DFTC1 2007.12 LabGuide

CUSTOMER EDUCATION SERVICES
DFT Compiler 1
Workshop
Lab Guide
30-I-011-SLG-012 2007.12
Synopsys Customer Education Services

700 East Middlefield Road
Mountain View, California 94043
Workshop Registration: 1-800-793-3448
www.synopsys.com
Copyright Notice and Proprietary Information
Copyright  2008 Synopsys, Inc. All rights reserved. This software and documentation contain confidential and
proprietary information that is the property of Synopsys, Inc. The software and documentation are furnished under a
license agreement and may be used or copied only in accordance with the terms of the license agreement. No part of the
software and documentation may be reproduced, transmitted, or translated, in any form or by any means, electronic,
mechanical, manual, optical, or otherwise, without prior written permission of Synopsys, Inc., or as expressly provided by
the license agreement.
Destination Control Statement

All technical data contained in this publication is subject to the export control laws of the United States of America.
Disclosure to nationals of other countries contrary to United States law is prohibited. It is the reader’s responsibility to
determine the applicable regulations and to comply with them.
Disclaimer
SYNOPSYS, INC., AND ITS LICENSORS MAKE NO WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, WITH
REGARD TO THIS MATERIAL, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
Registered Trademarks (®)

Synopsys, AMPS, Cadabra, CATS, CRITIC, CSim, Design Compiler, DesignPower, DesignWare, EPIC, Formality, HSIM,
HSPICE, iN-Phase, in-Sync, Leda, MAST, ModelTools, NanoSim, OpenVera, PathMill, Photolynx, Physical Compiler,
PrimeTime, SiVL, SNUG, SolvNet, System Compiler, TetraMAX, VCS, Vera, and YIELDirector are registered trademarks
of Synopsys, Inc.
Trademarks (™)
AFGen, Apollo, Astro, Astro-Rail, Astro-Xtalk, Aurora, AvanWaves, Columbia,Columbia-CE, Cosmos, CosmosEnterprise,
CosmosLE, CosmosScope, CosmosSE, DC Expert, DC Professional, DC Ultra, Design Analyzer, Design Vision,
DesignerHDL, Direct Silicon Access, Discovery, Encore, Galaxy, HANEX, HDL Compiler, Hercules, Hierarchical
Optimization Technology, HSIMplus, HSPICE-Link, iN-Tandem, i-Virtual Stepper, Jupiter, Jupiter-DP, JupiterXT,
JupiterXT-ASIC, Liberty, Libra-Passport,Library Compiler, Magellan, Mars, Mars-Rail, Milkyway, ModelSource, Module
Compiler, Planet, Planet-PL, Polaris, Power Compiler, Raphael, Raphael-NES,Saturn, Scirocco, Scirocco-i, Star-RCXT,
Star-SimXT, Taurus, TSUPREM-4, VCS Express, VCSi, VHDL Compiler, VirSim, and VMC are trademarks of Synopsys,
Inc.
Service Marks (SM)

MAP-in, SVP Café, and TAP-in are service marks of Synopsys, Inc.
SystemC is a trademark of the Open SystemC Initiative and is used under license. ARM and AMBA are registered
trademarks of ARM Limited. Saber is a registered trademark of SabreMark Limited Partnership and is used under license.
All other product or company names may be trademarks of their respective owners.
Document Order Number: 30-I-011-SLG-012

DFT Compiler 1 Lab Guide
Synopsys Customer Education Services

Creating Test
4 Protocols
Learning Objectives
After completing this lab, you should be able to:

• Provide enough information about the test features of
your design that DFTC can create a test protocol.
• Verify that you have told DFTC what it needs to know
• Verify that your protocol and design are compatible
• Verify that the design has been compiled to the correct
library and that all flip-flops have been replaced with
scan flip-flops
• Modify a test protocol to include an initialization
sequence for applying an internal test mode signal
Lab Duration:
90 minutes
Creating Test Protocols Lab 4-1

Synopsys 30-I-011-SLG-012
Lab 4
Part A Overview
The files for Part A of this lab are located in the directory lab4a_protocol.
Directory Structure
Lab4a_protocol Current working directory
analyzed Intermediate acs_read_hdl files
logs Session log files
unmapped Unmapped protocol
mapped Gate-level netlist
reports DFTC reports
tmax Files for downstream tools
scripts Constraint and run scripts
../ref/rtl/RISC_CORE/vhdl Design files
../ref/db Library files
If you need help…
dc_shell> help prev*

dc_shell> printvar *scan*
dc_shell> preview_dft -help
dc_shell> man compile
When working interactively, you do not need to type the entire command or option.
dc_shell> read_test_p -v
dc_shell> all_inputs -cl
Answers & Solutions

This lab guide contains answers and solutions to all questions. If you need some
help with answering a question, consult the back portion of this lab for help.
Lab 4-2 Creating Test Protocols

Synopsys DFT Compiler 1 Workshop
Lab 4
Instructions
During this lab, you will use the Design Vision GUI for debugging dft_drc
problems related to test protocols, write a script to create and verify a test protocol
and explore a dc_shell log file.
Task 1. Set Up Your Environment For Reading

1. Change directories to the project directory lab4_protocol.
Look at the “dot” files in that directory.
UNIX% cd lab4a_protocol
UNIX% pwd
UNIX% ls –a
2. Specify the libraries and append to the search path by editing the
.synopsys_dc.setup file with a text editor.
Add the following lines to the .synopsys_dc.setup file.
# for reading and linking

set target_library "sc_max.db"
set link_library "* sc_max.db dw_foundation.sldb"
lappend search_path ./scripts
The pound (#) at the beginning introduces a TCL comment.
Question 1. You use the set command to assign values to ‘things’ such
as target_library and link_library; what are these
‘things’ called in TCL?
....................................................................................................
Question 2. When you write out a design to a .ddc file, what is the
difference in how values defined with the set command are
treated, compared to how values defined with commands
starting with set_, such as set_scan_configuration or
set_dft_signal, are treated?
....................................................................................................
....................................................................................................

Lab 4
3. Create shortcuts by editing the .synopsys_dc.setup file.
a. Create the following commonly-used alias for the frequently-used

history command.
b. Then turn on the capability to edit commands in dc_shell.
c. Then save the .synopsys_dc.setup file and quit the editor.
# shortcuts
history keep 100
alias h "history"
set sh_enable_line_editing true ;# wow - read man page
Question 3. To use the history command in the Design Compiler

environment (design_vision/dc_shell/etc.) what would you
type at the command line after specifying: alias h
history?
....................................................................................................
4. Further examine the .synopsys_dc.setup file. Note the two variables

sh_output_log_file and sh_command _log_file. By setting these two
variables, you control where a command and an output log file are created.
The two files are created using the current time stamp and pid. This sequence
of commands can be inserted in most Synopsys tools.
5. Invoke Design Vision from the UNIX prompt in the lab4_protocol directory:
UNIX% design_vision
Note: You must open Design Vision from the same directory where the
.synopsys_dc.setup exists,because, as shown in the previous
steps, variables such as search_path are defined relative to the
directory containing the setup file.
Note: If you see any errors while invoking DV, refer to the golden solution at
./.solutions/dot.synopsys_dc.setup
6. List the contents of the logs directory. You should see the two log files
discussed above.
Task 2. Create a test protocol with a script
Utilize the flowchart below and the protocol test attributes specification to its right
as a guide to write and verify a DFTC script to create a test protocol.

Lab 4
Test Protocol Definition Flow

Apply protocol test attributes on the design
Run test attribute reports
no Test attributes
correctly applied?
yes
Create a test protocol

Protocol Test Attributes
Test Clock Port Clk
Test Clock 45 ns
no Resets and clocks Active Time
correctly applied? Test Clock 55 ns
yes
Inactive Time
Test Mode Port TEST_MODE
Run DFT DRC check
Reset port Reset
(asynchronous active low)
yes
Violations?
no
Save the test protocol file
1. Open the script called scripts/unmapped.tcl.

The tasks are defined in the flow chart above; the values for the attributes in
the chart beside it. Your task is to write the DFTC commands to perform the
tasks. For this lab, set the values for the test attributes in the script.

Synopsys DFT Compiler
iler 1 Workshop
Lab 4
2. Execute your script by running the unmapped.tcl script in batch mode.

