Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK...
Transcript of Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK...
![Page 1: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/1.jpg)
DTEKTM
Quantitative Optical Inspection
Presented to ERAI Executive Conference, Las Vegas, NV
Leonard Nelson
May 18, 2012
Presented to ERAI Executive Conference, Las Vegas, NV
![Page 2: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/2.jpg)
• Story of two plane crashes
• Quick intro to DTEK surface analysis
• An important change in our design approach
• Gaps addressed by DTEK
Agenda
• Validation of the method
• Questions posed by ERAI members
2
![Page 3: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/3.jpg)
US Airways Flight 1549
• January 15, 2009
3
![Page 4: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/4.jpg)
DTEK Offering Summary
• Non-destructive tool to aid
inspection of electronic components
through surface analysis
• Cost: $15K / year subscription
+ implementation+ implementation
– Ongoing part updates
– Customized reporting
– Algorithm improvements
• Time: 15 seconds per component tested,
under 5 min lot cycle time for trained user
4
DTEK 2.1 System
![Page 5: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/5.jpg)
Anti-Counterfeit Technology Background
• Supply chain security technology funded by the US Army Research Office
by ChromoLogic LLC
• ChromoLogic is a diversified research & development firm base in
Pasadena, CA. Covisus is a wholly owned subsidiary of ChromoLogic LLC
– Capabilities in optics, information processing,
diagnostic tools, and spectroscopy
– 30 employees, 12 PhD Scientists
5Section 1: The Counterfeit Problem & Resurfacing
“Counterfeiting, theft and diversion of military equipment are
significant issues within the Army especially during times of
warfare. A rapid, unambiguous tool for identifying a variety of
materiel for both military and civilian locations in which
verification of the identity of that materiel is critical.”
- US Army Research Office
![Page 6: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/6.jpg)
Early Research in Counterfeit Mitigation
Secure Tracking Tags
Intrinsic Surface Traceability
6
Comparative Surface Analysis
![Page 7: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/7.jpg)
Product Progression
Version
TRL Level
Date
Version 0.2
TRL 6
Dec 2010
DTEK 1.1
TRL 7
July 2011
DTEK 2.0
TRL 7
Dec. 2011
DTEK 2.1
TRL 7/8
May 2012
Hardware
7
Data Entry Time
(min per lot)6 minutes 5 minute
15 seconds with
barcode integration
15 seconds with
barcode or instant
(0 sec) with ERP
integration
Scan & Analysis
Time (min per
component)
4 minutes 1 minute
Per component:
30 seconds full batch
time, < 1s scan time
Per component:
20 seconds full batch
time, < 1s scan time
Key Features
Field validation on
“blacktopped”
counterfeits
Enhanced precision,
drastically enhanced
scan speed.
Custom glancing angle
illumination, barcode
integration, five
component tray, ESD
safety.
Modular design, ESD
enhancements,
increased precision.
API for ERP integration
in development
![Page 8: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/8.jpg)
• Original Goal: Remove the human element from the inspection process
– Fool-proof tool
– Reject or accept part based on result
DTEK Design Change
• Lesson Learned: The human element will likely never be removed
• Current Goal: To augment and enable the visual inspector
– Efficient, effective, adaptable
8
![Page 9: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/9.jpg)
• Old paradigm:
– Sales organization with sufficient quality systems
– Some tradeoff between throughput and quality tolerated
• New paradigm:
– Independent distributors as hybrid test labs
Observed Industry Shift
– Independent distributors as hybrid test labs
– Demand zero tradeoff between throughput and quality
– Yet: same demands for short lead times, cost, throughput, and availability
• Question:
– What happens when information, documentation, and testing requests overwhelm your people and systems?
