PBQ Online Help Home

Introduction to PBQ The Agilent Pushbutton Q-

Interpreting Failure Paretos
Interpreting Analog Histograms STATS Tutorial
Interpreting Production Reports
Interpreting Repair Paretos © Copyright 1993,2001 Agilent Technologies,
Interpreting Yield Reports Inc. All rights reserved.
The Reports Screen
Index of PBQ Topics This on-line help documentation for PBQ has been
programatically translated from the original content (in
"vue help" format) to HTML for viewing by a web
browser. Only minor editing has been done on the
results. Some of the topics and explanations given in
this documentation will only apply to PBQ on the HP-UX
platform (not the PC). This content has also not yet
been updated to conform to the standard format used
by other on-line help documention with the Agilent 3070
software. We apologize for any inconvenience this may
cause you.

Welcome to the Agilent Pushbutton Q-STATS Tutorial, which provides an introduction to the
Agilent Pushbutton Q-STATS Quality Management Software (PBQ).

This text can also be accessed through the VUE Help Index by traversing the index until the
item "The Agilent Pushbutton Q-STATS Tutorial" can be selected. Thus its various topics are
also available as examples when you are accessing to change the real data.

Additional information is available through the related topics under the "See Also" heading of
the Help Viewer window. If one of these is selected the user can return to the previous topic
by selecting the "Previous View" button.

Tutorial Topics Available:

• What can Agilent Pushbutton Q-STATS do for you?
• Purpose of this Tutorial
• Conventions used in this Tutorial
• Introduction to using a Mouse and Windows
• What will be accomplished in this Tutorial
• The Production Manager Scenario
• The Test Engineer Scenario
• The SMT Manufacturing (with Agilent PR PLUS) Scenario
• Selecting the Yield Report
• Interpreting the Yield Results
• Selecting HPdemoBrd1 for the Production Report
• Selecting HPdemoBrd4 for the Production Report
• Production Results for HPdemoBrd1
• Production Results for Non-Serialized HPdemoBrd4
• Selecting HPdemoBrd1 Failures Report
 • Failure Report results for HPdemoBrd1
• Selecting Histograms for HPdemoBrd1
• Histogram Report results for HPdemoBrd1
• Selecting the Production Report for HPdemoBrd2
• Interpreting HPdemoBrd2's Production Report
• Selecting Failures for HPdemoBrd2
• Interpreting HPdemoBrd2's Failure Report
• Selecting the Histogram Report for HPdemoBrd2
• Handling Special Names
• Interpreting the Histograms of HPdemoBrd2
• Selecting the Production Report for HPdemoBrd3
• Interpreting HPdemoBrd3's Production Report
• Selecting the Failure Report for HPdemoBrd3
• Interpreting HPdemoBrd3's Failure Report
• Selecting the Repair Report for HPdemoBrd3
• Interpreting the Repair Report for HPdemoBrd3
• The Database Administrator Scenario
• Deleting Data of Tests Removed from the Testplan
• Deleting Board Types from the Database
• Deleting Old Data from the Database
• Restoring Report Mode without Deleting
• Administrating Output Devices
• Adding an Item to the Device Prefix List
• Changing the Attributes of a Device Prefix List Item
• Removing a Device Prefix from the List
• Shipped Prefix Settings

WHAT CAN AGILENT PUSHBUTTON Q-STATS DO FOR YOU?

Data from your Agilent 3070 board test system is automatically collected for analysis when
datalogging is "turned on" in the testplan. Agilent Pushbutton Q-STATS transforms
manufacturing test data into informative Statistical Quality Control (SQC) reports. Its SQC
reports can help you analyze your production process and identify process problems. And,
after you have taken the appropriate actions, it can measure the effectiveness of your
solutions.

Agilent Pushbutton Q-STATS provides the following types of reports:

• Yield - Shows the yield per board type.

• Production - Shows an overview, by board type, of your production process.
• Failure - Shows which tests are failing for a selected board type.
• Repair - (Requires Agilent PR+) Shows the most frequent causes of repairs.
• Histogram - Shows the test measurement distribution for sampled analog components
of a selected board type.
PURPOSE OF THIS TUTORIAL

This tutorial gives you 'hands on' experience with Agilent Pushbutton Q-STATS Reports. The
reports are produced from sample data in a separate tutorial database, so you never need to
worry about the right data being present.

In this tutorial you will be shown:

• The 'mechanics' of using Agilent Pushbutton Q-STATS

• Three simplified, yet realistic, scenarios at a hypothetical manufacturing facility to show
how SQC reports are used to analyze production processes.
• A scenario for the database administration functions (this requires the user to log in as
qsys).

The manufacturing scenarios are:

THE PRODUCTION MANAGER SCENARIO:

You are the Production Manager and need to know how your production area is doing.
Are there any problems? If so, what are they? Where should you begin looking for
solutions?
THE TEST ENGINEER SCENARIO:
You are a Test Engineer who has just released the test plan for a new circuit board to
production. How effective are the tests? Are there any problems? If so, what are they?
THE SMT MANUFACTURING (& REPAIR) SCENARIO:
The production area for which you are the Production Manager manufactures a circuit
board that uses Surface Mount Technology (SMT). Using repair data collected by the
optional Agilent PR PLUS Repair Loop Management Software, you need to determine
why that board's first pass yield is so low and where to focus your efforts to improve
the production process.

Each scenario describes an actual production situation and tells how to generate a set of
SQC reports whose data is used to identify problems.

CONVENTIONS USED IN THIS TUTORIAL

Places where you must perform an action are indicated with an arrow, like this:

>>> some action to perform

Hints or other supplemental information appear in parentheses:

(This is a hint.)

As you work your way further into the tutorials, the explanations for frequent tasks will
become less detailed. It is therefore suggested you start with the first tutorial. Each scenario
is self-contained and can be exited at any point and restarted.

When directed to 'select' something, the label of the item to be selected will be contained in
quotes ("). Quoted Items should appear in one of the windows.
OTHER FEATURES

Agilent Pushbutton Q-STATS provides on-line help for many of its features. You can get
started with on-line help by selecting "Help" on the main menu bar, selecting "About this
Application" from the menu and then following the instructions that appear.

(NOTE: Selecting any Help Item will replace this text with the appropriate help text. To
reactivate this text select the "Previous View" button in this, the "Help Viewer", window.)

WHAT WILL BE ACCOMPLISHED IN THIS TUTORIAL?

• We will learn the "mechanics" of generating SQC reports with HP Pushbutton Q-

STATS. This includes how to use the mouse to specify report options from convenient
menus and screens.
• We will examine all of the report types generated by Agilent Pushbutton Q-STATS and
have some idea of how each is used.
• We will see how the reports work together to identify problems and show their relative
severity.
• After further investigation we will be able to zero in on the most probable cause of the
problems.
• Database administrator can learn how to enable output devices for usage by PBQ and
how to remove various data items from the database.

MONITORING THE PROCESS

Keep in mind that Agilent Pushbutton Q-STATS reports can also monitor the effectiveness of
the process fixes. The perfect fix, of course, is one where the problem goes away and never
returns. But you must continue to analyze your process in order to know how well the fix
worked.

Used properly, Agilent Pushbutton Q-STATS reports are powerful SQC tools. But you will
never get more out of your SQC efforts than you put into them. While you can certainly use
SQC techniques to find and fix specific 'one-time-only' problems, SQC is most beneficial as
a continuing effort, constantly monitoring your process for better quality and producibility. As
your understanding of SQC grows, so will your successes with HP Pushbutton Q-STATS

WHERE TO FIND MORE INFORMATION

For more information about Agilent Pushbutton Q-STATS refer to Chapter 1 in the
'Information Management' book that is part of your Agilent 3070 or HP82000FT/MCM users'
manual set.

To directly access any tutorial topic, select "Index" in the "Browse" menu of the "Help
Viewer" window. The "Help Index" window will appear with a list of topics. Double clicking
(pressing the mouse button twice in rapid succession) will make that the current topic in the
"Help Viewer" window.
To continue, a scenario must be selected. The various scenarios can be accessed by
selecting them from the "See Also" list in the "Help Viewer" window.

See Also:

• The Database Administrator Scenario

• The Production Manager Scenario
• The SMT Manufacturing (with Agilent PR PLUS) Scenario
• The Test Engineer Scenario
• Introduction to using a Mouse and Windows

What can Agilent Pushbutton Q-STATS do for you?

Data from your Agilent 3070 board test system is automatically collected for analysis when
datalogging is "turned on" in the testplan. Agilent Pushbutton Q-STATS transforms
manufacturing test data into informative Statistical Quality Control (SQC) reports. Its SQC
reports can help you analyze your production process and identify process problems. And,
after you have taken the appropriate actions, it can measure the effectiveness of your
solutions.

Agilent Pushbutton Q-STATS provides the following types of reports:

• Yield - Shows the yield per board type.

• Production - Shows an overview, by board type, of your production process.
• Failure - Shows which tests are failing for a selected board type.
• Repair - (Requires Agilent PR+) Shows the most frequent causes of repairs.
• Histogram - Shows the test measurement distribution for sampled analog components
of a selected board type.

Purpose of this Tutorial

This tutorial gives you 'hands on' experience with Agilent Pushbutton Q-STATS Reports. The
reports are produced from sample data in a separate tutorial database, so you never need to
worry about the right data being present.

In this tutorial you will be shown:

• The 'mechanics' of using Agilent Pushbutton Q-STATS

• Three simplified, yet realistic, scenarios at a hypothetical manufacturing facility to show
how SQC reports are used to analyze production processes.
• A scenario for the database administration functions (this requires the user to log in as
qsys).

The manufacturing scenarios are:

THE PRODUCTION MANAGER SCENARIO: You are the Production Manager and need to
know how your production area is doing. Are there any problems? If so, what are they?
Where should you begin looking for solutions?
THE TEST ENGINEER SCENARIO: You are a Test Engineer who has just released the test
plan for a new circuit board to production. How effective are the tests? Are there any
problems? If so, what are they?

THE SMT MANUFACTURING (& REPAIR) SCENARIO: The production area for which you
are the Production Manager manufactures a circuit board that uses Surface Mount
Technology (SMT). Using repair data collected by the optional Agilent PR PLUS Repair Loop
Management Software, you need to determine why that board's first pass yield is so low and
where to focus your efforts to improve the production process.

Each scenario describes an actual production situation and tells how to generate a set of
SQC reports whose data is used to identify problems.

Conventions used in this Tutorial

Places where you must perform an action are indicated with an arrow, like this:

>>> some action to perform

Hints or other supplemental information appear in parentheses:

(This is a hint.)

As you work your way further into the tutorials, the explanations for frequent tasks will
become less detailed. It is therefore suggested you start with the first tutorial. Each scenario
is self-contained and can be exited at any point and restarted.

When directed to 'select' something, the label of the item to be selected will be contained in
quotes ("). Quoted Items should appear in one of the windows.

OTHER FEATURES

Agilent Pushbutton Q-STATS provides on-line help for many of its features. You can get
started with on-line help by selecting "Help" on the main menu bar, selecting "About this
Application" from the menu and then following the instructions that appear.

(NOTE: Selecting any Help Item will replace this text with the appropriate help text. To
reactivate this text select the "Previous View" button in this, the "Help Viewer", window.)

Introduction to using a Mouse and Windows

You can use a "mouse" to interact with items that appear in "windows" on the screen.
Moving the mouse around causes an indicator called the "mouse cursor" to move across the
screen. This indicator might change its appearance as it enters different areas and at
different times. For instance the normal mouse cursor for Agilent Pushbutton Q-STATS is the
crosshair, it is shaped like +.
If you ever come to the edge of the work area (run out of room to roll the mouse) and need
to move the mouse cursor even further on the screen, simply pick up the mouse and
reposition it so it has additional room to move.

The mouse has buttons which you press to perform an action. When you are told to select
something, one way is to place the mouse cursor on the specified object and then press and
immediately release the left button.

Menu item selection may require a press on the menu selection button and to hold it down
and release it on the desired menu item. Note: The menu item showing the shadows is the
one which will be selected when the mouse button is released.

When the mouse cursor is in the shape of a watch it is indicating the application is busy and
can not currently respond to any button clicks (repeatedly selecting an item might result in
duplicate or unexpected results).

When you type on the keyboard, your input goes to the active window (if it will accept that
key press). A window can be made active by pressing the left mouse button when that
window contains the mouse cursor.

Most items are also keyboard accessible. Once the Agilent Pushbutton Q-STATS window
has been selected, a yellow border should appear around one of the items. To move
between items you can use the & or && (between text fields) and sometimes the various
arrow keys (usually between buttons). To 'select' a highlighted button, press the spacebar,
"Return", or "Select" key. If a menu is visible you can select an item from it by pressing the
key that corresponds to the underlined character. To access the menubar items, "File" and
"Help", you must press f10 (the second blank key from the left over the numeric keypad) and
then use the arrow keys to navigate through the menus. When a key sequence appears on
the right side of the button this indicates that a keyboard accelerator is available. For most
buttons this will be a hold of the "CTRL" button while pressing the underlined character key.
These accelerators allow menu item selection from the keyboard without having to make the
menu visible. An example of this would be using the& & sequence to start the tutorial (if it is
not already running). The text fields also change their background color when you have
selected them and they are 'editable'. Editable fields appear blue when they are inactive and
black when they are active. This is the only time that the keyboard key presses will go
directly into the selected field. Such fields start out in "replace" mode, that is the key press
will replace the character to the right of the character cursor, an item similar to I (note this
cursor is different than the mouse cursor). Use of the "Insert char" key will allow the user to
enter the "insert" mode which has the key presses' character inserted before the character
cursor. An additional press of the "Insert char" key will return the user to "replace" mode. The
mouse can be used to position the character cursor by clicking the mouse button when the
mouse cursor is positioned in the field where the character cursor is desired.

Some of the editable fields have a scrollbar next to them. The scrollbar consists of two arrow
buttons (one each end) and a rectangular piece. When the rectangular piece completely fills
the area between the arrows the corresponding window is displaying the entire contents.
The size of the rectangle is proportional to the percentage of information displayed.
Manipulating the scrollbar allows the different sections to become visible. There are several
ways to do this: 1. place the mouse cursor on one of the arrows and click until the desired
items appear or you have reached the end of the list. If you reached the end without finding
them, use the other arrow. 2. place the mouse cursor on the bar and hold the mouse button
down while moving the mouse up and down until the items appear and then release the
mouse button. 3. Use the keyboard by first using the "Tab" key until the yellow highlight is on
the editable field and then use the "Prev" and/or "Next" keys to cause scrolling. The arrow
keys can also cause scrolling, but only one item at a time. The home key (the arrow at an
angle) will cause the window to jump to the beginning or if used with the "Shift" key to the
end.

NOTE: Occasionally a report or message will appear and obscure part of this tutorial
window. If this happens, simply move the mouse cursor into the tutorial window and click the
mouse's left button. This will bring the tutorial window (or most other windows in which you
place the mouse cursor) to the foreground.

If you entered this topic from the "See Also" section of the "Help Viewer", you may use the
"Previous View" button at the bottom of that window to obtain a path to the scenarios.

