Tetra Max

DFT Compiler & TetraMAX
Kate, Yu-Jen Kate Yu Jen Huang Dec 17 2009
Outline
VLSI Testing
I t d ti Introduction Fault modeling T t generation Test ti
Design for Testability (DFT) Fault Simulation (TetraMAX) Lab time
Advanced Reliable Systems (ARES) Lab.
Yu-Jen Huang
Definitions
Design synthesis
Gi an I/O function, develop a procedure to Give f ti d l d t manufacture a device using known materials and processes
Verification
Predictive analysis to ensure that the synthesized design will perform the given I/O function
Testing
A manufacturing step that ensures that the physical device, device manufactured from the synthesized design design, has no manufacturing defect
Yu-Jen Huang
Why Testing?
Economy
P d t quality Product lit Product reliability
The purpose of testing is to weed out all bad products before they are shipped to users
The number of bad products heavily affect the price of good products
A profound understanding of the principles of manufacturing and test is essential for a designer to design a quality product
Yu-Jen Huang
Defect, Fault, and Error

Defect
A d f t is an unintended diff defect i i t d d difference b t between th the implemented hardware and its intended design It is caused during manufacture or the use of devices
Fault
A representation of a physical defect at the abstracted f f function level
E Error
A wrong output signal produced by a defective circuit It is caused by a fault or a design error
Yu-Jen Huang
Testing Problem
What faults to test
F lt modeling Fault d li
How are test pattern obtained

Test pattern generation
How is test quality

Fault simulation
How are test vectors applied and results evaluated

ATE/BIST
Yu-Jen Huang
Modern IC Testing
STIL 1 0; 110001001 000111000 101000011 111110001
Test program
Passed the test Devise under test (DUT) Automatic test equipment (ATE) Failed the test
Yu-Jen Huang
Cost of Test
Design for testability (DFT)
A Area overhead and yield reduction h d d i ld d ti Performance overhead
Software processes of test

Test generation and fault simulation Test programming and debugging
Manufacturing test g
Automatic test equipment (ATE) capital cost Test center operational cost p
Yu-Jen Huang
Outline
VLSI Testing
Yu-Jen Huang
Defect
Real defects are too numerous and hard to be analyzed
Yu-Jen Huang
Fault
Fault is a physical defect in a circuit/system
P Permanent Fault: a fault th t i continuous and stable, t F lt f lt that is ti d t bl whose nature do not change before, during, and after testing
Hard fault or solid fault
Temporary fault: a fault that is present only part of the time, occurring at random moments and affecting the system for finite, but unknown, intervals of time
Transient fault, soft error
Intermittent fault: caused by non-environmental conditions diti
Yu-Jen Huang
Fault Model and Error

Fault model is a logic effect of a fault
St t l fault Structural f lt
Stuck-at-faults Bridging fault Open fault Transition fault Delay fault
Functional fault
RAM coupling and pattern-sensitive faults
Error is manifestation of a fault that results in an incorrect module output or system state
Yu-Jen Huang
Failure
Failure is deviation of a system from its specified behavior
Fault -> error -> failure
F il Failure mechanism i a physical or chemical h i is h i l h i l process that causes devices to malfunction Failure mode is the cause of rejection of failed device (effect of failure mechanism), such as open/short interconnections, or degraded parameter values
Yu-Jen Huang
Defect Level, Fault Coverage and Yield

Defect Level
Th fraction of devices that pass all the tests but still The f ti fd i th t ll th t t b t till contain faults DL = 1 Y(1-FC) 1-Y(1 FC)
Fault Coverage (FC)

The measure of the ability of a test set T to detect a f f given set of faults FC = N of d t t d f lt / No. of possible f lt No. f detected faults N f ibl faults Can be determined by fault simulation
Yi ld (Y) = No. of good dies per wafer/No. of di Yield N f d di f /N f dies per wafer
Yu-Jen Huang
Defect Level and Fault Coverage

