-shooters with that of
specialized IC-placers and multi-functional placement systems. To maintain consistency when utilizing slugs, all
slugs in the pickup trays shall have the same orientation.

Other types of SMT components can be studied using the same basic methods defined in the standard. Glass slug
components that represent micro-BGA and Flip-Chips component types are obvious extensions of the standard that
suppliers or users may wish to consider. Rules for adding claims about additional component types are presented
later in this standard.

1.4.4 Test Panels The evaluation methods of this standard specify component placement into sticky media
dispensed on clear glass panels. This approach is advantageous for two reasons. The first is the dimensional stability
of the glass panel, as opposed to epoxy/glass PWB materials that are much more susceptible to shrinking and
wrapping. The second reason is that it permits use of a standard Optical Coordinate Measurement Machine (CMM)
to illuminate the outline of the components. CMMs are able to rapidly measure a large number of components with
very high precision.

In order to streamline the evaluation methodology, a single glass panel type is specified that permits placement
evaluation and verification of a variety of component types. This panel is referred to in this standard as the
Placement Verification Panel (PVP). The PVP has fiducial markings at specified locations with spacing accuracy
traceable to a NIST certification standard. These fiducials are used as references by both the placement and
measurement machines.

The panels accommodate the following component groups (one component group at a time) for this standard:
a. 36 QFP-100,
b. 30 QFP-208
c. 100 BGA-228
d. 80 SOIC16
e. 400 1608C (Capacitors)

Application of the sticky media requires experience. There must be sufficient amount of sticky media to hold the
component in place, but not too much as to cause interference with the back light intensity. Appendix C has
guidelines on the application of the sticky media.

1.4.5 Measurement This standard has been developed with emphasis on assuring that the measurement tools are
capable of properly characterizing the process. Capable measurement tools provide the customer with a
methodology for verifying vendor claims. They also assure the vendor that the customer is properly evaluating the
machine performance. For an Optical CMM to be a capable gauge for evaluating the surface mount process, its
accuracy and repeatability must be significantly higher that the accuracy and repeatability of the surface mount
equipment. The measurement capabilities required by the CMM depend on the type of components being evaluated
and on the specification limits that the supplier is committing to.

The accepted method for evaluating measurement system performance is called Gauge Repeatability and
Reproducibility (GR&R). This methodology determines the measurement system’s ability to consistently measure
the same product over and over again. The methodology requires that the measurement uncertainty (6x the GR&R
error) be at most 25% of the specification range of the product being measured. GR&R studies do not assess the
accuracy of the measurements, just their consistency. An additional check is needed to verify that the CMM
measurement routine is producing accurate results. See Appendix G for discussion.

The accuracy of the Optical CMM can only be evaluated against a certified gauge. The gauge selected for this
standard is a NIST traceable glass verification panel with etched images in the shapes of the selected components.
This panel is measured on the CMM, and the reported positions are compared to the actual positions of the
component images provided from the panel certification.

5
 IPC-9850
Official Proposal
May 2001
1.4.5.1 Equipment Measurement equipment that meets or exceeds the required GR&R and accuracy specification
limits under the specified evaluation procedures for a specified component type is acceptable.

Automated Optical CMM machines are currently favored by many placement equipment vendors and users due to
their measurement speed. The rapid measurement speed is required due the large number of measurements required
to attain statistically meaningful characterization results. Although this method has been demonstrated to work well,
significant limitations are recognized. The limitations restrict interpretation of the results to the actual SMT process.

Since CMMs were originally developed for a machine shop environment, their lighting systems and measurement
tools are not optimized for measuring surface mount components. The machines are applied to the placement
evaluation process by utilizing clear PVPs that make available high quality images on the CMM. The components
are mounted onto the PVP, are illuminated, and then their positions are measured.

Limitations of the CMMs are their cost and lack of mobility. CMMs are expensive since they utilize many features
that provide structural rigidity, precise position, and high quality imaging. They require delicate calibration and are
sensitive to changes in the environmental conditions, which makes them non-portable.

1.4.5.2 Reporting Measurement results are recorded on standardized forms. The Placement Performance Metric is
form IPC-9850-F1. This form serves two distinct functions. The first is to report the general performance by
machine model type and the second is to validate the performance of a specific machine by serial number. Values
for speed and accuracy that are simultaneously achievable are listed. Form IPC-9850-F1 is supported by the CMM
Capability to Evaluate Metric Parameters form, IPC-9850-F3, which is used to demonstrate the measurement
validity of the parameters presented in the Placement Performance Metric.

The Reliability Performance form (IPC-9850-F2) provides an indication of typical results determined from values
reported by users. Unlike the Placement Performance Metric, which cites guaranteed performance, the Reliability
Performance form merely conveys typical reliability levels for the machine model type based upon a nominal
amount of data collected by the placement machine supplier’s customers in their factories. But this reliability
performance form does establish a foundation upon which the industry can raise the reliability, availability and
maintainability of the surface mount equipment infrastructure.

1.5 Forms Reproducible copies of all required forms are provided in Section 7.

1.6 Data Methods

With regards to the statistical treatment of placement deviation data, we have assumed that deviations generally
follow normal (Gaussian) distributions. This may not always be the case. There is nothing inherently wrong with
data having some other distribution. However, depending upon the circumstances, it might mean that the
specification limits at the two prescribed Cpk limits (1.33 and 2.0) contain some percentage of data less than the
99.9968% and 99.9999% percentages normally associated with these Cpk limits. It would be prudent for the user to
verify that the machine has data that reasonably approximates a normal distribution if it is crucial to the user that the
machine's level of inaccurate placements precisely matches those associated with Cpks of 1.33 or 2.0.

1.7 Terms and Definitions

Additional terms and definitions are provided in Table 4-4 and 4.1.1 through 4.1.11.

Fiducials are markings on substrate materials --such as PWBs or glass-- that are utilized by the placement
equipment vision systems to locate the position and orientation of the substrate.

Component refers to surface mount parts which are utilizes in the construction of electronic devices.

Placement Error is defined as the physical distance between where the component was actually placed and where
the surface mount equipment was specified to place the component by the placement program data.

Slug refers to a NIST certified glass panel with etched image of a single Component and reference fiducials.

Xdev refers to the placement error in the X-direction (parallel to the “9850 Verification” label found on the
Placement Verification Panel).

6
 IPC-9850
Official Proposal
May 2001
Ydev refers to the placement error in the Y-direction direction (perpendicular to the “9850 Verification” label found
on the Placement Verification Panel).

θ dev refers to the rotational placement error (about the component’s X-Y area centroid location).

Overhang refers to the portion of the lead’s width, at the lead’s tip, that extends off of the land’s edge. Overhang I
is due to placement error jointly caused by the X, Y, and θ deviations.

Assist is defined as an unplanned interruption that occurs during an equipment cycle where all three of the following
conditions apply:
• The interrupted equipment cycle is resumed through external intervention (e.g., by an operator or user, either
human or host computer).
• 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.
• There is no further variation from specifications of equipment operations.

Build Time - The amount of time it takes the machine to pick and place all the components. This time period
includes fiducial read time and nozzle change time if any.

Failure is defined as any unplanned interruption or variance from the specifications of equipment operation other
than assists. Specifically, some part of the placement machine has to be replaced or the machine had to be turned off
and then back on in order to continue production.

Net Throughput - The number of Components Per Hour (CPH) the machine can place on the verification PVP.

Preventive Maintenance (PM) is a machine stop required by the supplier's published PM schedule.

Repeatability is defined as one standard deviation of the placement error when placing multiple components upon
multiple boards.

Tact Time - The average time required to place a single component while maintaining the specified placement
process capability. Excludes transfer time, fiducial time and nozzle change time.

Total Tact Time - The time required to place all components on verification glass verification panel while
maintaining the specified placement process capability. Excludes Transfer Time, Fiducial Time and Nozzle Change
Time.

Transfer time – The total transport time in and out of the machine, excluding the time the panel spends in the work-
area.
• Movement in and out of work area
• Time to clamp/release the board in the work area

Land The termination areas on the PWB referred to by the IPC-T-50 Terms and Definitions as a portion of a
conductive pattern usually, but not exclusively, used for the connection and/or attachment of components, also
commonly referred to as pads.

1.8 Units of Measurement All dimensions in this document are provided in hard metric (SI) followed by soft
English/Imperial in brackets. All components are referenced as a metric definition, i.e. 1608C is equivalent to the
0603 (60milx30mil) component type.

2. Referenced Documents

IPC-A-610 Acceptability of Electronic Assemblies

IPC-SM-782 Surface Mount Design and Land Pattern Standard

IPC-T-50 Terms and Definitions for Interconnecting and Packaging Electronic Circuits
7
 IPC-9850
Official Proposal
May 2001

SEMI E10-0699E Standard for Definition and Measurement of Equipment Reliability (Semiconductor Equipment
and Materials International 805 East Middlefield Road Mountain View, CA 94043-4080)

MSA-2 Measurement Systems Analysis Second Edition February 1995 published by AIAG (Automotive Industry
Action Group, 26200 Lahser Road Suite 200, Southfield MI 48034

IPC/EIA J-STD-001 Soldered Electrical and Electronic Assemblies standard (sometimes called the National
Soldering Standard)

3 Placement Performance Metric

3.1 Machine Performance Form IPC-9850-F1

This form (see 7 Forms) serves two distinct functions. One is to report the general performance by machine model
type. The second is to validate the performance of each machine shipped, as identified by serial number.
The top of the form is divided into two sections. At the top left is a region for identifying the vendor and machine
model. At the top right there is a region for use only when the form is utilized for performance validation:
• Serial Number
• Build Date and Time

The following information is reported in the body of the form, which is divided into four sections.

Section I informs of the Test Conditions during the build that generated the data reported in form Sections II, III and
IV. The conditions are:
• Number of Heads/Spindles
• Type of Heads/Spindles
• Type of Camera
• Number of Feeders/Trays
• Type of Nozzles
• Number of Nozzles
• Number of Panels Built
• Number of Parts Per Panel
and describe the hardware and software setup of the equipment for the reported results.

Section II presents the time-based parameters measured during the build cycle as components are placed. The
parameters are
• Build Time, seconds
• Transfer Time, seconds
• Tact Time, seconds

The Net Throughput (in CPH) also is attained from the time-based parameters.

Section III and IV present CMM measured and analyzed performance parameters attained for the panels populated
during the data collection for Section II while under the equipment setup of Section I:
• Repeatability (one standard deviation)
• Accuracy Spec. Limits for Cpk=1.33 for
• Accuracy Spec. Spec. Limits for Cpk=2.0
• Cpk for Termination-to-Land 50% Coverage
• Cpk for Termination-to-Land 75% Coverage

The Repeatability and Accuracy parameters are individually calculated for the X-axis, Y-axis, and θ rotation, while
the Termination-to-Land calculations integrate the X-axis, Y-axis, and θ rotation placement errors. See Section
3.4.3.2.2 (Cpk for Termination-to-Land Coverage) for details.

Appendix D provides a guide for two possible methods of measuring the center locations of the components to attain
the measurement error.
8
 IPC-9850
Official Proposal
May 2001

3.1.1 General Performance When IPC-9850 is used, Form IPC-9850-F1 shall be used to present the placement
capability of a specific machine model when the metric is derived using the methodologies of this standard.
Completed copies of this form shall be used as part of a placement machine model’s documentation and
performance package (i.e. it applies to all machines of the model type listed).

3.1.2 Performance Validation Form IPC-9850-F1 shall be used to validate the performance of a specific individual
machine prior to customer delivery. The vendor shall provide the customer form IPC-9850-F1 with data for at least
one component type placed by that specific machine and derived using the methodologies of this standard. The
vendor shall provide data for the most representative component placed by that machine, or as otherwise agreed
upon between the customer and the supplier. When form IPC-9850-F1 is used to validate a specific machine, the
serial number of the machine and the date of the build shall be provided.

A specific exception is provided for the performance verification using the 1608C component type only. The
exception is that the supplier may choose to measure only rows 1, 5, 9, 13, 17, and 20 (where row 1 is defined to be
the first horizontal line of components above the ‘IPC 9850 Verification Panel’ text) per panel when verifying the
performance of the shipping machine. This exception is granted in order to reduce the time suppliers spend in
preparation to ship a machine, specifically to reduce the CMM measurement time and to reduce the amount of tape
stretch the components on the panels waiting to be measured are exposed to after placement but prior to
measurement. However, the panel must still be populated with all 400 components in according to the procedures of
this standard.

