EE382V: System-on-a-Chip (SoC) Design

2y ago
15 Views
2 Downloads
769.08 KB
22 Pages
Last View : 14d ago
Last Download : 3m ago
Upload by : Roy Essex
Transcription

EE382V: System-on-Chip (SoC) DesignLecture 21EE382V:System-on-a-Chip (SoC) DesignLecture 21 – SoC TestingSources:Jacob A. AbrahamAndreas GerstlauerElectrical and Computer EngineeringUniversity of Texas at Austingerstl@ece.utexas.eduLecture 21: Outline SoC Manufacturing Test The testing problem SoC testing costs Design for Test (DFT) SoC Testability Features Boundary Scan P1500 standard Built-In Self Test Functional Test Access Mechanism (TAM)EE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A. Gerstlauer21

EE382V: System-on-Chip (SoC) DesignLecture 21Cost : Cents / 10,000 TransistorsThe Manufacturing Test Problem1000.00IC MfgCost100.0010.001.00Cost ofTestMixed Signal0.10Digital0.011982 1985 1988 1991 1994 1997 2000 2003 2006 2009EE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham3Partitioning for SoC Test Partition according to test methodology: Logic blocks Memory blocks Analog blocks Provide test access: Boundary scan Analog test bus Provide test-wrappers for cores Design for Test (DFT)EE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham42

EE382V: System-on-Chip (SoC) DesignLecture 21DFT Architecture for ModuleNwrapperTestsinkTestTestFunctionalinputsUser defined test access mechanism (TAM)wrapperTestsourceInstruction register controlTDOTRSTTMSTDISOC inputsTCKTest access port (TAP)Serial instruction dataSOC outputsSource: Bushnell and AgrawalEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham5Scan Convert each flip-flop to a scan register Only costs one extra multiplexerNormal mode: flip-flops behave as usualScan mode: flip-flops behave as shift register Contents of flopsSCANcan be scannedSIDout and newvalues scanned inCLKFlop 2V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham63

EE382V: System-on-Chip (SoC) DesignLecture 21Boundary Scan Testing boards is also difficult Need to verify solder joints are good– Drive a pin to 0, then to 1– Check that all connected pins get the values Through-hold boards used “bed of nails” SMT and BGA boards cannot easily contact pins Build capability of observing and controlling pins intoeach chip to make board test easierEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham7Boundary Scan ExamplePackageInterconnectCHIP BCHIP CSerial Data OutCHIP ACHIP DIOpad andBoundary ScanCellSerial Data InEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham84

EE382V: System-on-Chip (SoC) DesignLecture 21Boundary Scan (IEEE 1149.1, JTAG)EE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham9Boundary Scan Interface Boundary scan is accessed through five pins TCK:test clock TMS:test mode select TDI:test data in TDO:test data out TRST*: test reset (optional) Chips with internal scan chains can access the chainsthrough boundary scan for unified test strategy.EE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham105

EE382V: System-on-Chip (SoC) DesignLecture 21Additional DFT Components Test source: Provides test vectors via on-chip LFSR,counter, ROM, or off-chip ATE. Test sink: Provides output verification using on-chipsignature analyzer, or off-chip ATE. Test access mechanism (TAM): User-defined test datacommunication structure; carries test signals fromsource to module, and module to sink; tests moduleinterconnects via test-wrappers; TAM may contain bus,boundary-scan and analog test bus components. Test controller: Boundary-scan test access port (TAP);receives control signals from outside; serially loads testinstructions in test-wrappers.Source: H. KerkhoffEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham11Test Wrapper for a Core Logic added around a core to provide test access to theembedded core Test-wrapper provides for each core input terminal An external test mode – Wrapper element observes coreinput terminal for interconnect test An internal test mode – Wrapper element controls state ofcore input terminal for testing the logic inside core For each core output terminal A normal mode – Host chip driven by core terminal An external test mode – Host chip is driven by wrapperelement for interconnect test An internal test mode – Wrapper element observes coreoutputs for core testSource: H. KerkhoffEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham126