Create a log file for the run:
unix% dc_shell -f scripts/unmapped.tcl | \

tee logs/run_unmapped.log
3. Determine if there were any problems in the unmapped flow by looking at

logs/run_unmapped.log.
Question 4. What command is used to see which DFT signals have been
applied to the design?
....................................................................................................
Question 5. a. What value is the TEST_MODE port set to?

b. What is the name of the test clock port?
c. What is the period of the test clock?
d. When does the test clock go high?
e. When does the test clock go low?
f. What is the name of the reset port?
g. What type of reset has been assumed?
....................................................................................................
....................................................................................................
....................................................................................................
....................................................................................................
....................................................................................................
....................................................................................................
Question 6. Look at the results of running the dft_drc command, are

there any errors or warnings?
....................................................................................................

Lab 4
Task 3. Verify that the Design was Compiled Properly
Question 7. What command tells whether the entire design was compiled
to the target library?
....................................................................................................
1. Verify that the entire design was compiled to the target library.
Question 8. What target library was it compiled to?
....................................................................................................
Question 9. Now that you know the design has been properly compiled,
how do you determine if the flip-flops have been scan
replaced?
....................................................................................................
2. Determine if the design has been properly scan replaced.
Task 4. Part A Review Questions
Question 10. What is the purpose of the .synopsys_dc.setup file?
....................................................................................................
Question 11. How do you verify the library variables are set up correctly?
....................................................................................................
Question 12. How do you read VHDL or Verilog code into Design Vision?
....................................................................................................
Question 13. What is the function of the target_library variable?
....................................................................................................
3. Exit from DFTC.

Lab 4
Part B Overview
All files for Part B of this lab except for designs and libraries are located in the
directory lab4b_init.
Directory Structure
Lab4b_init Current working directory
mapped_scan Gate-level scan design netlist
../rtl/ORCA_init/vhdl Design files
../libs Library files

Lab 4
Design Introduction
The design used in this task requires a custom test initialization protocol. You
obtained the following schematic from the ORCA design documentation that
describes the process for asserting the internal test_mode signal.
CLOCK_GEN
pclk F4
int_clk
test_mode = 1
CONFIG
Status Bits
Initialization
101 Sequence
1 0 1
conf F1 F2 F3
1
conf_ena
pclk

Lab 4
Instructions
Since DFTC cannot infer a test protocol of this type you must develop one for this
design. Answer the following questions in preparation for doing the subsequent lab
tasks.
Task 1. Internally Generated test_mode Signal Using

the CONFIG Block
Question 14. What value must be placed on the conf_ena signal to initialize
the configuration register flip-flops F1, F2 and F3?
....................................................................................................
Question 15. How many cycles must you pulse pclk before test_mode is
asserted?
....................................................................................................
Question 16. What values should conf_ena and pclk have the cycle after
test_mode is asserted?
....................................................................................................
Question 17. Write in STIL format the Vector statements that would go
into the “test_setup”?
. . .
MacroDefs {
"test_setup"
{
. . .
. . .
}
You will use this information later when you develop the initialization
protocol for ORCA.

Lab 4
Instructions
During this lab you will, develop a custom test initialization protocol for a design
that has an internal test_mode signal.
Task 2. Read a Mapped Design and Create a Test

Protocol
1. Change directories to the project directory lab4b_init.
UNIX% cd lab4b_init
2. Invoke DFTC.
dc_shell
3. Create a shortcut to make running the scripts easier.
alias s "source -echo -verbose"
4. A mapped design has been prepared for you in the interest of saving time
during lab (Reading the RTL and compiling).
Examine the contents of the script that reads the saved ddc file for the ORCA
design and then source the script to read in ORCA.
exec cat scripts/1read_design.tcl
s scripts/1read_design.tcl
Question 18. Where is the design read from? Is the design RTL or mapped
gates?
....................................................................................................
Before you create the starting test protocol you must specify all your
clocks, resets, constants, etc. The clocks and the resets are already
defined in the 2create_test_protocol.tcl script.
Question 19. After the test initialization asserts the internal test_mode
signal what must you do to keep the ORCA design in that
mode throughout the rest of the scan testing?
....................................................................................................
5. Add the appropriate command to the 2create_test_protocol.tcl script.

6. Create the test protocol for ORCA by running the script.
Lab 4
s scripts/2create_test_protocol.tcl
Review the results of the dft_drc run on the ORCA RTL design.
Question 20. Are there few or many testability problems reported?
....................................................................................................
Task 3. Create an Initialization Sequence
Developing the initialization test protocol from scratch is not

practical for most engineers. A better method is to modify the
protocol files that DFTC or TetraMAX create for you.
1. Save out the current test protocol for the design.
write_test_protocol –o orca_mapped.spf
vi orca_mapped.spf
2. Add the required Vector statements (“V {}”) to the orca_mapped.spf.

Question 21. Since you only changed the “test_setup” section of the
STIL test protocol, which command should you use to read it
back into DFTC?
....................................................................................................
3. Verify that your initialization vectors are correct by reading in your custom
protocol and rerunning dft_drc.
Question 22. What happens when you first try to read orca_mapped.spf
back into DFTC?
....................................................................................................
4. Solve the problem by removing the previous protocol and reading in your
custom initialization.
remove_test_protocol
read_test_protocol -section test_setup \
orca_mapped.spf
Note: The “test_setup” or initialization section is only part of

a complete test protocol.

Lab 4
5. Re-create the rest of the test protocol before attempting to run dft_drc.
create_test_protocol –capture_procedure multi_clock

dft_drc
Question 23. How do the results of dft_drc compare to your first run?
....................................................................................................
Congratulations! You have completed the lab.

Answers / Solutions Lab 4
Answers / Solutions
Task 1. Set Up Your Environment For Reading

Question 1. You use the set command to assign values to ‘things’ such
as target_library and link_library. What are
these ‘things’ called in TCL?
Variables
Question 2. When you write out a design to a .ddc file, what is the
difference in how values defined with the set command are
treated, compared to how values defined with commands
starting with set_, such as set_scan_configuration
or set_dft_signal, are treated?
When a variable is applied to a design, it does not “stick” to

the design when the design is saved in a .ddc format. This is
different from a “set_” command that can be applied to a
.ddc file and is saved with the .ddc file. A variable must be
reapplied each time the design is brought back into Design
Compiler.
Question 3. To use the history command in the Design Compiler
environment (design_vision/dc_shell-t/etc.) what would you
type at the command line after specifying: alias h
history?
Task 2. Create a Test Protocol with a script

1. Modify the script called unmapped.tcl.
If you run into problems, consult .solutions/unmapped.tcl
This flowchart shows you the command flow of the
create_test_protocol.tcl file as well as the decision points.

Test Protocol Definition Flow

set_dft_signal ... –port Clk
set_dft_signal ... –port Reset
Set_dft_signal ... –port TEST_MODE
report_dft_signal –view existing_dft
no Test attributes
correctly applied?
yes
create_test_protocol
no Resets and clocks

correctly applied?
yes
dft_drc
yes
Violations?
no
write_test_protocol –output RISC_CORE.spf
Question 4. What command is used to see which DFT signals have been
applied to the design?
report_dft_signal
Question 5. a. What value is the TEST_MODE port set to?
b. What is the name of the test clock port?
c. What is the period of the test clock?
d. When does the test clock go high?
e. When does the test clock go low?
f. What is the name of the reset port?
g. What type of reset has been assumed?
a. TEST_MODE = 1

Lab 4 Answers / Solutions
b. Test clock port name is: Clk
c. Period of the test clock is:100 ns
d. Test clock high: 45 ns
e. Test clock low: 55 ns
f. Reset port name is: Reset
g. Reset type is: Asynchronous, active low
Question 6. Looking at the results of running the dft_drc command,

are there any errors or warnings?
No
Task 3. Verify that the Design was Compiled Properly

Question 7. What command tells whether the entire design was
compiled to the target library?
Use the report_hierarchy command. All cells should

be cb13fs120_tsmc_max cells (target library).
Question 8. What target library was it compiled to?
Run the report_hierarchy command from the

command line. All cells were cb13fs120_tsmc_max cells,
so the design was properly compiled to the target library
Question 9. Now that you know the design has been properly compiled,
how do you determine if the flip-flops have been scan
replaced?
Run report_scan_state after compile –scan has

occurred. Design was properly scan replaced:
dc_shell> report_scan_state
****************************************
Report : test
-state
Design : RISC_CORE
****************************************

Scan state : scan cells replaced with loops.
Task 4. Part A Review Questions

Question 10. What is the purpose of a .synopsys_dc.setup file?
The .synopsys_dc.setup file is a script file, which is

executed upon invocation of either Design Vision or
dc_shell. It is a useful file for commands that are to be
executed each time DC is invoked. For example, it is useful
to initialize library variables in this file.
Question 11. How do you verify the library variables are set up correctly?
acs_read_hdl does an automatic link, which produces

warnings or errors if the variables are not set correctly.
Question 12. How do you read VHDL or Verilog code into Design
Compiler?
To read individual files in, use the read_vhdl or

read_verilog command. To read files that have
subblocks, use the acs_read_hdl command
Question 13. What is the function of the target_library variable?
The target_library variable defines the technology

library used during compile mapping phase.
Task 1. Internally Generated test_mode Signal Using

the CONFIG Block
Question 14. What value must be placed on the conf_ena signal to
initialize the configuration register flip-flops F1, F2 and F3?
conf_ena must be logic 1 to initialize the flip-flops.

