Page 1 Material Created By: Jeremy Pemberton-Pigott • Technical Marketing Engineer for AXI (Singapore)
Page 1
Material Created By:
Jeremy Pemberton-Pigott
• Technical Marketing Engineer for AXI (Singapore)
Page 2
Agenda
1. Software release 8.50 preview
2. User Contributed Tools
• Build Operator Menu Demo
• codeNexus – Defect Reporter Demo
3. Q & A
Page 3
Agilent Medalist 5DXv8.50 Software Release
Preview
Page 4
8.5 Software Release
• Release Content
– Algorithm Enhancements
• Connector Joint Type
– Modifications for very heavily shaded components by using gray level analysis.
• Chip Joint Type
– T-Shirt Cap updates to reduce false failures.
• FET (Field Effect Transistor) Joint Type
– Allows testing of FET type components for Open / Insufficient / Voiding.
– Uses a special Voiding algorithm that compares a learned image vs. the currently inspected joint. This is superior to the Paste – Voiding algorithm.
– Defect Enhancements
• Critical / high priority defects as well as some minor new enhancements.
Page 5
8.5 Software Release
• Targeted Release Date: August – September 2008
Page 6
Connector SPC Algorithm Enhancements
Page 7
Connector SPC Algorithm Enhancements
• Based on the standard Connector joint type.
• Special features in the SPC algorithm allow for the inspection of heavily shaded solder joints, where other methods would normally fail. It is based on gray level analysis instead of just solder thicknesses to allow analysis of joints that may not even be visible by the naked eye (without image enhancement).
• Sample application: Ventura connectors.
– The 5DX is currently the only automated test system in world, that can detect opens on the bridged-ground pins, for this connector. ICT and AOI are unable to detect opens on the bridged-ground pins in the Ventura connectors, due to pin access limitations.
• The enhanced Connector joint type is specifically designed to test for Open / Insufficient solder joints on these heavily shaded types of components.
Page 8
Sample Ventura Connector (female)
Large amount of metal in the connector causes
shading issues for x-ray
Page 9
Ventura Connector Pins and PCB
Replaceable wafer
Bridged
Ground Pins
Signal
Pin
Damaged
Pins
SignalPins
Bridged Ground Pins
Page 10
Sample Ventura Connector Defect Image
• Image is heavily shaded.
• Maximum camera brightness and integrations are used.
• Special averaging technique used to further reduce noise.
• The algorithm’s special features
allow accurate measurement of
the resultant image for insufficient
and open defects.
Fail O
pen
Fail O
pen
Fail S
hort
Pass S
hort
2 Bridged-Ground Rows
Page 11
Good Solder Joint vs. Bad Open Solder Joint Comparison
Good Solder Joints:
- Heel, toe, and side fillets well defined.
- Consistent fillet length.
Suspect Solder Joints:
- Heel, toe, and side fillets not well defined.
- Inconsistent fillet length.
Page 12
Good Solder Joint vs. Bad Cold Solder Joint
BadGood
Page 13
Good Solder Joint Profile
Profile of a good solder joint includes the following:
- Well defined heel / toe fillets.
- Consistent fillet length.
- Steep heel and toe fillet slopes.
Page 14
Open Solder Joint Profile
Profile of an open solder joint includes the following:
- No heel / toe fillets, raised center.
- Inconsistent fillet length.
- Shallow heel and toe fillet slopes.
Page 15
Chip Enhancements for T-Shirt Capacitors
Page 16
Side Fillets of Chip T-Shirt Capacitors
Adds capability for discrete devices with solder fillets on the sides of the pin not just on the end of the pin.
Enhancements to the Chip joint type provide detection of:
- Side fillet insufficients
- Side fillet opens
- Misalignments
- Partially lifted bodies
- Lower false calls than previous releasesSide fillets with solder
Side fillets with solder
Body of package
Page 17
New Chip Open Test
Uses Maximum Pad Thickness Delta
• Creates a ratio of pad 1 thickness over pad 2 thickness.
• Joints are failed when they exceed this ratio:
Pad 1 test = (Pad 1 / Pad 2) * 100
Pad 2 test = (Pad 2 / Pad 1) * 100
• Note
– This test will run on either T-Shirt Cap or on normal Cap packages.