EE382V: System-on-Chip (SoC) DesignLecture 21A Test-Wrapperfrom/toExternalTest pinsScan chainFunctionalcore outputsCoreScan chainScan chainFunctionalcore inputsWrapperelementsWrappertestcontrollerto/from TAPSource: H. KerkhoffEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham13Goals of IEEE P1500 Core test interface between embedded core and systemchip Test reuse for embedded cores Testability guarantee for system interconnect and logic Improve efficiency of test between core users and coreprovidersSource: H. KerkhoffEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham147

EE382V: System-on-Chip (SoC) DesignLecture 21Set-up of P1500 ArchitectureSource: H. KerkhoffEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer15Core including Wrapper CellsSource: H. KerkhoffEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham168

EE382V: System-on-Chip (SoC) DesignLecture 21Wrapper Registers for P1500Source: H. KerkhoffEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham17Lecture 21: Outline SoC Manufacturing Test The testing problem SoC testing costs Design for test (DFT) SoC Testability Features Boundary Scan P1500 standard Built-In Self Test Functional Test Access Mechanism (TAM)EE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A. Gerstlauer189

EE382V: System-on-Chip (SoC) DesignLecture 21Built-In Self Test (BIST) Increasing circuit complexity, tester cost Interest in techniques which integrate some testercapabilities on the chip Reduce tester costs Test circuits at speed (more thoroughly) Approach: Compress test responses into “signature” Pseudo-random (or pseudo-exhaustive) patterngenerator (PRG) on the chip Integrating pattern generation and response evaluationon chip – BISTEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham19Pseudo-Random Sequences Linear Feedback Shift Register Shift register with input taken from XOR of state Pseudo-Random Sequence GeneratorDQ[1]DFlopQ[0]FlopDFlopCLKCan also be used tocompress test responsesEE382V: SoC Design, Lecture 21 2014 A. 1(repeats) 2014 A.J.GerstlauerA. Abraham2010

EE382V: System-on-Chip (SoC) DesignLecture 21Example of BISTTechnique calledSTUMPS(from IBM)EE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham21Why is Conventional Test Successful? Two innovations have allowed test to keep up withcomplex designs The stuck-at fault model– The model allows structural test generation, with a number of faults whichis linear in the size of the circuit Partitioning the circuit– Partitioning the circuit (with scan latches for example), alleviates the testproblem so that test generation does not have to deal with the entirecircuit Do these two assumptions hold for Deep SubMicron(DSM) circuits?EE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham2211

EE382V: System-on-Chip (SoC) DesignLecture 21IC TechnologyEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham23Features Smaller than WavelengthsSource: Raul Camposano, SynopsysEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham2412

EE382V: System-on-Chip (SoC) DesignLecture 21Increased LeakageEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham25Mean Number of Dopant AtomsRandom Dopant ogy Node (nm)EE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham2613

EE382V: System-on-Chip (SoC) DesignLecture 21Defects in DSM Technologies Experiments on real chips (e.g., Stanford) Stuck-at tests do not detect some defects unless they areapplied at speed Resistive opens comprise the bulk of test escapes in oneproduction line Likely in copper interconnect – cause delay faults Delay faults identified as the cause of most test escapeson another line Speed differences of up to a factor of 1.5 can existbetween fast and slow devices - problems with “speedbinning” Increasing possibility of shorts and crosstalkEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham27Effects on Chip? Change in delays of paths Effects could be distributed across pathsAt-Speed FunctionalStructuralAt-speed functionaltests better for delaydefectsSolution: At-Speed tests Tester Cost? Apply “NativeMode”?Stuck-atOpenShortSource: GelsingerEE382V: SoC Design, Lecture 21 2014 A. GerstlauerResistiveOpenLeakageResistiveShort Can use lowcost testers 2014 A.J.GerstlauerA. Abraham2814