Supplier commitments regarding all component types shown in IPC-9850-F1 that are within the capability claims of
the equipment must be included in the Machine Model type evaluation. Additionally, columns for other component
types may be added at vendor discretion. If this is done, these guidelines shall be followed:
1) at least 30 components/ slugs, or as many as will fit on the 8” x 8” Placement Verification Panel.
2) All 4 orientations must be utilized equally (unless the placement machine is not capable of placing at multiple
orientations).
3) The entire width and length of the area inside the Placement Verification Panel’s fiducials should be used.
4) The placement pattern should be reasonably balanced, density-wise, from left-to-right and top-to-bottom.
5) For boards shall be run.
6) If the pattern contains less than 100 parts, all parts shall be measured.
7) If the pattern contains at least 100 parts, then if orientation is identical throughout a row, a systematic sample of
columns must be measured for accuracy, while if orientation is identical throughout a column, then a systematic
sample of rows must be measured for accuracy.

3.2 Characterization Methodology

3.2.1 Background Many methods for evaluating the performance capabilities of placement equipment have been
employed in the past. The preferred method for this standard utilizes a non-contact optical CMM for measuring the
location of components with respect to panel fiducials. This method was selected because many vendors and users
have extensive knowledge of the method and have already developed evaluation and diagnostic tools utilizing
optical CMM equipment.

The characterization methodology for obtaining numeric values for the specified parameters was designed to be
repeatable and reproducible, and independent of a specific user product. To meet this goal, a standard PVP is
specified as a common test substrate for all procedures (see Section 6.1). This test vehicle is laminated with a layer
of adhesive that is used to capture and hold the mounted components.

The characterization procedure yields a set of performance parameters. These parameters are evaluated through the
population of four PVPs. The four panels must be populated consecutively by the machine, as though they were four
adjacent PWBs in a production environment. Machines that buffer PWBs pre and post population may utilize
additional panels to obtain the appropriate pulse rate, however only four consecutive panels are used for the analysis.
PVP carriers may be used at vendor’s discretion, as long as the carrier supports exactly one PVP. See Appendix I for
a PVP carrier designed by committee members.

3.2.1.1 Component-to-Component Variability Variability due to component-to-component physical differences is

reduced through the use of glass slugs to represent the QFP-100, QFP-208 and BGA-256 fine-pitch components.
9
 IPC-9850
Official Proposal
May 2001
The glass slugs are specified in Section 6.3. Due to the high quantity and packaging issues of discrete components,
placement test methods utilize production quality SOIC-16 and 1608C components that are specified in Section 6.5.

3.2.1.2 Machine’s Component Accommodation While the standard PVP is designed to accept a variety of
component types, placement patterns are prescribed only for a single type of component per PVP. This is done to
avoid a situation where certain machines are not capable of placing all of the specified components. Specific
examples include high-speed placement machines that are not designed to place QFP or BGA components and fine-
pitch machines that are not designed to rapidly place 1608C components. See Appendix E for the placement
locations to be used for each component type.

Each supplier may apply placement optimization procedures to the order by which components are placed, which
nozzle places a specific component, and which camera is utilized. However, it is expected that all spindles/ nozzles
shall be utilized as equally as possible, and that within the placements made by a particular spindle /nozzle, all
nozzle & rotation combinations shall be used as equally as possible (e.g. no manual manipulation to optimize the
pattern to exclude certain nozzles is permitted). At the very minimum, the population of the panel only requires the
use of one head and camera combination. The user has the responsibility to determine whether single or multiple
head and camera combinations optimize the balance between speed and accuracy results for a model type.

3.2.1.3 Panel-to-Panel Variability To captures both “within board” and “between boards” sources of variability for
each component, four-PVP are populated. The four panels are utilized to capture panel-to-panel placement
variation due to the fiducial finding process. The fiducial finding process accounts for imaging system errors in
reading the fiducials, mathematical analysis in accounting for panel orientation, etc.

3.3 Machine Performance Parameters

3.3.1 Test Conditions The Test Condition parameters section of IPC-9850-F1 report the machine conditions during
the execution of the speed and repeatability/ accuracy procedures. These conditions are selected to provide sufficient
information for a user to interpret and reproduce the documented performance.

3.3.1.1 Number of Heads/Spindles For this standard, this is the number of heads/spindles utilized by the machine
during the evaluation of performance. Each head/spindle should be used approximately equally. The placement
program is not to be manually optimized to avoid certain spindles or rotations. (A spindle moves up and down to
pick and place components. Nozzles are attached to the end of spindles and adapt spindles to a particular range of
component types. Some machines have multiple nozzle spindles. Some machines have multiple spindle heads. Other
machines have multiple heads.) Report the total number of heads/spindles used.

3.3.1.2 Type of Heads/Spindles This refers to kind of heads/spindles used to align the components. (Some
machines use one kind of head/spindle for fine pitch components and another kind for other components.) Report
the type of heads/spindles used.

3.3.1.3 Type of Camera

For this standard, this refers to the kind of camera used to align the components. (Some machines use one kind of
camera for fine pitch components and another kind for other components. Others are capable of aligning a particular
component type with more than one kind of camera.)

3.3.1.4 Number of Feeders/Trays

For this standard, this refers to the number of component feeders or matrix trays utilized by the machine during the
performance evaluation. (The number of feeders used affects the speed of some types of machines. Sometimes
multiple feeders per part number are required to maximize the throughput.)

3.3.1.5 Type of Nozzles

For this standard, this refers to the type of nozzle utilized by the spindles during the performance evaluation. (Some
machines may use more than one type of nozzle to pick and place a particular component type.)

3.3.1.6 Number of Nozzles

10
 IPC-9850
Official Proposal
May 2001

For this standard, this refers to the number of nozzle utilized by the heads/spindles during the performance
evaluation.

3.3.1.7 Number of Panels Built

For this standard, this refers to the number of panel populated during the evaluation of performance parameters.

Calculation Method - The number of panels is specified in the procedure to be four.

3.3.1.8 Number of Parts Per Panel

For this standard, this refers to the number of components placed on each panel during the evaluation Performance
parameters.

Calculation Method - The number of component placed on each panel is specified by the placement program
corresponding to the specific part type.

3.3.2 Time-Based Parameters

These parameters describe the defined time periods that a board endures during a complete cycle of placing
components on a PWB (glass verification panel in this instance). Figures 3-1 and 3-2 show how these parameters are
defined.

A Single Panel Cycle

First Last placement
placement

Fiducial Panel
Panel Clamping Nozzle Unclamping
Read Transfer
Transfer Exchange
Out of
into Total Movement
Movement into Work
Work Tact Out of
Buffer Zone Area
Area Time Buffer Zone

Transfer Time Build Time Transfer Time

Figure 3-1: Performance Parameter Description for a Single Panel

A Four Panel Cycle

Panel #4 Panel #3 Panel #2 Panel #1
Transfer Time Transfer Time Transfer Time Transfer Time

Total Tact Time Total Tact Time Total Tact Time Total Tact Time

BuildTime BuildTime BuildTime BuildTime

11
 IPC-9850
Official Proposal
May 2001
Figure 3-2: Performance Parameter Description for a Four-Panel Build

These measured parameters -- build time, transfer time and tact time -- must be obtained for the same panel
population build for which the repeatability and accuracy performance values are obtained.

The measurement procedures for these measured time-based parameters require the use of a stopwatch and/or an
oscilloscope. Stopwatches are utilized to measure the duration of a cycle. Oscilloscopes are usually utilized to
monitor hardware transitions in order to identify the start and end of a cycle for transitions that are not be easily (i.e.
repeatable and/or accurately) detectable with the naked eye.

The ability to measure a time-based parameter must be examined by the user to assure that measurements are
precise, i.e. both repeatability and accuracy are acceptable. The stopwatch and/or oscilloscope must have resolution
of at least 0.01 second or better. A GR&R study must be performed to verify that the repeatability of the
measurement is less than 0.01 seconds at a precision/tolerance ratio of better than 25% (see GR&R Discussion
Appendix G)

3.3.2.1 Build Time

Build time for this standard is the average time required to assemble each standard panel. It includes the time
required to align the panels as well as the time to place the components and change nozzles. Build time excludes the
time required to transfer the board into and out of the workstation. Some machines overlap the fiducial alignment
operation with the placement operation. Other machines overlap nozzle changing and pickups with the transfer
operation. The build time ignores these factors.

Measurement Procedure - Use a oscilloscope or stopwatch to measure the amount of time a panel is in the work-
area. Start the timing cycle when the clamp closes. Stop the timing cycle when the clamp opens. Populate four
panels and average the four measurements to obtain the build time.

Calculation Method - Build four panels and average the four measurements to compute the value of the build time
metric.

3.3.2.2 Transfer Time (Tt)

Transfer time for this standard includes the time required to move the board into the workstation, clamp the board,
release the board and move the board out of the workstation. It represents the overhead associated with transporting
the board when production is flowing normally.

Measurement Procedure - Use a oscilloscope or stopwatch to measure the time from the entry of the first panel to
entry of the fifth panel (not required to be a verification glass verification panel), less the entire Build Time of the
four panels. To minimize measurement error, some easily detectable and clearly defined point in the cycle should be
utilized. Machines that buffer PWB's pre and post population may utilize additional panels to obtain a sustainable
pulse rate, but only the set of four consecutive glass verification panels is acceptable for a proper characterization
procedure.

Calculation Method - Take the time from the entry of the first panel to entry of the fifth panel (not required to be a
verification glass verification panel) minus the entire build time of the four panels divided by four to compute the
transfer time.

3.3.2.3 Total Tact Time

For this standard, Total Tact Time is the required time to place all components on verification glass verification
panels while maintaining the specified placement process capability. It excludes transfer time, fiducial time and
nozzle change time.

Measurement Procedure – Start the oscilloscope or stopwatch to measure the time at which the first component is
placed, and stop the oscilloscope or stopwatch at the time the last component is placed for each of the four panels.

3.3.2.4 Tact Time

12
 IPC-9850
Official Proposal
May 2001

For this standard, tact time is the average time required to place each of the standard components at the standard
CAD coordinates on four standard glass verification panels. Tact time excludes the time during which the board is
being moved into position, clamped and aligned as well as the time required for the machine to change nozzles. The
standard CAD coordinates cause the machines to place components across the surface of a 200-mm2 board and
preclude tact times where components are placed artificially close together to minimize motion time. To the extent
that a machine is able to overlap pickup operations with board transfer operations, the tact time of this standard is
somewhat optimistic. This standard permits the supplier to define the feeder configuration and the pick and place
sequence used during the measurement. The number of feeders and heads are reported so the reviewers of the
standard tact time data understand the conditions required to attain it. The tact time is measured during the same
runs used to gather the process capability data. This ensures that placement process parameters are optimized to
achieve the best balance of speed and accuracy.

Calculation Method - Average the four total tact time measurements and divide by the number of components minus
one on a single panel to compute the tact time.

3.3.2.5 Net Throughput

For this standard, it is the number of components per hour (CPH) the machine can place on the verification PVPl.
This definition is implemented to provide a measure of equipment capability in a term that is common in the
industry. Although the value obtained for this parameters does not correlate to a production PWB throughput (since
each panel only has one type of component), it is a useful parameter for the comparison between equipment types.

Calculation Method – The number of components placed on a single glass verification panel, divided by the sum of
the build time and the transfer time parameters divided by 3600.

Example for QFP-100:

Net Throughput = 36 * 3600

(Build Time + Transfer Time)

3.3.3 CMM Measured Parameters

The following is the measurement procedure utilized for collecting data for repeatability and accuracy. Four
standard panels are treated with adhesive (see Appendix C for guidelines for adhesive application). The four panels
are consecutively populated with the placement program specified for the part type. The panels are then placed on a
capable Optical CMM and the component placement error along the X, Y and θ axes are measured.

The measurement system verification for these parameters is discussed in section Measurement Capability
Evaluation Section 5.