Question 15. How many cycles must you pulse pclk before test_mode is
asserted?
pclk must pulse three times.

Question 16. What values should conf_ena and pclk have the cycle after
test_mode is asserted?
Both should be logic 0.

Question 17. Write in STIL format the Vector statements that would go
into the “test_setup”?
. . .
MacroDefs {
“test_setup”
{
. . .
V { “conf_ena”=1; “conf”=1; “pclk”=P; }
V { “conf_ena”=0; “pclk”=0; }
. . .
}
Task 2. Read a Mapped Design and Create a Test
Protocol
Question 18. Where is the design read from? Is the design RTL or
mapped gates?
mapped/ORCA.ddc
The design is mapped gates.
Question 19. After the test initialization asserts the internal test_mode
signal what must you do to keep the ORCA design in that
mode throughout the rest of the scan testing?
You must hold the conf_ena at logic 0 by using the

following command:
set_dft_signal -view existing_dft -type Constant \

-active_state 0 -port conf_ena
Question 20. Are there few or many testability problems reported?
Many.
4010 PRE-DFT VIOLATIONS
2828 Uncontrollable clock input of flip-flop violations (D1)
1133 DFF set/reset line not controlled violations (D3)
32 Clock feeding both clock and data input violations (D11)
6 Clock feeding multiple clock/set/reset inputs violations (D12)
1 Clock path affected by clock captured by clock in trailing edge
clock_port violation (D16)
10 Clock_port not capable of capturing data violations (D17)
Task 4. Create an Initialization Sequence
8. Add the required vector statements to the unmapped_flow.spf.
V { "conf"=1; "conf_ena"=1; "pclk"=P; }



V { "conf_ena"=0; "pclk"=0; }
Question 21. Since you only changed the “test_setup” section of the
STIL test protocol, which command should you use to read
it back into DFTC?
read_test_protocol –section test_setup
Question 22. What happens when try to read orca_mapped.spf back

into DFTC?
dc_shell> read_test_protocol -section

test_setup orca_mapped.spf
Warning: a protocol already exists in mode
'all_dft.' Please use
remove_test_protocol command to delete the
existing test protocol. (TEST-1402)
Question 23. How do the results of dft_drc compare to your first run?
Fewer PRE-DFT violations, but new “cell is constant”

violations that reflect the 1-0-1 initialization sequence.
1 DLAT clock line not controlled
violation (D4)
6 Clock feeding multiple
clock/set/reset inputs violations (D12)
20 Data path affected by clock captured
by clock in trailing edge clock_port
violations (D14)
20 Clock_port not capable of capturing
data violations (D17)
3 OTHER VIOLATIONS
1 Cell is constant 0 violation (TEST-
504)
2 Cell is constant 1 violations (TEST-
505)

This page was intentionally left blank.

5 DFT Design Rule
Checks
Learning Objectives

• Perform DRC checks on RTL, a pre-scan netlist, and a
post-scan netlist
• Interpret the results of the DRC run from the dft_drc
report
• Obtain an ATPG test coverage estimate
Lab Duration:
45 minutes
DFT Design Rule Checks Lab 5-1

Lab 5
Overview
All files for this lab except for designs and libraries are located in the directory
lab5_drc.
Directory Structure
lab5_drc Current working directory
mapped_scan Will contain final netlist
../ref/db Library db files
Answers & Solutions

Lab 5-2 DFT Design Rule Checks

Lab 5
Instructions
During this lab, you will perform DFT design rule checks and various points in the
scan insertion flow.
Task 1. RTL DFT DRC
1. Invoke DC/DFTC.
UNIX% cd lab5_drc
UNIX% dc_shell | tee logs/rtl_drc.log
2. Read the RTL design into DFTC and create the test protocol.
s scripts/task1_setup.tcl
Question 1. What command verifies that the RTL design and the protocol
are compatible?
....................................................................................................
Question 2. What variable needs to be set to enable file and line number
tracking for RTL DFT DRC?
....................................................................................................
3. Run dft_drc
dft_drc
4. Analyze the dft_drc report

Question 3. Were there any violations? What type?
....................................................................................................
5. The TEST_MODE signal (active high) was not declared prior to creating the
test protocol. Declare the signal with set_dft_signal and rerun
dft_drc
set_dft_signal ...
create_test_protcol –capture_procedure multi_clock
dft_drc

Lab 5
Question 4. Are there any violations now?
....................................................................................................
6. Exit dc_shell
exit

Lab 5
Task 2. Pre-DFT DRC
1. Invoke DC/DFTC.
UNIX% dc_shell | tee logs/gate_drc.log
2. Read the design and perform a test-ready compile.
Question 5. What command is used to perform a test-ready compile?
....................................................................................................
Question 6. From looking at the logfile, can you tell if automatic shift
register identification was performed?
....................................................................................................
Question 7. What command verifies that the gate-level design and the
protocol are compatible?
....................................................................................................
3. Run pre-DFT DRC
dft_drc
Question 8. Are the results the same as when dft_drc was run on the
RTL? Is the report format the same?
....................................................................................................
4. Enable enhanced dft_drc reporting and re-run dft_drc
set test_disable_enhanced_dft_drc_reporting FALSE

dft_drc
Question 9. What is the difference in the dft_drc report?
....................................................................................................

Lab 5
Task 3. Post-DFT DRC
1. Specify scan constraints and preview the scan architecture that will result.
Question 10. How many scan chains does preview_dft report will be
inserted?
....................................................................................................
2. Insert the scan chain(s).
insert_dft
Question 11. What command verifies that the inserted scan chain follows
design rules?
....................................................................................................
3. Perform DFT DRC on the post-inserted netlist
dft_drc
Question 12. Were there any DRC violations?
....................................................................................................
Question 13. What section of the dft_drc report is now included when
performing DRC check on a post-scan inserted netlist?
....................................................................................................
Question 14. What dft_drc command option is used to report an ATPG test
coverage estimate after scan insertion?
....................................................................................................
4. Get a coverage estimate.
dft_drc ...
Question 15. What is the estimated coverage?
....................................................................................................

Lab 5

Answers / Solutions
Task 1. RTL DFT DRC

Question 1. What command verifies that the RTL design and the
dft_drc
Question 2. What variable needs to be set to enable file and line number
tracking for RTL DFT DRC?
set hdlin_enable_rtldrc_info true

Question 3. Were there any violations? What type?
Yes. 310 D1’s, 90 D3’s, 1 D15

310 Uncontrollable clock input of flip-flop violations (D1)
90 DFF set/reset line not controlled violations (D3)
1 Clock path affected by clock captured by clock in level
sensitive clock_port violation (D15)
set_dft_signal -view existing_dft -type Constant -

active_state 1 -port TEST_MODE
create_test_protcol –capture_procedure multi_clock
dft_drc
Question 4. Are there any violations now?
No.
Task 2. Pre-DFT DRC

Question 5. What command is used to perform a test-ready compile?
compile_ultra –scan (or “compile –scan”)

Question 6. From looking at the logfile, can you tell if automatic shift
register identification was performed?
Yes, it is enabled.
Note: Automatic shift-register identification is enabled for scan. Not all registers
will be scan-replaced (OPT-467)

Question 7. What command verifies that the gate-level design and the
dft_drc
Question 8. Are the results the same as when dft_drc was run on the
RTL? Is the report format the same?
Yes. Still 0 violations.