– For T-Shirt Cap packages, the average thickness is taken from the T-Shirt search regions
– For normal packages, the average thickness is taken from the pad 1
and pad 2 regions of interest.
Page 18
Side Fillets
Enabling the Side Fillet testing
• Turns on new regions of interest that generate a profile across a joint.
Pad Region Width
• Controls the width of the region of interest.
Page 19
IdealAcceptable
Reject
Sample T-Shirt Caps
• Cap is between the BGA pins
Page 20
Sample T-Shirt Caps
OpenInsufficient
Page 21
The New FET Joint Type
Page 22
Why a new joint type?
• We required a new family of algorithms to improve the detection of defects in non-standard joint shapes.
• Some example defect conditions:
– Insufficient solder and no solder
– Un-reflowed solder or cold solder
– Excessive voiding in the solder joint
• Sample applications: PolarPAKTM and DirectFETTM devices.
– The FET joint type can be used on any FET type device but it is not restricted to only these devices and can also be used to analyze many other oddly shaped joints.
Page 23
Sample PolarPAK™ FET device
Page 24
Sample FET Layout
Drain
Land Pattern layout- Single large pad for source pin
Gate
Source
Drain
Page 25
FET Overview – The SPC Algorithm
Along Pad Profile
Measures changes in slopes along the region:
– Generates a 2nd derivative profile (which identifies a change in slopes).
– Changes are summed.
– Finds Opens / Insufficient.
• Gap Along Measurements
• Activates 2 search regions which identifies & measures ‘gaps’ in the solder.
• Gap Across Measurements
• Same as along, but used for going across the joint.
• Span Size
• Number of pixels to smooth the joint profiles with. Eliminates noise from the
joint profile to allow for more accurate insufficients testing.
Page 26
FET Overview – SPC Algorithm
Gap Across
Gap Along
Page 27
Sample FET DefectsFirst Device Type
Insufficient solder on Source / Drain / Gate pins
Large Gap
Page 28
Sample FET DefectsSecond Device Type
Insufficient solder on Drain / Gate pins
Page 29
Sample Insufficient Defects (both types of devices)
Page 30
FET Overview – Voiding Algorithm
The Voiding algorithm uses a special technique of comparing a known, previously learned, good image to the actual joint.
The previously learned image is called an ‘Expected Image’.
By analyzing the differences, between the images, the voids in the joint can be found – no matter what shape the joint is.
You do have to take some care not to learn too many voids or
defects into the ‘Expected Image’.
Page 31
FET Voiding – Expected Image Analysis
How it works?
• The FET Voiding algorithm learns a set of images which are used to create an overall ‘Expected Image’ that does not contain any voids.
1. The region of each joint that contains solder is identified.
2. Within the soldered region, each pixel is compared to it’s corresponding pixel in the ‘Expected Image’.
3. Pixels that fall into a void will have a lower gray level value than the current
‘Expected Image’.
4. A minimum area of pixels is required to count the area as being a void.
5. The pixels in each area that are identified as being a void are counted and
compared to the total number of pixels within the soldered region.
6. The total percentage of voiding is then used to pass or fail the joint.
Page 32
How Expected Image Works (visually)?
Expected Image Pin 1
Component 1
Pin 1
Component 2
Pin 1
Component 3
Pin 1
Expected Image Pin 2
Component 1
Pin 2
Component 2
Pin 2
Component 3
Pin 2
Set 1 Set 2
+
+
+
+
+ +
Page 33
How Expected Image Works (visually)?
Calculated voids are marked in light blue(small voids omitted)
Pin to be analyzed
for voiding
The
‘Expected Image’−
> Threshold?
Page 34
Sample FET Voiding
In the next slides we will get a representative ‘step by step’approach to how the machine looks for the voids and calculates
the voiding area of only the ‘Region of Interest’.