What will be accomplished in this Tutorial

• We will learn the "mechanics" of generating SQC reports with HP Pushbutton Q-
STATS. This includes how to use the mouse to specify report options from convenient
menus and screens.
• We will examine all of the report types generated by Agilent Pushbutton Q-STATS and
have some idea of how each is used.
• We will see how the reports work together to identify problems and show their relative
severity.
• After further investigation we will be able to zero in on the most probable cause of the
problems.
• Database administrator can learn how to enable output devices for usage by PBQ and
how to remove various data items from the database.

The Production Manager Scenario

INTRODUCTION

You are now the Production Manager at a company that manufactures four board types:
"HPdemoBrd1", "HPdemoBrd2", "HPdemoBrd3", and "HPdemoBrd4".

The first question to ask is: "How are we doing?"

We can break this overall question down into several smaller questions:

• 1 How well is the manufacturing process doing?

• 2 What is the most serious problem?
• 3 What are the possible causes of that problem?
THE YIELD REPORT

The Yield Report is the best place to begin because it can tell us the board yield for each
board type.
>>> Select "Report Type:".

(The >>> means that this is an action for you to perform. In this case if the button has
a yellow border highlighting it,you may press the spacebar, "Return", or "Select" key.
Alternatively you may place the mouse cursor on the "Report Type" button and press
the mouse button.)

>>> When the menu appears, select the area labeled "Yield".

(Yield is the first item in the list, but try moving the mouse cursor up and down just to
see what happens.)

The menu disappears and the field beneath "Report Type" now says "Yield". Thus, you have
selected Yield as the current report.
>>> Verify that value under "Output Device:" is "Screen".

(Screen is the default, so you shouldn't have to change this. You might want to select
"Output Device:" to see what devices are available to this system. If you do, select
either Screen or press the "ESC" key to exit the menu. If nothing happens, it is
possible that no legal output device was found for this system, the interface therefore
will not allow you to change the device.)

>>> Verify that the editable fields beneath "Time Interval ..." show "12/01/88" at "00:00" as
the values to the right of the "Start" label, and "12/12/88" at "17:00" as the values to
the right of the "End" label.

(These are the defaults, so you shouldn't have to change them.)

>>> Verify that the editable field beneath the "4/4 Boards" button shows "HPdemoBrd1",
"HPdemoBrd2", "HPdemoBrd3", and "HPdemoBrd4" as the values since we want all
board types reported.

(Again, these are the defaults for this tutorial. Otherwise you would have to specify
which board types are of interest. The 4/4 on the button indicates that 4 out of 4 board
types are currently selected. Frequently all the board types selected may not be able to
appear at once in the field, so the scroll bar allows the ability to view the other values.)

(Note there are three fields with grey backgrounds. These fields are not currently
editable and require some button action to change their values. Also note that the
button next to the "4/4 Boards ..." button is not easily readable. This indicates that the
button is 'inactive' or not 'selectable'.)

The option fields in the main screen should now contain the following values:
 Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBrd4
Report Footer = TUTORIAL database report

>>> Select "Run Report" to generate the report.

(The "Run Report" button appears at the bottom left of the main window. Selecting it
means "Generate a report using the options shown in this window.")

Status messages appears briefly in a new window to show that Agilent Pushbutton Q-STATS
is producing the report. Eventually the actual report appears in this window since the "Output
Device:" selection was "Screen". When the report appears the background color of the test
section will change to yellow.

(Select the report window and use the "Prev" and "Next" keys or arrow keys to scroll
through the entire report, which is longer than the window. You might wish to resize the
window to get a more complete view. Two additional things to note are that most keys
have no affect when pressed and this window is active and that the arrow keys do not
always appear to have an effect. Actually the arrow keys effect the test cursor which is
not visible so there is no visible result unless the cursor tries to move outside the
visible text.)

WHAT DOES THE YIELD REPORT TELL US?

The Yield Report shows that:

• 50 HPdemoBrd2's were tested.

• 51 HPdemoBrd4's were tested.
• 250 HPdemoBrd1's were tested.
• 100 HPdemoBrd3's were tested.
• The first pass yield (1st Pass Pct.) was low for all board types, except HPdemoBrd4
which does not have a value because it contains non-serialized boards.

("1st. Pass Pct." is the yield the first time that board type was tested, while "Pct. Yield"
is the final yield after all boards of that type were tested, repaired, retested, scrapped,
etc. A high "1st. Pass Pct." is desirable because that means you will spend less time
and money fixing boards.)

(If your system has a printer connected, you can copy the report to it by selecting the
"Print ..." button in the report window's "File" menu and selecting the "OK" button on
the window which appears.)

>>> When you have finished examining the Yield Report, select the "Exit" button in the
report window's "File" menu. (Accelerator: <CTRL><x>.)

(This removes the current report.)

THE PRODUCTION REPORT

Let's take a closer look at the HPdemoBrd1 board type to determine why its first pass yield
was low. Using the Production Report is the next step.
>>> Select "Report Type:".
>>> When the menu appears, select "Production".
The menu disappears and the field beneath "Report Type:" now contains "Production".

Since HPdemoBrd1 is the board type of interest, you must change the values in the field
beneath "4/4 Boards ..." button as follows:
>>> Select "4/4 Boards ...".
When the "Agilent PBQ List Selection" screen appears, note that there are entries for all
valid board types, which in this case means HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, and
HPdemoBrd4. If the board type is lit (a bright color), then data for that board type will appear
in the report. Since none of the board types are selected and you're only interested in
HPdemoBrd1, you must 'select' that entry.
>>> Select the box labeled "HPdemoBrd1" and verify that it is the only item lit.

(Each box is a 'toggle', so selecting it reverses its current status. Upon entering the
screen, the "OK" button is highlighted, using the tab key will transfer the highlight and
focus onto the first item in the list. Keyboard selection of the item can be accomplished
by using the spacebar, "Select", or the "Return" key.)

When you have successfully completed this step the field above the selection list should
read "1 of 4 Boards Selected".
>>> Select "OK" at the bottom left of the screen.

(In effect, this says 'OK, I'm done. Save the changes I made and return to the main
screen.' Had you selected the "Cancel" button instead, you would have returned to the
main screen without saving any of the changes you made.)

The "Agilent PBQ List Selection" screen disappears and the field beneath the "1/4 Boards
..." button now contains only "HPdemoBrd1".

The fields in the main screen should now contain the following values:
Report Type = Production
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
1/4 Boards = HPdemoBrd1
Report Footer = TUTORIAL database report

>>> Generate the Production Report for HPdemoBrd1 by selecting "Run Report".
This will produce the "Agilent Pushbutton Q-STATS Report" status window which contains
information about how the report is running. A report can be canceled before completion by
selecting the "Exit" button in the "File" menu of the status window. Once the report software
determines the report is is running, it also changes the label on the status window. The
status window will be relabeled as "Agilent PBQ Yield for 4 Board Types" and eventually
contain the actual report (when the background color changes in the text portion.)

WHAT DOES THE PRODUCTION REPORT TELL US?

The Production Report gives a more detailed look at the test history of the HPdemoBrd1
board type. Examining this reports shows that:

• Although the first pass yield ("1st Pass Pct.") is low, almost all of the boards (243)
eventually passed (after repairs and retests).
(When more than one board type appears in the Production Report, the data is
sorted by increasing yield in left to right order.)
• The average test time ("Avg. Test Time"), which is 11.30 seconds, looks reasonable for
this board type. This is the average time required to test the entire board. This will be
dependent on the number of component tests run and the complexity of each.
• Each failing board test averaged 1.01 failing component tests ("Avg. Test Fails"). So for
the most part the failing boards only had a single failing component. This can be
misleading when the testplan quits testing after the first failure or section failure occurs.
For example failing the shorts test does not usually allow other tests to run so that
damage to the board can be avoided.
• There were no fixture contact (CHECK-POINT) problems, as indicated by 0 for "Avg.
Ckpt. Fails". A large number of fixture contact failures is generally a good indication
that fixture maintenance is required.

These statistics indicate that the problem does not reside in any of these areas. To further
isolate the cause of the problems, we need a report type that provides more in-depth
information than the Production Report does.

Before removing this report let us contrast this against a board which is not serialized -

Let's take a closer look at the HPdemoBrd4 board type and examine the difference not
serializing makes in the Production Report.
>>> Select "Report Type:".
>>> When the menu appears, select "Production".
The menu disappears and the field beneath "Report Type:" now contains "Production".

Since HPdemoBrd4 is the board type of interest, you must change the values in the field
beneath "1/4 Boards ..." as follows:
>>> Select "1/4 Boards ...".
When the "Agilent PBQ List Selection" screen appears, note that there are entries for all
valid board types, which in this case means HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, and
HPdemoBrd4. If the board type is lit (a bright color), then data for that board type will appear
in the report. Since none of the board types are selected and you're only interested in
HPdemoBrd4, you must 'select' that entry.
>>> Select the box labeled "HPdemoBrd4" and verify that it is the only item lit.
When you have successfully completed this step the field above the selection list should
read "1 of 4 Boards Selected".
>>> Select "OK" at the bottom left of the screen.
The "Agilent PBQ List Selection" screen disappears and the field in the main window now
contains only "HPdemoBrd4" and the button label reads "1/4 Boards ...".

The fields in the main screen should now contain the following values:
Report Type = Production
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
1/4 Boards = HPdemoBrd4
Report Footer = TUTORIAL database report

>>> Generate the Production Report for HPdemoBrd4 by selecting "Run Report".

We can contrast the serialized and non-serialized production reports now -

The Production Report gives a less detailed look at the test history of HPdemoBrd4 than it
provided for HPdemoBrd1 since without serialization a board identity can not be determined.
Examining this reports shows that:

• First pass yield ("1st Pass Pct.") is not available and "Pct. Yield" is reflecting not the
percent of passing boards, but the percent of passing board tests. It indicates that 28
boards finally passed (assuming once a board passes it is not retested) but we do not
know how many different boards were tested only that 50 board tests were done. The
number of different boards of type HPdemoBrd4 through the system during the
selected time period was somewhere between 28 and 50
• The average test time ("Avg. Test Time") looks reasonable for this board type.
• Each failing board test averaged 1.27 failing components ("Avg. Test Fails").
• There were no fixture contact (CHECK-POINT) problems, as indicated by 0 for "Avg.
Ckpt. Fails".

With the unserialized boards of type HPdemoBrd4 it is not possible to determine the amount
of rework required or the effectiveness of the repairs.

(Additionally, further contrast on the Production Report may be obtained by running it

with both HPdemoBrd1 and HPdemoBrd4 selected. Note what happens in the 1st pass
entries.)

>>> When you have finished examining the Production Report, select the "Exit" button in
each of the report windows' "File" menus.

THE FAILURE REPORT

A typical SQC strategy is to tackle the most serious problems first. Also, we would like to
know whether the problem affects many different parts or only a few. Because it lists
problems by severity, a Pareto chart (such as shown by the Failure Report) is a good way to
determine what our most serious problems are.
>>> Select "Report Type:".
>>> Select "Failure".
The menu disappears and the field beneath "Report Type:" now contains "Failure". The
board designation label also changes from "1/4 Boards ..." to "Board ..." to indicate the
report is only run on a single board type at a time.

The option fields in the main screen should now contain the following values:
Report Type = Failure
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd1
Report Footer = TUTORIAL database report

(If the field under the "Board ..." button does not contain "HPdemoBrd1" the simplest
way to change it may be to place the mouse cursor on the last character and click the
mouse button. If the character cursor appears to the left of the last character simply
type "1". If the cursor is to the right of the last character press "Backspace" and then
"1".)

>>> Select Run Report.

This will produce the "Agilent Pushbutton Q-STATS Report" status window which will later
change its name to "Agilent PBQ HPdemoBrd1 Test Failures" and later another window
containing the report graphics. The name change will help you associate the status and
graphics windows (This is important when several graphics reports have been run.) If the
status window is obscured, you may have to move (iconify or lower) the graphics window to
have access to it.

WHAT DOES THE FAILURE REPORT TELL US?

The Failure Report is a Pareto chart of test failures for the selected board type. Vertical bars
indicate the number of failures of the individual tests; the taller the bar, the greater the
number of failures of the corresponding test. The bars are sorted from left to right, with the
tallest bars (the most frequently failing tests) appearing at the left side of the report.

Notice that associated with each bar is a tick mark (along the green cumulative line), and
near each tick mark is the cumulative failure percentage represented by that bar.

Examining the Failure Report shows that component "C5" has the highest failure rate. By
itself, C5 accounts for over 60% of the failures for this board type. Thus if we can determine
the problem associated with C5, we may be able to remove a major impediment to good
throughput.
>>> When you have finished examining the Failure Report, select the "Exit" button in the
"Agilent PBQ HPdemoBrd1 Test Failures" window's "File" menu.
This will remove the "Agilent PBQ HPdemoBrd1 Test Failures" status window and the
window containing the report graphics.
THE HISTOGRAM REPORT

We now know that the test for C5 is the most severe failure, but we don't know how it is
failing. Since C5 is an analog part (the C is the usual way of labeling capacitor tests), a
histogram can help.
>>> Select "Report Type:".
>>> When the menu appears, select "Histogram".
The menu disappears and the field beneath "Report Type:" now contains "Histogram". The
field beneath "Tests ..." is now active (notice that it has changed color) and it contains the
designators for all the analog component tests for which the database contains the test limits
information for the current board type. The button label now reads "34/34 Tests ...".

Since the amount of time it takes to produce the report will depend on the amount of data
that is processed and we are only interested in one test, we will run the report only for C5.
>>> Select "34/34 Tests ...".
When the "Agilent PBQ List Selection" screen appears, it shows all of the tests that can
appear in a Histogram Report for the currently selected board type. If the test designator is
lit, then data for that test will appear in the report. Since you are only interested in the test for
C5, you must select that test.
>>> Select the "C5" item in the list.

(This selects the test for C5 to appear in the report.)

When you have successfully completed this step the label above the selection list should
read "1 of 34 Tests Selected".
>>> Select "OK".
The test selection screen disappears, the main screen reappears, and the field beneath the
"1/34 Tests ..." button contains only "C5".

The fields in the main screen should now contain the following values:
Report Type = Histogram
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd1
1/34 Tests = C5
Report Footer = TUTORIAL database report

>>> Generate a Histogram Report for HPdemoBrd1 by selecting "Run Report".

WHAT DOES THE HISTOGRAM REPORT TELL US?

Examining the Histogram Report shows that:

• The tests that passed, which are shown as green bars, are closely arranged around the
mean (the small blue "X" beneath the graph). Thus, the passing tests look good statistically
and indicate that the test is probably valid.

(Interpreting the Histogram Report is explained in more detail in the second scenario.)
• The tests that failed are shown as blue bars to indicate they are so far outside the test
limits that they constitute an unusual (basically non-functioning) failure.

(Had there been any moderate failures, they would have appeared as red bars. An
example of these appears in the second scenario.)

The fact that some tests for C5 pass well within the test limits while others fail grossly might
mean that the wrong parts are being inserted. With the statistical results pointing to this, the
next step is to investigate the problem on the manufacturing floor.
>>> When you have finished examining the Histogram Report, select the "Exit" button in
the report window's "File" menu.

WHAT ARE OUR CONCLUSIONS?

Our investigation revealed the following:

A recent change of vendors produced a mix of similar looking part reels for different parts.
Operators sometimes picked up the wrong auto-insertion machine reel for capacitor C5.
Thus, wrong parts were being inserted into the circuit boards.