3.3.3.1 Repeatability

For this standard, repeatability is defined as one standard deviation of the placement error when placing multiple
components upon multiple PVPs. A deviation is defined as the placement error experienced when the machine
places a component. The placement error is defined as the distance between the actual center location of the
component to the specified CAD location with reference to the board fiducials. Appendix D provides a guide for two
possible methods of measuring the center locations of the components. The errors are defined where Xdev refers to
the placement error in the x-direction. For use of this document, the x-direction is defined as parallel to, and the y-
direction is perpendicular to, the lettering on the placement verification panel . θdev refers to the rotational
placement error (about the component’s X-Y area centroid).

13
 IPC-9850
Official Proposal
May 2001
Calculation Method
1 n 1 n
The average of the Xdev’s, X ,= ∑ X i and Standard Deviation of the Xdev’s, sx , =
n i =1
∑
n − 1 i =1
( X i − X )2 ,

where i refers to the slug or component number,

Xi = refers to the x error of the ith slug or component,
and n refers to the total number of slugs or components placed over all boards and all locations. Example, when the
placement of QFP-100 or BGA-228 slugs is evaluated, 36 slugs are placed per PVP on a total of 4 panels; thus
n=36*4 = 144.

3.3.3.2 Accuracy

In industry, many suppliers specify performance against their own unique specification limits and capability index
values (e.g. ±40 um with Cpk ≥ 1.5). One of the aspects of this IPC-9850 standard is that, instead of having a
different specification limit and a different sigma level for each supplier, results from each shall be provided in
terms the specification limits required to sustain Cpk’s of 1.33 and 2.0. Specification limits are directly comparable
from one machine to another when machines are reported at the same Cpk levels.

As long as surface mount placement equipment has been on the market, the manufacturers’ claims and guarantees
regarding placement accuracy levels have been cited in terms of X, Y, and θ (rotational) axes, separately. This
traditional method of reporting performance, where performance is considered individually per axis considered, is
covered in Sections 3.4.3.1 and 3.4.3.2.

A relatively recent method that considers the collective effect of all axes is also included in this standard (see
Section 3.4.3.2.2). The essential advantage of this method is that it relates more directly to the soldering process than
the traditional method, because there are combinations of the X, Y, and θ deviations that separately may fall within
spec, but when combined, sometimes prevent the formation of an adequate solder joint.

3.3.3.2.1 Spec limits for Cpk

For this standard, spec limits for Cpk are the specification limits for which the machine is capable of placing the
specific component type with reference to the panel fiducials. This parameter indicates the machine's ability to
provide a centered placement for a given value of the process capability index, Cpk. Limit for a Cpk of 1.33 implies
a capability of 64 PPM, and a Cpk of 2.0 implies a capability of 0.002 PPM. Users who wish to use some other level
of defect rate (besides Cpk’s of 1.33 or 2.0) can easily convert the information provided here to evaluate machine
performance against the preferred capability level.

Calculation Method -

SL = 3S x Cpk+ |avg|

See Appendix A for discussion of Capability indices and Appendix B for discussion of the Specification Limits
associated with Cpk Values and an example of this calculation.

3.3.3.2.2 Cpk for Termination-to-Land Coverage

Many defects are due to the combination of moderately large X, Y, and θ placement errors, rather than just one
prevailing X, or Y, or θ placement error. This approach considers the combined effects of X, Y, and θ placement
errors and is referred to as overhang. Overhang is utilized to determine the amount of overlap between the
termination shape and the land pattern. In a broader definition, termination refers to lead, end-cap, ball, or column
(to name a few common terminations), depending upon the type of component being placed.

Two performance parameters chosen for Form IPC-9850-F1 evaluate the termination-to-land percentage. These two
parameters are the machine’s Cpk against the class 1 and 2 specification limits (50% max overhang of lead’s width),
as well as class 3 (25% max overhang of lead’s width), based on IPC-SM-782 and IPC A-610. These parameters are

14
 IPC-9850
Official Proposal
May 2001
evaluated with mathematical equations that quantify total error, i.e. the combined impacts of the X, Y, and θ
deviations. The ‘Termination-to-Land Calculations.xls’ spreadsheet provides an instruction guide and a sample lead-
to-land spreadsheet that supports these equations.

For leaded components, total error is referred to as maximum lead tip error (MLTE) . The word maximum is a part
of the metric name because the metric refers to the tip error experienced by the lead most impacted by the particular
combination of the X, Y, and θ deviations as shown in Figure 3-3. By definition, Sx refers to the component span
(lead tip to lead tip) in one direction (X) while Sy refers to the component’s span in an orthogonal direction (Y). For
area array components, total error is referred to as maximum ball error (MBE). Total error is then used to calculate
termination-to-land. Termination-to-land evaluation for the 1608 component is not performed, since the end-cap to
land proportions do not necessarily provide meaningful results. In this standard, the widely embraced IPC/EIA J-
STD-001 and IPC-A-610 standards are utilized.

Figure 3-3

Total error (for leaded and area array components) is then used to calculate termination-to-land, based on the
nominal component dimensions and the minimum land dimensions (IPC-SM-782). For leaded devices, termination-
to-land is based on the width of the lead. For area array components, termination-to-land is based on the percentage
of the land’s area that is in contact with the ball or column based on a two-dimensional mathematical model (not the
percentage of the ball that is on land since the land typically is designed to have a smaller surface than the ball). The
mean and standard deviation of these values for a group of placements are computed. The mean and standard
deviation are used to attain the Cpk’s values.

For this standard, Cpk for termination-to-land coverage is the parameter that quantifies the placement machine’s
capability of placing the component terminations on the associated lands. That is lead-to-land (LTL), ball-to-land
(BTL), depending upon the component type, in terms of the amount of termination that ends up on the land as a
result of the machine’s placement.

15
 IPC-9850
Official Proposal
May 2001
Land sizes are defined by IPC-SM-782. Termination sizes used here are the nominal values (average of the
minimum and maximum values) (see Table 3-1). The Cpk is calculated relative to the 50% and 75% coverage called
for in the assembly standards.

Table 3-1 NEEDS TITLE

Component Component Termination Land Spec Limits for Spec Limits for
Type Length x Width Length x Total Error Total Error
Width Width (Class 1,2) (Class 3)
SOIC-16 8.89 x 6.0 0.42 NA x 0.60 0.3001 0.1951
QFP-100 16.0 x 16.0 0.20 NA x 0.30 0.1502 0.1002
QFP-208 32.0 x 32.0 0.20 NA x 0.30 0.1502 0.1002
BGA-228 15.0 x 15.0 0.50 (Dia.) 0.45 (Dia.) 0.2073 0.1143

Notes
1. Requiring Total Error (for leaded this is MLTE) to be ≤ 0.195 for this SOIC is equivalent to requiring Lead-to-
Land to be ≥ 75%. The Total Error limit of 0.3 corresponds to LTL ≥ 50%.
2. Requiring Total Error (for leaded this is MLTE) to be ≤ 0.100 for this QFP is equivalent to requiring Lead-to-
Land to be ≥ 75%. The Total Error limit of 0.15 corresponds to LTL ≥ 50%.
3. Requiring Total Error (for area array this is MBE) to be ≤ 0.114 is equivalent to requiring Lead-to-Land to be ≥
75%. The Total Error limit of 0.207 corresponds to LTL ≥ 50%.

Calculation Method Calculation methods are provided for three component types: two leaded, one area array.
Termination-to-land calculations are based on IPC-SM-782 designed termination and designed land dimensions,
rather than actual ones. At time of publication of this standard, there is no standard for ball to land ratios for area
array components. A 0.5 mm diameter ball with a 0.45 mm land for the BGA228 was selected for calculation of this
metric.

Lead-to-land (LTL): is based on the percentage of the lead’s width that is placed on land. This measurement
quantifies the error for a component’s lead that experiences the most offset (of all the leads) due to the joint x, y, and
θ errors.

Calculation for a component with leads on four sides is shown here:

  MLTE − LandWidth−2 LeadWidth 

LTL= 100 −  ⋅ 100%  , where MLTE is defined below:
 
 LeadWidth

 
For a rectangular or square component with leads on four sides:

 Sx
 ydev + ⋅ sin( θdev )
2
Maximum Lead Tip Error (MLTE) = max 
 Sy
 xdev + ⋅ sin( θdev)
 2

For a rectangular or square component with leads on its two longest sides:

If the component is placed so that the long sides are along the board’s X-axis,
max( S x , S y )
Maximum Lead Tip Error (MLTE) = xdev + ⋅ sin( θdev)
2

16
 IPC-9850
Official Proposal
May 2001
If the component is placed so that the long sides are along the board’s Y-axis,
max( S x , S y )
Maximum Lead Tip Error (MLTE) = ydev + ⋅ sin( θdev)
2

Ball-to-Land (BTL) refers to the percentage of the BGA ball or column of an area array components that is placed
on its (round) land. BTL quantifies the error for the “ball” that experiences the most offset (of all the balls) due to
the joint x, y, and θ errors.

 A 
BTL= 1 −  • 100 , where
 πR12 
R2  c − ∆r  2  R2  −1  c − ∆r   R 22 c  c 
A = 2 1 Sin −1 (1) − 
R2
 R 1 − (c − ∆ r) +  1 sin   − 2 sin −1 (1) −  R 22 − c 2 + 2 sin −1  
2

 2  2   2   R 1   2 2 2  R 2 

where:
Sx Sx Sy
∆X = |Xdev| + | -( · cos (θdev) - · sin (θdev)) |
2 2 2

Sy Sy Sx
∆Y = |Ydev| + | -( · cos (θdev) + · sin (θdev)) |
2 2 2

∆r = ∆X 2 + ∆Y 2

R 2L + ∆r 2 − R 2B
c=
2∆r
RL = radius of the land
RB= radius of the ball
R1= min(Radius of Ball, Radius of Pad)
R2= max(Radius of Ball, Radius of Pad)
Xdev = Offset of component along the X-axis from the target placement
Ydev = Offset of component along the Y-axis from the target placement
θdev = Rotational offset of component from the target placement
∆X = Maximum offset along the X-axis of the ball from its target placement
∆Y = Maximum offset along the Y-axis of the ball from the target placement
∆r = Radial offset of ball center to pad center. Synonymous with MBE in this document.

Computational note: the values of Sin −1 (x) need to be expressed in radians (rather than degrees).
Example of Computing Cpk limits for Termination-to-Land Coverage: After the x,y, and theta errors have been
determined by the CMM (for a four board run using one particular component type), and the individual components’
Termination-to-Land computations have been made as well (using the equations above), the next step is to calculate
the mean and standard deviation of this group of Termination-to-Land values. The Cpk equation should then be
applied, using first the 50% specification limit and then the 75% limit. These 2 Cpk values then go in the
Performance Reporting Form. Suppose the mean of the 144 Termination-to-Land values in a QFP100 run is 85%
85 − 50
and the standard deviation of these values is 5%. Then Cpk versus 50% requirement is = 2.33 .
3• 5

17
 IPC-9850
Official Proposal
May 2001
4 Attribute Defect Rate and Reliability Performance Metric-- Form IPC-9850-F2

Complete evaluation of an SMT placement system must not only consider placement capability performance
covered by Form IPC-9850-F1, but also include information about the yield and reliability of the system. Like the
general performance use of Form IPC-9850-F1, Form IPC-9850-F2 shall be used to present the expected attribute
defect rate and reliability of a specific machine model. However, unlike Form IPC-9850-F1 there shall be no
requirement for a validation for a specific individual machine and there shall be no guarantee for the warranty
period associated with IPC-9850-F2.

The selected reliability metric parameters for this standard fall into three categories -- reliability, availability, and
maintainability. With the exception of Attribute Defect Rate, parameters are based on the SEMI E10-0699E
standard.

The information presented in this form shall be gathered and reported according to this standard, and shall reflect
the performance of the same machine model as in the Placement Performance Metric. The following information is
reported on this form:

1. Manufacturer Name, Machine Model

2. Attribute Defect Rate (ppm)
3. Mean Placements Between Assists (MPBA)
4. Mean Time To Recover from assists (MTTRa )
5. Mean Placements Between Failures (MPBF)
6. Mean Time To Repair failures (MTTRf)
7. Equipment Dependent Uptime
8. Amount of Preventative Maintenance per 6000 Hours
9. Mispick Rate (ppm)
10. Data Collected From:
a. Number of Factories
b. Number of Machines
c. Total Number of Placements

4.1 Attribute Defect Rate

Although production yield depends on the solder paste printing quality of the PWBs, sticky media is used for this
procedure instead solder paste. It is recognized that the attribute defect level from placement operations on sticky
tape are likely to be lower than defect levels from placement on solder paste or conductive adhesives in production
runs. However use of solder paste introduces many additional variables such as paste viscosity, tackiness and height.
The use of sticky tape provides a method to compare machines with a minimum number of variables. In addition, it
eliminates the need for vendors to have costly screen-printing equipment and expertise in utilizing stencil printing,
board cleaning processes, and controlled disposal of residual solder paste.