Question 9. What is the difference in the dft_drc report?
The results are now summarized in a table

--------------------------------------------------------------------------
Sequential Cell Report: Sequential Core Core
Cells Segments Segment Cells
--------------------------------------------------------------------------
Sequential Elements Detected: 303 0 0
Clock Gating Cells : 0
Synchronizing Cells : 0
Non-Scan Elements : 0
Excluded Scan Elements : 0 0 0
Violated Scan Elements : 0 0 0
(Traced) Scan Elements : 303 (100.0%) 0 0 ( 0.0%)
--------------------------------------------------------------------------
Task 3. Post-DFT DRC

Question 10. How many scan chains does preview_dft report will be
inserted?
One.
Question 11. What command verifies that the inserted scan chain follows
design rules?
dft_drc
Question 12. Were there any DRC violations?
No.

Question 13. What section of the dft_drc report is now included when
performing DRC check on a post-scan inserted netlist?
The “Traced Scan Chains” section
-------------------------------------------------------------------------
Traced Scan Chains
-------------------------------------------------------------------------
Chain "1" traced through 303 flip-flops
Question 14. What dft_drc command option is used to report an ATPG

test coverage estimate after scan insertion?
help –v dft_drc
The option is -coverage_estimate

Question 16. What is the estimated coverage?
99.77%

Introduction to
6 Design Vision
Learning Objectives

• Read a design and related files into Design Vision, the
graphical interface to Design For Test Compiler
(DFTC)
• Compile that design into a test-ready design
• Explore that design in Design Vision
• Save the test-ready design
Lab Duration:
45 minutes
Introduction to Design Vision Lab 6-1

Lab 6
Overview
lab6_gui.
Directory Structure
lab6_gui Current working directory
mapped_scan Will contain final netlist
../ref/db Library db files
Answers & Solutions

Lab 6-2 Introduction to Design Vision

Lab 6
Instructions
Task 1. Read a design into Design Vision

1. Invoke Design Vision.
UNIX% cd lab6_gui
UNIX% design_vision | tee logs/gui.log
Question 1. What is the difference between Design Vision and

dc_shell?
....................................................................................................
This is what is in your log

file of the current Design
Vision session.
Commands can be entered from here
Once in Design Vision, you can take advantage of its graphical interface for
debugging and troubleshooting problems, or work in dc_shell, taking advantage
of the speed of a shell.

Lab 6
2. Read the file containing the RISC_CORE entity or module, and read all
associated subfiles, by applying the following command:
Before executing the command, click on the Log tab in the lower left corner
of the GUI, and make the Session Window as large as possible, so that you
can follow the progress of the read.
acs_read_hdl -f vhdl \
-hdl_source ../ref/rtl/RISC_CORE/vhdl RISC_CORE
Question 2. What does the acs_read_hdl command do in addition to

reading in the model you specify?
....................................................................................................
Task 2. Compile and Save the Design

1. Constrain the design using the constraints file provided for you in the scripts/
directory called: constraints.tcl.
Design timing and area constraints are needed to direct the next step:
compile.
Note: These constraints drive the standard Design Compiler logic synthesis, and
are not the focus in this class. Do not spend your lab time trying to learn
them.
source -e -v constraints.tcl
Question 3. The constraint file is in the scripts subdirectory; you did not
have to specify the directory when you sourced the
constraints file, why not?
....................................................................................................
2. Compile the design using the following command:
compile_ultra -scan
The compile command synthesizes the design to the target library by taking
the GTECH (Generic Technology) library parts and translating them into the
target library parts (in this case, a 0.13μ training library).
Compile maps to ordinary D flip-flops by default. The –scan switch directs
DC to translate the GTECH sequential elements directly to special test
flip-flops called scan flops. This is called “test-ready compile” or “1-Pass
Scan Synthesis”.
3. Save the design using the command:

Lab 6
write -format ddc -hierarchy \

-output mapped/RISC_CORE.ddc
This command saves your design in Synopsys internal (DDC) format into the
mapped subdirectory. The name of the file is RISC_CORE.ddc. Saving all
the sub blocks in the design is enabled by the –hierarchy switch.

Lab 6
Task 3. Explore the Design

When inserting scan into your design, problems may occur. Being able to explore
your design graphically is a powerful debugging aid.
1. Resize your Session Window, making it smaller so that you can focus on the
graphical display at the top portion of the GUI.
The following figure shows some key points for navigating in Design Vision.
Display a schematic. Click on the orange AND

gate to see a schematic of the design. To look
at a schematic for a particular sub-block, select
that sub-block in the Design Overview window
and then click on the orange gate.
Current Design. In this

case, RISC_CORE is the
current design.
Design Overview Window. Once the designs have been read into
Design Vision, this window shows the top-level block and all of the
subblock names.
A + sign indicates subblocks exist beneath
For instance, select the + sign next to I_STACK_TOP to see
the subblocks of the STACK_TOP block
Transcript Window: Here

it shows the results of the
link command. Many
command return 1 and 0 to
Use the Options
indicate success and
pulldown to search the
Command failure, respectively. Be
log for Errors, Warnings,
history log aware that other
Informational messages,
commands return
or commands. You can
collections or other values.
also use F3.
Question 4. What is the current design?
....................................................................................................

Lab 6
2. Display the current design from the TCL shell.
You can also display the current design from the TCL shell by using the
current_design command. Try it:
current_design
3. Display all the designs from the TCL shell by using the list_designs
command.
list_designs
4. Display all the designs graphically.
Select the pull down arrow to the right of the current design; the results match
those from the list_designs command.
5. Display a schematic of the entire RISC_CORE ASIC.
Specify which part of the ASIC you want to see.

In this case, select RISC_CORE from the Design Overview Window.
Notice a list of the subdesigns below RISC_CORE is displayed
Notice the Cell name and ref name of each subdesign are shown. Cells are
also called instances. Ref names are also called designs.

Lab 6
6. Display a schematic of the RISC_CORE ASIC.
Select the orange AND gate in the toolbar. You should see a schematic of the
RISC_CORE ASIC.
You can zoom in and out to see things more clearly.
7. Zoom in to examine the schematic in more detail.
Select the view pull-down menu to zoom in and out
Note: You may first have to select an area of the schematic which has nothing
in it with the mouse to be able to use the Zoom features.
Once in Zoom In mode, you may need to click on the Selection Tool
before changing to another mode, such as Zoom Out. The escape key
(ESC) is the shortcut for the Selection Tool.
You can also use the middle mouse button to zoom and pan, using so
called strokes, e.g. depress the middle mouse button then move the mouse
from SW to NE. You should have zoomed in.
8. Display the symbol view (input and output ports) of the RISC_CORE ASIC.
Select the green IC chip icon next to the orange AND gate.

Lab 6
9. Verify that your entire design is mapped to the gates.
The report_hierarchy command shows you the entire hierarchy of the

design in a textual format, as well as the cells in the design and libraries to
which they belong. In this case, you are using the tsmc library. If any cells
belong to the gtech library, you can conclude that there are unmapped cells in
your design.
Run the command:
report_hierarchy
10. Compare the previous result to an unmapped design.
Remove all of the designs, bring in the unmapped design once more, and run
the report_hierarchy command:
remove_design –design
–hdl_source ../ref/rtl/RISC_CORE/vhdl RISC_CORE
report_hierarchy
Notice the RPT-7 message and that all the cells are GTECH library cells
because the design is unmapped.
11. Verify that all the library variables have been defined correctly.
printvar *_library
12. Exit Design Vision
exit

Lab 6
Task 4. Debug Violations using the GUI
13. Invoke Design Vision
UNIX% design_vision | tee logs/debug.log
14. Read the mapped design saved in the previous task
read_ddc mapped/RISC_CORE.ddc
15. Run a design rule check.
dft_drc
Question 5. What error message do you get?
....................................................................................................
16. Resolve DFTC’s complaint; that is, create a test protocol.
17. Run a design rule check again.
dft_drc
Question 6. Is the design free of DFT problems or does the report contain
many violations?
....................................................................................................
18. Look at the GUI. You should see a Violation Browser showing the same
errors shown by the command output of dft_drc. You can bring up the
violation browser any time using the menu Test Æ Browse Violations.
19. Click on the “+” infront of “PRE-DFT”, then click on the D1 category.
Select the first violation on the right. On the bottom, you will see an
explanation of the error.
20. Right click on the first D1 violation, then select “Inspect Violation”, or click
on the “Inspect Violation” button on the lower right. A schematic should
appear.
Question 7. What logic value is present on the clock?
....................................................................................................
Lab 6
Question 8. What DFTC command is needed to specify this test clock?
....................................................................................................
21. Type the command that you just noted (check the back for help).
22. Recreate the test protocol, then switch back to the Violation Browser, and
click on “Run DFT DRC”. The schematic window should disappear.
Note: The test protocol is recreated by simply re-running the
create_test_protocol command.
Question 9. Was the number of violations affected at all?
....................................................................................................
23. Once again, “Inspect Violation” for the first D1 violation shown.
Question 10. What value is shown for TEST_MODE?
....................................................................................................
Question 11. What DFTC command is needed to specify this test mode?
....................................................................................................
24. Type the command that you just noted (check the back for help).
click on “Run DFT DRC”.
....................................................................................................
26. Analyze the first D3 violation.
Question 13. What value is shown for Reset?
....................................................................................................
Question 14. What DFTC command is needed to specify this asynchronous

reset (the reset is active low)?
....................................................................................................
....................................................................................................
27. Type the command you just noted down (check the back for help).
click on “Run DFT DRC”.