The fix was to clearly label the reels so that it was easier to distinguish one from the other.

WAS IT THE RIGHT FIX?

The 'right fix' is one that continues to work. Once a problem is fixed, you should generate
additional reports to monitor the results and make sure the fix is effective.

This ends the Production Manager Scenario. To continue with the next scenario select
"Previous View" in the "Help Viewer" window.

See Also:

• Selecting HPdemoBrd1 for the Production Report

• Selecting HPdemoBrd4 for the Production Report
• Selecting HPdemoBrd1 Failures Report
• Selecting the Histogram Report for HPdemoBrd2
• Selecting Histograms for HPdemoBrd1
• Introduction to using a Mouse and Windows
• Selecting the Yield Report

The Test Engineer Scenario

(This scenario focuses on HPdemoBrd2's unresolved yield problems from the previous
scenario.)
INTRODUCTION

You are now a Test Engineer (or Test Developer) at the same company as in the previous
scenario. The first production run of "HPdemoBrd2" has just been tested, and it's your job to
determine why the first pass yield is so low. The board type contains 203 components with
the majority being digital parts and only 30 analog parts.

THE PRODUCTION REPORT

Let's begin by taking a look at the Production Report.

>>> Select "Report Type:".
>>> When the menu appears, select "Production".
The menu disappears and the field beneath "Report Type" now says "Production". Since
HPdemoBrd2 is the board type of interest, you must change the value in the field beneath
the "*/4 Boards" button as follows.
>>> Select "*/4 Boards".
When the "Agilent PBQ List Selection" screen appears, note that there are entries for all
valid board types. Since we only desire "HPdemoBrd2" do the following.
>>> Edit the blue field next to the "For items matching:" label to read "*2".

(The field will change to a dark background and have a yellow border when you can
type in it. NOTE: DO NOT ACTUALLY TYPE THE QUOTE MARKS IN - IT WILL
CAUSE ALL MATCHES TO FAIL.)

>>> Select the "Add" button.

The list will be searched for all items ending in 2, which in this case is only HPdemoBrd2.
When you have successfully completed this step the label above the selection list should
read "1 of 4 Boards Selected".

(If it reads "0 of 4 Boards Selected" check the edit field to make sure it only contains
the two characters - asterisk and two.)

>>> Select "OK" at the bottom left of the screen.

The "Agilent PBQ List Selection" screen disappears and the main screen reappears. The
field beneath "1/4 Boards ..." now contains only "HPdemoBrd2".

Since we know that the first production run took place in one day, we want to change the
time interval to examine a specific set of data.
>>> Select "Time Interval ...".
When the "Agilent PBQ Time Selection" screen appears, the current values for "Start" and
"End" each appear in their own sections with the month number replaced by a three letter
abbreviation.

Let's leave the value for "Start" as-is ("Dec 1, '88" at "00:00"), but modify "End" so that the
next report generated will show data for only one day.
In order to examine data for a single day, you need to set "End" to "Dec 2, '88" at "00:00".
Since it is easier to go up from "17:00" to "00:00" -
>>> Select the upward pointing arrow button in the "End" section under the "Hour" label
seven times until the time reads "00:00".

(Note that this changes the end date field to "Dec 13, '88".)

>>> Select the downward pointing arrow button in the "End" section under the "Day" label
eleven times until the date reads "Dec 2, '88".
>>> Select "OK".
When the main screen appears, the values for "Start" and "End" are "12/01/88" at "00:00"
and "12/02/88" at "00:00" respectively.

The fields in the main screen should now contain the following values:
Report Type = Production
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/02/88 00:00
1/4 Boards = HPdemoBrd2
Report Footer = TUTORIAL database report

>>> Generate the Production Report for HPdemoBrd2 by selecting "Run Report".

WHAT DOES THE PRODUCTION REPORT TELL US?

The Production Report shows that:

• 50 HPdemoBrd2's were tested.

• The first pass yield (1st Pass Pct.) and overall yield are the same, since no retests or
repairs were done during the time interval.
• The yield was only 56%; that is, 28 boards passed of the 50 that were tested.
• The average test time (Avg. Test Time) looks reasonable.
• The number of failures per board test (Avg. Test Fails) is 1.05.
• Since Avg. Ckpt. Fails is 0, there were no fixture contact problems.

>>> When you have finished examining the Production Report, select the "Exit" button in
the report window's "File" menu.

THE FAILURE REPORT

Let's generate a Failure Report to find out which tests are failing.
>>> Select "Report Type:".
>>> Select "Failure".
The "Select Report Type" menu disappears and the option field beneath "Report Type:" now
contains "Failure".

The option fields in the main screen should now contain the following values:
 Report Type = Failure
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/02/88 00:00
Board = HPdemoBrd2
Report Footer = TUTORIAL database report

>>> Select "Run Report".

WHAT DOES THE FAILURE REPORT TELL US?

The Failure Report for HPdemoBrd2 shows that:

• Several tests are failing: R25, R10, R15, SHORTS, U17, and R20, but none of them are
failing every time. We know this because none of the tests show 22 failures. The test failing
most often is R25, but at this point (initial turn on), it may be prudent to gather as much
information as possible on all of them.
>>> When you have finished examining the Failure Report, select the "Exit" button in the
report status window's "File" menu.

THE HISTOGRAM REPORT

A Histogram Report will show us how the analog components are failing (R10, R15, R20,
and R25). The other failing tests, SHORTS and U17, can not have histograms produced
since they do not obtain any measured values. Also, let's include a couple of passing resistor
tests for reference so we can see what a passing test looks like; for example: R5, R6, and
R7.
>>> Select "Report Type:".
>>> When the menu appears, select "Histogram".
The menu disappears and the field beneath "Report Type:" now contains "Histogram". The
field beneath "30/30 Tests ..." is now active (notice that it has changed color) and it contains
the designators for all the analog component tests. This may not always be the case since
this information is only available after the limits have been learned. The database contains
limits information for the current board type. The button label now reads "30/30 Tests ...".
>>> Select "30/30 Tests ...".
When the "Agilent PBQ List Selection" screen appears, it shows all of the tests that can
appear in a Histogram Report for the currently selected board type. If the test designator is
lit, then data for that test will appear in the report.
>>> Edit the blue field next to the "For items matching:" label to read "R*5".
>>> Select the "Add" button.
The list will be searched for all items starting with R and ending in 5, which in this case is
R5, R15, and R25. When you have successfully completed this step the label above the
selection list should read "3 of 30 Tests Selected".
>>> Edit the blue field again this time to read "R*0".
>>> Select the "Add" button.
The list will be searched for all items starting with R and ending in 0, which in this case is
R10 and R20. When you have successfully completed this step the label above the selection
list should read "5 of 30 Tests Selected".
>>> Select the "R6" and "R7" items in the list.

(You may have to move the list scrollbar to see R6 and R7 first.)

When you have successfully completed this step the label above the selection list should
read "7 of 30 Tests Selected".
>>> Select "OK".
The test selection screen disappears, the main screen reappears, and the field beneath
"7/30 Tests ..." now contains "R10", "R15", "R20", "R25", "R5", "R6", and "R7". The Tests
label now reads "7/30 Tests ...".

(The order in which the test designators appear is not important; the histograms for the
tests will automatically be sorted by quality when the Histogram Report is generated.)

The fields in the main screen should now contain the following values:
Report Type = Histogram
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/02/88 00:00
Board = HPdemoBrd2
7/30 Tests = R10, R15, R20, R25, R5, R6, R7
Report Footer = TUTORIAL database report

>>> Generate a Histogram Report for HPdemoBrd2 by selecting "Run Report".

WHAT DOES THE HISTOGRAM REPORT TELL US?

Because you selected more than six tests (which appear six per page), the Histogram
Report is two pages long. Each new page appears in a window that overlaps the previous
page, but the windows are offset so that they do not entirely cover each other.

If the page you need to look at is hidden behind a foreground page, you can bring it to the
front by placing the mouse cursor on its exposed edge and then clicking the mouse's left
button.

Each page has a page number at its lower right corner; for example "1|2" means "page 1 of
2".

Each test is in its own section and is ordered by "% passing" first and when that is equivalent
the value in "C.P".

The Histogram Report uses the following symbol conventions in its graphics section:

• Red tick marks indicate the high and low test limits for passing tests; thus,
measurements within the red tick marks are within the specified limits for the test.
(On the histogram for "R25" these marks represent the position of the values
"1.87K" and "2.148K", which appear on the "Limits" line of the test portion.)
• The green tick mark is the position of the nominal value.
(If the nominal tick mark ever fails to appear, it is because the nominal and the
mean coincide; that is, the mean tick mark is obscuring the nominal tick mark.
 Note that it is visible in the histogram for "R10" and "R15", but appears as part of
the x-bar symbol in the rest of the graphs on page 1.)
• Blue tick marks indicate the high and low limits for tests that failed but were not gross
failures.
(The area between the red and blue tick marks is the expected failure range.
That is where measurements for functioning components that were not up to
specifications are expected to be.)
• The mean is shown by a 'barred-X' symbol beneath the graph.
(The mean and nominal almost coincide in the case of "R25". Looking at the
values for the Limits and Mean we can determine that this is probably a 2K
resistor.)
• Green bars represent tests that passed; they were between the high and low test
limits. Green bars appear inside the red tick marks.
(The passing range is divided up into seven equal divisions. The number above
the graph indicates the number of measurements in the highest bar.)
• Red bars represent tests that failed but were within approximately 50% of the high or
low test limits (moderate failures). Red bars appear outside the red tick marks but
inside the blue tick marks.
(Each expected failure range is divided up into four equal divisions. "R25" has
failures on both the high and low ends, "R10" failed only on the low end, while
"R15" and "R20" both failed only on the high end.)
• Blue bars represent grossly failing measurements; they appear outside the blue tick
marks.

(This report has no grossly failing measurements. See the Histogram for HPdemoBrd1
for an example.)

Examining the Histogram Report shows that:

• All the tests consist of 50 measurements (the value on the "Total" line) which is a large
enough sample to be statistically significant (it should be at least 25). This number may
not correspond to the number of board tests run for several reasons. First histograms
are only run on the sampled tests which are determined by the sampling rate selected
in the test plan. This sampling rate should be set so that a statistically significant
number of measurements will be made during the normal interval the histogram report
displays (one day for this report). In this example the sample rate must have been set
to 1.00 (100%) to record every board test. When a board type reaches its normal
production state, the sample rate should be decreased so that the PBQ database does
not grow too fast. A second influence on the number of tests sampled is the form of the
testplan. If the testplan is written such that the failure of some tests prohibit other from
running, some test may not be run every time. Normally tests are run in groups or
sections so that damage is not done as when a short exists on a board.
• The histograms are in order of increasing "% passing" values going down the columns.
For tests with the same "% passing" value they are ordered by the "CPk", coefficient of
Producibility, value. This produces a worst to best type ordering and shows the largest
problems first. The "% passing" values represented here are only from the sampled
tests even though unsampled failures may exist in the database. Therefore the test
which shows up as the worst on the failure report may not be the worst on the
histogram.
 • "R25" has the highest failure rate, which might be caused by:
◦ Noisy measurements, which means that the test needs further enhancements.
◦ Components that are out of tolerance, perhaps due to a faulty batch of
components.
◦ A test whose test limits are too tight, and could be widened with no adverse
effects on the quality or the performance of the board.
◦ An incorrectly specified component (the design is wrong), which means that a
closer tolerance component is needed on this board.
(These speculations are supported by broadness of the graph. The standard
deviation, "Std. Dev.", represents 5% of the nominal value.)
• "R10"'s measurements are skewed low (i.e., the mean value is less than the nominal
value), which might be caused by:
◦ A test where additional guarding is needed because a component is in parallel
with this resistor.
◦ Some kind of production problem.
• "R15"'s measurements are skewed high, which could be caused by:
◦ Source loading, which means that the test requires modification.
◦ (If "R15" is a metal-film resistor) Overheating in the wave solder process, which
is permanently raising the resistor's value.
◦ Some other production problem.
• "R20"'s measurements are centered, but where we would expect the highest bar (the
greatest number of measurements) in the middle at the nominal value (the mean), it is
actually the smallest bar. Perhaps the vendor is removing the one percent resistors
from the batches we receive.

Obviously, any of these possibilities requires further investigation. But the causes are typical
of what you might expect to find in a real production environment.

The remaining resistors ("R5", "R6", and "R7") are good examples of histograms for passing
measurements; that is, all measurements are within the limits, and most of the
measurements are tightly distributed about the mean. Note that the tighter the grouping the
higher the "CPk" value.
>>> When you have finished examining the Histogram Report, select the "Exit" button in
the report status window's "File" menu. This will remove both the graphics and status
windows.

ADVANCED ITEM SELECTION

Generating a Histogram Report for HPdemoBrd4 will show us how the advanced selection
mechanism can be used when names contain special characters and how very long names
are handled.
>>> Select a "Time Interval ..." of "12/01/88" at "00:00" to "12/12/88" at "17:00".

(Try using the keyboard this time by selecting the erroneous fields and just typing into
them. Note the separators to be consistent. Use the arrow keys to skip over the correct
characters and "Tab" / "Shift" "Tab" to change the selected field.)

>>> Select "Board ..." "HPdemoBrd4".

 >>> Select "Report Type:" "Histogram".
The menu disappears and the field beneath "Report Type:" now contains "Histogram". The
field beneath "33/33 Tests ..." is now active (notice that it has changed color) and it contains
the designators for all the analog component tests. The first test name might not be
completely visible (it is over 64 characters long). To be able to see the entire name two
methods are possible. First we will select the field and move the cursor through the name.
>>> Select the test field containing "A_VERY_LONG_ ...".
>>> If the blinking cursor is not to the left of the first character press the home key (a hollow
arrow which points up and to the left).
>>> Use the right arrow key to move the blinking cursor. When it reaches the right side of
the field, additional presses make the field shift left exposing the hidden characters.
When you reach the end of "A_VERY_LONG_ANALOG_TEST_NAME_THAT_EXCEEDS
_DISPLAY_LIMITS" the cursor will jump to the second test name.

A way to see the entire name at once is to stretch the interface window until the entire name
shows up. To do this:
>>> Use the mouse and place its cursor on the left edge of the window (it should appear
something like |&<- ). Hold the left mouse button down while moving the mouse to the
left until the entire name is visible. (This may not be possible on small displays).
>>> Restore the window to the original size.
Other than being cumbersome, long test names do not effect the interface.