Since products assembled by surface mount equipment vary greatly by the type of component mix and size of the
PWB, a special kit is provided. This kit is designed to be representative of the product of a ‘typical user’. This
‘typical user’ kit is made of 4400 components and shall be mounted onto the PWB with sticky media. Table 4-1
provides a list of the 4400 component types and quantities. Because some equipment is not designed to place all
SMT parts, it may not capable of placing this specific set of components. If it is necessary to change the mix of
components, the test shall be run using a similar mix of component types that are within the machine’s capability.
The selected mix of components shall be clearly stated on form IPC-9850-F2. The components shall be presented to
the machine using standard feeders or trays manufactured by a vendor-approved supplier.

18
 IPC-9850
Official Proposal
May 2001
Table 4-1 Component Types and Quantities for ‘Typical kit’
Quantity Component
Type
440 SOT-23
440 SOIC-8
880 1608C
880 1608R
880 1005C
880 1005R

The components are to be placed onto an adhesive laminated PWB. The land pads for the PWB shall meet IPC-SM-
782. (Specific board size and layout are not defined by this standard.) After the parts have been placed on the sticky
PWB, a manual visual inspection is made to determine how many placement defects have occurred. While manual
inspection is less than ideal due to its subjectivity, the investment in Automated Optical Inspection (AOI) equipment
for such a limited application cannot be justified.

In this standard, attribute defect rate information differs significantly from the performance metric defined by the
repeatability and accuracy parameters. Whereas the repeatability and accuracy performance parameters are defined
through continuous data -i.e. numerically how much is the component placement error-, the attribute defect rate
parameter is only a function of discrete defects i.e. components which are known to be placed such that a proper
solder joint will not form. Generally these discrete defects occur at such small rates that they are not measured in
percentages of the total number of placements but in defective parts per million (PPM) placements. These discrete
defects are referred to as Attribute Defects, and the anticipated frequency of their appearance is referred to as the
Attribute Defect Rate.

Attribute defects are defined as components placed upside down, tombstone, on side, missing, extra part, damaged
lead(s), damaged part, completely off land, and wrong polarity. An exception is that upside down components are
not to be counted as defects when using bulk feeders. Partially off-land and askew type placement defects are not
included, since the performance level for these types of defects is accounted for by the repeatability and accuracy
parameters. For this parameter evaluation, each component equals one opportunity for a defect and can have a
maximum of only one defect, regardless of the associated anomalies –such as the number of leads, etc. Things other
than the placement machine can cause attribute defects, but they shall be attributed to the machine in the absence of
any other clear cause, such as poor board fiducials, etc.

Estimating the average number of defects that will occur in each million placements is no small task. Estimating
defect rates less than 50 PPM requires very large sample sizes. Different sampling plans exist, each with it's own set
of characteristics and sample sizes.) This standard requires the placement of 88,000 parts. Due to the time and cost
associated with a machine supplier placing this many actual components with each machine manufactured, the
88,000 placements may be spread out over 20 consecutive machines in builds of 4,400 components. As a result, the
confidence level applies only to that group of machines rather than any specific member of that group of machines.
The data may be used as an estimate of the performance of the model type. When a new model machine is launched,
where only a single machine of its kind is available for testing, this single machine shall be utilized for the 88,000
placements.

Calculation Method The PPM level is the total number of observed attribute defects during 88,000 placements of
various component types, divided by 88,000, multiplied by 1,000,000. A minimum total of 88,000 components is to
be placed over 20 tests, where each test run places a component mix of at least 4400 components. The number of
defects is then calculated as a moving average of the last 20 tests.

i + 20
Attribute Defect Rate = (1,000,000/88,000) * ∑ AttributeDefects i
test =i

The following example refers to Table 4-3 where two defects over the 88K parts yields Attribute Defect Rate of 23
ppm. For the first and second machines, the attribute defect rate is obtained by dividing the three attribute defects –
found during run series 2, 7, and 17-- divided by 88K parts, for a 34 PPM. Once the third machine is shipped, the
attribute defect level is reduced to 23 PPM, since the single attribute defect of run series 2 is replaced by the zero
attribute defect of machine 3, run series 22.
19
 IPC-9850
Official Proposal
May 2001

Table 4-3 Calculation Example

Machine 1 (Note 1) 2 3
Number
Run 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Series 1
Run 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Series 2
Run 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Series 3
Number 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
of
Attribute
Defects
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 χ2 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 χ2 ) 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

20
 IPC-9850
Official Proposal
May 2001
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

300,000 + 500,000
= 160,000 cycles.
2 +3
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:

21
 IPC-9850
Official Proposal
May 2001

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.

22
 IPC-9850
Official Proposal
May 2001

4.2.1 Mean Placements Between Assists (MPBA) –This is the average number of component placements between
assists.

Calculation Method The total component placements divided by the number of assists during those component
placements.
Total Equipment Placements
MPBA =
# of Assists
4.2.2 Mean Time To Recover from Assists (MTTRa ) –This is the average time to correct an assist and return the
equipment to a condition where it can perform its intended function.

Calculation Method The sum of all assist recovery time (elapsed time, not necessarily total man hours) incurred
during a specified period (including equipment and process test time, but not including maintenance delay), divided
by the number of assists during that period.

Total Recovery Time

MTTRa =
# of Assists
4.2.3 Mean Placements Between Failures (MPBF) –This is the average number of component placements
between failures.

Calculation Method The total component placements divided by the total number of failures during those component
placements.

For statistical credibility, the MPBF claim must be no higher than one third of the number of placements completed.
Total Equipment Placements
MPBF =
# of Failures
4.2.4 Mean Time To Repair Failures (MTTRf ) –This is the average time to correct a failure and return the
equipment to a condition where it can perform its intended function

Calculation Method The sum of all repair time (elapsed time, not necessarily total man hours) incurred during a
specified period (including equipment and process test time, but not including maintenance delay), divided by the
number of failures during that period.

Total Repair Time

MTTR f =
# of Failures
4.2.5 Total Time- This is all time (at the rate of 24 hours/day, 7 days/week) during the period being measured.

4.2.6 Non-Scheduled Time -This is a period when the equipment is not scheduled to be utilized in production.

4.2.7 Operations Time -This is total time minus non-scheduled time.

4.2.8 Equipment Dependent Uptime –This is the percentage of operations time when a machine is in a condition to
perform its intended function. It includes productive, standby, and engineering time, and does not include any
portion of non-scheduled time.

Calculation Method The Equipment Dependent Uptime multiplied by 100 and divided by operations time.

23
 IPC-9850
Official Proposal
May 2001
4.2.9 Preventative Maintenance (PM) Time
This is the amount of time required for preventative maintenance as specified by the machine supplier. This metric
will be based on the maintenance routines that are required each year in order to keep the machine in a warranted
condition and how long those routines take for a 6000 hour period.

Calculation Method Multiply the amount of time required for each PM procedure by the number of times it is
required in 6000 hours and add these up for a grand total.

4.2.10 Mispick Rate (in PPM) This is the frequency with which a component was not picked or picked incorrectly,
even if recovery handled the problem. Mispicks include both pickup failures and vision failures, but excludes
components exhaust.

Pickup defects will be handled in a manner similar to placement defects. Pickup defects are categorized as vacuum
failures or vision failures. While many suppliers have various auto-recovery routines that prevent a pickup defect
from resulting in a missing component placement defect, this type of defect has still been deemed worthy of tracking
and reporting due to its impact on the assembly process. Pickup defects cause a reduction in net CPH and the
unnecessary rejection and scrapping of good parts.

The number of mispicks that occur during the 4400 part placements made to determine placement defect rate will be
recorded on Form IPC-9850-F2. The number of pickup defects made during this process for the past 20 machines
will be reported, from which a point estimate for pickup defects in PPM can be made.

Measurement Procedure Clear the Mispick counters in the machine. Empty the reject bin. Record the relative
humidity. Run the attribute defect test (place 4,400 components). Record the vacuum failure, and vision failure
counter values. Examine the components found in the reject bin and count the defective components.

Calculation Method The total pickup attempts minus total number of placements minus the number of defective
components divided by total pickup attempts minus defective components, multiplied by one million to obtain ppm.

Mispick Rate =1,000,000 * (Total Pickup Attempts)- (Total Number of Placements) –(Number of Defective Components)
(Total Pickup Attempts) –(Number of Defective Components)

4.3 Reporting Sites This is the number of machine sites included in the reliability metric calculations.

4.3.1 Number of Machines This is the number of individual machines included in the reliability metric calculations.

4.3.2 Total Number of Placements This is the number of individual machine placements included in the reliability
metric calculations.

4.3.3 Total Pickup Attempts This is the number of times the nozzles attempted to pick components.

5 Measurement Capability Verification –Form IPC-9850-F3

Form IPC-9850-F3 functions as a tool for demonstrating the measurement capability of the measurement equipment
used to generate the parameter values for the Placement Performance Metric (form IPC-9850-F1). The measurement
capability information presented on the form shall be gathered and reported according to this standard, and shall
reflect the performance of the measurement equipment used in the machine performance analysis. The reported
information should be acquired no more than 90 days prior to the machine performance study.

The parameters of this metric shall evaluate the measurement system’s ability to accurately measure the test
vehicles produced during the machine performance evaluation analysis in a repeatable and reproducible manner.
This evaluation shall be made for each component type and yields results about the measurement capabilities of the
CMM along the same axes as reported by the placement performance metric. exceed The specification limits
reported in the performance reporting form shall never be smaller than the larger of the GR&R and accuracy CMM
specification limits per component type. Suppliers that add additional component types to the Placement
Performance Metric shall also report their measurement capabilities for the same component types.

24
 IPC-9850
Official Proposal
May 2001
Suppliers may include procedures to measure placements with the placement machine itself. This is acceptable for a
particular placement machine as long as that particular machine (not just a sample machine from that particular
model type) passes the gauge requirements laid out in this section which determine whether or not its measurement
capability is acceptable.

5.1 Gauge Repeatability and Reproducibility Capability

This verification procedure determines the CMM's ability to repeat and reproduce the component location and
rotation mounted onto the PVP. This evaluation shall be made for each component type.

While multiple GR&R methodologies are available, this standard uses the traditional average-range method and
recommends the use of the spreadsheet discussed in Appendix H. The results are reported in terms of the upper and
lower specification limits that yield a precision-to-tolerance ratio of better than 25%.

Although a precision-to-tolerance (P/T) ratio of less than 20% is usually considered acceptable, 25% is utilized in
this standard to accommodate existing CMM machines. However this reduced P/T ratio specification if somewhat
offset by the fact that the analysis is performed using a more stringent requirement of six sigma (+/-3 sigma or
99.7%) versus the customary value of 5.15 sigma (+/- 2.575 sigma or 99%) recommended by the Automotive
Industry Action Group (AIAG).

Measurement Procedure Obtain a single glass verification panel with a specific part type populated according to the
procedures described in the machine placement performance metric, Section 3. Measure the populated PVP three
consecutive times. The PVP is to be completely removed and replaced from the CMM for each run. Only 36
components need to be measured for each run. Appendix F outlines the component locations for the specific
component panel layout. After the three measurements are obtained the CMM must be shutdown and restarted.

The same PVP plate shall be subjected to two additional measurement cycles (two more sets of 3 measurements)
using an identical procedure. At least two operators should be utilized to obtain the three sets of data. The entire
GR&R evaluation shall be performed in a time period no longer than it takes to perform the measurement of the
consecutive four PVPs for the placement performance evaluation.

For the SOIC-16 and 1608C components, the center of the component shall be measured using the lead outline and
component outline, respectively. The same measurement procedure applies to the PVP populated with slugs.
However, since the CMM utilizes the fiducial markings on all glass slugs, regardless of the specific component
image, it is only necessary to evaluate the CMM’s ability to measure the slugs using the fiducial markings and not
individually for the QFP and BGA component types. Appendix D provides suggested methods for measurement
methods.