Lab 6
....................................................................................................
29. Exit Design Vision.

Answers / Solutions
Task 1. Read a design into Design Vision

Question 1. What is the difference between Design Vision and
dc_shell?
Design Vision (design_vision) is the GUI to the Synopsys

logic synthesis tools; it is built on dc_shell, which is
the command line interface.
Question 2. What does the acs_read_hdl command do in addition to

reading in the model you specify?
It finds the files containing all the designs and subdesigns,

reads them in, and links them, setting the
current_design to the top-level design.
Task 2. Compile and Save the Design

Question 3. The constraint file is in the scripts subdirectory; you did not
have to specify the directory when you sourced the
constraints file, why not?
The “scripts” directory was appended to the

search_path variable in .synopsys_dc.setup. You can
verify this by using ”printvar *path”.
Task 3. Explore the Design

Question 4. What is the current design?
The current design is: RISC_CORE (top-level of the ASIC)
Task 4. Debug Violations using the GUI

Question 5. What error message do you get?
Error: No model found on design

0

Question 6. Is the design free of DFT problems or does the report

contain many violations?

303 Uncontrollable clock input of
flip-flop violations (D1)
1 DFF set/reset line not
controlled violation (D2)
90 DFF set/reset line not
controlled violations (D3)
Question 7. What logic value is present on the clock?
X meaning unknown or unspecified in this case.
Question 8. What DFTC command is needed to specify this test clock?
set_dft_signal -view exist –type ScanClock \

-timing {45 55} -port Clk
Yes, the number of D1 violations goes down to 300.
Question 10. What value is shown for TEST_MODE?
Question 11. What DFTC command is needed to specify this test mode?
set_dft_signal -view exist -type Constant \

-active 1 -port TEST_MODE
All D1 violations are gone now. 1 D2 and 90 D3 violations

remain.
Question 13. What value is shown for Reset?
Question 14. What DFTC command is needed to specify this

asynchronous reset?
set_dft_signal -view exist -type Reset \

-active 0 -port Reset
All violations should be gone now.
Question 16. How do you optimize, map, and scan-replace the flip-flops
in a design with Design Vision?
Use the compile –scan command
Command Sequence For This Lab
–hdl_source ../ref/rtl/RISC_CORE/vhdl RISC_CORE
source -e -v constraints.tcl
compile_ultra -scan
write –format ddc –hier -output mapped/RISC_CORE.ddc
set_dft_signal -view exist -type ScanClock \
-timing {45 55} -port Clk
set_dft_signal -view exist -type Constant \
-active 1 -port TEST_MODE
set_dft_signal -view exist -type Reset \
-active 0 -port Reset
dft_drc

Appendix: Searching Synopsys Online

Documentation using Indexes
1. Synopys stores its product documentation in a collection called Synopsys

Online Documentation (SOLD). In addition to the command man pages,
SOLD is a good place for you to look for more information about Synopsys
tools. To access SOLD type sold from the UNIX command line followed by
an ampersand (&).
UNIX> sold &
sold is a UNIX alias which invokes Adobe Acrobat reader and opens up a
file called top.pdf, which is the top-level view of all documents.
2. To search for information in SOLD and look only in specific documents, you
need to set up indexes to customize the searches. SOLD consists of a host of
documentation for the various Synopsys tools. Each set of documentation is
accessed through its corresponding index. For instance, there is an index for
the Man Pages and one for Test Automation, as well as many others. When
the indexes are set up properly, you can use SOLD to search in one or many
indexes to control how much information you will receive from any search
you perform.
3. For information on how to setup indexes consult this SolvNet article
https://solvnet.synopsys.com/retrieve/903898.html “Script for building SOLD
search indexes”.
4. To search for information in a SOLD index, first select the binoculars icon
(the binoculars icon with documents behind it is the Search button).
Select the second

Binocular Icon to
perform a search

Debug and Fix DFT
7 DRC Violations
Learning Objectives

• Debug dft_drc problems with clocks and resets
• Propose a manual solution that could be implemented by
you or by another designer
• Use AutoFix to implement a fix
Lab Duration:
45 minutes
Debug and fix DFT DRC Violations Lab 7-1

Lab 7
Overview
lab7_fixing.
Directory Structure
Lab7_fixing Current working directory
../ref/rtl/RISC_CORE_nodft/vhdl Design files
./.solutions Solution files
Answers & Solutions

Lab 7-2 Debug and fix DFT DRC Violations

Lab 7
Instructions
During this lab you will debug Uncontrollable Clock and Reset DFT problems using
the Design Vision GUI, propose a fix and implement a fix using AutoFix.
Task 1. Read in Mapped Design

1. Change directories to the project directory lab7_fixing.
2. Invoke DC in the GUI mode.
UNIX% design_vision
3. Read in the mapped design. Remember file name completion!
source scripts/4read_gate_and_protocol.tcl
Task 2. Debug Uncontrollable Clocks and Resets

1. From information in the log window, answer the following:
Question 1. How many violations are there?
What type of violations are they?
...................................................................................................
...................................................................................................
...................................................................................................
2. In the Violation Browser, analyze the first D1 violation and you should
observe a schematic similar to the one shown below.
Figure 1.

Lab 7
Question 2. What is the netlist name for the flip-flop on the left, the
cause for the D1 violation?
...................................................................................................
Question 3. Draw a schematic “fix” for this uncontrollable clock

problem.
...................................................................................................
3. Analyze some of the other D1 violations. Select the Violation Browser tab,
then select some of the other D1 violations and examine the schematic.
Question 4. What can you conclude about the D1 violations in the
RISC_CORE design?
...................................................................................................
Analyze one of the D3 violations in the same manner as you did the D1
violations. To expand the second X input to I_RST/q2_reg, double click on
the D input.
Question 5. Based on your analysis draw a schematic “fix” for the

uncontrollable reset problem.
.. ..................................................................................................

Lab 7
4. Select some of the other D3 violations and analyze them.

RISC_CORE design?
....................................................................................................
Task 3. Autofix The D1, D2, and D3 violations

You have now analyzed both the uncontrollable clock and uncontrollable reset DFT
violations in the RISC_CORE design. Your next task is to fix the violations. As
you learned in the lecture there are three choices.
1) You could re-code the RTL design and check the design once again using
the unmapped DFTflow. When you have no more violations you can
continue on to the mapped flow.
2) You could draw the solution and have someone else re-code the design for
you.
3) You can use the AutoFix capability of DFTC.
Question 7. What kind of DFT problems can Autofix solve?
...................................................................................................
...................................................................................................
1. The 5preview_dft.tcl file is setup to source the autofix.tcl script after

sourcing the settings_insert_dft.tcl file.
Set up to run AutoFix by modifying a script called scripts/autofix.tcl.
Utilize the lecture material and the man pages to modify the script called
autofix.tcl in the scripts/ directory. Enable AutoFix to fix uncontrolled
clocks, uncontrolled resets, and uncontrolled sets.
Here is a list of commands related to AutoFix to get you started:
• set_dft_configuration
• set_autofix_configuration
• set_dft_signal
2. Preview what AutoFix will add by running step 5 of the Mapped Flow.
source –e –v scripts/5preview_dft.tcl
3. Evaluate the previewed AutoFix results.

Lab 7
Take a careful look at the results of running the preview_dft command.

Notice the sections of the report titled Test Point Plan Report, and DFT
signals. These sections relate directly to the AutoFix portion of the flow.
Compare your results carefully with those in the Answers / Solutions
section—it is easy to make a mistake and not realize it.
Question 8. How many AutoFix test points are there? Why?
...................................................................................................
Question 9. What do you think would happen if a register had an

unclontrolled reset and and uncontrolled set and the same
top-level signal was specifyed as the TestData?
...................................................................................................
4. Perform both AutoFix and scan insertion by running step 6 of the DFT flow.
source –e –v scripts/6insert_dft.tcl
5. Look at the results of runing dft_drc –coverage_estimate.
Question 10. What were the number of violations?