There are special characters that must be handled differently in the selection interface when
they appear in test or board names. To see the effect of these do the following:
>>> Select "33/33 Tests ...".
>>> Edit the blue field next to the "For items matching:" label to read "C*T1".
>>> Select the "Add" button.
The list will be searched for all items starting with C and ending in T1, which in this case is
both C*T1 and C-T1. Even though the pattern exactly matches "C*T1" the add does not
recognize this as a special case, but this is generally not a problem. The problem comes
when you want to select an item which contains one of the special characters by using
pattern matching. The way to do this is to use the escape character (&\) in the pattern just
before the special character which is no to be treated specially.
>>> Select "Undo Action".
>>> Edit the blue field next to the "For items matching:" label to read "C&\*T1".
>>> Select the "Add" button.
The list will be searched and this time only C*T1 is selected. This unfortunately is not very
useful so we are going to generalize this example and find all tests which start with the
characters 'C*'.
>>> Select "Undo Action".
>>> Edit the blue field again this time to read "C&\**".
>>> Select the "Add" button.
Once again only C*T1 is selected, since it is the only item that matches this pattern. This
method also works for the other special characters - '?' the single match character, '&\' the
escape character, and '[' the match set start character. For more information see the
"PATTERN MATCHING NOTATION" section of the "REGEXP(5)" entry of the HP-UX
manual.
>>> Select the "A_VERY_LONG_..." and "C-T1" items in the list.
When you have successfully completed this step the label above the selection list should
read "3 of 33 Tests Selected".
>>> Select "OK".
The test selection screen disappears, the main screen reappears, and the field beneath
"3/33 Tests ..." now contains "A_VERY_LONG_ANALOG...", "C*T1", and "C-T1". The Tests
label now reads "3/33 Tests ...".

The fields in the main screen should now contain the following values:
Report Type = Histogram
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd4
3/33 Tests = A_VERY_LONG_ANALOG_TEST_NAME_THAT_EXCEEDS_DISPLAY_LIMITS,
C*T1, and C-T1.
Report Footer = TUTORIAL database report

>>> Generate a Histogram Report for HPdemoBrd4 by selecting "Run Report".

Notice that the second item has "*** 1 ***" instead of a test name. When a test name of
excessive length would appear in a graph, a legend index number is used instead. To see
the name the index refers to:
>>> Scroll the status window back until the line containing "*** 1 ***" appears.
This shows that
"A_VERY_LONG_ANALOG_TEST_NAME_THAT_EXCEEDS_DISPLAY_LIMITS"
corresponds to "*** 1 ***".
>>> When you have finished examining the Histogram Report, select the "Exit" button in
the report window's "File" menu.
The legend will also appear for a failure report when the name length exceeds 32
characters.
>>> Select "Failure".
The fields in the main screen should now contain the following values:
Report Type = Failure
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd4
Report Footer = TUTORIAL database report

>>> Select "Run Report".

A three page Failure Report is generated with "*** 1 ***" as the label for the fourth bar of the
first page (1|3). Another noteworthy item is the label on the first bar of the second page. This
label is cut into two pieces so that no more than sixteen characters are printed together. The
retrieval of the legend is the same:
>>> Scroll the status window back until the line containing "*** 1 ***".
This shows that
"A_VERY_LONG_ANALOG_TEST_NAME_THAT_EXCEEDS_DISPLAY_LIMITS"
corresponds to "*** 1 ***".
>>> When you have finished examining the Failures Report, select the "Exit" button in the
report window's "File" menu. All three graphics windows are destroyed with the status
window.

This ends the Test Engineer Scenario. Use "Previous View" button on the "Help Viewer" to
access other scenarios.

See Also:

• Selecting Failures for HPdemoBrd2

• Selecting the Histogram Report for HPdemoBrd2
• Selecting the Production Report for HPdemoBrd2
• Introduction to using a Mouse and Windows

The SMT Manufacturing (with Agilent PR PLUS) Scenario

INTRODUCTION

The production area for which you are the Production Manager manufactures a circuit board,
"HPdemoBrd3", that uses Surface Mount Technology (SMT). The production line is using the
optional Agilent PR PLUS Repair Loop Management Software to gather repair information
that is sent to Agilent Pushbutton Q-STATS for process analysis.

The first pass yield for "HPdemoBrd3" is low, and it is your job to find out why.

THE PRODUCTION REPORT

Let's begin the investigation by generating a Production Report for HPdemoBrd3.

>>> Select "Production".

(As mentioned earlier, the further you progress in this tutorial, the less detailed the
explanations become. From now on, we expect you to know what steps are required to
select an item in a menu or a screen.)

>>> Use "Agilent PBQ List Selection" to select "HPdemoBrd2" and "HPdemoBrd3".

(Unselect any other board types.)

>>> Select a time interval of "12/01/88" at "00:00" to "12/12/88" at "17:00".

(If you cannot remember how to select the time interval, Select "Help" on the "Agilent
PBQ Time Selection" screen and follow the instructions that appear. If necessary, refer
back to 'Selecting the Production Report for HPdemoBrd2' for a detailed explanation.)

The fields in the main screen should now contain the following values:
 Report Type = Production
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
2/4 Boards = HPdemoBrd2, HPdemoBrd3
Report Footer = TUTORIAL database report

>>> Generate the Production Report by selecting "Run Report".

WHAT DOES THE PRODUCTION REPORT TELL US?

The Production Report shows that:

• 100 HPdemoBrd3's were tested, of which 98 passed during the specified time interval.
• Only 25 of the 100 boards passed the first time they were tested. The other 75 boards
were sent to be repaired.
• At the repair operation, 2 boards were scrapped and the other 73 boards were
successfully repaired. The average repair time was 12.75 minutes.

(Notice that there are places where 'N/A', which means 'Not Applicable,' appears in the
Production Report; for example, "Boards Scrapped" for "HPdemoBrd2" is N/A. N/A
means there was no repair data available for that board type. If you are using the
optional Agilent PR PLUS Repair Loop Management Software to collect repair data,
your Production Reports will contain N/A whenever repair statistics were not collected.)

(You can remove the production report.)

THE FAILURE REPORT

Now that we know something about the history of HPdemoBrd3, let's generate a Failure
Report to see which tests failed.
>>> Select "Failure".
>>> Confirm "Board ..." contains "HPdemoBrd3" - change if necessary.
The fields in the main screen should now contain the following values:
Report Type = Failure
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd3
Report Footer = TUTORIAL database report

>>> Select "Run Report".

WHAT DOES THE FAILURE REPORT TELL US?

The Failure Report for HPdemoBrd3 shows that:

• The main failure is "SHORTS" (>71%), which tests for both shorts and opens on the
circuit board.
• The remaining test failures are mostly digital devices ("U12", "U9", "U3"), as well as one
analog component ("R7").

(You can remove the failure report.)

The Repair Report can now help in determining the cause of the shorts as observed by the
repair operators, thus leading to a possible solution of the problem.

THE REPAIR REPORT

The Repair Report shows what caused the failures, as determined by the repair operator
(who entered the repair information in Agilent PR PLUS).

Generate a Repair Report for HPdemoBrd3 as follows:

>>> Select "Repair".
>>> Enter the edit field beneath the "2/4 Boards ..." button and delete the line containing
"HPdemoBrd2".

(The label "2/4 Boards ..." will not change until the field is exited. This is accomplished
by selecting another item. To do this press the Tab key.)

The fields in the main screen should now contain the following values:
Report Type = Repair
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
1/4 Boards = HPdemoBrd3
Report Footer = TUTORIAL database report

>>> Select Run Report.

WHAT DOES THE REPAIR REPORT TELL US?

The Repair Report shows what caused the shorts and opens failures shown collectively as
SHORTS test failures in the Failure Report we examined earlier.

Among the causes of the SHORTS test failures revealed by the Repair Report are:

• "INSUFFICIENT_SOLDER", which caused the most failures (24).

• "BENT_LEADS", which can touch and cause shorts, caused the second largest
number of failures (12).
• "SOLDER_BRIDGE" caused some shorts.
• Opens caused by solder on the pads but not touching a part's leads, denoted by
"POOR_WETTING_OF_LEAD".
• "TOMBSTONE" parts, which means that one end of an SMT component is lifted (an
open circuit).

Also among the problems shown are:

 • Parts are being inserted backwards, "BACKWARD_PART"; for example, digital ICs are
reversed end for end.
• "WRONG_PART"s are being inserted on the boards.
• Parts missing from the boards appear as "MISSING_PART".

Two boards had "OPEN_TRACE". These are probably the two scrapped boards which were
shown in the Production Report.

(Note that when a cause label is over 16 characters it is printed in two parts. If it is over
32 characters it will show up as a legend index *** number *** and you may need to
scroll the status window to make it appear. This is also true for the test names on the
failure report.)

(You can remove the repair report.)

As the Production Manager, you now know enough to begin correcting the problems.
Concentrate on the biggest problems first and you will see the greatest improvement in first
pass yields.

See Also:

• Selecting the Failure Report for HPdemoBrd3

• Selecting the Production Report for HPdemoBrd3
• Selecting the Repair Report for HPdemoBrd3
• Selecting the Production Report for HPdemoBrd2
• Introduction to using a Mouse and Windows

Selecting the Yield Report

The Yield Report is the best place to begin because it can tell us the board yield for each
board type.
>>> Select "Report Type:".

(The >>> means that this is an action for you to perform. In this case if the button has
a yellow border highlighting it,you may press the spacebar, "Return", or "Select" key.
Alternatively you may place the mouse cursor on the "Report Type" button and press
the mouse button.)

>>> When the menu appears, select the area labeled "Yield".

(Yield is the first item in the list, but try moving the mouse cursor up and down just to
see what happens.)

The menu disappears and the field beneath "Report Type" now says "Yield". Thus, you have
selected Yield as the current report.
>>> Verify that value under "Output Device:" is "Screen".
 (Screen is the default, so you shouldn't have to change this. You might want to select
"Output Device:" to see what devices are available to this system. If you do, select
either Screen or press the "ESC" key to exit the menu. If nothing happens, it is
possible that no legal output device was found for this system, the interface therefore
will not allow you to change the device.)

>>> Verify that the editable fields beneath "Time Interval ..." show "12/01/88" at "00:00" as
the values to the right of the "Start" label, and "12/12/88" at "17:00" as the values to
the right of the "End" label.

(These are the defaults, so you shouldn't have to change them.)

>>> Verify that the editable field beneath the "4/4 Boards" button shows "HPdemoBrd1",
"HPdemoBrd2", "HPdemoBrd3", and "HPdemoBrd4" as the values since we want all
board types reported.

(Again, these are the defaults for this tutorial. Otherwise you would have to specify
which board types are of interest. The 4/4 on the button indicates that 4 out of 4 board
types are currently selected. Frequently all the board types selected may not be able to
appear at once in the field, so the scroll bar allows the ability to view the other values.)

(Note there are three fields with grey backgrounds. These fields are not currently
editable and require some button action to change their values. Also note that the
button next to the "4/4 Boards ..." button is not easily readable. This indicates that the
button is 'inactive' or not 'selectable'.)

The option fields in the main screen should now contain the following values:
Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBrd4
Report Footer = TUTORIAL database report

See Also:

• Introduction to using a Mouse and Windows

Interpreting the Yield Results

The Yield Report shows that:

• 50 HPdemoBrd2's were tested.

• 51 HPdemoBrd4's were tested.
• 250 HPdemoBrd1's were tested.
• 100 HPdemoBrd3's were tested.
• The first pass yield (1st Pass Pct.) was low for all board types, except HPdemoBrd4
which does not have a value because it contains non-serialized boards.
 ("1st. Pass Pct." is the yield the first time that board type was tested, while "Pct. Yield"
is the final yield after all boards of that type were tested, repaired, retested, scrapped,
etc. A high "1st. Pass Pct." is desirable because that means you will spend less time
and money fixing boards.)

See Also:

• Introduction to using a Mouse and Windows

• Selecting the Yield Report

Selecting HPdemoBrd1 for the Production Report

Let's take a closer look at the HPdemoBrd1 board type to determine why its first pass yield
was low. Using the Production Report is the next step.
>>> Select "Report Type:".
>>> When the menu appears, select "Production".
The menu disappears and the field beneath "Report Type:" now contains "Production".

Since HPdemoBrd1 is the board type of interest, you must change the values in the field
beneath "4/4 Boards ..." button as follows:
>>> Select "4/4 Boards ...".
When the "Agilent PBQ List Selection" screen appears, note that there are entries for all
valid board types, which in this case means HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, and
HPdemoBrd4. If the board type is lit (a bright color), then data for that board type will appear
in the report. Since none of the board types are selected and you're only interested in
HPdemoBrd1, you must 'select' that entry.
>>> Select the box labeled "HPdemoBrd1" and verify that it is the only item lit.

(Each box is a 'toggle', so selecting it reverses its current status. Upon entering the
screen, the "OK" button is highlighted, using the tab key will transfer the highlight and
focus onto the first item in the list. Keyboard selection of the item can be accomplished
by using the spacebar, "Select", or the "Return" key.)

When you have successfully completed this step the field above the selection list should
read "1 of 4 Boards Selected".
>>> Select "OK" at the bottom left of the screen.

(In effect, this says 'OK, I'm done. Save the changes I made and return to the main
screen.' Had you selected the "Cancel" button instead, you would have returned to the
main screen without saving any of the changes you made.)

The "Agilent PBQ List Selection" screen disappears and the field beneath the "1/4 Boards
..." button now contains only "HPdemoBrd1".

The fields in the main screen should now contain the following values:
Report Type = Production
Output Device = Screen
 Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
1/4 Boards = HPdemoBrd1
Report Footer = TUTORIAL database report

>>> Generate the Production Report for HPdemoBrd1 by selecting "Run Report".
This will produce the "Agilent Pushbutton Q-STATS Report" status window which contains
information about how the report is running. A report can be canceled before completion by
selecting the "Exit" button in the "File" menu of the status window. Once the report software
determines the report is is running, it also changes the label on the status window. The
status window will be relabeled as "Agilent PBQ Yield for 4 Board Types" and eventually
contain the actual report (when the background color changes in the text portion.)

See Also:

• Introduction to using a Mouse and Windows

Selecting HPdemoBrd4 for the Production Report

Let's take a closer look at the HPdemoBrd4 board type and examine the difference not
serializing makes in the Production Report.
>>> Select "Report Type:".
>>> When the menu appears, select "Production".
The menu disappears and the field beneath "Report Type:" now contains "Production".

Since HPdemoBrd4 is the board type of interest, you must change the values in the field
beneath "1/4 Boards ..." as follows:
>>> Select "1/4 Boards ...".
When the "Agilent PBQ List Selection" screen appears, note that there are entries for all
valid board types, which in this case means HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, and
HPdemoBrd4. If the board type is lit (a bright color), then data for that board type will appear
in the report. Since none of the board types are selected and you're only interested in
HPdemoBrd4, you must 'select' that entry.
>>> Select the box labeled "HPdemoBrd4" and verify that it is the only item lit.
When you have successfully completed this step the field above the selection list should
read "1 of 4 Boards Selected".
>>> Select "OK" at the bottom left of the screen.
The "Agilent PBQ List Selection" screen disappears and the field in the main window now
contains only "HPdemoBrd4" and the button label reads "1/4 Boards ...".

The fields in the main screen should now contain the following values:
Report Type = Production
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
1/4 Boards = HPdemoBrd4
Report Footer = TUTORIAL database report
>>> Generate the Production Report for HPdemoBrd4 by selecting "Run Report".