5.2 Accuracy Capability

This verification procedure determines the Optical CMM's ability to accurately measure the placement of
components on the glass verification panel. This is accomplished by evaluating the CMM’s ability to measure the
NIST certified glass verification panel specified in Figure 5-1. This accuracy verification panel (AVP) was designed
to evaluate the CMM’s ability to measure three component types required for compliance with Section 3.

To evaluate the measurement capabilities of the 1608C and SOIC-16 components, their images have been etched
onto the AVP. Since slugs are utilized for the QFP and BGA components, only slug fiducials marking are etched
onto the AVP. Some of the etched images are translated and rotated from their specified CAD location, in order to
test the CMM’s ability to measure components with measurement error. Some of the 1608C and SOIC-16
component images on the AVP are translated in a range of + 50 micrometers, and rotated in a range of +3º range.
Some of the four fiducial image sets will be translated in a range of + 25 micrometers, and rotated in a range of +1º
range. The ‘CMM Accuracy Eval.xls’ file provides an instruction guide and a sample CMM accuracy evaluation
spreadsheet.

This methodology is implemented using a measurement program on the CMM, which searches for the components
on the glass panel at their CAD locations. These CAD locations are the centers of the components as if no
translation or rotations were applied to the images on the glass panel. The CAD locations of the images can be found
in Appendix F. The CMM measurement results are compared to the locations and rotations of the component images
25
 IPC-9850
Official Proposal
May 2001
provided by the NIST certification for the unique glass panel utilized. The closer the CMM measurement values are
to the certified locations the better the measurement capability of the CMM.

The results are reported in terms of the upper and lower specification limits that yield process capability indexes
equal to 2.0. See Appendix B for discussion.

Measurement Procedure Measure the center location and orientation of the component images on the certified
accuracy verification panel a single time. Measure the center location and orientation of each group of four fiducials
that represent the slug component types. Use the same image analysis procedures that are utilized for the Placement
Performance Metric evaluation. Appendix D provides a guide for the proper methods of measuring the center
locations of the components.

Calculation Method Subtract the CMM measured image locations from the NIST certified image locations along the
X, Y and θ axes, for each of the 1608C, SOIC-16, and slug features. This provides a distribution of the measurement
errors of the CMM. Use these distributions to calculate the average and standard deviation for each axis. Calculate
the process specification limits for each axis using the following equations, when Cpk = 2.0. See Appendix B for a
discussion of these equations.

SL=3*SD Error *Cpk+|Avg Error |

1 18
where Avg Error, = ∑ (MeasuredLo cationi − CertifiedLocationi )
18 i =1

1 18
18 − 1 ∑
and SD Error, ( MeasuredLo cationi − CertifiedLocationi ) 2 ,
i =1

Figure 5-1 Accuracy Verification Panel

26
 IPC-9850
Official Proposal
May 2001
6 Test Vehicles

Users of this standard shall utilize standardized materials and methods to ensure that users and suppliers of SMT
placement equipment are able to observe the same placement performance results within the tolerance of their
measurement gauge’s repeatability, reproducibility and accuracy.

Table 6-1 is the material list for the execution of the placement performance evaluation. Check the IPC website for
suggested suppliers of evaluation material.

Table 6-1 Tool Kit Materials List For Performance Evaluation

Quantity Material
4 Placement Verification Panel (see Figure 6-2)
1 CMM Measurement Verification Panel
150 QFP-100 slugs with no background or with white background
130 QFP-208 slugs with no background or with white background
150 BGA-228 slugs with no background or with white background

1600+ 1608C Components

240+ SOIC 16 Components
Optional
4 PVPl Carrier
- Sticky tape

6.1 Placement Verification Panel Specifications

Glass verification panels with anti-reflective chrome metallization are used because they are dimensionally stable
and can be made very accurately. The glass board fiducials create stable high-contrast images on both placement
machines and CMMs. The transparent glass allows the use of back lighting on the CMM to increase the gauge
repeatability of the measurements.

The IPC-9850-P1 placement verification panel (Figure 6-1) was designed to be small enough to measure on
relatively affordable vision CMMs and large enough to represent a typical sized PWB.

27
 IPC-9850
Official Proposal
May 2001
Figure 6-1 Placement Verification Panel

6.2 Placement Verification Panel Carrier Specifications

The standard does not specify or require the use of a carrier for the PVP. A recommended carrier design is shown in
Appendix I.

6.3 Glass Slug Specifications

Slugs must meet the dimensional specifications of Figures 6-2, 6-3, 6-4, 6-5, 6-6, and 6-7.

Implication of the glass slug weight on placement acceleration and speed was considered. While there may be cases
where the machine performance will need to be adjusted to accommodate the slug, this type of slugs are currently
being used by most vendors without performance issues.

The patterns on the slugs will be made of black chrome. Since some suppliers use front lighting and some use back
lighting the slugs may be either white background for leads on chrome background or chrome leads with no
background.

1 2 3 4 5 6 7 8
RevNo Signature

A A

B B
10.00 10.00

0.500 0.50mm R (4 PLCS) 8.500 8.500

5.000 5.000
2.000
1.5mm R-WIDTH of LINES 0.050mm 10.00
0.200
8.000
7.000
C 1.00mm X 1.00mm CROSS
6.000
2.000 C
10.00
8.500

1.5mm LONG X .050mm WIDE LINE 5.000

(4 PLCS) ON CENTER LINE
0.000

5.000

D 6.000
8.500
10.00
D
EQUILATERAL TRIANGLE 1mm SIDES 7.000
8.000

10.00

E FRONT SIDE NOTES: E

BACK SIDE
HATCHED AREA ANTI-REFLECTIVE CHROME PLATED
WHITE EPOXY IN HATCHED AREA
CHROME SIDE UP
25 LEADS EACH SIDE 0.2mm THICK 1mm LONG

F Material / Finish F
’ Draft

Sheet
Unless otherwise stated

1 2 3 6 7 8

Figure 6-2 QFP 100 Slug for White Background for Leads on Chrome Background (Negative)
Drawing File Name: ipc_100QFP_glass_NEGATIVErevb.eps

28
 IPC-9850
Official Proposal
May 2001
1 2 3 4 5 6 7 8
RevNo Signature

A A

20.00

B 0.500
12.50 B
5.000 5.000
1.5mm R-WIDTH of LINES 0.050mm 10.00
0.200
8.000
7.000
1.00mm X 1.00mm CROSS
6.000

C 5.000
C
1.5mm LONG X .050mm WIDE LINE
(4 PLCS) ON CENTER LINE
0.000
12.50 20.00

5.000

6.000
EQUILATERAL TRIANGLE 1mm SIDES 7.000
D 8.000 D
10.00

0.25mm R (4 PLCS)

FRONT SIDE NOTES:

HATCHED AREA ANTI-REFLECTIVE CHROME PLATED
E E
CHROME SIDE UP
25 LEADS EACH SIDE 0.2mm THICK 1mm LONG

F Material / Finish F
’ Draft

Sheet
Unless otherwise stated

1 2 3 6 7 8

Figure 6-3 QFP 100 Slug for Chrome Leads With No Background (Positive)
Drawing File Name: ipc_100QFP_glass_POSITIVErevb.eps

1 2 3 4 5 6 7 8
RevNo Signature

A 34.00 A
17.00

15.80 15.80
0.500

0.200

B B

17.00 17.00

C C
0.000 33.00

1.00mm X 1.00mm CROSS

D 15.80 D

EQUILATERAL TRIANGLE
1mm SIDES CHROME FILLED

E E
1.5mm LONG X .050mm WIDE LINE BACK SIDE
HATCHED AREA ANTI-REFLECTIVE CHROME PLATED (4 PLCS) ON CENTER LINE WHITE EPOXY IN HATCHED AREA
CHROME SIDE UP
52 LEADS EACH SIDE 0.2mm THICK 1.3mm LONG

F Material / Finish F
Draft
’
Sheet
Unless otherwise stated

1 2 3 6 7 8

Table 6-4 QFP 208 Slug for White Background for Leads on Chrome Background (Negative)
Drawing File Name: ipc_QFP208_glass_NEGATIVE.eps

29
 IPC-9850
Official Proposal
May 2001

1 2 3 4 5 6 7 8
RevNo Signature

A 34.00 A

17.00
0.500
12.00 12.00
0.200

B B

17.00

12.00

C C
0.000
34.00

1.00mm X 1.00mm CROSS

12.00

D D

EQUILATERAL TRIANGLE
1mm SIDES CHROME FILLED

E E
1.5mm LONG X .050mm WIDE LINE BACK SIDE
HATCHED AREA ANTI-REFLECTIVE CHROME PLATED (4 PLCS) ON CENTER LINE WHITE EPOXY IN HATCHED AREA
CHROME SIDE UP
52 LEADS EACH SIDE 0.2mm THICK 1.3mm LONG

F Material / Finish F
Draft
’
Sheet
Unless otherwise stated

1 2 3 6 7 8

Table 6-5 QFP 208 Slug for Chrome Leads With No Background (Positive)
Drawing File Name: ipc_QFP208_glass_POSITIVE.eps

Figure 6-6 BGA228 Slug for White Background for Leads on Chrome Background

30
 IPC-9850
Official Proposal
May 2001

Figure 6-7 BGA228 Slug for Chrome Leads With No Background

31
 IPC-9850
Official Proposal
May 2001
6.4 Sticky Media Application

To maintain the components in location on the glass panel, a double-sided sticky media adhesive is applied to the
glass verification panel area between the fiducial markings. There are two important requirements for the sticky
media. The first is the consistency by which the sticky material is applied to the glass verification panel surface. It is
important that the adhesive be applied evenly across the entire surface of the glass while providing sufficient clarity
for the CMM back light. The second requirement is the stability of the sticky material once it is applied to the glass
media. The adhesive must maintain the components in position during the placement cycle, where the glass
verification panel may be exposed to rapid accelerations. It must also keep the components in position throughout
the evaluation procedure as time and environmental conditions change from the placement machine location to the
CMM location.

The application of the sticky media to the glass verification panel requires significant care and experience. The
application procedure of the sticky media is not specified (it is not possible to improve test results by improper
application--only worse results). It is left up to the user to determine the proper application of the adhesive.

Details and suggestions are provided in Appendix C.

6.5 Placement Program

6.5.1 1608C Component

The placement layout for 1608C components was designed to exercise the placement machine's capability at placing
components at various distances from each other, so all machine axis are exercised, see Figure 6-8. The components
are placed at the four most common placement angles -- 0, 90, 180 270 degrees -- and the placement order and the
number of supporting feeders is to be determined by each individual machine vendor with the goal of optimizing
placement rate and accuracy for a specific machine type. Appendix E presents the positions of the components with
respect of the fiducials in numeric format.

32
 IPC-9850
Official Proposal
May 2001

Figure 6-8 Component Location Layout for 1608C Components

33
 IPC-9850
Official Proposal
May 2001
6.5.2 SOIC-16 Component

Figure 6-9 presents the placement layout for the SOIC-16 component on the PVP. Appendix E presents the positions
of the components with respect of the fiducials in numeric format.

Figure 6-9 Component Location Layout for SOIC 16 Components

File Name: SOIC-16 Placement Layout.pcx

34
 IPC-9850
Official Proposal
May 2001
6.5.3 QFP-100 and BGA-228 Slugs

The panel is populated with glass slugs for the corresponding component type. Fine pitch placement programs for
the QFP-100, and BGA-228 have the position and angular rotation pattern shown in Figure 6-10. Appendix E
presents the positions of the components with respect of the fiducials in numeric format.

Figure 6-10 Component Location Layout for QFP-100 and BGA 228 Slug Components

6.5.4 QFP-208 Slug

Figure 6-11 presents the placement layout for the QFP-208 slug component on the PVP. Appendix E presents the
positions of the components with respect of the fiducials in numeric format.