DL is measured in terms of DPM (defects per million), million) and typical values claimed are less than 200DPM, or 0.02% R Required FC f DL = 200DPM i d for
Y(%) FC(%) 10 99.99 50 99.97 90 99.8 99 98
Yu-Jen Huang
Stuck-at-fault
Single stuck-at-fault: line has a constant value M lti l stuck-at-fault: several single SAF occur Multiple t k t f lt l i l SAFs at the same time
a b c
a 0 0 1 1
b 0 1 0 1
c 0 0 0 1
c(a/0) c(a/1) c(b/0) c(b/1) 0 0 0 0 0 0 0 1 0 1 0 0 0

Yu-Jen Huang
c(c/0) 0 0 0 0
c(c/1) 1 1 1 1
0 1 1
Single Stuck-at Fault Assumption

ATPG tools assume at most one SA fault is present on the chip under test This assumption is made to simplify the analysis
Detecting multiple faults would complicate ATPG
The SSF model disregards the possible presence of any other faults affecting the test for a target fault SSF assumes no chance of another fault masking the target fault, making it impossible to g g , g p detect
Yu-Jen Huang
Number of Single Stuck-at Faults

Number of faults in a Boolean gate circuit
N of PI + No. of gates + No. of fanout branches No. f N f t N ff tb h
Example: XOR gate

24 SAFs
c d j 0/1
1 0
a b
g 1
h i
s/0
z 1/0
k 1
Yu-Jen Huang
Bridging Faults
Two or more normally distinct points (lines) are shorted together Two types of bridging faults
Input bridging
Can form wired logic or voting model
F db k b id i Feedback bridging
Can introduce feedback
x1 x2 xn . . . F Y x1 x2 xn x1 . . . F Y x2 xn . . . F Y
Yu-Jen Huang
Bridging Fault behavior

I1 I2 O1 O2
AND bridges
The bridge acts like an AND gate
OR bridges
The bridge acts like an OR gate
Dominant bridges
Th dominating net always controls th value output The d i ti t l t l the l t t
Yu-Jen Huang
Single Cell Fault

Cells can have any implementation All possible cell faults are allowed ibl ll f lt ll d C-testability: constant number of test patterns, independent of circuit size
X1 Y1 X2 Y2 X3 Y3 X4 Y4
C0
FA
FA
FA
FA
C4
Z1
Z2
Yu-Jen Huang
Z3
Z4
Transistor Faults
MOS transistor is considered an ideal switch and two types of faults are modeled
Stuck-open: a single transistor is permanently stuck in the open state Stuck-on: a single transistor is permanently shorted irrespective of its gate voltage
Detection of a tuck-open fault requires two vectors Detection of a stuck-on fault requires the measurement of quiescent current (IDDQ) t f i t t
Yu-Jen Huang
Timing-Related or Delay Fault

Delay fault
G t delay fault: A gate delay f lt occurs when a gate Gate d l f lt t d l fault h t operates more slowly than expected
Path delay fault: A path delay fault assumes that a logic transition is delayed along an entire path
Because delays refers to differences in behavior over time delay faults focus on transition in logic time, values
Yu-Jen Huang
Outline
VLSI Testing
Yu-Jen Huang
Testing
Testing = test generation + test application + output evaluation FC can be determined by fault simulation Cost of test generation (TG) depends on
Complexity of the fault model Complexity of the TG algorithm Complexity of the DUT y
A test set for a class of faults F is a set of tests T such that for any fault f F , there exists t T such that t detects f
Yu-Jen Huang
Test Generation by Truth Table