See Also:

• Introduction to using a Mouse and Windows

Production Results for HPdemoBrd1

The Production Report gives a more detailed look at the test history of the HPdemoBrd1
board type. Examining this reports shows that:

• Although the first pass yield ("1st Pass Pct.") is low, almost all of the boards (243)
eventually passed (after repairs and retests).
(When more than one board type appears in the Production Report, the data is
sorted by increasing yield in left to right order.)
• The average test time ("Avg. Test Time"), which is 11.30 seconds, looks reasonable for
this board type. This is the average time required to test the entire board. This will be
dependent on the number of component tests run and the complexity of each.
• Each failing board test averaged 1.01 failing component tests ("Avg. Test Fails"). So for
the most part the failing boards only had a single failing component. This can be
misleading when the testplan quits testing after the first failure or section failure occurs.
For example failing the shorts test does not usually allow other tests to run so that
damage to the board can be avoided.
• There were no fixture contact (CHECK-POINT) problems, as indicated by 0 for "Avg.
Ckpt. Fails". A large number of fixture contact failures is generally a good indication
that fixture maintenance is required.

These statistics indicate that the problem does not reside in any of these areas. To further
isolate the cause of the problems, we need a report type that provides more in-depth
information than the Production Report does.

See Also:

• Selecting HPdemoBrd1 for the Production Report

• Introduction to using a Mouse and Windows

Production Results for Non-Serialized HPdemoBrd4

The Production Report gives a less detailed look at the test history of HPdemoBrd4 than it
provided for HPdemoBrd1 since without serialization a board identity can not be determined.
Examining this reports shows that:

• First pass yield ("1st Pass Pct.") is not available and "Pct. Yield" is reflecting not the
percent of passing boards, but the percent of passing board tests. It indicates that 28
boards finally passed (assuming once a board passes it is not retested) but we do not
know how many different boards were tested only that 50 board tests were done. The
 number of different boards of type HPdemoBrd4 through the system during the
selected time period was somewhere between 28 and 50
• The average test time ("Avg. Test Time") looks reasonable for this board type.
• Each failing board test averaged 1.27 failing components ("Avg. Test Fails").
• There were no fixture contact (CHECK-POINT) problems, as indicated by 0 for "Avg.
Ckpt. Fails".

With the unserialized boards of type HPdemoBrd4 it is not possible to determine the amount
of rework required or the effectiveness of the repairs.

See Also:

• Selecting HPdemoBrd4 for the Production Report

• Introduction to using a Mouse and Windows

Selecting HPdemoBrd1 Failures Report

A typical SQC strategy is to tackle the most serious problems first. Also, we would like to
know whether the problem affects many different parts or only a few. Because it lists
problems by severity, a Pareto chart (such as shown by the Failure Report) is a good way to
determine what our most serious problems are.
>>> Select "Report Type:".
>>> Select "Failure".
The menu disappears and the field beneath "Report Type:" now contains "Failure". The
board designation label also changes from "1/4 Boards ..." to "Board ..." to indicate the
report is only run on a single board type at a time.

The option fields in the main screen should now contain the following values:
Report Type = Failure
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd1
Report Footer = TUTORIAL database report

(If the field under the "Board ..." button does not contain "HPdemoBrd1" the simplest
way to change it may be to place the mouse cursor on the last character and click the
mouse button. If the character cursor appears to the left of the last character simply
type "1". If the cursor is to the right of the last character press "Backspace" and then
"1".)

>>> Select Run Report.

This will produce the "Agilent Pushbutton Q-STATS Report" status window which will later
change its name to "Agilent PBQ HPdemoBrd1 Test Failures" and later another window
containing the report graphics. The name change will help you associate the status and
graphics windows (This is important when several graphics reports have been run.) If the
status window is obscured, you may have to move (iconify or lower) the graphics window to
have access to it.
See Also:

• Introduction to using a Mouse and Windows

Failure Report results for HPdemoBrd1

The Failure Report is a Pareto chart of test failures for the selected board type. Vertical bars
indicate the number of failures of the individual tests; the taller the bar, the greater the
number of failures of the corresponding test. The bars are sorted from left to right, with the
tallest bars (the most frequently failing tests) appearing at the left side of the report.

Notice that associated with each bar is a tick mark (along the green cumulative line), and
near each tick mark is the cumulative failure percentage represented by that bar.

Examining the Failure Report shows that component "C5" has the highest failure rate. By
itself, C5 accounts for over 60% of the failures for this board type. Thus if we can determine
the problem associated with C5, we may be able to remove a major impediment to good
throughput.
>>> When you have finished examining the Failure Report, select the "Exit" button in the
"Agilent PBQ HPdemoBrd1 Test Failures" window's "File" menu.
This will remove the "Agilent PBQ HPdemoBrd1 Test Failures" status window and the
window containing the report graphics.

See Also:

• Selecting HPdemoBrd1 Failures Report

• Introduction to using a Mouse and Windows

Selecting Histograms for HPdemoBrd1

We now know that the test for C5 is the most severe failure, but we don't know how it is
failing. Since C5 is an analog part (the C is the usual way of labeling capacitor tests), a
histogram can help.
>>> Select "Report Type:".
>>> When the menu appears, select "Histogram".
The menu disappears and the field beneath "Report Type:" now contains "Histogram". The
field beneath "Tests ..." is now active (notice that it has changed color) and it contains the
designators for all the analog component tests for which the database contains the test limits
information for the current board type. The button label now reads "34/34 Tests ...".

Since the amount of time it takes to produce the report will depend on the amount of data
that is processed and we are only interested in one test, we will run the report only for C5.
>>> Select "34/34 Tests ...".
When the "Agilent PBQ List Selection" screen appears, it shows all of the tests that can
appear in a Histogram Report for the currently selected board type. If the test designator is
lit, then data for that test will appear in the report. Since you are only interested in the test for
C5, you must select that test.
>>> Select the "C5" item in the list.

(This selects the test for C5 to appear in the report.)

When you have successfully completed this step the label above the selection list should
read "1 of 34 Tests Selected".
>>> Select "OK".
The test selection screen disappears, the main screen reappears, and the field beneath the
"1/34 Tests ..." button contains only "C5".

The fields in the main screen should now contain the following values:
Report Type = Histogram
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd1
1/34 Tests = C5
Report Footer = TUTORIAL database report

>>> Generate a Histogram Report for HPdemoBrd1 by selecting "Run Report".

See Also:

• Introduction to using a Mouse and Windows

Histogram Report results for HPdemoBrd1

Examining the Histogram Report shows that:
• The tests that passed, which are shown as green bars, are closely arranged around the
mean (the small blue "X" beneath the graph). Thus, the passing tests look good statistically
and indicate that the test is probably valid.

(Interpreting the Histogram Report is explained in more detail in the second scenario.)

• The tests that failed are shown as blue bars to indicate they are so far outside the test
limits that they constitute an unusual (basically non-functioning) failure.

(Had there been any moderate failures, they would have appeared as red bars. An
example of these appears in the second scenario.)

The fact that some tests for C5 pass well within the test limits while others fail grossly might
mean that the wrong parts are being inserted. With the statistical results pointing to this, the
next step is to investigate the problem on the manufacturing floor.
>>> When you have finished examining the Histogram Report, select the "Exit" button in
the report window's "File" menu.
See Also:

• Selecting the Histogram Report for HPdemoBrd2

• Selecting Histograms for HPdemoBrd1
• Introduction to using a Mouse and Windows

Selecting the Production Report for HPdemoBrd2

Let's begin by taking a look at the Production Report.
>>> Select "Report Type:".
>>> When the menu appears, select "Production".
The menu disappears and the field beneath "Report Type" now says "Production". Since
HPdemoBrd2 is the board type of interest, you must change the value in the field beneath
the "*/4 Boards" button as follows.
>>> Select "*/4 Boards".
When the "Agilent PBQ List Selection" screen appears, note that there are entries for all
valid board types. Since we only desire "HPdemoBrd2" do the following.
>>> Edit the blue field next to the "For items matching:" label to read "*2".

(The field will change to a dark background and have a yellow border when you can
type in it. NOTE: DO NOT ACTUALLY TYPE THE QUOTE MARKS IN - IT WILL
CAUSE ALL MATCHES TO FAIL.)

>>> Select the "Add" button.

The list will be searched for all items ending in 2, which in this case is only HPdemoBrd2.
When you have successfully completed this step the label above the selection list should
read "1 of 4 Boards Selected".

(If it reads "0 of 4 Boards Selected" check the edit field to make sure it only contains
the two characters - asterisk and two.)

>>> Select "OK" at the bottom left of the screen.

The "Agilent PBQ List Selection" screen disappears and the main screen reappears. The
field beneath "1/4 Boards ..." now contains only "HPdemoBrd2".

Since we know that the first production run took place in one day, we want to change the
time interval to examine a specific set of data.
>>> Select "Time Interval ...".
When the "Agilent PBQ Time Selection" screen appears, the current values for "Start" and
"End" each appear in their own sections with the month number replaced by a three letter
abbreviation.

Let's leave the value for "Start" as-is ("Dec 1, '88" at "00:00"), but modify "End" so that the
next report generated will show data for only one day.

In order to examine data for a single day, you need to set "End" to "Dec 2, '88" at "00:00".
Since it is easier to go up from "17:00" to "00:00" -
>>> Select the upward pointing arrow button in the "End" section under the "Hour" label
seven times until the time reads "00:00".

(Note that this changes the end date field to "Dec 13, '88".)

>>> Select the downward pointing arrow button in the "End" section under the "Day" label
eleven times until the date reads "Dec 2, '88".
>>> Select "OK".
When the main screen appears, the values for "Start" and "End" are "12/01/88" at "00:00"
and "12/02/88" at "00:00" respectively.

The fields in the main screen should now contain the following values:
Report Type = Production
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/02/88 00:00
1/4 Boards = HPdemoBrd2
Report Footer = TUTORIAL database report

>>> Generate the Production Report for HPdemoBrd2 by selecting "Run Report".

See Also:

• Introduction to using a Mouse and Windows

Interpreting HPdemoBrd2's Production Report

The Production Report shows that:

• 50 HPdemoBrd2's were tested.

• The first pass yield (1st Pass Pct.) and overall yield are the same, since no retests or
repairs were done during the time interval.
• The yield was only 56%; that is, 28 boards passed of the 50 that were tested.
• The average test time (Avg. Test Time) looks reasonable.
• The number of failures per board test (Avg. Test Fails) is 1.05.
• Since Avg. Ckpt. Fails is 0, there were no fixture contact problems.

>>> When you have finished examining the Production Report, select the "Exit" button in
the report window's "File" menu.

See Also:

• Selecting the Production Report for HPdemoBrd2

• Introduction to using a Mouse and Windows
Selecting Failures for HPdemoBrd2
Let's generate a Failure Report to find out which tests are failing.
>>> Select "Report Type:".
>>> Select "Failure".
The "Select Report Type" menu disappears and the option field beneath "Report Type:" now
contains "Failure".

The option fields in the main screen should now contain the following values:
Report Type = Failure
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/02/88 00:00
Board = HPdemoBrd2
Report Footer = TUTORIAL database report

>>> Select "Run Report".

See Also:

• Introduction to using a Mouse and Windows

Interpreting HPdemoBrd2's Failure Report

The Failure Report for HPdemoBrd2 shows that:
• Several tests are failing: R25, R10, R15, SHORTS, U17, and R20, but none of them are
failing every time. We know this because none of the tests show 22 failures. The test failing
most often is R25, but at this point (initial turn on), it may be prudent to gather as much
information as possible on all of them.
>>> When you have finished examining the Failure Report, select the "Exit" button in the
report status window's "File" menu.

See Also:

• Selecting Failures for HPdemoBrd2

• Introduction to using a Mouse and Windows

Selecting the Histogram Report for HPdemoBrd2

A Histogram Report will show us how the analog components are failing (R10, R15, R20,
and R25). The other failing tests, SHORTS and U17, can not have histograms produced
since they do not obtain any measured values. Also, let's include a couple of passing resistor
tests for reference so we can see what a passing test looks like; for example: R5, R6, and
R7.
>>> Select "Report Type:".
 >>> When the menu appears, select "Histogram".
The menu disappears and the field beneath "Report Type:" now contains "Histogram". The
field beneath "30/30 Tests ..." is now active (notice that it has changed color) and it contains
the designators for all the analog component tests. This may not always be the case since
this information is only available after the limits have been learned. The database contains
limits information for the current board type. The button label now reads "30/30 Tests ...".
>>> Select "30/30 Tests ...".
When the "Agilent PBQ List Selection" screen appears, it shows all of the tests that can
appear in a Histogram Report for the currently selected board type. If the test designator is
lit, then data for that test will appear in the report.
>>> Edit the blue field next to the "For items matching:" label to read "R*5".
>>> Select the "Add" button.
The list will be searched for all items starting with R and ending in 5, which in this case is
R5, R15, and R25. When you have successfully completed this step the label above the
selection list should read "3 of 30 Tests Selected".
>>> Edit the blue field again this time to read "R*0".
>>> Select the "Add" button.
The list will be searched for all items starting with R and ending in 0, which in this case is
R10 and R20. When you have successfully completed this step the label above the selection
list should read "5 of 30 Tests Selected".
>>> Select the "R6" and "R7" items in the list.

(You may have to move the list scrollbar to see R6 and R7 first.)

When you have successfully completed this step the label above the selection list should
read "7 of 30 Tests Selected".
>>> Select "OK".
The test selection screen disappears, the main screen reappears, and the field beneath
"7/30 Tests ..." now contains "R10", "R15", "R20", "R25", "R5", "R6", and "R7". The Tests
label now reads "7/30 Tests ...".

(The order in which the test designators appear is not important; the histograms for the
tests will automatically be sorted by quality when the Histogram Report is generated.)

The fields in the main screen should now contain the following values:
Report Type = Histogram
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/02/88 00:00
Board = HPdemoBrd2
7/30 Tests = R10, R15, R20, R25, R5, R6, R7
Report Footer = TUTORIAL database report

>>> Generate a Histogram Report for HPdemoBrd2 by selecting "Run Report".

See Also:

• Introduction to using a Mouse and Windows

Handling Special Names

Generating a Histogram Report for HPdemoBrd4 will show us how the advanced selection
mechanism can be used when names contain special characters and how very long names
are handled.
>>> Select a "Time Interval ..." of "12/01/88" at "00:00" to "12/12/88" at "17:00".

(Try using the keyboard this time by selecting the erroneous fields and just typing into
them. Note the separators to be consistent. Use the arrow keys to skip over the correct
characters and "Tab" / "Shift" "Tab" to change the selected field.)

>>> Select "Board ..." "HPdemoBrd4".

>>> Select "Report Type:" "Histogram".
The menu disappears and the field beneath "Report Type:" now contains "Histogram". The
field beneath "33/33 Tests ..." is now active (notice that it has changed color) and it contains
the designators for all the analog component tests. The first test name might not be
completely visible (it is over 64 characters long). To be able to see the entire name two
methods are possible. First we will select the field and move the cursor through the name.
>>> Select the test field containing "A_VERY_LONG_ ...".
>>> If the blinking cursor is not to the left of the first character press the home key (a hollow
arrow which points up and to the left).
>>> Use the right arrow key to move the blinking cursor. When it reaches the right side of
the field, additional presses make the field shift left exposing the hidden characters.
When you reach the end of "A_VERY_LONG_ANALOG_TEST_NAME_THAT_EXCEEDS
_DISPLAY_LIMITS" the cursor will jump to the second test name.

A way to see the entire name at once is to stretch the interface window until the entire name
shows up. To do this:
>>> Use the mouse and place its cursor on the left edge of the window (it should appear
something like |&<- ). Hold the left mouse button down while moving the mouse to the
left until the entire name is visible. (This may not be possible on small displays).
>>> Restore the window to the original size.
Other than being cumbersome, long test names do not effect the interface.