[new drawing pending]

Figure 6-11 Component Location Layout for QFP-208 Slug Components

35
 IPC-9850
Official Proposal
May 2001
7 Forms and Diagrams

The forms and diagrams on the following pages may be freely copied and reproduced. Contact IPC for electronic
copies of the forms. [web download after this is published]

Placement Performance Form

IPC-9850-F1

Manufacturer Name When form is used for ship

verification
Machine Model Serial Number
Build Date and Time

Component Type QFP-100 QFP-255 BGA-228 1608C SOIC-16

Test Conditions During Build
Number of Heads/Spindles
Type of Heads/Spindles
Type of Camera
Number of Feeders/Trays
Type of Nozzles
Number of Nozzles
Number of Panels Built 4 4 4 4 4
Number of Parts Per Panel 36 30 36 400 (120 for 80
CMM meas.
validation)
Test Results
Build Time, seconds
Transfer Time, seconds
Tact Time, seconds
Net Throughput, CPH
Repeatability (One Standard Deviation)
X, micrometers
Y, micrometers
Axis
θ, degrees

Accuracy
+X, micrometers
Spec. Limits 1 for +Y, micrometers
Cpk=1.33 +θ, degrees

+X, micrometers
Spec. Limits 1 for
+Y, micrometers
Cpk=2.0
+θ, degrees
Cpk for 50% Coverage N/A2
Termination-to-Land 75% Coverage N/A2
Note 1 : To be valid, reported Spec. Limits must not exceed the Spec. Limits of the CMM capability reported in
form IPC-9850-F3.
Note 2: Termination-to-Land evaluation for 1608 components is not required. See Section 3.4.3.2.2 for details.

Form 9850-F1 Performance, issue date

36
 IPC-9850
Official Proposal
May 2001
Reliability Performance Form
IPC-9850-F2
Attribute defect rate and Reliability numbers are not to be construed as guarantees in any way. They simply indicate
actual levels of reliability for this equipment achieved at customer locations.

Manufacturer Name
Machine Model

Attribute Defect Rate (ppm)

Specified Components Alternate Components Used
Component Quantity Component Quantity
Type Type
SOT-23 440
SOIC-8 440
1608C 880
1608R 880
1005C 880
1005R 880

Mean Placements Between Assists (MPBA)

Mean Time To Recover from assists (MTTRa), minutes
Mean Placements Between Failures (MPBF)
Mean Time To Repair failures (MTTRf), hours
Equipment Dependent Uptime, %

Amount of Preventative Maintenance per 6000 Hours, hours

Mispick Rate (ppm)

Data Collected From:

Number of Factories
Number of Machines
Total Number of Placements

Form 9850-F2 Reliability, issue date

37
 IPC-9850
Official Proposal
May 2001
CMM Capability to Evaluate Metric Parameters
IPC-9850-F3

Date of GR&R Study Machine Repeatability & Reproducibility,

Accuracy, Spec. Limits 1
Axis GR&R Limits 1
__________ Glass Glass 1608
1608 Cap SOIC-16 SOIC-16
Slugs2 Slugs2 Cap
X,
micrometers
Placement Verification Y,
Glass Panel micrometers
θ, deg

X,
micrometers
Measurement Capability Y,
Glass Panel micrometers

θ, deg

Note 1: To be valid, reported Spec. Limits in form IPC-9850-F1 must not exceed worse case Spec. Limits of the CMM
capability per component type.
Note 2: Since the CMM utilizes the fiducial markings on all glass slugs, regardless of the specific component image, it is only
necessary to evaluate the CMM’s ability to measure the slugs using the fiducial markings.

Form 9850-F3 CMM Capability, issue date

38
 IPC-9850
Official Proposal
May 2001
Appendix A: Capability Indexes.

Cp can be computed when a process specification limit is stated:

UpperSpecLimit − Lower SpecLimit

Cp =
6⋅s
For a process with target of zero, the upper and lower specs have the same magnitude, the equation can be rewritten
as:

UpperSpecLimit
Cp =
3⋅s
Cpk quantifies how many sets of 3 standard deviations lie (fit) between the process mean and the closest
specification limit. Cpk is intended for use with distributions that are normal or near normal, and is valuable because
it can be used to quantify the number of placements per million that will occur outside of the specification limits due
to the lack of accuracy or repeatability of the placement machine.

min ( x − Lower Spec Limit , Upper Spec Limit − x )= Distance between the mean and the nearest spec
Cpk =
3⋅s 3⋅s
where s = the standard deviation of the sample.

LSL USL

Mean

Cpk Numerator

Cpk Denominator

-4 - - - -3 - - - -2 - - - -1 - - - - 0. 0. 0. 1 1. 1. 1. 2 2. 2. 2. 3 3. 3. 3. 4
3. 3. 3. 2. 2. 2. 1. 1. 1. 0. 0. 0. 1. 25 5 75 25 5 75 25 5 75 25 5 75
75 25 75 25 75 25 75 25 42
1E
-

Figure A-1, title

For a centered process the Specification Limits (SL) are obtained by using the following equation, where SD is the
standard deviation and Avg is the average of the data set.

39
 IPC-9850
Official Proposal
May 2001
Appendix B: Specification Limits for Cpk Values

Definitions:

Process Spread = Standard Deviation of Data Set = SD

Process Mean Average of Data Set = Avg
Process Target = Target
Upper Specification Limit =USL
Lower Specification Limit =LSL

For a centered process Target =0 and thus USL=LSL

Define Specification Limit for a Centered Process = SL = USL=-1*LSL.

Define Specification Width SW = USL-LSL, thus for a Centered Process

SW=SL-(-1*SL)= 2*SL

Define a General Process Capability Index Cp and Cpk as

Cp= SW/(6*SD)

and

Cpk=Cp*(1-K)

Where,

K= (Target-Avg)/(SW/2)

For a centered Process

Cp=2*SL/(6*SD)=SL/(3*SD)

Cpk=(SL/(3*SD))*(1-K) Where K=(0-Avg)/(2*SL/2))=-Avg/SL

Substituting for Cp and K in the Cpk expression yields

Cpk=(SL/(3*SD))*(1--Avg/SL)=SL/(3*SD)+SL*Avg/(3*SD*SL)=SL/(3*SD)+Avg/(3*SD)

Solving for the SL in the Cpk expression yields

SL=3*SD*(Cpk-Avg/(3*SD))=3*SD*Cpk+|Avg|

Example:

X Average=-9.7 micrometers
X Std Dev = 15.9 micrometers

Y Average = 7.6 micrometers

Y Std Dev = 21.7 micrometers

Thus for a Cpk of 2 the SL for the X-axis is

SL=3*15.9*2+|9.7|=105 micrometers

And the SL for the Y-axis is 122 micrometers.

40
 IPC-9850
Official Proposal
May 2001
Appendix C
Guidelines for Adhesive Used for Machine Capability Testing (Selection and Application)
C-1 Background
Based on experience by the participating members of the IPC Assembly Equipment Subcommittee, it was
recommended that a guideline be written for the selection and application of adhesive used for machine capability
testing. It was the Subcommittee's preference to not specify the exact adhesive to use but to establish a guideline that
would give some direction for users to select and apply adhesive while performing machine capability tests.
Improper selection and application of adhesive could result in a machine performance to be misrepresented. In other
words, the data will show a machine to look worse than it actually is.

C-2Purpose
The purpose of this guideline is to provide a user wishing to perform a machine capability test some direction for the
selection and application of adhesive. The adhesive is used to hold the test parts onto the glass plate during machine
placement. The adhesive also holds the parts in place until an accurate measurement can be taken using a verified
CMM (Coordinate Measurement Machine). The adhesive should perform such that the parts do not move more dm a
specified amount in actual conditions in which the glass plate is subjected to. The actual conditions being time,
temperature and movement of the glass plate.

C-3 Materials / Equipment

It is the intent of the Subcommittee not to specify the adhesive to be used. Based on experience of the members of
this Subcommittee, a spray adhesive is not stable enough and can be very inconsistent. This guideline is intended for
use with tape or sheet adhesive. This guideline will provide information to select an adequate adhesive and
instructions on how to best apply it. At some point in the future, the Subcommittee may elect to set up a process or
system to acquire a specific adhesive through IPC if the members see benefit to do so and EPC is in agreement.

The CMM to be used is assumed to be calibrated and in proper working order. The CMM must have been verified to
be acceptable using the GR & R procedure that this Subcommittee has set forth. The procedure specifies the glass
plate, parts, placement patterns and conditions for performing the GR&R study.

The glass plate and parts that are used to test a particular adhesive should be the same glass plate and part types that
will be used to conduct the placement tests needed to report machine capability per the IPC specification.

C-4 Procedure
1. Clean the glass plate and parts (if applicable) using alcohol free glass cleaner and lint free cloth.

2. Tape the area of placement with the double side adhesive. It is recommended that the adhesive be pre-cut to size.
The biggest problem with applying adhesive is stretching the adhesive and applying is under stress. The adhesive
then moves after the parts have been placed. To avoid this, purchase the adhesive pre-cut or dispense and cut the
adhesive avoiding stretch instead of pulling and tearing.

3. For test parts other that chips, strips of adhesive smaller than the test parts should be used versus a full sheet. This
will promote removal of test parts, particularly, glass test parts.

4. While applying tape, ensure minimal air bubbles.

5. After the adhesive is applied, verify it is in the correct position and flat.

6. Carefully remove the backing from the adhesive so as not to lift the adhesive from the glass board.

7. Run the glass plate using the same glass plate and part types that will be used to conduct the placement tests. The
placement tests are those which have been defined by the Subcommittee.

8. Take the glass plate with parts placed to the CMM for measurement immediately. The glass plate should be
transported in a horizontal position. Movement should be minimized as well as conditions such as temperature.

9. Measurements should be taken and output generated for later comparison.

41
 IPC-9850
Official Proposal
May 2001
10. The glass plate should be re-measured at some time period after the first measurement. It is recommended that
the glass plate be re-measured for comparison at worst case intervals at which are expected by the user. The worst
case scenario would be the longest period at which would be expected to elapse between placement and
measurement. As a minimum, two hours should elapse between the first measurement and re-measurement.

II. Other re-measurements can be taken after simulating movement conditions or temperature changes.

12. Some adhesives are capable of being re-used. If this is planned for the application, Ns same series of tests should
be repeated for each re-use of the adhesive to establish the frequency at which the adhesive should be re-applied.
The obvious best case would be to re-apply after each test. As a reminder, poor adhesive will only worsen end
results.

13. Measurements should be compared from the initial measurement to all subsequent measurements to look for
changes or movement of parts after placement.

14. For the adhesive to be acceptable, the maximum movement that can be observed between the initial
measurement and re-measurements should be a maximum of two (2) microns.

C-5 Results / Summary

Poor adhesive, improper application of adhesive, poor measurement re-use of adhesive and many other factors affect
test results for machine capability measurement. By following these guidelines, only better results can be obtained.
Failure to do so may degrade the perceived performance of a placement machine.

42
 IPC-9850
Official Proposal
May 2001
Appendix D
Suggested Methodologies for Measuring Components using an Optical CMM
D-1 Introduction

In the IPC-9850 standard the following components are used to prove the accuracy of placement machines:

• Glass QFP-100, QFP-208

• Glass BGA-228
• Real 1608C
• Real SOIC16

However it is not described how those components must be measured on the CMM machine. The center of the
components can be constructed in different ways. Every method has its own accuracy.

This report investigates the different methods to measure the components and the sensitivity of the estimate of the
center location of the component to measuring and component variances.
Finally a recommendation is given how the components should be measured.

D-2 Slugs (glass components)

The glass QFP-100, QFP-208 and BGA-228 are provided with 4 fiducials, one at each corner of the component.
There are several ways to construct the center and the angle of the component using 2 or all 4 fiducials.

D-2.1 Methods to measure the X- and Y-location of the slug

The different methods to measure the center location of the component are described below.

D-2.1.1 Using 2 fiducials

Measure two diagonal fiducials and take the average x and y

x1 + x2
xcomponent =
2
y + y2
ycomponent = 1
2

D-2.1.2 Using 4 fiducials

Method 1

Measure all fiducials and take the average x and y:

x1 + x2 + x 3 + x 4
x component =
4
y + y2 + y3 + y4
y component = 1
4
Method 2
Take the intersection point of two diagonal lines through the fiducials

D-2.2 Methods to measure the angle of the slug

43
 IPC-9850
Official Proposal
May 2001
The different methods to measure the angle of the component are described below.

D-2.2.1 Using 2 diagonal fiducials

Take the angle of the line through the 2 diagonal fiducials and subtract the nominal angle (in this case the
component is square so the nominal angle is 45 degrees).