EE382V: System-on-Chip (SoC) DesignLecture 21Native-Mode Built-In Self Test Functional capabilities of processors can be used toreplace BIST hardware – [UT Austin, ITC’1998] Application to self-test of processors at Intel – FRITSmethod applied to Pentium 4, Itanium [ITC’2002]D1D2D1CnQ1D3D2Cn-1Dn. . .Q2Cn-2DnC1Hardware for MISREE382V: SoC Design, Lecture 21Qnfor each data value Di {Shift Right Through Carry(S);if (Carry) S XOR(S, polynomial);S XOR(S, Di);}Software implementationof MISR 2014 A.J.GerstlauerA. Abraham29Native-Mode Self Test for Processors Random instructions can be run from cache and resultscompressed into a signature Implementation in Intel FRITS system showed benefits forreal chips (Pentium 4, Itanium) Technique can be used for self-test of an embeddedprocessor in a System-on-Chip Is it possible to now use this processing capability to testother modules (digital, analog/mixed-signal and RF) onthe SoC? First, can the processor test be improved to detect realisticdefects, e.g., small delays?EE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham3015

EE382V: System-on-Chip (SoC) DesignLecture 21Are Random Tests Sufficient? Intel implementation involved code in the cache whichgenerated random instruction sequences Interest in generating instructions targeting faults Possible to generate instruction sequences which will testfor an internal stuck-at fault in a module [Gurumurthy,Vasudevan and Abraham, ITC 2006] In order to deal with defects in DSM technologies, need totarget small delay defects Recent work: automatically generate instruction sequenceswhich will target small delay defects in an internal module[Gurumurthy, Vemu, Abraham and Saab, European TestSymposium (ETS) 2007]EE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham31Test Access Mechanisms (TAMs) Non-functional access Uses a kind of access to core not allowed during thenormal functional operation Generally based on scan chains or other design for test(DFT) structures Can also use the embedded processor as the testsource/sink Needs wrappers around the core under test Functional access Embedded processor is the test source/sink No DFTstructures or wrappers around the coresEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham3216

EE382V: System-on-Chip (SoC) DesignLecture 21Non-Functional TAMs Boundary scan based Uses the JTAG/boundary scan mechanism to load/capturethe tests Slow since the access is serial Direct access based Direct access to core test pins given through external pins Faster High overhead to route the access pins and also multiplepins requiredEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham33Functional TAMs Software-Based Self Test (SBST): Use the intelligence ofthe embedded processor to test the SOC At-speed tests are possible Cores in the SOC can be of three kinds1. White box -- internals visible, structure changeable2. Grey box – all the internals visible, but structure of thecore cannot be changed3. Black box – no internals visible, no change can be madeon the core Any methodology for testing black box cores should notdepend on knowledge of the core’s internalsEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham3417

EE382V: System-on-Chip (SoC) DesignLecture 21Approach to Testing Cores Uses functional TAM Uses pre-existing vectors Generates software to beloaded on to the embeddedprocessor Reverse driver that producesgiven test vectors for coreTest stimuliReverse driverData valuesGeneratesoftwarecode[Gurumurthy, Sambamurthy and Abraham,Int'l Test Synthesis Workshop (ITSW) 2008]EE382V: SoC Design, Lecture 21Software to beloaded intothe processor 2014 A.J.GerstlauerA. Abraham35Pre-Existing Vectors If using a core bought from vendor Vectors might also be provided by the vendor Reusing a core Vectors from the previous use Newly designed core Validation vectors Only constraint: these vectors must be functional testpatterns for the coreEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham3618

EE382V: System-on-Chip (SoC) DesignLecture 21Reverse Driver Parses the vector sequence to generate the data set to besent to the core being tested Is specific to each core – as many as the number of driverprograms Only overhead involved Generates the output in a format readable by the driverprogramEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham37Reverse Driver Illustration Peripheral core communicating with externalenvironment (send/receive 32-bit data) Five 8-bit registers addresses 0 – 4 Register 0 – Control Registers 1 to 4 – er0x030x710x040x78EE382V: SoC Design, Lecture 21 2014 A. GerstlauerSend at speed rate 1Data 0x0754DF7178 2014 A.J.GerstlauerA. Abraham3819

EE382V: System-on-Chip (SoC) DesignLecture 21Software Generation Use the driver program associated with each core beingtested Driver programs Software code that actually talks with the non-processorcores Know about the bus protocol Generally able to take in the data to be sent to the core orread back data from the core Developed as part of designing the SOCEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham39Coverage Measurement Simulate the SOC using the software generated Platform used SOC validation can be used Monitor the core boundaries to capture the pin data Fault simulate the core with the captured dataEE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham4020