There are special characters that must be handled differently in the selection interface when
they appear in test or board names. To see the effect of these do the following:
>>> Select "33/33 Tests ...".
>>> Edit the blue field next to the "For items matching:" label to read "C*T1".
>>> Select the "Add" button.
The list will be searched for all items starting with C and ending in T1, which in this case is
both C*T1 and C-T1. Even though the pattern exactly matches "C*T1" the add does not
recognize this as a special case, but this is generally not a problem. The problem comes
when you want to select an item which contains one of the special characters by using
pattern matching. The way to do this is to use the escape character (&\) in the pattern just
before the special character which is no to be treated specially.
>>> Select "Undo Action".
>>> Edit the blue field next to the "For items matching:" label to read "C&\*T1".
>>> Select the "Add" button.
The list will be searched and this time only C*T1 is selected. This unfortunately is not very
useful so we are going to generalize this example and find all tests which start with the
characters 'C*'.
>>> Select "Undo Action".
>>> Edit the blue field again this time to read "C&\**".
>>> Select the "Add" button.
Once again only C*T1 is selected, since it is the only item that matches this pattern. This
method also works for the other special characters - '?' the single match character, '&\' the
escape character, and '[' the match set start character. For more information see the
"PATTERN MATCHING NOTATION" section of the "REGEXP(5)" entry of the HP-UX
manual.
>>> Select the "A_VERY_LONG_..." and "C-T1" items in the list.
When you have successfully completed this step the label above the selection list should
read "3 of 33 Tests Selected".
>>> Select "OK".
The test selection screen disappears, the main screen reappears, and the field beneath
"3/33 Tests ..." now contains "A_VERY_LONG_ANALOG...", "C*T1", and "C-T1". The Tests
label now reads "3/33 Tests ...".

The fields in the main screen should now contain the following values:
Report Type = Histogram
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd4
3/33 Tests = A_VERY_LONG_ANALOG_TEST_NAME_THAT_EXCEEDS_DISPLAY_LIMITS,
C*T1, and C-T1.
Report Footer = TUTORIAL database report

>>> Generate a Histogram Report for HPdemoBrd4 by selecting "Run Report".

Notice that the second item has "*** 1 ***" instead of a test name. When a test name of
excessive length would appear in a graph, a legend index number is used instead. To see
the name the index refers to:
>>> Scroll the status window back until the line containing "*** 1 ***" appears.
This shows that
"A_VERY_LONG_ANALOG_TEST_NAME_THAT_EXCEEDS_DISPLAY_LIMITS"
corresponds to "*** 1 ***".
>>> When you have finished examining the Histogram Report, select the "Exit" button in
the report window's "File" menu.
The legend will also appear for a failure report when the name length exceeds 32
characters.
>>> Select "Failure".
The fields in the main screen should now contain the following values:
Report Type = Failure
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd4
Report Footer = TUTORIAL database report

>>> Select "Run Report".

A three page Failure Report is generated with "*** 1 ***" as the label for the fourth bar of the
first page (1|3). Another noteworthy item is the label on the first bar of the second page. This
label is cut into two pieces so that no more than sixteen characters are printed together. The
retrieval of the legend is the same:
>>> Scroll the status window back until the line containing "*** 1 ***".
This shows that
"A_VERY_LONG_ANALOG_TEST_NAME_THAT_EXCEEDS_DISPLAY_LIMITS"
corresponds to "*** 1 ***".
>>> When you have finished examining the Failures Report, select the "Exit" button in the
report window's "File" menu. All three graphics windows are destroyed with the status
window.

See Also:

• Introduction to using a Mouse and Windows

Interpreting the Histograms of HPdemoBrd2

Because you selected more than six tests (which appear six per page), the Histogram
Report is two pages long. Each new page appears in a window that overlaps the previous
page, but the windows are offset so that they do not entirely cover each other.

If the page you need to look at is hidden behind a foreground page, you can bring it to the
front by placing the mouse cursor on its exposed edge and then clicking the mouse's left
button.

Each page has a page number at its lower right corner; for example "1|2" means "page 1 of
2".

Each test is in its own section and is ordered by "% passing" first and when that is equivalent
the value in "C.P".

The Histogram Report uses the following symbol conventions in its graphics section:

• Red tick marks indicate the high and low test limits for passing tests; thus,
measurements within the red tick marks are within the specified limits for the test.
 (On the histogram for "R25" these marks represent the position of the values
"1.87K" and "2.148K", which appear on the "Limits" line of the test portion.)
• The green tick mark is the position of the nominal value.
(If the nominal tick mark ever fails to appear, it is because the nominal and the
mean coincide; that is, the mean tick mark is obscuring the nominal tick mark.
Note that it is visible in the histogram for "R10" and "R15", but appears as part of
the x-bar symbol in the rest of the graphs on page 1.)
• Blue tick marks indicate the high and low limits for tests that failed but were not gross
failures.
(The area between the red and blue tick marks is the expected failure range.
That is where measurements for functioning components that were not up to
specifications are expected to be.)
• The mean is shown by a 'barred-X' symbol beneath the graph.
(The mean and nominal almost coincide in the case of "R25". Looking at the
values for the Limits and Mean we can determine that this is probably a 2K
resistor.)
• Green bars represent tests that passed; they were between the high and low test
limits. Green bars appear inside the red tick marks.
(The passing range is divided up into seven equal divisions. The number above
the graph indicates the number of measurements in the highest bar.)
• Red bars represent tests that failed but were within approximately 50% of the high or
low test limits (moderate failures). Red bars appear outside the red tick marks but
inside the blue tick marks.
(Each expected failure range is divided up into four equal divisions. "R25" has
failures on both the high and low ends, "R10" failed only on the low end, while
"R15" and "R20" both failed only on the high end.)
• Blue bars represent grossly failing measurements; they appear outside the blue tick
marks.

(This report has no grossly failing measurements. See the Histogram for HPdemoBrd1
for an example.)

Examining the Histogram Report shows that:

• All the tests consist of 50 measurements (the value on the "Total" line) which is a large
enough sample to be statistically significant (it should be at least 25). This number may
not correspond to the number of board tests run for several reasons. First histograms
are only run on the sampled tests which are determined by the sampling rate selected
in the test plan. This sampling rate should be set so that a statistically significant
number of measurements will be made during the normal interval the histogram report
displays (one day for this report). In this example the sample rate must have been set
to 1.00 (100%) to record every board test. When a board type reaches its normal
production state, the sample rate should be decreased so that the PBQ database does
not grow too fast. A second influence on the number of tests sampled is the form of the
testplan. If the testplan is written such that the failure of some tests prohibit other from
running, some test may not be run every time. Normally tests are run in groups or
sections so that damage is not done as when a short exists on a board.
• The histograms are in order of increasing "% passing" values going down the columns.
For tests with the same "% passing" value they are ordered by the "CPk", coefficient of
 Producibility, value. This produces a worst to best type ordering and shows the largest
problems first. The "% passing" values represented here are only from the sampled
tests even though unsampled failures may exist in the database. Therefore the test
which shows up as the worst on the failure report may not be the worst on the
histogram.
• "R25" has the highest failure rate, which might be caused by:
◦ Noisy measurements, which means that the test needs further enhancements.
◦ Components that are out of tolerance, perhaps due to a faulty batch of
components.
◦ A test whose test limits are too tight, and could be widened with no adverse
effects on the quality or the performance of the board.
◦ An incorrectly specified component (the design is wrong), which means that a
closer tolerance component is needed on this board.
(These speculations are supported by broadness of the graph. The standard
deviation, "Std. Dev.", represents 5% of the nominal value.)
• "R10"'s measurements are skewed low (i.e., the mean value is less than the nominal
value), which might be caused by:
◦ A test where additional guarding is needed because a component is in parallel
with this resistor.
◦ Some kind of production problem.
• "R15"'s measurements are skewed high, which could be caused by:
◦ Source loading, which means that the test requires modification.
◦ (If "R15" is a metal-film resistor) Overheating in the wave solder process, which
is permanently raising the resistor's value.
◦ Some other production problem.
• "R20"'s measurements are centered, but where we would expect the highest bar (the
greatest number of measurements) in the middle at the nominal value (the mean), it is
actually the smallest bar. Perhaps the vendor is removing the one percent resistors
from the batches we receive.

Obviously, any of these possibilities requires further investigation. But the causes are typical
of what you might expect to find in a real production environment.

The remaining resistors ("R5", "R6", and "R7") are good examples of histograms for passing
measurements; that is, all measurements are within the limits, and most of the
measurements are tightly distributed about the mean. Note that the tighter the grouping the
higher the "CPk" value.
>>> When you have finished examining the Histogram Report, select the "Exit" button in
the report status window's "File" menu. This will remove both the graphics and status
windows.

See Also:

• Selecting the Histogram Report for HPdemoBrd2

• Introduction to using a Mouse and Windows
Selecting the Production Report for HPdemoBrd3
Let's begin the investigation by generating a Production Report for HPdemoBrd3.
>>> Select "Production".

(As mentioned earlier, the further you progress in this tutorial, the less detailed the
explanations become. From now on, we expect you to know what steps are required to
select an item in a menu or a screen.)

>>> Use "Agilent PBQ List Selection" to select "HPdemoBrd2" and "HPdemoBrd3".

(Unselect any other board types.)

>>> Select a time interval of "12/01/88" at "00:00" to "12/12/88" at "17:00".

(If you cannot remember how to select the time interval, Select "Help" on the "Agilent
PBQ Time Selection" screen and follow the instructions that appear. If necessary, refer
back to 'Selecting the Production Report for HPdemoBrd2' for a detailed explanation.)

The fields in the main screen should now contain the following values:
Report Type = Production
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
2/4 Boards = HPdemoBrd2, HPdemoBrd3
Report Footer = TUTORIAL database report

>>> Generate the Production Report by selecting "Run Report".

See Also:

• Selecting the Production Report for HPdemoBrd2

• Introduction to using a Mouse and Windows

Interpreting HPdemoBrd3's Production Report

The Production Report shows that:

• 100 HPdemoBrd3's were tested, of which 98 passed during the specified time interval.
• Only 25 of the 100 boards passed the first time they were tested. The other 75 boards
were sent to be repaired.
• At the repair operation, 2 boards were scrapped and the other 73 boards were
successfully repaired. The average repair time was 12.75 minutes.

(Notice that there are places where 'N/A', which means 'Not Applicable,' appears in the
Production Report; for example, "Boards Scrapped" for "HPdemoBrd2" is N/A. N/A
means there was no repair data available for that board type. If you are using the
 optional Agilent PR PLUS Repair Loop Management Software to collect repair data,
your Production Reports will contain N/A whenever repair statistics were not collected.)

(You can remove the production report.)

See Also:

• Selecting the Production Report for HPdemoBrd3

Selecting the Failure Report for HPdemoBrd3

Now that we know something about the history of HPdemoBrd3, let's generate a Failure
Report to see which tests failed.
>>> Select "Failure".
>>> Confirm "Board ..." contains "HPdemoBrd3" - change if necessary.
The fields in the main screen should now contain the following values:
Report Type = Failure
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd3
Report Footer = TUTORIAL database report

>>> Select "Run Report".

See Also:

• Introduction to using a Mouse and Windows

Interpreting HPdemoBrd3's Failure Report

The Failure Report for HPdemoBrd3 shows that:
• The main failure is "SHORTS" (>71%), which tests for both shorts and opens on the
circuit board.
• The remaining test failures are mostly digital devices ("U12", "U9", "U3"), as well as one
analog component ("R7").

(You can remove the failure report.)

See Also:

• Selecting the Failure Report for HPdemoBrd3

• Introduction to using a Mouse and Windows
Selecting the Repair Report for HPdemoBrd3
The Repair Report shows what caused the failures, as determined by the repair operator
(who entered the repair information in Agilent PR PLUS).

Generate a Repair Report for HPdemoBrd3 as follows:

>>> Select "Repair".
>>> Enter the edit field beneath the "2/4 Boards ..." button and delete the line containing
"HPdemoBrd2".

(The label "2/4 Boards ..." will not change until the field is exited. This is accomplished
by selecting another item. To do this press the Tab key.)

The fields in the main screen should now contain the following values:
Report Type = Repair
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
1/4 Boards = HPdemoBrd3
Report Footer = TUTORIAL database report

>>> Select Run Report.

See Also:

• Introduction to using a Mouse and Windows

Interpreting the Repair Report for HPdemoBrd3

The Repair Report shows what caused the shorts and opens failures shown collectively as
SHORTS test failures in the Failure Report we examined earlier.

Among the causes of the SHORTS test failures revealed by the Repair Report are:

• "INSUFFICIENT_SOLDER", which caused the most failures (24).

• "BENT_LEADS", which can touch and cause shorts, caused the second largest
number of failures (12).
• "SOLDER_BRIDGE" caused some shorts.
• Opens caused by solder on the pads but not touching a part's leads, denoted by
"POOR_WETTING_OF_LEAD".
• "TOMBSTONE" parts, which means that one end of an SMT component is lifted (an
open circuit).

Also among the problems shown are:

• Parts are being inserted backwards, "BACKWARD_PART"; for example, digital ICs are
reversed end for end.
• "WRONG_PART"s are being inserted on the boards.
 • Parts missing from the boards appear as "MISSING_PART".

Two boards had "OPEN_TRACE". These are probably the two scrapped boards which were
shown in the Production Report.

(Note that when a cause label is over 16 characters it is printed in two parts. If it is over
32 characters it will show up as a legend index *** number *** and you may need to
scroll the status window to make it appear. This is also true for the test names on the
failure report.)

(You can remove the repair report.)

See Also:

• Selecting the Repair Report for HPdemoBrd3

• Introduction to using a Mouse and Windows

The Database Administrator Scenario

INTRODUCTION

You are the Database Administrator of the Agilent Pushbutton Q-STATS database. If you did
not log in as "qsys" you will not be able to perform the administrative functions, so log out
and log back in as "qsys" and restart the tutorial with this scenario.

WARNING! RUNNING TUTORIALS AS "qsys" MAKES THE SYSTEM COPY THE

TUTORIAL DATABASE FOR SAFE KEEPING. IT WILL BE RESTORED WHEN ANY
PBQ USER EXITS THE SYSTEM. THIS KEEPS THE TUTORIAL DATABASE
INTACT, BUT REQUIRES SOME TIME. IT IS IMPERATIVE THAT NO OTHER PBQ
USERS ARE RUNNING WHILE QSYS IS RUNNING THE TUTORIAL OR POSSIBLE
CORRUPTION OF THE TUTORIAL DATABASE MIGHT OCCUR

THE Agilent Pushbutton Q-STATS INTERFACE MIGHT TAKE SOME TIME TO

BECOME ACTIVE AGAIN. THE MESSAGES ABOUT THE EFFECTS BEING
IRREVERSIBLE DO NOT HOLD.

DELETING OLD DEVICE OR FUNCTIONAL TESTS FROM THE DATABASE

To be able to see the difference that the deletion makes, we need to provide before and after
examples. The "Failure" report on "HPdemoBrd1" will provide us with the comparison.
>>> Set the screen fields to the following:

Report Type = Failure

Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd1
Report Footer = TUTORIAL database report
 >>> Select Run Report.
It will not be necessary for the report to complete before deletion setup can start, although
the active window might change and you will have to be aware of that and correct it when
necessary.