D-2.2.2 Using 4 fiducials

Method 1
Take the average of the angle of the two diagonal lines through the fiducials

Method 2
Take the angle of the line through the vertical midpoints minus 90 degrees:

Method 3
Take the angle of the line through the horizontal midpoints:

Method 4
Take the average angle of all “sides” of the component:

Method 5: Take the average of the angles of method 2 and 3

D-2.3 Simulation

By simulating measurement errors on the X- and Y-locations of the fiducials, the sensitivity of the methods above is
determined.

The measurement errors on the fiducials are assumed to be distributed with a standard deviation of 0.0005 mm.
Then 1000 measurements are simulated (using MS Excel). With those measurements, the x, y, and theta locations of

44
 IPC-9850
Official Proposal
May 2001
the component are calculated with the different methods above. The mean and standard deviation of the center
location using the different methods is given in the tables below.

Using 2 fiducials

Table D-1 Mean and stdev using 2 fiducials (nominal means are: x=0, y=0, theta=0)
x y theta
(mm) (mm) (mm)
Mean 0.000 0.000 0.000
Stdev 62.10-5 61.10-5 243.10-5

Using 4 fiducials

Table D-2 Mean and stdev per method using 4 fiducials (nominal means are: x=0, y=0, theta=0)
Method 1 Method 2 Method 1 Method 2 Method 3 Method 4 Method 5
x y x y theta theta theta theta theta
(mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm)
Mean 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Stdev 42.10-5 44.10-5 62.10-5 60.10-5 174.10-5 228.10-5 230.10-5 167.10-5 166.10-5

D-2.4 Conclusion

4 C 3

D B

1 A 2

From the results above can be concluded that using 4 fiducials to estimate the center location of the component is
more reliable. From table 2 is concluded that the best methods to calculate x,y and theta are method 1 for x,y and
method 5 for theta.

So the glass QFP and BGA are recommended to be measured as follows:

Measure the X- and Y-location of the 4 (black) fiducials. Construct with these results 4 (white) midpoints at each
side of the component. The theta of the component is constructed by taking the average angle of the 2 lines. Taking
the average x and y of the measured fiducials determine the X- and Y-location of the component. Beware that this is
NOT the same as taking the intersection point of the 2 lines through the midpoints.

45
 IPC-9850
Official Proposal
May 2001
In formulas:
x1 + x 2 + x3 + x 4
x component =
4
y + y 2 + y3 + y 4
y component = 1
4
thetaline1 + thetaline2 − 90 tan
−1
( y 2 + y 3 − y1 − y 4
x2 + x 3 − x1 − x 4
) + tan (
−1 x1 + x2 − x3 − x4
y1 + y2 − y3 − y 4
)
thetacomponent = =
2 2

D-3 1608C component

The methods to measure the 1608C component are almost the same as is described for the slugs.
First all four sides of the component are measured (black lines). With these lines, the intersection points are
constructed (black points). Now the same methods as described for the 4 fiducials of the slugs can be applied.
Therefore also for the 1608C component it is recommended to use method 1 and 5 for the component- x, y and theta
respectively.

D-4 SOIC16 component

Method 1
The leads of the SOIC-16 component are measured as follows. The right, left and bottom sides of the lead are
measured. Two intersection points are constructed from these measurements. The average x and y of intersection
points are the coordinates of the lead.
T

3 7 11 15 16 12 8 4

1 5 9 13 14B 10 6 2

This can be done for all leads, but also for only a few leads. In the section below, measurements are simulated to
determine the influence of the number of measured leads on the center-estimate of the component. Notice that
always the most outer leads are measured to get a correct and reliable calculation of the x,y and angle (e.g. Using 8
leads, at both sides the two outer leads are measured).
The position of the center of the component is the average x and y of all measured leads.

The theta of the component is

1, Average the location for both XT and YT on the top side leads of the component
2, Average the location for both XB and YB on the bottom side leads of the component
3. Use the points (XT ,YT) and (XB,YB) to determine a line
4, Then measure the angle of this line minus 90 degrees to determine the angle of the component.
46
 IPC-9850
Official Proposal
May 2001
Method 2

3 7 11 15 16 12 8 4

1 5 9 13 14 10 6 2

The leads of the SOIC16 component are measured as follows. The right, left and bottom sides of the lead are
measured. Two intersection points are constructed from these measurements. The average x and y of those
intersection points are the coordinates of the lead.

This can be done for all leads, but also for only a few leads. In the section below, measurements are simulated to
determine the influence of the number of measured leads on the center-estimate location of the component. Notice
that always the most outer leads are measured to get a correct and reliable calculation of the x,y and theta location
(e.g. Using 8 leads, at both sides the two outer leads are measured).

The position of the center location of the component is the average X- and Y-location of all measured leads. The
theta of the component is determined by fitting a line through the leads at the north side and one through the leads at
the south side. The average angle of those two lines is the estimated theta location of the component.

In formulas:
n

∑x i

x comp = i =1

n
n

∑y i

y comp = i =1

n
     
 n
∑x y −  ∑ xi ∑ y i    n
∑x y −  ∑ xi ∑ yi  
−1 
i i
 south leads south leads   + tan −1 
i i
 north leads northleads  
2 2
south leads north leads
tan    
   


n
2 ∑ i  south∑leadsi 
x 2
−  x  



n
2 ∑ i  north∑leadsi 
x 2
−  x  

   
=
south leads north leads
thetacomp
2

47
 IPC-9850
Official Proposal
May 2001
D-4.1 Simulation

The variation on the intersection points of the leads comes from two error causes:

1. The measurement error, distribution assumed with stdev of 0.0005

2. The component variation, distribution assumed with stdev of 0.0017 in x and 0.0127 in y

With these estimates of the intersection points the lead location is calculated. And with those leads, the center- and
angle location of the component is estimated, using 4,8,12 or 16 leads.
In this way ten thousand measurements are simulated. In the table below are given the estimates of the center- and
angle location of the component using 4,8,12 or 16 leads.

Nr. of leads Mean Stdev

used x y theta x y theta
4 0.000 0.000 0.000 0.0014 0.0081 1.0325
8 0.000 0.000 0.005 0.0010 0.0057 0.8420
12 0.000 0.000 0.000 0.0008 0.0047 0.8028
16 0.000 0.000 0.001 0.0007 0.0040 0.7989
Table D-3: Mean and stdev vs. nr. of measured leads (nominal means are: x=0, y=0, theta=0)

D-4.2 Conclusion
From the results above, can be concluded that using all 16 leads, gives the best estimate of the center- and angle
location of the component.

D-5 Recommendations
The results in this report show that the way the center- and angle location of a component are measured influences
the accuracy. Therefore it is recommended to include general directions for measuring the components which are
used in the IPC-9850 standard.

In the figures below, the measurement recommendations given in this Appendix are summarized.

Glass QFP-100, QFP-208 and BGA-228

Measure the fiducials (1,2,3,4) 4 C 3

Construct intersection points (A,B,C,D)

x,y: Average of x resp. y of the fiducials (1,2,3,4)

theta: Average angle of lines AC and BD D B

1 A 2
1608C

Measure sides of component (thick lines)

Construct intersection points (1,2,3,4) 4 C 3
Calculate midpoints (A,B,C,D)

x,y: Average of x resp. y of the fiducials (1,2,3,4)

D B
theta: Average angle of lines AC and BD

1 A 2

48
 IPC-9850
Official Proposal
May 2001
SOIC16

Measure left, right and bottom of lead. T

Construct intersection points.
Construct midpoint (*).
Draw fines through northern and
southern midpoints.

x,y. Average of all leads

Method 1

The theta of the component is

1, Average the location for both XT and
YT on the top side leads of the component
2, Average the location for both XB and
YB on the bottom side leads of the
component
3. Use the points (XT ,YT) and (XB,YB) to determine a line
4, Then measure the angle of this line minus 90 degrees to determine the angle of the component.

Method 2

theta: Average of angles of northern and southern line.

49
 IPC-9850
Official Proposal
May 2001
Appendix E
Component Locations for Panel Population

IPC 9850 1608C CHIP GRID LAYOUT

5 7.5 10.7 14.75 19.5 25 31.25 38.25 46 54.5
X1 X2 X3 5 X4 X5 X6 X7 X8 X9 X10
X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ
5 5 5 0 7.5 5 90 10.75 5 180 14.75 5 270 19.5 5 0 25 5 90 31.25 5 180 38.25 5 270 46 5 0 54.5 5 90
Y1
7.5 5 7.5 0 7.5 7.5 90 10.75 7.5 180 14.75 7.5 270 19.5 7.5 0 25 7.5 90 31.25 7.5 180 38.25 7.5 270 46 7.5 0 54.5 7.5 90
Y2
10.75 5 10.75 0 7.5 10.75 90 10.75 10.75 180 14.75 10.75 270 19.5 10.75 0 25 10.75 90 31.25 10.75 180 38.25 10.75 270 46 10.75 0 54.5 10.75 90
Y3
14.75 5 14.75 0 7.5 14.75 90 10.75 14.75 180 14.75 14.75 270 19.5 14.75 0 25 14.75 90 31.25 14.75 180 38.25 14.75 270 46 14.75 0 54.5 14.75 90
Y4
19.5 5 19.5 0 7.5 19.5 90 10.75 19.5 180 14.75 19.5 270 19.5 19.5 0 25 19.5 90 31.25 19.5 180 38.25 19.5 270 46 19.5 0 54.5 19.5 90
Y5
25 5 25 0 7.5 25 90 10.75 25 180 14.75 25 270 19.5 25 0 25 25 90 31.25 25 180 38.25 25 270 46 25 0 54.5 25 90
Y6
31.25 5 31.25 0 7.5 31.25 90 10.75 31.25 180 14.75 31.25 270 19.5 31.25 0 25 31.25 90 31.25 31.25 180 38.25 31.25 270 46 31.25 0 54.5 31.25 90
Y7
38.25 5 38.25 0 7.5 38.25 90 10.75 38.25 180 14.75 38.25 270 19.5 38.25 0 25 38.25 90 31.25 38.25 180 38.25 38.25 270 46 38.25 0 54.5 38.25 90
Y8
46 5 46 0 7.5 46 90 10.75 46 180 14.75 46 270 19.5 46 0 25 46 90 31.25 46 180 38.25 46 270 46 46 0 54.5 46 90
Y9
54.5 5 54.5 0 7.5 54.5 90 10.75 54.5 180 14.75 54.5 270 19.5 54.5 0 25 54.5 90 31.25 54.5 180 38.25 54.5 270 46 54.5 0 54.5 54.5 90
Y10
63.75 5 63.75 0 7.5 63.75 90 10.75 63.75 180 14.75 63.75 270 19.5 63.75 0 25 63.75 90 31.25 63.75 180 38.25 63.75 270 46 63.75 0 54.5 63.75 90
Y11
73.75 5 73.75 0 7.5 73.75 90 10.75 73.75 180 14.75 73.75 270 19.5 73.75 0 25 73.75 90 31.25 73.75 180 38.25 73.75 270 46 73.75 0 54.5 73.75 90
Y12
84.5 5 84.5 0 7.5 84.5 90 10.75 84.5 180 14.75 84.5 270 19.5 84.5 0 25 84.5 90 31.25 84.5 180 38.25 84.5 270 46 84.5 0 54.5 84.5 90
Y13
96 5 96 0 7.5 96 90 10.75 96 180 14.75 96 270 19.5 96 0 25 96 90 31.25 96 180 38.25 96 270 46 96 0 54.5 96 90
Y14
108.25 5 108.25 0 7.5 108.25 90 10.75 108.25 180 14.75 108.25 270 19.5 108.25 0 25 108.25 90 31.25 108.25 180 38.25 108.25 270 46 108.25 0 54.5 108.25 90
Y15
121.25 5 121.25 0 7.5 121.25 90 10.75 121.25 180 14.75 121.25 270 19.5 121.25 0 25 121.25 90 31.25 121.25 180 38.25 121.25 270 46 121.25 0 54.5 121.25 90
Y16
135 5 135 0 7.5 135 90 10.75 135 180 14.75 135 270 19.5 135 0 25 135 90 31.25 135 180 38.25 135 270 46 135 0 54.5 135 90
Y17
149.5 5 149.5 0 7.5 149.5 90 10.75 149.5 180 14.75 149.5 270 19.5 149.5 0 25 149.5 90 31.25 149.5 180 38.25 149.5 270 46 149.5 0 54.5 149.5 90
Y18
164.75 5 164.75 0 7.5 164.75 90 10.75 164.75 180 14.75 164.75 270 19.5 164.75 0 25 164.75 90 31.25 164.75 180 38.25 164.75 270 46 164.75 0 54.5 164.75 90
Y19
180.75 5 180.75 0 7.5 180.75 90 10.75 180.75 180 14.75 180.75 270 19.5 180.75 0 25 180.75 90 31.25 180.75 180 38.25 180.75 270 46 180.75 0 54.5 180.75 90
Y20

