optimal testing: data management for enhanced quality, reliability and yield improvement
TRANSCRIPT
1 10 July 2012
“Optimal Testing”
Data Management for
Highest Yields & Best Quality
‘OptimalTest: The information backbone of the Semiconductor & Electronics supply chain’
Presented at… By Dan Glotter, CEO
OptimalTest
2 10 July 2012
●OptimalTest has established a ~90% presence at the Fabless supply chain which consists of all the major:
● Subcontractors sites
● Foundry sites
● Enabling extremely fast deployment for a new Fabless company
OptimalTest has paved the Semiconuctor Information Highway
In 2011: Over 10M units run on OT per day…
~4B per year…. and growing
3 10 July 2012
• Was a “Back-End Operation”: Necessary evil to separate good from bad
• Then, became “Rear-View Mirror”: Where was the Fab positioned?
• Modern, “World-Class Test” Needs to support the escalating challenges of the Sub 28nm processes.
• Test today is the Central Hub Within a sophisticated supply chain and supports enhancements in: Design, Fab Process, Yields, Quality, Reliability, Operations, TTM & NPI
Device Test: Rapidly Changing Role
4 10 July 2012
• It’s all about the real-time data... It’s all about the off-line data... It’s all about ALL the data… Capture, Store, Move, Mine, Stack, Analyze AND Learn From
The Keystone : Leveraging Data
• Its all about Data Exchange Its all about Data Feed Forward (DFF) & Data Feed Backward (DFB) It must occur seamlessly across the enterprise
8/13/201
2
4
5 10 July 2012
• OT-Detect automatically tracks after ALL your products for ANY changes in BASELINE production
• OT-Detect creates on the fly hundreds of BASELINES for your products' yields, bins, soft-bins, failing parameters and more.
Leveraging Massive Data: OT-Detect
Lot level
Analysis Prod level
Analysis
Bin level
Analysis
Param. level
Analysis
Equip level
Analysis
Facility level
Analysis
Note: The ability to do so many baselines on the fly is a huge technological breakthrough since it requires optimized algorithms to enable super fast computations
What is a “Baseline”?
Preforming a “baseline” means that the system automatically identify the incoming product, scan the last 20-40 lots of that product and determine if the current lot signature significantly exceeds the value of the historical baseline that was created on the fly - Either in Real-Time or in Off-line
• And that's not all, once triggered it provides a step-by-step ROOT-CAUSE ANALYSIS. This means that any extreme change in the products' manufacturing, test or assembly processes will be tracked, captured and assessed.
6 10 July 2012
Correlate results across multiple operations based on:
• Mapping E-Test sites with WS dies locations
• Use die ID / Unit-Level Traceability to track units into FT, SLT & Board test
Combine all the above into cross operational unit level analysis and single view graphics based on Wafer Map Reconstruction
Leveraging Massive Data: Cross Operations
E-Test 1 E-Test 2
WS 1 WS 2
Wafer Map Reconstruction
7 10 July 2012
Reference Die/Units - For Quality Control
• Few units that are tested first with augmented test program for:
• Quality Assurance
• Health Monitor
• Reliability screens
• Additionally, the augmented testing on the reference dies/units creates BASELINE for yield learning without Sacrificing overall test-times
Note-1: in wafer-sort it’s a few dies and if unit ID exist then its in Final-Test the units are tested with augmented TP too
Note-2: The ability to augment the testing in real-time on the fly without touching the test-program I s powered by OptimalTest’s TTR engine in reverse
All w
afe
rs, A
ll lots
, All P
RO
DU
CT
8 10 July 2012
1. Next to E-test structures
2. Within wafer Geographies
Reference Dies Selection
Ring 1
Ring 2
Ring 3
Upper Left Corner
Lower Right Corner
+ Etest Structure Litho:
3. Within yield failure areas
4. Within lithography locations
Testing content augmentation may vary for different baseline dies
9 10 July 2012
Summary
“Hop on” OptimalTest’s Semiconductor Information Highway… for measurable results
• Turn piles of data into actionable data
• Use OT-Detect to catch rare & extreme quality issues
• Use cross operations data – all across from EDABoard
• Use Reference dies/unit to enhance your overall Quality & Yield
• Past: Test Was A Back-End-Of-Line Operation “Disconnected Activity”
• Future: Test is the central hub for Product Optimization