>>> Select one of the "Delete" buttons in the "Agilent Pushbutton Q-STATS" window's "File"
menu.

(Note that several of the buttons' labels are less readable - this indicates that the
button can not be selected. Several fields now contain "N/A" which is short for Not
Applicable. The button previously labeled "Run Report" is now labeled "Delete".)

>>> Select the "Agilent Pushbutton Q-STATS" window's "File" menu.

(Note that now the "Run Report" button is readable - this indicates that the button may
be selected.)

>>> Select the "Run Report" button.

All the fields should now return to the values they had before the delete button was
selected.

>>> Select the "Delete Test &T" button in the "Agilent Pushbutton Q-STATS" window's
"File" menu.

(Note that several of the buttons' labels are less readable - this indicates that the
button can not be selected. Several fields now contain "N/A" which is short for Not
Applicable. The button previously labeled "Run Report" is now labeled "Delete".)

>>> Select the "47/47 Tests ..." button.

>>> Use the pattern matching capabilities to select all test names ending in 3 and return to
the main screen.
The fields in the main screen should now contain the following values:
Report Type = Delete Test
Output Device = N/A
Time Interval (Start) = N/A N/A
Time Interval (End) = N/A N/A
Board = HPdemoBrd1
6/47 Tests = C-T3, Q3, R13, R23, R3, U3
Report Footer = N/A

(Note that several of the selected tests did not show up in the Failure Report. Only U3
and R23 do.)

>>> Select the "Delete" button.

A "Agilent Pushbutton Q-STATS ERROR" window appears with a warning message
indicating what will be deleted if you proceed. Remember when you are running through the
tutorial you can ignore the "Process is irreversible ..." message. Make sure the Failure
Report is complete before executing the next step.
>>> Select the "Delete" button on the error message window.
When the deletion is completed the "Agilent Pushbutton Q-STATS ERROR" window will
disappear and the main interface screen will contain the following:
Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = TUTORIAL database report

We will now re-run the Failure Report to see some of the results of the deletion.
>>> Set the screen fields to the following:

Report Type = Failure

Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
Board = HPdemoBrd1
Report Footer = TUTORIAL database report

>>> Select "Run Report".

You can see that the new pareto only has 5 bars and the one created before deletion had 7
(It included U3 and R23). The new report only lists "Failures: 113" versus the old one's
"Failures: 146" event though they still agree on the numbers by "Boards:" and "Tests:".
Another quick check of the deletion can be made by doing:
>>> Select "Delete Test".
Careful examination of the "Tests ..." button reveals that now there are 41 tests for
"HPdemoBrd1".
>>> Remove the "Agilent PBQ HPdemoBrd1 Test Failures" Reports
To return to a state from which you can run reports do:
>>> Select "Delete".
and when the warning appears:
>>> Select "Cancel".

DELETING OLD DATA FROM A BOARD TYPE -

This is the most commonly used deletion mechanism. It allows the administrator to control
the database size by removing any data which exists before the selected cutoff date.

To be able to see the difference that the deletion makes, we need to provide before and after
examples. The "Production" report on "HPdemoBrd1", "HPdemoBrd2", "HPdemoBrd3", and
"HPdemoBrd4" will provide us with the comparison.
>>> Set the screen fields to the following:

Report Type = Production

Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
 4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = TUTORIAL database report

>>> Select Run Report.

It will not be necessary for the report to complete before deletion setup can start, although
the active window might change and you will have to be aware of that and correct it when
necessary.

>>> Select the "Delete before Date" button in the "Agilent Pushbutton Q-STATS" window's
"File" menu.

(Note that several of the buttons' labels are less readable - this indicates that the
button can not be selected. Several fields now contain "N/A" which is short for
Not Applicable. The button previously labeled "Run Report" is now labeled
"Delete".)
>>> Set the "Time Interval" "End" fields to "12/06/88" "00:00"
The fields in the main screen should now contain the following values:
Report Type = Delete before Date
Output Device = N/A
Time Interval (Start) = N/A N/A
Time Interval (End) = 12/06/88 00:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = N/A

>>> Select the "Delete" button.

A "Agilent Pushbutton Q-STATS ERROR" window appears with a warning message
indicating what will be deleted if you proceed. Remember when you are running through the
tutorial you can ignore the "Process is irreversible ..." message. Make sure the Production
Report is complete before executing the next step.
>>> Select the "Delete" button on the error message window.
When the deletion is completed the "Agilent Pushbutton Q-STATS ERROR" window will
disappear and the main interface screen will contain the following:
Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/06/88 00:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = TUTORIAL database report

We will now re-run the Production Report to see some of the results of the deletion.
>>> Set the "Time Interval" "End" fields to "12/12/88" "17:00"
>>> Set the screen fields to the following:

Report Type = Production

Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = TUTORIAL database report
 >>> Select "Run Report".
You can see that the new production report no longer lists any boards of types HPdemoBrd2
or HPdemoBrd4 tested in the interval. The number of boards of type HPdemoBrd1 and
HPdemoBrd3 have also decreased. You may also note that for the specified time interval
(the second report is now effectively from 12/6/88 to 12/12/88) HPdemoBrd3 performed
better than HPdemoBrd1.
>>> Remove the "Agilent PBQ Production for 4 Board Types" Reports

DELETING OBSOLETE BOARD TYPES -

Since HPdemoBrd4 is no longer in production we will remove it from the database so that it
no longer appears in our Yield and Production Reports showing nothing but zeros and N/A
as entries.

To have a basis of comparison do the following:

>>> Select "Report Type:" "Yield".
If the board type section button does not read as "4/4 Boards ..." do the following:
>>> Select "*/4 Boards ...".
>>> Select "Select All".
>>> Select "OK".
The fields in the main screen should now contain the following values:
Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = TUTORIAL database report

>>> Select Run Report.

After the "board data retrieved" message appears in the report status window -

>>> Select the "Delete Board Type" button in the "Agilent Pushbutton Q-STATS" window's
"File" menu.
The fields in the main screen should now contain the following values:
Report Type = Delete Board Type
Output Device = N/A
Time Interval (Start) = N/A N/A
Time Interval (End) = N/A N/A
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = N/A

>>> Select "HPdemoBrd4".

>>> Select the "Delete" button.

after the warning appears, once again -
>>> Select the "Delete" button.
When the deletion is completed the "Agilent Pushbutton Q-STATS ERROR" window will
disappear and the main interface screen will contain the following:
Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
3/3 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3
Report Footer = TUTORIAL database report

If the board type section button does not read as "3/3 Boards ..." do the following:
>>> Select "*/3 Boards ...".
>>> Select "Select All".
>>> Select "OK".
>>> Select Run Report.
Note that the only difference between the reports is that there is no longer a line for
HPdemoBrd4 (if we had not already delete the data the total lines would have been different
in each report).
>>> Remove the "Agilent PBQ Yield for 4 Board Types" and the "Agilent PBQ Yield for 3
Board Types" Reports
ADDING/CHANGING/DELETING OUTPUT DEVICES RECOGNIZED BY Agilent PBQ -

Select "Administering Output Devices" from the "See Also" list.

See Also:

• Administrating Output Devices

• Introduction to using a Mouse and Windows

Deleting Data of Tests Removed from the Testplan

>>> Select the "Delete Test &T" button in the "Agilent Pushbutton Q-STATS" window's
"File" menu.

(Note that several of the buttons' labels are less readable - this indicates that the
button can not be selected. Several fields now contain "N/A" which is short for Not
Applicable. The button previously labeled "Run Report" is now labeled "Delete".)

>>> Select the "47/47 Tests ..." button.

>>> Use the pattern matching capabilities to select all test names ending in 3 and return to
the main screen.
The fields in the main screen should now contain the following values:
Report Type = Delete Test
Output Device = N/A
Time Interval (Start) = N/A N/A
Time Interval (End) = N/A N/A
Board = HPdemoBrd1
 6/47 Tests = C-T3, Q3, R13, R23, R3, U3
Report Footer = N/A

(Note that several of the selected tests did not show up in the Failure Report. Only U3
and R23 do.)

>>> Select the "Delete" button.

A "Agilent Pushbutton Q-STATS ERROR" window appears with a warning message
indicating what will be deleted if you proceed. Remember when you are running through the
tutorial you can ignore the "Process is irreversible ..." message. Make sure the Failure
Report is complete before executing the next step.
>>> Select the "Delete" button on the error message window.
When the deletion is completed the "Agilent Pushbutton Q-STATS ERROR" window will
disappear and the main interface screen will contain the following:
Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = TUTORIAL database report

Deleting Board Types from the Database

>>> Select the "Delete Board Type" button in the "Agilent Pushbutton Q-STATS" window's
"File" menu.
The fields in the main screen should now contain the following values:
Report Type = Delete Board Type
Output Device = N/A
Time Interval (Start) = N/A N/A
Time Interval (End) = N/A N/A
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = N/A

>>> Select "HPdemoBrd4".

>>> Select the "Delete" button.

after the warning appears, once again -
>>> Select the "Delete" button.
When the deletion is completed the "Agilent Pushbutton Q-STATS ERROR" window will
disappear and the main interface screen will contain the following:
Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/12/88 17:00
3/3 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3
Report Footer = TUTORIAL database report
Deleting Old Data from the Database
>>> Select the "Delete before Date" button in the "Agilent Pushbutton Q-STATS" window's
"File" menu.

(Note that several of the buttons' labels are less readable - this indicates that the
button can not be selected. Several fields now contain "N/A" which is short for
Not Applicable. The button previously labeled "Run Report" is now labeled
"Delete".)
>>> Set the "Time Interval" "End" fields to "12/06/88" "00:00"
The fields in the main screen should now contain the following values:
Report Type = Delete before Date
Output Device = N/A
Time Interval (Start) = N/A N/A
Time Interval (End) = 12/06/88 00:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = N/A

>>> Select the "Delete" button.

A "Agilent Pushbutton Q-STATS ERROR" window appears with a warning message
indicating what will be deleted if you proceed. Remember when you are running through the
tutorial you can ignore the "Process is irreversible ..." message. Make sure the Production
Report is complete before executing the next step.
>>> Select the "Delete" button on the error message window.
When the deletion is completed the "Agilent Pushbutton Q-STATS ERROR" window will
disappear and the main interface screen will contain the following:
Report Type = Yield
Output Device = Screen
Time Interval (Start) = 12/01/88 00:00
Time Interval (End) = 12/06/88 00:00
4/4 Boards = HPdemoBrd1, HPdemoBrd2, HPdemoBrd3, HPdemoBr4
Report Footer = TUTORIAL database report

Restoring Report Mode without Deleting

>>> Select one of the "Delete" buttons in the "Agilent Pushbutton Q-STATS" window's "File"
menu.

(Note that several of the buttons' labels are less readable - this indicates that the
button can not be selected. Several fields now contain "N/A" which is short for Not
Applicable. The button previously labeled "Run Report" is now labeled "Delete".)

>>> Select the "Agilent Pushbutton Q-STATS" window's "File" menu.

(Note that now the "Run Report" button is readable - this indicates that the button may
be selected.)
>>> Select the "Run Report" button.
All the fields should now return to the values they had before the delete button was selected.

Administrating Output Devices

The Database Administrator now has more control over PBQ's output devices. This includes
the ability to restrict PBQ's access as well as allowing PBQ to use new devices as they
become available.
>>> Select the "Set Up Device ..." button in the "Agilent Pushbutton Q-STATS" window's
"File" menu.
The "Agilent PBQ Output Device Administration" window should appear. The main section
consists of three columns. The left column is a list called "System Devices". This is a list of
printer (interface script) names obtained directly from the system. THIS IS REALLY WHAT
THE SYSTEM CURRENTLY SEES AS PRINTER DEVICES VIA THE LPSTAT COMMAND
(see HP-UX Reference Vol 1: Section 1 for information on LPSTAT.) The 3070 is shipped
with four scripts installed: lp, lp1602a, plt7550, and plt. Updated systems might also include
lp3630a, which has been discontinued. These and/or others may be currently available in
the "System Devices" list. This list will include both accepting and non-accepting printers and
therefore might contain more choices than the "Output Device:" menu of the main interface.
(The "Output Device:" menu only contains some of the accepting devices.) Use the HP-UX
system's "sam" program to install and change the status of system devices; you must be
logged on as superuser (root) to run "sam". The only way to add or remove items to this list
is through the use of "sam". Agilent Pushbutton Q-STATS will not recognized a change in the
available system devices until it is restarted.

The second column contains the "Device Prefix" list, which lists the printer classes PBQ will
recognize. Each class instructs PBQ on how to format the data for output. This list probably
contains printer models you do not have and will never get (several that were available when
the 3070 series was introduced are now obsolete). Selecting any "Device Prefix" item will
cause the third column to show the attributes Agilent PBQ uses with this printer type. It will
also cause items in the first column to highlight to show the actual "System Devices" (printer
available to this system) associated with the prefix if such a device exists. Conversely,
selecting an item from the "System Device" list shows, by highlighting, the associated
"Device Prefix" used (if any) and its attributes, along with any other "System Devices" using
the prefix.
>>> Select each item from the "System Devices" list and observe the reaction.
When no item in the "Device Prefix" list highlights for a "System Device" selection, this
device will not be allowed by Agilent PBQ since the format of the output can not be
determined.

Conversely the "Device Prefix" list can be checked -

>>> Select each item from the "Device Prefix" list and observe the reaction.
If no item in the "System Devices" list is highlighted, the formatting attributes of the "Device
Prefix" class do not apply to any of the current system devices.

ADDING A DEVICE PREFIX.

A new item can be added to the "Device Prefix" list with or without having an associated item
in the "System Devices" list. The difference will show up only in the "System Devices" list.
The first scenario is to add a new prefix for an already existing set of attributes.
>>> Select the "lp" item from the "System Devices" list.

(If there is no "lp" item in the "System Devices" list, someone may have already
removed it via "sam" and you will not be able to complete this section. Conversely if
there is a "lp" item in the "Device Prefix" list that became highlighted, it must be
removed with the remove button before you will be able to continue this example. Do
not be alarmed about removing the item since it will automatically be restored upon
exit of the tutorial.)

If this is the device "lp" that is shipped with the system, then we can set the association. First
determine whether "lp" has a Laserjet I, II, or III installed. If it has a Laserjet I or II then -
>>> Select the "lp2686" item from the "Device Prefix" list.
The third column will contain the following:
Associated Prefix = lp2686
Format: = Text & Graphics
Device Type: = B & W PCL
Resolution = 100

Otherwise, if it is a Laserjet III, or not really installed, then -

>>> Select the "lp33449" item from the "Device Prefix" list.
The third column will contain the following:
Associated Prefix = lp33449
Format: = Text & Graphics
Device Type: = B & W PCL
Resolution = 96

(Note that the only difference in attributes of the two possibilities is the "Resolution",
which is the number of dots per inch the printer is set for. This must be correctly set for
graphics output to fill the page correctly.)

Since we want to match the "System Device", "lp", do the following:

>>> Change the "Associated Prefix" to "lp".
>>> Press the "Tab" key to exit the field.
The "System Devices" list should now have "lp" highlighted. If any other items are
highlighted they will also be treated as having the same device attributes as "lp"

(Note that if more than "lp" is highlighted in the "System Devices" list and the other
devices are not the same type, another device prefix needs to be added.)