Were there any cells with violations?
What was the fault coverage?
...................................................................................................
...................................................................................................
Question 11. What kind of fault coverage would you expect to get, if you
had not AutoFixed the design?
...................................................................................................
6. Try it and compare coverages.

7. Exit DFTC.

Answers / Solutions
Task 2. Debug Uncontrollable Clocks and Resets

Question 1. How many violations are there? What type of violations are
they?

300 Uncontrollable clock input of flip-
flop violations (D1)
1 DFF set/reset line not controlled
violation (D2)
88 DFF set/reset line not controlled
violations (D3)
Question 2. What is the netlist name for the flip-flop on the left, the
cause for the D1 violation?
I_DIV_CLK/q_reg
Question 3. Draw a schematic “fix” for this uncontrollable clock
problem.
RISC_CORE
DIV_CLK ALU
TEST_MODE
Carry_Flag_Reg
Q D Q D Q
....
q_req 0 SE
F0
SE
FN
1
Clk
Note: The image above only shows I_DIV_CLK/q_req/Q going

to the ALU block in RISC_CORE, but it is actually
connected to several other blocks (a total of 310 FFs).

RISC_CORE design?
All the D1 violations are coming from the same launch

point and have the same solution.

Question 5. Based on your analysis draw a schematic “fix” for the

uncontrollable reset problem.
RISC_CORE
RST ALU
TEST_MODE +Vdd
Carry_Flag_Reg
D Q D Q D Q D Q
q1_req q2_req F0 .... FN

0 SE SE
RST

RISC_CORE design?
The D3 violations all come from the same launch point and
the solution in Question 5 is adequate to fix all of the D3
violations.
Task 3. AutoFix The D1, D2, and D3 violations

Question 7. What kind of problems can AutoFix solve?
Uncontrollable clocks and sets/resets, contention issues

during scan shift on internal tristate busses and bidirectional
ports.
Set up to run Autofix by modifying the 5preview_dft.tcl file.
Contents of file autofix.tcl (also available in the .solutions
directory):

set_dft_configuration -fix_set enable \

-fix_reset enable -fix_clock enable
set_dft_signal -view spec -type TestMode \
-port TEST_MODE
# use the clock Clk as the autofix clock

set_dft_signal -view spec -port Clk -type TestData
set_autofix_configuration -type clock \
-control TEST_MODE -test_data Clk
# Use the Reset line as the reset fix

set_dft_signal -view spec -port Reset -type TestData
set_autofix_configuration -type reset –method mux \
-control TEST_MODE -test_data Reset
# Use the same Reset line to fix the sets

set_autofix_configuration -type set -method mux \
-control TEST_MODE -test_data Reset
************ Test Point Plan Report ************

Total number of test points : 2
Number of Autofix test points: 2
Number of Wrapper test points: 0
Number of test modes : 1
Number of test point enables : 0
Number of data sources : 2
Number of data sinks : 0
**************************************************
Dft signals:
TestMode: TEST_MODE (no hookup pin)
TestData: Clk (no hookup pin)
TestData: Reset (no hookup pin)
Question 8. How many AutoFix test points are there? Why?
2 AutoFix test points. One to fix the uncontrollable clocks

and one to fix the uncontrollable resets and sets.
Question 9. What do you think would happen if a register had an

unclontrolled reset and uncontrolled set and the same top-
level signal was specifyed as the TestData?
This situation would create a race condition if the same

signal controlled both the reset and set pin of scan register.
Therefore, DFTC will not allow this condition and will
create a new port to fix the sets.
If in this situation and you don’t want to use a separate port

to fix the sets, you have a couple options:
- Use a the “gate” method to fix all sets and/or resets
- Use set_autofix_element to specify the “gate” method to
fix uncontrolled sets or resets for only the specific
registers that have both and uncontrolled reset and set.
Question 10. What were the number of violations?

Were there any cells with violations?
What was the fault coverage?
0 violations.
No cells with violations.
99.73%
Question 11. What kind of fault coverage would you expect to get, if you
had not AutoFixed the design?
Very low (0.36%) because almost every flip-flop in the

design will not be included on the scan chain.

8 Top-Down Scan Insertion
Learning Objectives

• Use the preview_dft command to evaluate different scan
architectures and iterate by revising scan specifications
until you produce balanced top-level scan chains.
• Use the preview_dft command to confirm your scan
specification for desired scan signals was applied
correctly.
Lab Duration:
45 minutes
Top-Down Scan Insertion Lab 8-1

Synopsys 30-I-011-SSG-012
Lab 8
Overview
During this lab you will:

1. Preview the results of different scan specifications before implementing one that
yields well-balanced top-level chains.
2. Perform the Mapped DFT flow for a test-ready design through scan insertion and
test coverage estimation.
Unmapped Flow:
1. Read the RTL design into DFT.
2. Create and save a test protocol; verify that the design and the test protocol are
compatible.
3. Compile and save the gate-level design; exit.
Mapped Flow:
4. Read the gate-level design and test protocol into DFTC.
5. Specify scan constraints and preview the scan architecture that will result from
applying them.
6. Insert the scan chain.
7. Hand off the design and related files for use by downstream tools; exit.
Lab 8-2 Top-Down Scan Insertion

Lab 8
File Locations
lab8_topdown.
Directory Structure
Lab8_topdown Current working directory
../ref/rtl/ORCA/vhdl Design files

Lab 8
Instructions
Your goal is to complete and run the first three steps of the Mapped Flow portion of
the DFT Insertion Flow. The three you will modify are for reading the test-ready
design and test protocol, previewing the scan architecture, inserting scan chains and
finally, estimating test coverage.
The three scripts:
Unmapped Flow
1read_design.tcl
2create_test_protocol.tcl
3compile_and_save.tcl
Mapped Flow
4read_gates_and_protocol.tcl
5preview_dft.tcl
6insert_dft.tcl
7handoff.tcl
Task 1. Read Gate-level Design and its Test Protocol
1. Invoke DFTC.
cd lab8_topdown
dc_shell | tee logs/lab8.log
2. Examine and execute the script that performs step 4 of the DFT flow: reading
in the gate-level design and the saved protocol.
exec cat scripts/4read_gate_and_protocol.tcl
s scripts/4read_gate_and_protocol.tcl

Lab 8
3. Check the scan state of the current design using this command:
report_scan_state
Question 1. If the Unmapped Flow ended with a compile –scan

command, what should the scan state of the design you just
read in be? Is that the design’s current scan state?
....................................................................................................
At the end of the 4read_gate_and_protocol.tcl, a DFT check was performed.

4. Examine the last several lines in your transcript to review the DFT check
results for the test-ready design.
Question 2. What percentage of the total number of flip-flops are valid
scan cells?
....................................................................................................
Question 3. Considering just that percentage and no other factors, what

range of estimated fault coverage would you predict for this
design?
....................................................................................................
5. Enable enhanced DRC reporting and re-run dft_drc
set test_disable_enhanced_dft_drc_reporting FALSE

dft_drc
Question 4. Does the percentage reported by the enhanced dft_drc

reporting match your expectations?
....................................................................................................
Note: Step 4 of the DFT flow reads the gate-level design and the test
protocol. The dft_drc command used the saved test protocol for DFT
design rule checking. The test protocol is one set of information and
the scan specifications that may have been used to originally create the
test protocol is another set of information.
6. Check this by reporting which dft signals are currently defined.
report_dft_signal

Lab 8
Question 5. How many ScanClocks and Resets are reported?
....................................................................................................
Task 2. Specify and Preview Scan Architectures
1. Display the default scan architecture that will be implemented for this design.
preview_dft
Question 6. How many scan chains will be inserted?

What are the names of the preview scan chains?
Will the chains be well-balanced?
....................................................................................................
....................................................................................................
....................................................................................................
Question 7. Will scan insertion use existing functional pins as the scan
signals (scan in, scan out for each scan chain) or will it create
new ones (test_si, test_so ports)?
....................................................................................................
2. Display more scan architecture detail about the test clocks used for each scan
chain.
preview_dft -show scan_clocks
Note: The ORCA design has only 3 clocks in test mode: pclk, sys_clk and
sdr_clk.
Question 8. Why are four scan chains previewed by default?
....................................................................................................