a b c
a 0 0 1 1
b 0 1 0 1
c 0 0 0 1
c(a/0) c(a/1) c(b/0) c(b/1) 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 1
c(c/0) 0 0 0 0
c(c/1) 1 1 1 1
Ta/0 = {11}; Ta/1 = {01}; Tb/0 = {11}; Tb/1={10} Tc/0 = {11}; Tc/1 = {00} or {01} or {10} T = {01, 10, 11}
Yu-Jen Huang
Undetectable Fault
No pattern can be devised to detect fault U2 SA0
a U1
b 1/0 U2 c U3 Y
Yu-Jen Huang
D-algorithm
Select a primitive cube to activate fault f S Sensitize all possible paths f iti ll ibl th from th f lt site t the fault it to POs (fault propagation or D-drive)
Continued until a PO has a D or D
Develop a consistent set of primary input (PI) values that will account for all lines set to 0 or 1during D-drive. If not consistent, try another path
Yu-Jen Huang
Example
a c e b a d SA1 h g i f
a 1 1 1 1 1
b 1 1 1 1 1
d D D D
e 1 1 1 1
Yu-Jen Huang
g D D
0 0 0
D D D D D
1 1
D D
D D D D
Outline
VLSI Testing
Yu-Jen Huang
Design for Testability (DFT)

A fault is testable if there is a well-specified procedure to expose it which can be it, implemented with a reasonable cost using current technique DFT
A class of design methodologies which put constraints on the design process to make test generation and diagnosis easier
Yu-Jen Huang
Sequential Logic - I
clk
Yu-Jen Huang
Sequential Logic - II
Harder to test
S Sequential circuit h memory i addition t ti l i it has in dditi to combinational logic It takes more clock cycles to activate the fault and propagate the fault effect
Example
x x 1 x x 0 a b 1/0 z H x x x x 0 x x 0 c d
Yu-Jen Huang
Scan Design Approaches

MUX scan
Shift register modification i t difi ti
Clock scan
LSSD
Partial scan Boundary scan

1149.1 (JTAG) 1149.4 1149 5 1149.5
Yu-Jen Huang
Mux Scan
Primary Inputs (PIs) Combinational Logic Primary Outputs (POs)
Sequential Logic
Primary inputs
Combinational Logic
Primary outputs
Pseudo Primary Outputs (PPOs) Scan i S input (SI) t
Pseudo Primary inputs (PPIs) Scan output (SO) S t t
Sequential Logic
Controllability and observability

Yu-Jen Huang
Scan Cell
PI Combinational Logic
PO
PPO SI T clk lk
mux PPI
..
SO
Yu-Jen Huang
Sequential Logic Full Scan

Scan_en Scan_in
Scan_out
clk
Yu-Jen Huang
Operation
1. Switch to SR mode 2 I iti li SR 2. Initialize
Load the first pattern
3. Return to normal mode

Apply test pattern
4. Switch to SR mode
Shift out the final state Setting the starting state for the next test Go back to the previous operation 3
Yu-Jen Huang
Advantages and Disadvantages

Advantages
S Scan paths allow use of combinational ATPG th ll f bi ti l Make high coverage possible in reasonable ATPG time and test program length
Disadvantages
Area overhead
Scan-in pin (SI), scan-out pin (SO), test mode pin (T), and mux in front of each FF
For large design, back-end re-optimization to fix scan timing is tedious and time-consuming g g
Yu-Jen Huang
Scan Synthesis
Scan configuration
Number of scan chains Types of scan cells Storage elements to exclude from scan synthesis How scan cells are arranged within scan chains
Scan replacement
Replace original design to scannable design
Scan reordering
Reorder scan chains
Scan stitching
Connect all scan cells together to form scan chains
Yu-Jen Huang
1. Before DFT Synthesis

Add scan related I/O pins in the design
Add si, se, t t d as th i i testmode the input pins t i Add so as the output pin
Yu-Jen Huang
2. Set Scan Chain Type

Set in .synopsys_dc.setup file
set_scan_configuration -style multiplexed_flip_flop t fi ti t l lti l d fli fl
Yu-Jen Huang
DFT Compiler Flow

HDL Set constrains: scan style, speed, area
Scan-ready synthesis
check dft
Pre-scan DRC Constraint-based scan synthesis: Routing, balancing, g g, g, gate-level opt. p
insert dft
Insert scan
check dft
Post-scan DRC
Preview coverage
Yu-Jen Huang
Scan Synthesis Flow