(Note that there may be other items in the "System Devices" list which start with the
letters lp but are not highlighted. This is because a more complete match exists for
these items in the "Device Prefix" list. That is, they are matched by a longer prefix
item.)

>>> Select the "Add" button.

An item labeled "lp" should now appear in the "Device Prefix" list.

Since the new prefix "lp" encompasses the old prefix "lp2686" or "lp33449" we could remove
one of those after we add "lp" (unless another prefix such as "lp2", "lp26" or "lp268" exists
with different attributes and we used "lp2686"; or a prefix starting with "lp3" exists and we
used "lp33449").

For the extreme case, to uniquely class a system device, you might have to enter the
"System Devices" item label completely in the "Associated Prefix" field. PBQ will always try
to associate the system device with the longest "Device Prefix" available.

The second add scenario is to add a completely new printer type. This requires the system
administrator to know a few things about the printer being added; that information can
usually be obtained from the device's user guide. This includes such things as the device's:

• "Format:" attribute indicates the type of report the output device can accept. One that
is purely text such as the Yield and Production reports or one that is graphical such as
the Failure, Repair, and Histogram reports.
• "Device Type:" attribute indicates the printer language type a graphics device can
handle.
(Note that the printer language "postscript" is not supported)
• "Resolution" attribute indicates the number of dots-per-inch a PCL device uses when
producing its graphics.

(Note: We suggest you use the lowest resolution that you find acceptable since the
higher the resolution the larger the graphics file.)

Assume our new device is an LOVELYCOLOR printer. We must first decide what prefix to
use. (We have purchased three that will be set up on the network as lc1, lc2, and lc3.) So
first we -
>>> Change the "Associated Prefix" to "lc".
This will allow us to access all three, but only have to set up one prefix. From the manual we
find that the new printer will be able to produce both text and graphics. So -
>>> Select the "Text& Graphics" item from the "Format:" menu.
The printer language is HP PCL4 so we then -
>>> Select the "Color PCL" item from the "Device Type:" menu.
Finally we find that the printer can produce output in 103 or 300 dots per inch. For the sake
of memory conservation and network throughput we select the 103 dpi mode and -
>>> Change the "Resolution" to "103".
Since we now have our new prefix completely defined we must -
>>> Select the "Add" button.
We are now informed that the general manager has had "lc2" removed to in his office on the
far side of the plant. We now decide to remove "lc2" from the list of possible printers for
Agilent PBQ. So -
>>> Change the "Associated Prefix" to "lc2".
>>> Select the "None" item from the "Format:" menu.
This should set the attributes to the following
 Associated Prefix = lc2
Format: = N/A
Device Type: = N/A
Resolution = N/A

>>> Select the "Add" button.

This will keep PBQ from using lc2, but still allows access to lc1 and lc3.

REMOVING A DEVICE PREFIX.

Although PBQ does not prohibit the system administrator from removing a prefix that
matches a current system device, the user should be aware that the system will try to match
the device to another prefix. For example, assuming that lp has been entered as an item in
the "Device Prefix" list -
>>> Select the "Device Prefix" "lp".
This should highlight some of the items in the "System Devices" list.
>>> Select an item from the "System Devices" list which starts with the letters lp and has
an associated item in the "Device Prefix" list. (You might try lp3630a or lp1602a)
>>> Select the "Remove" button.

(Note that you are removing the associated "Device Prefix" item and not the "System
Device" item you selected.)

This should force the association of the "System Devices" item with another entry in the
"Device Prefix" list, probably "lp". This also means that PBQ will use the new attributes
associated with the new prefix for the device and results might be significantly different or
completely unusable.

The only danger in removing a prefix that is not currently being used by the system is the
possibility that it will be needed in the future. Otherwise the system administrator should feel
free to remove any unused prefix by:
>>> Select the item from the "System Devices" list to be removed.
>>> Select the "Remove" button.

CHANGING A DEVICE PREFIX.

If the system administrator decides to prohibit PBQ from using a device, he can do so by
changing the "Format:" attribute of the associated prefix. Suppose we wish to change "lp" to
be unusable by PBQ. We would do the following:
>>> Select the "Device Prefix" "lp".
This should set the attributes to the following
Associated Prefix = lp
Format: = Text & Graphics
Device Type: = B & W PCL

>>> Select "None" from the "Format:" menu.

This should set the attributes to the following
 Associated Prefix = lp
Format: = N/A
Device Type: = N/A
Resolution = N/A

This will keep PBQ from using any of the matching system devices, without restricting any of
the other HP3070 applications.

(Note that this change has not really been registered since the "Add" button has not
been pressed.)

The administrator has declared that we do not want graphics reports sent over the network.
This is because they overload the network with files that can require a megabyte per page.
To restrict the types of reports sent to the 'text only' type, we could do the following -
>>> Select the "Device Prefix" "lp".
This should set the attributes to the following
Associated Prefix = lp
Format: = Text & Graphics
Device Type: = B & W PCL
Resolution = 100

>>> Select "Text" from the "Format:" menu.

This should set the attributes to the following
Associated Prefix = lp
Format: = Text
Device Type: = B & W PCL
Resolution = N/A

To register the change -

>>> Select the "Add" button.
This will keep PBQ from using the associated items in the "System Devices" list to produce
graphical reports. The user will be advised of this when trying to select "Run Report" in the
main interface.

To make the change permanent, the user must exit via the "OK" button. Exiting via "Cancel"
will erase any changes made.

(Note that the changes made after "Run Tutorial" was selected will last only until the
session is exited.)

See Also:

• Printer Resolution
• Allowable Formats
• Printer Language Type
• Shipped Prefix Settings
Adding an Item to the Device Prefix List
A new item can be added to the "Device Prefix" list with or without having an associated item
in the "System Devices" list. The difference will show up only in the "System Devices" list.
The first scenario is to add a new prefix for an already existing set of attributes.
>>> Select the "lp" item from the "System Devices" list.

(If there is no "lp" item in the "System Devices" list, someone may have already
removed it via "sam" and you will not be able to complete this section. Conversely if
there is a "lp" item in the "Device Prefix" list that became highlighted, it must be
removed with the remove button before you will be able to continue this example. Do
not be alarmed about removing the item since it will automatically be restored upon
exit of the tutorial.)

If this is the device "lp" that is shipped with the system, then we can set the association. First
determine whether "lp" has a Laserjet I, II, or III installed. If it has a Laserjet I or II then -
>>> Select the "lp2686" item from the "Device Prefix" list.
The third column will contain the following:
Associated Prefix = lp2686
Format: = Text & Graphics
Device Type: = B & W PCL
Resolution = 100

Otherwise, if it is a Laserjet III, or not really installed, then -

>>> Select the "lp33449" item from the "Device Prefix" list.
The third column will contain the following:
Associated Prefix = lp33449
Format: = Text & Graphics
Device Type: = B & W PCL
Resolution = 96

(Note that the only difference in attributes of the two possibilities is the "Resolution",
which is the number of dots per inch the printer is set for. This must be correctly set for
graphics output to fill the page correctly.)

Since we want to match the "System Device", "lp", do the following:

>>> Change the "Associated Prefix" to "lp".
>>> Press the "Tab" key to exit the field.
The "System Devices" list should now have "lp" highlighted. If any other items are
highlighted they will also be treated as having the same device attributes as "lp"

(Note that if more than "lp" is highlighted in the "System Devices" list and the other
devices are not the same type, another device prefix needs to be added.)

(Note that there may be other items in the "System Devices" list which start with the
letters lp but are not highlighted. This is because a more complete match exists for
these items in the "Device Prefix" list. That is, they are matched by a longer prefix
item.)
>>> Select the "Add" button.
An item labeled "lp" should now appear in the "Device Prefix" list.

Since the new prefix "lp" encompasses the old prefix "lp2686" or "lp33449" we could remove
one of those after we add "lp" (unless another prefix such as "lp2", "lp26" or "lp268" exists
with different attributes and we used "lp2686"; or a prefix starting with "lp3" exists and we
used "lp33449").

For the extreme case, to uniquely class a system device, you might have to enter the
"System Devices" item label completely in the "Associated Prefix" field. PBQ will always try
to associate the system device with the longest "Device Prefix" available.

The second add scenario is to add a completely new printer type. This requires the system
administrator to know a few things about the printer being added; that information can
usually be obtained from the device's user guide. This includes such things as the device's:

• "Format:" attribute indicates the type of report the output device can accept. One that
is purely text such as the Yield and Production reports or one that is graphical such as
the Failure, Repair, and Histogram reports.
• "Device Type:" attribute indicates the printer language type a graphics device can
handle.
(Note that the printer language "postscript" is not supported)
• "Resolution" attribute indicates the number of dots-per-inch a PCL device uses when
producing its graphics.

(Note: We suggest you use the lowest resolution that you find acceptable since the
higher the resolution the larger the graphics file.)

Assume our new device is an LOVELYCOLOR printer. We must first decide what prefix to
use. (We have purchased three that will be set up on the network as lc1, lc2, and lc3.) So
first we -
>>> Change the "Associated Prefix" to "lc".
This will allow us to access all three, but only have to set up one prefix. From the manual we
find that the new printer will be able to produce both text and graphics. So -
>>> Select the "Text& Graphics" item from the "Format:" menu.
The printer language is HP PCL4 so we then -
>>> Select the "Color PCL" item from the "Device Type:" menu.
Finally we find that the printer can produce output in 103 or 300 dots per inch. For the sake
of memory conservation and network throughput we select the 103 dpi mode and -
>>> Change the "Resolution" to "103".
Since we now have our new prefix completely defined we must -
>>> Select the "Add" button.
We are now informed that the general manager has had "lc2" removed to in his office on the
far side of the plant. We now decide to remove "lc2" from the list of possible printers for
Agilent PBQ. So -
>>> Change the "Associated Prefix" to "lc2".
>>> Select the "None" item from the "Format:" menu.
This should set the attributes to the following
Associated Prefix = lc2
Format: = N/A
Device Type: = N/A
Resolution = N/A

>>> Select the "Add" button.

This will keep PBQ from using lc2, but still allows access to lc1 and lc3.

See Also:

• Printer Resolution
• Allowable Formats
• Printer Language Type
• Shipped Prefix Settings

Changing the Attributes of a Device Prefix List Item

If the system administrator decides to prohibit PBQ from using a device, he can do so by
changing the "Format:" attribute of the associated prefix. Suppose we wish to change "lp" to
be unusable by PBQ. We would do the following:
>>> Select the "Device Prefix" "lp".
This should set the attributes to the following
Associated Prefix = lp
Format: = Text & Graphics
Device Type: = B & W PCL

>>> Select "None" from the "Format:" menu.

This should set the attributes to the following
Associated Prefix = lp
Format: = N/A
Device Type: = N/A
Resolution = N/A

This will keep PBQ from using any of the matching system devices, without restricting any of
the other HP3070 applications.

(Note that this change has not really been registered since the "Add" button has not
been pressed.)

The administrator has declared that we do not want graphics reports sent over the network.
This is because they overload the network with files that can require a megabyte per page.
To restrict the types of reports sent to the 'text only' type, we could do the following -
>>> Select the "Device Prefix" "lp".
This should set the attributes to the following
Associated Prefix = lp
Format: = Text & Graphics
 Device Type: = B & W PCL
Resolution = 100

>>> Select "Text" from the "Format:" menu.

This should set the attributes to the following
Associated Prefix = lp
Format: = Text
Device Type: = B & W PCL
Resolution = N/A

To register the change -

>>> Select the "Add" button.
This will keep PBQ from using the associated items in the "System Devices" list to produce
graphical reports. The user will be advised of this when trying to select "Run Report" in the
main interface.

See Also:

• Shipped Prefix Settings

Removing a Device Prefix from the List

Although PBQ does not prohibit the system administrator from removing a prefix that
matches a current system device, the user should be aware that the system will try to match
the device to another prefix. For example, assuming that lp has been entered as an item in
the "Device Prefix" list -
>>> Select the "Device Prefix" "lp".
This should highlight some of the items in the "System Devices" list.
>>> Select an item from the "System Devices" list which starts with the letters lp and has
an associated item in the "Device Prefix" list. (You might try lp3630a or lp1602a)
>>> Select the "Remove" button.

(Note that you are removing the associated "Device Prefix" item and not the "System
Device" item you selected.)

This should force the association of the "System Devices" item with another entry in the
"Device Prefix" list, probably "lp". This also means that PBQ will use the new attributes
associated with the new prefix for the device and results might be significantly different or
completely unusable.

The only danger in removing a prefix that is not currently being used by the system is the
possibility that it will be needed in the future. Otherwise the system administrator should feel
free to remove any unused prefix by:
>>> Select the item from the "System Devices" list to be removed.
>>> Select the "Remove" button.
Shipped Prefix Settings
The HP3070 system is shipped with the following defaults for the prefixes:
| | |
Prefix | Format | Device Type | Resolution
---------+-----------------+-------------+------------
lp1602 | Text & Graphics | Color PCL | 180
---------+-----------------+-------------+------------
lp3630 | Text & Graphics | Color PCL | 90
---------+-----------------+-------------+------------
lp2276 | Text & Graphics | B & W PCL | 100
lp2686 | | |
---------+-----------------+-------------+------------
lp2225 | | |
lp2227 | Text & Graphics | B & W PCL | 96
lp2228 | | |
lp33449 | | |
---------+-----------------+-------------+------------
lp2235 | Text & Graphics | B & W PCL | 90
lp2934 | | |
---------+-----------------+-------------+------------
plt7550 | Graphics | HPGL | N/A
---------+-----------------+-------------+------------
plt | N/A (None) | N/A (None) | N/A
---------+-----------------+-------------+------------
dj1200c | Text & Graphics | Color PCL | 300
dj1600c | Text & Graphics | Color PCL | 300

This list shows the relationship of the shipped prefix (HP model number) to the model name.
lp1602 - HP Paintjet XL
lp2225 - HP Thinkjet
lp2227 - HP Quietjet
lp2228 - HP Quietjet
lp2235 - HP Ruggedwriter
lp2276 - HP Deskjet
lp2686 - HP Laserjet I & II
lp2934 - HP line printer
lp33449 - HP Laserjet III
lp3630 - HP Paintjet
plt - non-spooling plotter
plt7550 - HP spooling Graphics plotter
dj1200c - HP DeskJet 1200 C
dj1600c - HP DeskJet 1600 C

PBQ does not explicitly support certain use models, such as sending Postscript to a
Postscript printer, using European-standard sizes of paper, sending HPGL as well as PCL to
a printer that can automatically handle either format (such as the DeskJets), nor using every
other printer you might happen to try to use. However, you can usually find a way to provide
such support. For example, for paper of a different size, you can alter the Resolution so that
the image fits within the paper. You can also provide multiple lp spooling device entries for a
single printer device, each having its own properties. You could even provide your own
HPGL-to-Postscript converter program and have it automatically invoked by an lp spooler
script. The corresponding lp device could be configured in PBQ as a plotter so that PBQ
would send HPGL to it, even though the printer itself would be receiving postscript from your
converter program. Unfortunately, such customizations can become complicated and are
therefore not supportable by Hewlett-Packard.