Lab 8
3. Use another preview_dft option and compare the content of the former
and current preview_dft output.
preview_dft -show scan_summary
Question 9. Is there any information in the scan_clocks preview that is not

also reported in the scan_summary preview?
When might you want to use the format of the scan_summary
preview rather than the first?
....................................................................................................
....................................................................................................
.....................................................................................................
4. Try mixing the rising and falling sdr_clk scan flip-flops into the same chain.
set_scan_configuration -clock_mixing mix_edges

Question 10. How many scan chains are previewed now and are they now
balanced?
....................................................................................................
5. Allow mixing of different clock domain scan flip-flops into the same scan
chain.
set_scan_configuration -clock_mixing mix_clocks

Question 11. How many scan chains are previewed now and what do you
think the default set_scan_configuration
–chain_count value is?
....................................................................................................
Question 12. What specifications would yield 6 balanced scan chains?
....................................................................................................

Lab 8
6. Apply those specifications (6 chains) and confirm using preview_dft.
set_scan_configuration # specify your options

Question 13. What is the difference in scan chain length between the
longest and shortest chains now?
....................................................................................................
....................................................................................................
Question 14. Will DFTC insert any lock-up latches into this scan design?
Which chains will include them?
....................................................................................................
....................................................................................................
Note: The functional pins of ORCA to use for scan are specified in the
settings_insert_dft.tcl script.
7. Source that script and use preview_dft to answer the following questions.
s scripts/settings_insert_dft.tcl
preview_dft
Question 15. What are the names of the previewed scan chains now and
what command defined their new names?
Why were the chains given explicit names?
....................................................................................................
....................................................................................................
Question 16. Which pins are used for scan in, scan out and scan enable?
....................................................................................................
.....................................................................................................

Lab 8
Task 3. Insert Scan Chains and Estimate Test Coverage
One of the commands in the settings_insert_dft.tcl script avoids all the subdesigns
being renamed during the scan insertion process.
Question 17. What is the name of that command?
....................................................................................................
Note: Use the man pages if needed to find the option to this command that
disables all the gate-level optimization that occur by default during the
last phases of the insert_dft process.
Question 18. Which option to this command will disable gate-level

optimization?
....................................................................................................
1. Apply those two options whose effects will be to speed-up the runtime of the
subsequent insert_dft and to limit the number of changes to the existing
design.
set_dft_insertion_configuration . . .
2. Perform scan insertion using the script provided.
s scripts/6insert_dft.tcl
Question 19. What is the estimated test coverage for the ORCA scan
design?
....................................................................................................
Note: Actual ATPG runs in TetraMAX obtained about 99% test coverage for
this design. Doing so though required using functional RAM models
and both TetraMAX’s Fast Sequential and Full Sequential engines.
Using DFT techniques to avoid the S30 violation, for example, allows
that very high coverage much easier to obtain by merely using the Fast
Sequential engine.
Question 20. How does this estimate compare to your earlier prediction?
....................................................................................................
Note: When obtaining the estimated fault coverage for a scan inserted
design, dft_drc performs Post-DFT checking and reports the DFT
violations using the same ones that TetraMAX does.

Lab 8
Note: You can obtain detailed information for both the D*

PRE-DFT violations and the {C,S,Z,X}* Post-DFT violations using
the man command in DFTC.
3. Scroll back from the coverage report and answer the following questions.
Question 21. The S category of violations cannot be reported during
Pre-DFT checks, why?
....................................................................................................
Question 22. What does the S22 violation mean and does this explain why
2 Sequential Cells are reported as synchronization elements in
the Sequential Cells Without Violations summary?
....................................................................................................
....................................................................................................

Answers / Solutions
Task 1. Read gate-level Design and its Test Protocol

Question 1. If the Unmapped Flow ended with a compile –scan
command, what should be the scan state of the design you
just read in? Is that the design’s current scan state?
The scan state should be “test-ready” which is reported as

“scan cells replaced with loops”.
Question 2. What percentage of the total number of flip-flops are valid
scan cells?
2925 / (2925 + 33) equals 98.9%

Question 3. Considering just that percentage and no other factors, what
range of estimated fault coverage would you predict for this
design?
Close to 99% test coverage.

Question 4. Does the percentage reported by the enhanced dft_drc
reporting match your expectations?
Yes. It reports 98.9%.

Question 5. How many ScanClocks and Resets are reported?
Three ScanClocks (sys_clk, sdr_clk, and pclk) and one

Reset (prst_n) are reported. The DFT signals are defined
when the test protocol is read with the read_test_protocol
command.

Question 6. How many scan chains will be inserted?
What are the names of the preview scan chains?
Will the chains be well-balanced?
4 scan chains
Named ‘1’, ‘2’, ‘3’, and ‘4’.
No, the longest scan chain has 1074 cells while the shortest
chain has half that many, 512, a very unbalanced scan chain
architecture.

Question 7. Will scan insertion use existing functional pins as the scan
signals (scan in, scan out for each scan chain) or will it
create new ones (test_si, test_so ports)?
New test_si and test_so ports are previewed because you

have not yet specified your scan signals.
Question 8. Why are four scan chains previewed by default?
Default is to not mix clocks or edges on scan chains. There

are three clock domains. The sdr_clk domain has both
rising-edge and falling-edge triggered flip-flops.
Question 9. Is there any information in the scan_clocks preview that is
not also reported in the scan_summary preview?
When might you want to use the format of the
scan_summary preview rather than the first?
The previewed information is almost the same. The key

difference is that the scan_clocks preview gives the head
and tail of each scan chain, plus the first scan flip-flop
whenever the scan path crosses clock domains. The
scan_summary only previews the head scan flip-flop. The
scan_summary format of the preview report might be easier
to parse with Perl or TCL.
Question 10. How many scan chains are previewed now and are they now
balanced?
Three chains, still unbalanced with the longest chain having

1132 cells and the shortest chain having 719 cells.
Question 11. How many scan chains are previewed now and what do you
think the default set_scan_configuration \
–chain_count value is?
Only one chain, the default for –chain_count.

Question 12. What specifications would yield 6 balanced scan chains?
set_scan_configuration \
-chain_count 6 -clock_mixing mix_clocks
Question 13. What is the difference in scan chain length between the
longest and shortest chains now?
Only one, with the longest chains being 488 cells long and
the shortest 487.

Question 14. Will DFTC insert any lock-up latches into this scan design?
Which chains will include them?
Yes.
Chain ‘4’ has both pclk and sdr_clk rising-edge flops at time
45 and Chain ‘5’ has both sdr_clk and sys_clk rising-edge
flops at time 45.
Question 15. What are the names of the previewed scan chains now and
what command defined their new names? Why were the
chains given explicit names?
Chain names: chain0, chain1, chain2, chain3, chain4 and

chain5.
The set_scan_path command was used to define the scan
chain names.
To associate a given pair of ScanDataIn and ScanDataOut
scan signals to a specific chain. You need to specify the
name of chain when you apply the set_scan_path
command.
pads[0] to pad[5] are the scan in pins and sd_A[0] to

sd_A[5] are the corresponding scan out pins. The scan
enable port is called scan_en.
Task 3. Insert Scan Chains and Estimate Test Coverage

Question 17. What is the name of that command?
set_dft_insertion_configuration
Question 18. Which option to this command will disable gate-level
optimization?
set_dft_insertion_configuration -synthesis none \

-preserve_design_name true
Question 19. What is the estimated test coverage for the ORCA scan
design?
~95%

Question 20. How does this estimate compare to your earlier prediction?
This is probably lower than what you predicted, likely 99%
Question 21. The S category of violations cannot be reported during

Pre-DFT checks, why?
There are no scan chains then, and the S rules are related to
problems with scan chain shifting.
Question 22. What does the S22 violation mean and does this explain
why 2 Sequential Cells are reported as synchronization
elements in the Sequential Cells Without Violations
summary?
S22 warns that ORCA has scan chains that are driven by
more than one test clock. Scan chains chain3 and chain4
have lock-up latches and thus avoid the S22 violation. The
S22 warnings about a negative-edge triggered scan flop in
the sdr_clk domain being stitched directly to a rising-edge
triggered scan flop in the pclk domain.