1. Create Test Protocol 2. DFT Check 3. Test-Ready Compile Check Constraint 4. Design + Test Protocol DFT Check
Existing Scan Chain Flow

Top-Level Netlist
Unmapped pp DFT Flow
5. Specify Scan Chain 6. DRC & Preview 7. Scan Chain Synthesis
Mapped DFT Flow
8. Scan Chain Identification
9. DRC & Coverage 10. 10 Handoff Design

Yu-Jen Huang
3.Create Test Protocol

Set_dft_signal
set t t d f lt t test_default_period 100 i d set_dft_signal -view existing_dft -type ScanClock timing {45 55} -port clk port set_dft_signal -view existing -type Reset -active_state 1 -port rst set_dft_signal -view existing -type Constant active_state port test_mode active state 1 -port test mode create_test_protocol
Yu-Jen Huang
4.DFT Check
Pre-DFT DRC Ch k scan d i rule b f Check design l before scan chain h i synthesis dft_drc
Yu-Jen Huang
5.Test-Ready Synthesis
compile -scan
D clk
Q Replaced by DFF D se clk DFF
Yu-Jen Huang
6.Read Design & Test Protocol

Write out the test protocol and scan-ready design
write_test_protocol -output cpu.spf it t t t l t t f Write -format ddc -hierarchy -output cpu.ddc
Read design & test protocol

read_file -format ddc cpu.ddc current_design design link read_test_protocol cpu.spf
Yu-Jen Huang
7.Specify Scan Chain

Set global attributes for scan paths in the current design
set_scan_configuration -chain_count 1 set_scan_configuration -clock_mixing no_mix t fi ti l k i i i set_dft_signal -view spec -type ScanDataIn -port si set_dft_signal -view spec -type S t dft i l i t ScanDataOut -port so D t O t t set_dft_signal -view spec -type ScanEnable -port se active_state active state 1 set_scan_path chain1 -scan_data_in si scan_data_out scan data out so
Yu-Jen Huang
Memory Block
add memory wrapper
si so si so
RAM
RAM
Un-observable
Un-controllable
Un-observable
Un-controllable
Yu-Jen Huang
Memory Wrapper
Add in step 5
set test point element -type observe [get object name set_test_point_element type [get_object_name [get_pins RAM_64B/D*]] -clock_signal clk point_element -type observe [g _object_name yp [get j set_test_p [get_pins RAM_64B/A*]] -clock_signal clk set_test_point_element -type control_01 [get_object_name [get_pins RAM_64B/Q*]] l k i [ t i RAM 64B/Q*]] -clock_signal clk l lk report_test_point_element
Yu-Jen Huang
8.Scan Preview
Check scan-path consistency Determines th chain count D t i the h i t Allocates and orders scan cells Adds connecting hardware
preview dft -show all preview_dft show preview_dft -test_points all
Yu-Jen Huang
9.Scan Chain Synthesis

Scan replacement Ensures no contention E t ti Inserts test points Optimized the logic
insert dft insert_dft
Yu-Jen Huang
10.Scan Chain Identification

Use when you import an existing scan design in non-ddc non ddc netlist format
set_scan_state scan_existing
Yu-Jen Huang
11.DRC & Coverage

Post-DFT DRC
dft d -coverage_estimate dft_drc ti t
Yu-Jen Huang
12.Handoff Design
Report scan information
report_scan_path -view existing_dft -chain all t th i i ti dft h i ll report_scan_path -view existing_dft -cell all
Yu-Jen Huang
13.Handoff Design
Prepare TetraMax script
change_names -hierarchy -rule verilog h hi h l il write -format verilog -hierarchy -out cpu_dft.vg write -format dd -hierarchy -output cpu_dft.ddc it f t ddc hi h t t dft dd write_scan_def -output cpu_scan.def set test_stil_netlist_format verilog f write_test_protocol -output cpu.spf
Yu-Jen Huang
Outline
VLSI Testing
I t d ti Introduction Fault modeling T t generation for combinational and sequential Test ti f bi ti l d ti l circuits
D i f T t bilit (DFT) Design for Testability Fault Simulation (TetraMAX) Lab time
Yu-Jen Huang
Automatic Test Pattern Generation

