Design for Testability (DFT) - The University of Texas at Austin
406040
6080mm With scan chain ,problem of testing any the combinational logic
ECE Department,University of Texas at Austin Lecture 21.Design for Testability 40604060
80
mm Level-Sensitive Scan Design (LSSD)
Structured DFT developed at IBM
All internal storage implemented in latches (SRLs),part of a scan chain
Latches controlled by two or more with rules for clocking
All clock inputs to SRLs must be in primary inputs (PIs)are “o?”
Clock signal at any clock input to from one or more clock PIs
No clock can be ANDed with another Clock PIs cannot feed data inputs through combinational logic ECE Department,University of Texas at Austin Lecture 21.Design for Testability
60
80
Department,University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 9,20114/49406080100120
40
60
80
mm Scannable Flip-Flops
University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 9,
mm
406080100 of Texas at Austin Lecture21.Design for Testability J.A.Abraham,
40608010040
60
80
mm University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 406080100406080
mm Boundary Scan Interface
Boundary scan is accessed through ?ve 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 chains through boundary scan for uni?ed test strategy
University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November
mm
406080100 of Texas at Austin Lecture21.Design for Testability J.A.Abraham,November
60
80
Department,University of Texas at Austin Lecture21.Design for Testability
40608040
6080
mm ECE Department,University of Texas at Austin Lecture 21.Design for Testability 40604060
80mm Testability Issues –Custom Microprocessors
Chip complexity,large volumes,costs
Area/performance penalties
Fault (defect)coverage goals
Test
generation (fault simulation)costs Development time
Tester time (manufacturing costs)
ASICs generally use a “full scan”methodology Some of the ad-hoc techniques used in custom be described in the following slides ECE Department,University of Texas at Austin Lecture 21.Design for Testability
6080Department,University of Texas at Austin
Lecture 21.Design for 40604060
80
mm Microrom Test in MC68881Floating ROM physically split into two sections
In test mode,ROM is addressed command register Exhaustive addresses fed to ROM (address/address complement)
Outputs of ROM go to two 16-bit CCITT-16polynomial x 15+x 12Monitor both the quotient and test pin (Probability of aliasing Department,University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,
Austin Lecture 21.Design for Testability J.406080406080Entry PLA Test
PLAs,A0,A1,A2and A3
PLA contains the entry point reset vectors completely tested functionally
tested like the ROMs
Command register patterns and outputs through a bus to register
Test patterns generated test generator
PLA Inputs A16A213A35Response 25Department,University of Texas at Austin
Lecture 21.Design for Testability J.
406080Responses compressed using multiple-input Department,University of Texas at Austin Lecture 21.Design for Testability 40604060
80
mm Testing Cache Memory Arrays in MC68030
Cache cell layout design is resistant and capacitive coupling
Most likely bridging defect is between Memory fault model:
One or more cells stuck at 0or 1
Coupling between cells
11n March Test (Marinescu,1982)
Refresh and data retention tests for the dynamic memory cells
Department,University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November
mm
406080100
40
60
80
University of Texas at Austin Lecture21.Design for Testability J.A.Abraham,November
40608010040
60
80
mm University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 40608010040
60
80mm IBM Risc System/6000
with pseudo-random BIST
University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November Common Engineer-
Processor (COP),
on-card sequencer
engineering
processor (ESP)
406080100120
40
60
80mm Hardware for pseudo-random vector generation and result
compression (31-bit LFSR)
Department,University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 9,201128/49Small processor controls operation 406080
100120
406080
mm BIST in IBM/Motorola Power-PC
Variety of test techniques applied to the Power-PC 603
Full LSSD test of logic BIST of “large”embedded RAMS
Functional test of small RAMS
I DDQ tests BIST for cache and tag RAMs Functional vectors (good for data cache,not instruction cache)
and random BIST (size,complexity,test coverage)not
applicable
Use modi?ed march test of Dekker (1988)
log 2n pattern for data
Overhead:BIST is 2.9%of RAM array,0.58%of total chip
Performance impact:less than 100pS due to extra MUX
input leg All four RAMs tested in parallel,2.5mS at 80MHz
Department,University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 9,201129/49
4060
4060
80mm Source/Sink for test:Embedded processor
Test access mechanisms:Busses
Core test “wrapper”for Test Access
Interface (TAI)
Department,University of Texas at Austin Lecture 21.Design for Testability
4060406080mm Issues with Built-In Self Test
Technique can run test sequences at operating capture results within the chip
Pseudorandom pattern generators,Can also use weighted random
patterns or deterministic patterns
Example (Synopsys):(Deterministic
Logic BIST
Problems:
Hardware overhead
Test power
Non-functional modes during test Department,University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 9,201131/
406080100mm Pseudo-code for software signature analyzer general register or memory,holds signature each register R i to be compressed Shift Right Through Carry(R i );(Carry){S =XOR(S,feedback polynomial)}XOR(S,R i );of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 406080100Functional BIST
Functional Random Instruction Testing at Speed (FRITS)Itanium processor family and Pentium 4line (ITC 2002)(kernels)are instruction sequences
Kernels loaded into cache and executed in real time during
application
They generate and execute pseudo-random or directed
sequences of machine code
Pentium-4,FRITS
added 5%unique coverage to manual tests
screened 10%–15%of chips which passed wafer
sort/package tests,but failed system tests
enabled low-cost testers:40%increase in defect screening structural tester
Kernels execute 20loops in ≈8mSecs
of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November
406080100mm DAFCA
of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November 406080100406080
mm Analysis and Signal Capture
DEMON implements Signals Tracer
Software trace bu?er prepares FSDB ?les
Source:DAFCA
Department,University of Texas at Austin Lecture 21.Design for Testability J.A.Abraham,November