9
![Page 10: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/10.jpg)
• February 12, 2009
Continental / Colgan Flight 3407
10
![Page 11: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/11.jpg)
• Largest investment in quality
• Highest “ROI” for counterfeit mitigation
• Counterfeit mitigation = reputation
• Today, visual inspectors are increasingly critical:
Visual Inspectors
– MUST be efficient information conduits between
suppliers, salespeople, management, customers,
regulatory/certification bodies, and other
distributors
11
![Page 12: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/12.jpg)
• Conscientiousness
• Company culture (Pressure)
• Confirmatory bias
• Human perception bias
• Illusion of control
Visual Inspectors: Weaknesses
• Illusion of control
• Learning plateau
12
![Page 13: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/13.jpg)
Human Perception Bias
Human beings have a visual perception bias against random patterns.1,2
1. Characterizing the Limits of Human Visual Awareness. Huang, L. Science. Vol. 317 no. 5839 (2007)
2. Visual Perception of Texture. Landy, M. New York University. 2002 (source of image above)
13
![Page 14: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/14.jpg)
Human Perception Bias, Continued
14
1960s Today
![Page 15: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/15.jpg)
• Approximately 80% of identified counterfeit components have altered surfaces1
• Remarking or resurfacing is accomplished by counterfeiters through variety of different techniques include, but are not limited to:
– Blacktopping: Painting the surface of the component with a color matching the component packaging and then adding new part markings. The surface may be sanded prior to blacktopping to remove the old margins.
– Epoxy coatings: The surface is coated with a
Random Patterns are Important Because Component Surfaces are Important
– Epoxy coatings: The surface is coated with a compound resembling the original mold compound with a similar chemical composition prior to remarking
– Microblasting: A micro-etching tool is used to remove the part markings and superficial surface layers of the component prior to remarking.
– Lapping: A machine or manual flat lapping tool is used to resurface the top of a component, typicallyon ceramic ICs.
1Defense Industrial Base Assessment: Counterfeit Electronics, January 2010, US Department of Commerce Bureau of Industry and Security
Microblasting can be used to alter
component surfaces
15
![Page 16: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/16.jpg)
Component Set 2: SEM InspectionXilinx XC3030A
Authentic
FEI QUANTA 600F
Top BottomBottom Top
Authentic
1000XEpoxy-Coated Counterfeits
1000X
16Section 2: Quantitative Optical Inspection - Introduction
![Page 17: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/17.jpg)
DTEK Inspection: Xilinx XC3030A
DTEK Top Surface Quantitative Comparison: XC3030A Test Lot (Counterfeit) vs. Reference Sample (Authentic)
Counterfeit Authentic
Plot shows clear difference between re-marked counterfeit (red)
and original surface (green) values.
17Section 2: Quantitative Optical Inspection - Introduction
![Page 18: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/18.jpg)
• Tells inspector which components have the:
– Least consistent surfaces in the sample
– Most consistent surfaces in the sample
Lot Conformance
18
![Page 19: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/19.jpg)
• Counterfeit techniques
always changing
• New analysis techniques
Adaptive Learning & Updates
The Learning Plateau
• New analysis techniques
present themselves
• New visualization tools
and methods deployed
19
![Page 20: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/20.jpg)
Additional Questions
• Does QSA / DTEK replace other analytical
techniques?
• If a “golden sample” needed in order to gain usable
results?
20
• What types of components that can be screened?
• What are the required or recommended lot sizes
needed for analysis?
• Costs associated with buying, using and maintaining
the DTEK system
![Page 21: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/21.jpg)
Additional Questions
• Does QSA / DTEK replace other analytical
techniques?
• If a “golden sample” needed in order to gain usable
results?
21
• What types of components that can be screened?
• What are the required or recommended lot sizes
needed for analysis?
• Costs associated with buying, using and maintaining
the DTEK system
![Page 22: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/22.jpg)
Additional Questions
• Do lot-to-lot variation or different manufacturing
facilities can cause “false positive” results?
• What type of training needed to operate the
equipment and analyze results?
22
equipment and analyze results?
• What are the safety considerations for using DTEK
technology?
![Page 23: Presented to ERAI Executive Conference, Las Vegas, NV Session 2/QuantitativeInspecti… · DTEK Inspection: Xilinx XC3030A DTEK Top Surface Quantitative Comparison: XC3030A Test Lot](https://reader036.fdocuments.in/reader036/viewer/2022063010/5fc37da1f8b19d73cc36083d/html5/thumbnails/23.jpg)
Thank You & Contact Information
Contact:
Covisus Corporation
180 N. Vinedo Ave.
Pasadena, CA 91107
www.covisus.com
Leonard Nelson
DTEK Design Team (from left): Skylar Gauss, Lawrence
Yu, Jordan Crede, Naresh Menon, PhD, Dan Reiley,
PhD, Masha Belyi, Leonard Nelson, Orin Serviss,
Andrew Dyer. Not pictured – Theresa Nguyen,
George Kryboyan, Greg Bearman, PhD
23
Leonard Nelson
626.372.9791
Skylar Gauss
(805) 452-7025