Goal
G Generate the test patterns for target fault model and t th t t tt f t t f lt d l d keep the number of test pattern as small as possible
H ? How?
Use an ATPG tool which relies on proprietary techniques to speed up and extend the basic D algorithm
Fault list Test Generation Fault simulation Add pattern to test set
Remove all detected fault and select next fault

Test Set
Yu-Jen Huang
DFT Compiler to TetraMAX
design_dft.v write f verilog hierarchy

-output design_dft.v p g _
read netlist design_dft.v
Simulation Testbenches
Design Compiler
Simulation Library
read netlist library.v
TetraMax
ATE Vectors
write_test_protocol f stil _ _p -out design.spf
design.spf design spf
Fault Reports
run d d i drc design.spf f
Yu-Jen Huang
5 Fault Categories
During ATPG and functional fault simulation, TetraMAX classifies faults into 5 major categories
DT: Detected PT: Possibly d t t d PT P ibl detected UD: Undetected AU: AU ATPG untestable t t bl ND: Not detected
Test coverage
all faults - UD+ AU au_credit

DT DT+ PT posdet credit posdet_credit
Yu-Jen Huang
TetraMAX
Setup file (CSHRC)
source / / d/ /usr/cad/synopsys/CIC/tmax.csh /CIC/t h source /APP/cshbank/tmax.csh
Invoking TetraMAX
tmax&
command
Yu-Jen Huang
Command Input and Script Files

Command input may be specified in many ways
manus, GUI b tt buttons, di l b dialog boxes typed at the command input line read from command fil df d files
By default, TetraMAX aborts a script file when a command returns an error To continue executing scripts, use g
BUILD> set_command noabort
Yu-Jen Huang
Multi-line Input in the GUI

Multiple commands may also be entered by separating each command by a semicolon and white space
BUILD> b ild force; read net mylib.v; run build build f d t lib b ild
Yu-Jen Huang
Help Command
BUILD> help add
Add Atpg Constraints Add Cell Constaints Add Equivalent Nofaults Add Net Connections Add PI Constraints Add PO Masks Add Atpg Gates Add Clocks Add Faults Add Nofaults Add PI Equivalences
BUILD h l read netlist BUILD> help d tli t

REAd NEtlist [file_name] [-Format <Edif |VErilog | VHdl>] [-Sensitive | -INSensitive] [-Delete] [-Library] [-Master modules] [-Master_modules] [-Noabort] [-Verbose]
Yu-Jen Huang
MAN for Command Reference

Entering man and a command name, a message ID or a DFT rule ID or violation ID will ID, open up the on-line help to the reference page for that topic
BUILD> man getting_started BUILD> man add clock BUILD> man report faults BUILD> man z4-6 BUILD> man m68
// // // // //
a topic a command a command a violation a message ID
Yu-Jen Huang
Stop Process
Submit button changes to Stop while performing operation or CTRL+C and CTRL+Break
Yu-Jen Huang
Basic Flow
Read in the netlist B ild M d Build Mode
BUILD> read netlist xxx.vg BUILD> read netlist tsmc18.v BUILD> run build_model top_model
DRC mode
DRC> set drc xxx.spf p DRC> run drc
Yu-Jen Huang
1.Build Mode - Read Netlist