Total Un-Masked
Area of the ‘Expected Image’:
48,413 pixels
After processing:Voided Area (Un-Masked): 1,341 pixelsVoided Area (Un-Masked): 2.76%
Joint image to be analyzed
Mask created by setting a threshold using the
‘Expected Image’
Total Pad Area: 86,072 pixelsThis is how the CAD is setup
Total Un-Masked
Area of the ‘Expected Image’:
48,413 pixels
After processing:Voided Area (Un-Masked): 26,074 pixelsVoided Area (Un-Masked): 53.86%
Joint image to be analyzed
Mask created by setting a threshold using the
‘Expected Image’
Total Pad Area: 86,072 pixelsThis is how the CAD is setup
Page 37
Sample FET Voiding
Expected learned image
of the soldered region.Void identification made from the
comparison to the expected image.
Areas excluded
(masked) from voiding
the analysis
Page 38
Sample FET Voiding
Expected learned image
of the soldered region.Void identification made from the
comparison to the expected image.
Areas excluded from
voiding analysis
Page 39
Defect Fixes
– Issues reported by customers from v8.30 – v8.42 are being incorporated
into v8.50.
– Our focus is on customer reported critical and high priority defects.
There are over 35 fixes to the new release of software.
– Fixes and enhancements have just completed beta testing in Asia and the Americas. Final fixes are being implemented and tested before going out for release worldwide.
TD3
Slide 39
TD3 For the comment "list of fixes not available today" - I would replace this with our focus is on Customer Reported Critical/High defects. There are some defects already fixed and waiting on the 8.5 branch. George David should be able to give you this list, however it is really small, so it may not be worth mentioning here.Terri Devlin, 11/29/2007
Page 40
Defect Fixes
Short list of fixes:
• Test Link:
• Changing the family of a part resets all components’ test status
• 5DX / TDW:
• Excess algorithm added to the PressFitConnector joint type
• Short threshold “Test region” is not displaying the correct information when first displayed
• Review Measurements “Default Configuration” function is incorrect
• 5DX:
• AutoUI compile menu is limited to 500 programs
• Alignment does not fail all the time for a bad alignment point
• Duplicate barcodes not checked across multiple boards on a single panel
• Manual board alignment on multi-board panel allows wrong alignment points
• Review Defects BGA pin # mismatch between the review defect image and list
• Panel with x-out board last has no res/log file saved/sent to the ITF server
• SMEMA outputs incorrectly set in Automatic Review mode
Page 41
Defect Fixes
Short list of fixes continued:
• 5DX
• Software hangs when aborting manual alignment
• Review Defects shows the wrong slice
• Software reloads previous program during align & map
• Topside Reference Point from wrong panel is used
• Align and Map mis-focused images: Surface Map Alignment Views not working
• Aligning on two-row devices or single caps or resistors fails to work
• Automatic Reporter re-sends all of the test results to the ITF server upon logging into Windows
• 5DX Crashes:
• serialNumberListSelectionChanged error
• loadPanelAsynchronously error
• setSingleFamilyAllAlgorithms error
• setAlignmentTimeInMilliSeconds error
• getTopSideMeasuredZHeightNanoMeters error
• getBottomSideMeasuredZHeightNanoMeters error
Page 42
End of 8.50 Preview
Page 43
Questions?
Page 44
User Contributed Disclaimer
• Since this software is user contributed software it may be outdated or unmaintained in future releases.
• You may use it at your own risk.
• We are not responsible for any damages incurred from the use of this software.
• This software is offered as a free service to our customers.
• We take no responsibility for the infringement of any proprietary rights or of any export or security restrictions caused by the distribution or of the items contained in it.
• In addition, we may not have reviewed or modified these items and do not make any claims as to the suitability or fitness of this software for any purpose.
Page 45
Build Operator Menu Demo
• Multiple menu & user configuration control for the 5DX AutoUI
Page 46
BLDOPMNU DOS Command
The “Build Operator Menu” software replaces and extends the functionality found in the BLDOPMNU DOS command.
BLDOPMNU DOS Program
Page 47
Build Operator MenuFeatures
• Same features as the BLDOPMNU DOS command for enabling / disabling menu items in a single menu setup (default mode).
• Extends this functionality to allow for different menu items to be configured (enabled / disabled) by the group that a user belongs to.
• Security features like password protection for each user as well as generating security logs. Security logs track program startup, shutdown, user login/out, and wrong password entry attempts.