63.75 73.75 84.5 96 108.25 121.25 135 149.5 164.75 180.75

X11 X12 X13 X14 X15 X16 X17 X18 X19 X20

X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ
5 63.75 5 180 73.75 5 270 84.5 5 0 96 5 90 108.25 5 180 121.25 5 270 135 5 0 149.5 5 90 164.75 5 180 180.75 5 270
Y1
7.5 63.75 7.5 180 73.75 7.5 270 84.5 7.5 0 96 7.5 90 108.25 7.5 180 121.25 7.5 270 135 7.5 0 149.5 7.5 90 164.75 7.5 180 180.75 7.5 270
Y2
10.75 63.75 10.75 180 73.75 10.75 270 84.5 10.75 0 96 10.75 90 108.25 10.75 180 121.25 10.75 270 135 10.75 0 149.5 10.75 90 164.75 10.75 180 180.75 10.75 270
Y3
14.75 63.75 14.75 180 73.75 14.75 270 84.5 14.75 0 96 14.75 90 108.25 14.75 180 121.25 14.75 270 135 14.75 0 149.5 14.75 90 164.75 14.75 180 180.75 14.75 270
Y4
19.5 63.75 19.5 180 73.75 19.5 270 84.5 19.5 0 96 19.5 90 108.25 19.5 180 121.25 19.5 270 135 19.5 0 149.5 19.5 90 164.75 19.5 180 180.75 19.5 270
Y5
25 63.75 25 180 73.75 25 270 84.5 25 0 96 25 90 108.25 25 180 121.25 25 270 135 25 0 149.5 25 90 164.75 25 180 180.75 25 270
Y6
31.25 63.75 31.25 180 73.75 31.25 270 84.5 31.25 0 96 31.25 90 108.25 31.25 180 121.25 31.25 270 135 31.25 0 149.5 31.25 90 164.75 31.25 180 180.75 31.25 270
Y7
38.25 63.75 38.25 180 73.75 38.25 270 84.5 38.25 0 96 38.25 90 108.25 38.25 180 121.25 38.25 270 135 38.25 0 149.5 38.25 90 164.75 38.25 180 180.75 38.25 270
Y8
46 63.75 46 180 73.75 46 270 84.5 46 0 96 46 90 108.25 46 180 121.25 46 270 135 46 0 149.5 46 90 164.75 46 180 180.75 46 270
Y9
54.5 63.75 54.5 180 73.75 54.5 270 84.5 54.5 0 96 54.5 90 108.25 54.5 180 121.25 54.5 270 135 54.5 0 149.5 54.5 90 164.75 54.5 180 180.75 54.5 270
Y10
63.75 63.75 63.75 180 73.75 63.75 270 84.5 63.75 0 96 63.75 90 108.25 63.75 180 121.25 63.75 270 135 63.75 0 149.5 63.75 90 164.75 63.75 180 180.75 63.75 270
Y11
73.75 63.75 73.75 180 73.75 73.75 270 84.5 73.75 0 96 73.75 90 108.25 73.75 180 121.25 73.75 270 135 73.75 0 149.5 73.75 90 164.75 73.75 180 180.75 73.75 270
Y12
84.5 63.75 84.5 180 73.75 84.5 270 84.5 84.5 0 96 84.5 90 108.25 84.5 180 121.25 84.5 270 135 84.5 0 149.5 84.5 90 164.75 84.5 180 180.75 84.5 270
Y13
96 63.75 96 180 73.75 96 270 84.5 96 0 96 96 90 108.25 96 180 121.25 96 270 135 96 0 149.5 96 90 164.75 96 180 180.75 96 270
Y14
108.25 63.75 108.25 180 73.75 108.25 270 84.5 108.25 0 96 108.25 90 108.25 108.25 180 121.25 108.25 270 135 108.25 0 149.5 108.25 90 164.75 108.25 180 180.75 108.25 270
Y15
121.25 63.75 121.25 180 73.75 121.25 270 84.5 121.25 0 96 121.25 90 108.25 121.25 180 121.25 121.25 270 135 121.25 0 149.5 121.25 90 164.75 121.25 180 180.75 121.25 270
Y16
135 63.75 135 180 73.75 135 270 84.5 135 0 96 135 90 108.25 135 180 121.25 135 270 135 135 0 149.5 135 90 164.75 135 180 180.75 135 270
Y17
149.5 63.75 149.5 180 73.75 149.5 270 84.5 149.5 0 96 149.5 90 108.25 149.5 180 121.25 149.5 270 135 149.5 0 149.5 149.5 90 164.75 149.5 180 180.75 149.5 270
Y18
164.75 63.75 164.75 180 73.75 164.75 270 84.5 164.75 0 96 164.75 90 108.25 164.75 180 121.25 164.75 270 135 164.75 0 149.5 164.75 90 164.75 164.75 180 180.75 164.75 270
Y19
180.75 63.75 180.75 180 73.75 180.75 270 84.5 180.75 0 96 180.75 90 108.25 180.75 180 121.25 180.75 270 135 180.75 0 149.5 180.75 90 164.75 180.75 180 180.75 180.75 270
Y20

50
 IPC-9850
Official Proposal
May 2001

IPC 9850 SOIC16 GRID LAYOUT

IPC 9850 QFP-100 & BGA-228 GRID LAYOUT

X1 17 X2 47 X3 77 X4 107 X5 137 X6 167

X Y θ X Y θ X Y θ X Y θ X Y θ X Y θ

Y1 17 17 17 90 47 17 270 77 17 0 107 17 180 137 17 90 167 17 0

Y2 47 17 47 270 47 47 0 77 47 180 107 47 90 137 47 0 167 47 270

Y3 77 17 77 0 47 77 180 77 77 90 107 77 0 137 77 270 167 77 180

Y4 107 17 107 180 47 107 90 77 107 0 107 107 270 137 107 180 167 107 90

Y5 137 17 137 90 47 137 0 77 137 270 107 137 180 137 137 90 167 137 270

Y6 167 17 167 0 47 167 270 77 167 180 107 167 90 137 167 270 167 167 180

Rotation Chart

90 270 0 180 90 0

270 0 180 90 0 270

0 180 90 0 270 180

180 90 0 270 180 90

90 0 270 180 90 270

0 270 180 90 270 180

IPC 9850 QFP-208 GRID LAYOUT

51
 IPC-9850
Official Proposal
May 2001
Appendix F- Component Location for Accuracy Verification Panel
(to be developed)

52
 IPC-9850
Official Proposal
May 2001
Appendix G- How to Perform a GR&R Test
Instruction for Using the Gauge R&R Spreadsheet

G-1 Purpose Gauge R&R is performed to determine the reliability of the measurement process. The Gauge R&R
test sums all the possible sources of error related to having repeatable and reproducible measurements. A list of
possible sources of error are:

G-2 Inspection Measurement Machine – all inspection machines are specified to have limits on how accurately
and repeatably they can measure. This is due mostly to machine design and will vary depending quality and price of
CMM used.
G-3 Inspection Program – the user will write a program for the CMM machine that will measure different
components and different boards. These programs could possibly be a big source of error if they are not written
properly. Every program written will need to have a GR&R test done to ensure its accuracy. For this standard there
needs to be a minimum of 3 different GR&R tests:
• 1608 Capacitor Board
• SOIC16 Board
• Glass Part Board (Assumes the same program is run for both BGA and QFP part)
G-4 Error Contributed by Fixture – When the board is mounted on the CMM during measurement any movement
will cause an error. The goal of the fixture is to hold the board securely every time it is measured.
G-5 Operator Error – If different operators run the same test they should get the same results. Any variation
between the different operators is error. Note: The goal of a good inspection program and fixturing system is to
reduce this variable. If operator error is high the inspection program and/or the fixtures need to be changed.
G-6 Error Contributed by the Component
Glass Component - measurement uncertainty
Live Components
Chips - Resisters, the die cutting process. Capacitors cut with a die saw
Leaded - plastic package forming, lead finish etc.

G-7 Equipment Needed:

• CMM
• 1 Measurement Verification Panel and mounting fixture
• Parts (one of the following)
Ø 36 glass QFP-100 slugs
Ø 30 glass QFP-208 slugs
Ø 36 glass BGA slugs
• 80 SOIC16
• 400 1608C

G-8 Procedure:
1. Run one Measurement Verification Panel on any piece of equipment.

Note: This test is not testing the placement equipment in any way, so the machine that places the placement does not
matter; but the components should be placed on the board with a average amount of deviation.

2. Operator 1 then takes this board and places it on the inspection machine, run the board and logs the results. These
results are copied into the Operator 1; Test 1 area of the spreadsheet.
3. Operator 1 then removes the board from the inspection machine.
4. Operator 1 then places the board back onto the inspection machine, runs the board and gets the results. These
results are copied into the Operator 1; Test 2 area of the spreadsheet.
5. Operator 1 then removes the board from the inspection machine.
6. Operator 1 then places the board back onto the inspection machine, runs the board and gets the results. These
results are copied into the Operator 1; Test 3 area of the spreadsheet.
7. Operator 1 then removes the board and gives to Operator 2.
8. Operator 2 then runs steps 2 though 6 placing their data into the Operator 2 section of the spreadsheet.
9. When all of the data has been entered into the spreadsheet, switch screens to the analysis sheet. Press the get data
button. This will enter the data into the correct columns of the spreadsheet and calculate the result.

The calculation that is performed is the Average & Range method.

53
 IPC-9850
Official Proposal
May 2001
Appendix H- Instruction for Using the Gauge R&R Spreadsheet

G-9 Calculation Method: Average and Range Method

G-10 Definitions
XA -Bar = Average of all the points measured by operator A
RA -Bar = The average of ranges measured by operator A
XB -Bar = Average of all the points measured by operator B
RB -Bar = The average of ranges measured by operator B
R-DblBar = The average of the ranges from operator A and B
RDIF = This is the range between the averages for operators A and B
d2 = 1.693

G-11 Formulas (As seen in the spreadsheet above)

G-11.2 Range of Means

RDIF = ABS{(XA -Bar)-(XB -Bar)}

G-11.3 Repeatability Error

RPT = (R-DblBar)*(6.0/d2)

G-11.4 Reproducibility Error

RPD = (RDIF)*(6.0/d2)

G-11.5 Total Error

R&R Error = SQRT(RPT2 +PRD2 )

Gage R&R =100*[(R&R Error)/(USL-LSL)]

(Must be less then 10% for gage to be valid)

Lowest Specification Limits possible with current gage error.

+/-Specification Limits = ABS{(1/2)*(10)*R&R Error}

54
 IPC-9850
Official Proposal
May 2001

Appendix I - Placement Verification Panel Carrier

A photograph of a suggested suitcase for the Placement Verification Panel carrier. The suitcase is the packaging for
the set of 4 separate carriers. See ‘CarrierDocuments.zip’ for design details of the carrier.

55
 IPC-9850
Official Proposal
May 2001
Appendix J- Instruction for Using The ‘Termination-to-Land Calculations.xls’
spreadsheets

The workbook contains a spreadsheet for each of the 5 component types. In each worksheet, instructions are
included that explain where to place the x, y, and theta errors that were established by the CMM for all components
on the 4 boards. A workbook with termination-to-land calculation spreadsheets for each of the five component types
may be downloaded from the IPC web site. (Document reviewers contact Jack Crawford at IPC for this file.)

The spreadsheet automatically calculates the summary information that needs to be put into the Performance
Reporting Form when it is being used to verify specification compliance.

As an example, here is what the QFP100 spreadsheet looks like:

The equations used are identical to those presented in the body of this standard.

56