2.Run
1.Read cpu_dft.vg tsmc18.v RAM_64B.tv 1 Read cpu dft vg & tsmc18 v & RAM 64B tv
Yu-Jen Huang
2.Choose Top Module
1.Build 1 Build
2. Choose top module cpu 3.Run
Yu-Jen Huang
3. DRC Mode Choose spf file
1.DRC
2. Choose spf file cpu_dft,spf
3.Run 3R
Yu-Jen Huang
4.Test Mode
Test mode
TEST> report summaries f lt patterns t i faults tt TEST> add faults -all TEST report summaries f lt patterns TEST> t i faults tt TEST> run atpg -auto TEST> report summaries f S faults patterns Increase fault coverage
run atpg -fast seq. (or full seq.) f ( f ll )
Write ATPG patterns

write patterns xxx_atpg.v -internal -format verilog -serial -replace
Yu-Jen Huang
Test Coverage
Yu-Jen Huang
Reference
DFT compiler user guide T t M user guide TetraMax id
/APP/cad/synopsys/sold/2004.12/doc/online/test/
Yu-Jen Huang
Yu-Jen Huang

Tetra Max

Uploaded by

Copyright:

Available Formats

Tetra Max

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tetra Max

Uploaded by

Copyright:

Available Formats

What is the purpose of testing integrated circuits?

What is the purpose of testing integrated circuits?

What are the three main types of defects that can occur in integrated circuits?

What are the three main types of defects that can occur in integrated circuits?

DFT Compiler & TetraMAX

Kate, Yu-Jen Kate Yu Jen Huang Dec 17 2009

Design for Testability (DFT) Fault Simulation (TetraMAX) Lab time

Advanced Reliable Systems (ARES) Lab.

Defect, Fault, and Error

Advanced Reliable Systems (ARES) Lab.

How are test pattern obtained

How is test quality

How are test vectors applied and results evaluated

Advanced Reliable Systems (ARES) Lab.

Software processes of test

Advanced Reliable Systems (ARES) Lab.

Design for Testability (DFT) Fault Simulation (TetraMAX) Lab time

Advanced Reliable Systems (ARES) Lab.

Advanced Reliable Systems (ARES) Lab.

Intermittent fault: caused by non-environmental conditions diti

Advanced Reliable Systems (ARES) Lab.

Fault Model and Error

Advanced Reliable Systems (ARES) Lab.

Defect Level, Fault Coverage and Yield

Fault Coverage (FC)

Defect Level and Fault Coverage

Advanced Reliable Systems (ARES) Lab.

c(a/0) c(a/1) c(b/0) c(b/1) 0 0 0 0 0 0 0 1 0 1 0 0 0

Advanced Reliable Systems (ARES) Lab.

Single Stuck-at Fault Assumption

Number of Single Stuck-at Faults

Example: XOR gate

Advanced Reliable Systems (ARES) Lab.

Advanced Reliable Systems (ARES) Lab.

Bridging Fault behavior

Advanced Reliable Systems (ARES) Lab.

Single Cell Fault

Timing-Related or Delay Fault

Design for Testability (DFT) Fault Simulation (TetraMAX) Lab time

Advanced Reliable Systems (ARES) Lab.

Test Generation by Truth Table

c(a/0) c(a/1) c(b/0) c(b/1) 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 1

Advanced Reliable Systems (ARES) Lab.

Advanced Reliable Systems (ARES) Lab.

Advanced Reliable Systems (ARES) Lab.

Design for Testability (DFT) Fault Simulation (TetraMAX) Lab time

Advanced Reliable Systems (ARES) Lab.

Design for Testability (DFT)

Advanced Reliable Systems (ARES) Lab.

Advanced Reliable Systems (ARES) Lab.

Advanced Reliable Systems (ARES) Lab.

Scan Design Approaches

Partial scan Boundary scan

Advanced Reliable Systems (ARES) Lab.

Pseudo Primary Outputs (PPOs) Scan i S input (SI) t

Pseudo Primary inputs (PPIs) Scan output (SO) S t t

Controllability and observability

Advanced Reliable Systems (ARES) Lab.

Sequential Logic Full Scan

Advanced Reliable Systems (ARES) Lab.