EE382V: System-on-Chip (SoC) DesignLecture 21Experiment Implemented a SOC containingARM core, AES cryptographic coreand a Wishbone bus interface (Verilog) AES 128-bit data/keyencryption/decryptionfrom www.opencores.orgValidation vectors:Set of random values encrypted anddecryptedEE382V: SoC Design, Lecture 21 2014 A.J.GerstlauerA. Abraham41ResultsDetails about the synthesized AES coreNo. of inputs69No. of outputs33No. of sequential elements9225No. of combinational elements 1119No. of stuck-at faults64070ResultsSize(bytes)FaultOriginal No. ofcoverage coverage 96EE382V: SoC Design, Lecture 21 2014 A. Gerstlauer 2014 A.J.GerstlauerA. Abraham4221

EE382V: System-on-Chip (SoC) DesignLecture 21Lecture 21: Summary SoC Manufacturing Test Scan chains JTAG Boundary Scan Test wrappers for cores Built-in self test (BIST) Advanced SoC test topics Analog/Mixed-Signal (AMS) test RF test Micro-Electro-Mechanical Systems (MEMS) testEE382V: SoC Design, Lecture 18 2014 A. Gerstlauer 2014 A. Gerstlauer4322

EE382V: System-on-Chip (SoC) Design Lecture 21 2014 A. Gerstlauer 3 DFT Architecture for

Related Documents:

EE382V-ICS: System-on-Chip (SoC) Design Lecture 8 2010 A. Gerstlauer 3 EE382V-ICS: SoC Design, Lecture 8 2010 A. Gerstlauer 5 Complexity ForcesFile Size: 701KB

EE382V: Embedded Sys Dsgn and Modeling, Lecture 2 2008 A. Gerstlauer 6 Text vs. Graphics Models versus languages not to be confused with text versus graphics Text and graphics are just two types of languages – Text: letters, numbers – Graphics: circles, arrows (plus some letters, numbers) Source: T. Givargis, F. Vahid.

10.3.5 Connecting ADC and DAC IPs into a Cortex-M system 242 10.4 243Bring an SoC to life – Beetle test chip case study 10.4.1 Beetle test chip overview 243 10.4.2 Beetle test chip challenges 245 10.4.3 Beetle test chip system design 246 10.4.4 Implementation of the Beetle test chip 24

The parallel chip breaker, that the figure N 1 shows an example, is mainly used in finishing operations. Figure N 1 - Parallel chip breaker Inclined chip breaker The angular type is suitable for most machining operations and is the most difficult to do. The chip is flexed towards the area of the workpiece where is broken without damaging .

NEW CHIP BREAKER FGS TYPE FGS type chip breaker The FGS chip breaker is a positive ground type insert. Its sharp cutting edge generates low cutting resistance while guarantee-ing high precision machining. The chip breaker minimizes heat when machining heat resistant super alloys and the small dot located in the corner is effective for chip control.

Nomenclature bch Chip width (mm), (in) bγ1 Insert/chip restricted contact length (mm), (in) Ch Cutting ratio (the ratio of the undeformed chip thickness to the chip thickness) d Depth of cut (mm), (in) d0 The standard critical depth of cut under pre-defined standard cutting condition (mm), (in) dmax Maximum depth of cut in extra region of chip-breaking chart (mm), (in)

1.1 Chip breaker A chip breaker is the tool which has a groove or an obstacle placed on the incline face of the tool. A chip breaker can be used for increasing chip breakability which results in efficient chip control and improved productivity. It also decreases cutting resistance, and gives a better surface finish to the workpiece.

Text and illustrations 22 Walker Books Ltd. Trademarks Alex Rider Boy with Torch Logo 22 Stormbreaker Productions Ltd. MISSION 3: DESIGN YOUR OWN GADGET Circle a word from each column to make a name for your secret agent gadget, then write the name in the space below. A _ Draw your gadget here. Use the blueprints of Alex’s past gadgets on the next page for inspiration. Text and .