• Ability to run from a single network share directory and setup different users on different 5DXs or run in different configuration modes.
• Ability to set different ‘S’ user passwords while locking the ‘S’ user out from the Auto UI. This ensures that there is an audit trail for the ‘S’ user, if configured.
Page 48
Build Operator Menu2 Modes of Operation
• Single User Menu Mode
– Runs in tandem with the 5DX Auto UI interface.
– Does not change any system settings.
– Users can optionally log into the Build Operator Menu software for tracking purposes, user name is automatically entered into the Auto UI.
• Multiple User Menu Mode
– Runs in tandem with the 5DX Auto UI interface.
– Changes the system settings so that all users must log into the Auto UI through the Build Operator Menu interface.
– Allows the disabling of different menu items for different users depending on the group that they belong to.
– Optionally locks out the ‘S’ user if it is added to the ‘SUPERVISOR’group.
Page 49
Build Operator MenuHow It Works – Log in
Screen shot of the 5DX Auto UI interface and the Build Operator Menu about to enter a user name into Auto UI.
Page 50
Build Operator MenuHow It Works – Log in
Screen shot of the 5DX Auto UI interface and the Build Operator Menu after entering a user name into Auto UI.
1
2
Page 51
Build Operator MenuHow It Works – Log out
Log out of the 5DX before trying to log out of the Build Operator Menu software.
1
2
Page 52
Build Operator MenuMultiple User Menu Mode – Configure Users and Menus
• Double click a group or user to configure a group’s menu access in the AutoUI.
• The F2 key will enable / disable the menu item.
Groups and users.
Configured menu items
Page 53
Build Operator Menu InterfaceFile Menu Items – Multiple User Menu Mode
Enable ‘Single User Menu’ mode features
Set Administrator password for access to the
system’s configuration screen
View the security audit log file
Page 54
End of Build Operator Menu Demo
Page 55
What is the codeNexus – Defect Reporter?
A ‘User Contributed’ tool for the 5DX that replaces functionality found in:
• Defect Reporter (DOS based)
• Some Review Measurements graphing (DOS based)
Page 56
What does the codeNexus – Defect Reporter do?
It is a graphical user interface of the DOS Defect Reporter.
It provides some basic graphing / exporting functionality:
• Different graph types for the test results
• A ‘Defects Only’ view of the test results for production tuning
• Measurement filtering and graphing
• Exporting filtered defect / measurement data to CSV
• Basic chart printing, zooming, and copying functions
Page 57
What does the interface look like?
Select a result file from the list to view the results
Choose any result folderSwitch viewing mode
Details of the selected result file
Results List View
Optionally: Select any results file from
the graph to view the test results
Auto. refresh test results Remotely monitor results on a 5DXResults Plot View
Grouping by the selected attribute
Export Defects to CSV
Switch viewing mode
Defect Results List View
Load a graph of the data
Defect Results Plot View
Chart failures grouped by
‘Joint Type - Subtype’
(#) of failures in
‘Joint Type - Subtype’
Measurement filters to filter any measurement
Plot multiple slices at once
Measurements Filtering / Charting View
Export filtered data to CSV
Defective joints always
show up no matter
which measurement
filter is selected
Show all or only defective measurements
Menu for Copy / Print / Save Picture
Zoom In / Out
Mean+/- 1σ
+/- 3σ
Passing Range
UtilizationTest result chart for selected program name
Click the program name to view test result chart data for it
Machine ‘Operation Summary Report’
Page 65
End of codeNexus – Defect Reporter Demo
Page 66
FAQ on User Contributed Software
What does it cost to buy?
• Nothing.
What versions of 5DX software does it work on?
• All versions from v7.xx to v8.xx
What versions of Microsoft Windows does it work on?
• All versions of Windows NT to Windows XP
– Provided .NET 2.0 can be installed on it
Page 67
EUP (E-mail Update Program)
Please subscribe to the Agilent E-mail Update Program to receive the latest notices of
software updates / articles / application notes and other monthly technical content.
To subscribe by e-mail or RSS simply go to:
• www.agilent.com/find/axi and click on one of subscription links
Page 68
Questions?