Scan Design Handoff
9
Learning Objectives

• Name the 3 types of DFTC output files downstream
tools require specifically related to scan testing.
• Write out the final test protocol and a scandef file.
Lab Duration:
30 minutes
Scan Design Handoff Lab 9-1

Lab 9
Overview
Unmapped Flow:
1. Read the RTL design into DFT.
2. Create and save a test protocol; verify that the design and the test protocol are
compatible.
3. Compile and save the gate-level design; exit.
Mapped Flow:
5. Specify scan constraints and preview the scan architecture that will result
from applying them.
Answers & Solutions

help with answering a question, check the back portion of this lab for help.
Lab 9-2 Scan Design Handoff

Lab 9
File Locations
lab9_export.
Directory Structure
Lab9_export Current working directory
../rtl/ORCA_init/vhdl Design files
../libs Library files

Lab 9
Instructions
Given a mapped design, export from DFTC the files required by downstream tools
such as ATPG and P&R.
Task 1. Execute the Mapped Flow and Export Files
1. Change directories to the project directory lab9_export.
UNIX% cd lab9_export
2. Invoke DFTC.
dc_shell
4. Specify scan constraints and preview the scan architecture that will result
from applying them by using the script provided.
s scripts/settings_insert_dft.tcl
preview_dft
Question 1. Explain why the 3 configuration registers will not be placed

in the scan chains?
....................................................................................................
....................................................................................................
s scripts/6insert_dft.tcl
Question 2. What is the estimated test coverage for the scan design?
....................................................................................................

Lab 9
6. Complete the script that writes out the design and related files to be used by
downstream tools (scripts/7handoff.tcl).
Add to this script:
• The command to save the needed ATPG SPF file (to the ./tmax
directory) that is not currently being saved.
• The command to write out the scan chains to a SCANDEF file (to the
mapped_scan directory).
• The command to check the SCANDEF
unix% vi scripts/7handoff.tcl
7. Execute the handoff script and examine that additional file.
s scripts/7handoff.tcl
Question 3. Is the custom initialization sequence still there?

What has changed about it?
....................................................................................................
....................................................................................................
The test protocol for preview_dft contains your scan chain in and out
signal declarations.
Question 4. What additional information has DFTC added to the
ScanStructures section now that the protocol is for a scan
inserted design?
....................................................................................................
Question 5. How many scanchains does the SCANDEF file show, and
how many partitions were created (use grep or consult the
check_scan_def report)?
....................................................................................................
8. Exit dc_shell
exit
Task 2. Verify Exported Files in TetraMAX
1. Change directories to the TetraMAX directory tmax.

Lab 9
UNIX% cd tmax
2. Inspect the TetraMAX run script. Ensure that the paths for the scan inserted
netlist and STIL protocol file referenced in the script match the file written out
by DFT Compiler.
unix% vi orca_tmax.tcl
3. Run TetraMAX.
tmax –tcl orca_tmax.tcl
Question 6. What is the Test Coverage reported by TetraMAX?
....................................................................................................
Question 7. How does the Test Coverage compare to the estimated

coverage reported in DFTC? Can you explain the reason for
the difference?
....................................................................................................

Answers / Solutions
Solution scripts are available in the .solutions directory.
Task 1. Execute the Mapped Flow and Export Files

Question 1. Explain why the 3 configuration registers will not be placed
in the scan chains?
The PRE-DFT violations of “cell is constant 0/1” mean

DFTC cannot shift arbitrary logic 0s and 1s through these
cells as required for scan shift; they are always 0 or
always 1.
Question 2. What is the estimated test coverage for the scan design?
The test coverage is around 92%.

Question 3. Is the custom initialization sequence still there? What has
changed about it?
Yes. In the unmapped_flow.spf file, the test_setup

vectors you added may have specified the same values for
the pins for adjacent Vector statements. In the STIL file
DFTC wrote for TetraMAX, notice “pclk”=P is only done
once, for the first configuration cycle.
Question 4. What additional information has DFTC added to the
ScanStructures section now that the protocol is for a scan
inserted design?
The ScanStructures section now includes for each chain a

ScanLength specification that declares how many scan cells
are in each chain.
Question 5. How many scanchains does the SCANDEF file show, and
how many partitions were created?
Use “grep PARTITION mapped_scan/ORCA.scandef”.

9 scan chains and 4 partitions.
Task 2. Verify Exported Files in TetraMAX
Question 6. What is the Test Coverage reported by TetraMAX?
Over 96 %

Question 7. How does the Test Coverage compare to the estimated

coverage reported in DFTC? Can you explain the reason for
the difference?
The difference can be attributed to TetraMAX running fast

sequential ATPG (enabled by “set_atpg –
capture_cycles 4”). The test coverage estimate
provided by “dft_drc –coverage_estimate” only
runs basic scan ATPG (combination ATPG).

Improving Scan Insertion
10 Run-Time and Capacity
Learning Objectives

• Given existing scan insertion scripts, implement RSS in
a top-down scan insertion flow achieving well-balanced
scan chains
• Modify a bottom-up scan insertion script for full gate-
level designs to use test models with RSS and
run it
• Preview top-level chain balance using test models after
block level scan insertion and revise the block level scan
architecture as needed to improve top-level scan chain
balance
Lab Duration:
75 minutes
Improving Scan Insertion Run-Time and Capacity Lab 10-1

Lab 10
Overview
Lab 10 Tasks
Perform Default Top-Down

Scan Insertion
Revise Specification to
Achieve Balanced Chains
Modify Bottom-Up Script to

Use Test Models
Revise Block Level scan to

Improve Top-level Balance
Analyze Estimated versus

Actual Test Coverage
Answers & Solutions

help with answering a question, check the back portion of this lab for help.
Lab 10-2 Improving Scan Insertion Run-Time and Capacity

Lab 10
File Locations
All files for this lab are located in the directory lab10_hicap.
Directory Structure
Lab10_hicap Current working directory
dc_ilms ILMs with test models for subblocks
mapped ORCA design and its subblocks
mapped_scan Scan-inserted design and subblocks
reports DFT reports
scripts Scan insertion scripts
test_models Test models for subblocks
tmax TetraMAX data
.solutions/scripts Solutions for this lab
Relevant Files
scripts/
bottom_up.tcl Bottom-up scan insertion script.
Calls the following two scripts:
scan_blocks.tcl Inserts scan in to the subblocks
scan_top.tcl Inserts scan in to the top-level
top_down.tcl Top-Down scan insertion script

Lab 10
Instructions
During this lab you will perform a top-down and a bottom-up scan insertion flow on
the design ORCA. Some scripts are already provided. You will modify the scripts
for the various tasks.
These scripts contain almost all the DFTC commands you would need to use when
working on a real-world design back in your office. After completing this lab, you
will have practiced the steps needed to improve both the run-time and the capacity
of your scan insertion scripts. When you re-use the scripts back at your office, focus
on the design-specific items (names of clocks, resets, and so forth).
The lab starts with a top-down script since it is easier to understand and provides
reference data (number and length of top-level scan chains, estimated test coverage)
that you can later use to compare the results of the bottom-up scan insertion script.
Task 1. Top-Down Scan Insertion using RSS
1. Make sure your current directory is lab10_hicap:
UNIX% cd lab10_hicap
2. Open the file scripts/top_down.tcl in your favorite editor (vi, emacs, etc.).
The top_down.tcl script is used to insert scan in a top-down fashion on the
ORCA design.
UNIX% vi scripts/top_down.tcl
3. Modify this script to use the RSS features you learned about in lecture.
Compared to the original script in the lab directory, your revised script should:
a. Modify the original design as little as possible
b. Run scan insertion faster
c. Consume less memory
If needed, check the answers section for hints.
Once you are done, save the modified script file.

Lab 10
4. Start Design Compiler from within the lab10_hicap directory:
UNIX% dc_shell | tee logs/lab10.log
5. Execute the script you just edited:
s scripts/top_down.tcl
6. Open the report generated by preview_dft and analyze it.

Question 1. Are the scan chains well balanced? If not, why?
....................................................................................................
Note: It is important to balance the scan chains if tester time and the cost of
test are to be reduced.
7. Open the top_down.tcl file again and change it so that you can achieve
balanced top-level scan chains.
8. Remove the first pass of scan design files before improving the scan
architecture.
remove_design -design
9. Rerun the top_down.tcl script in DFTC.

Question 2. Are the scan chains more balanced now?
....................................................................................................
Question 3. What DFT structure did DFTC have to add to achieve these
balanced scan chains?
....................................................................................................
Note: The ORCA design has only 3 test clocks, but your
real-world design might have many more test clocks. The more test
clocks, the greater the probability for having more of these DFT
structures. If minimizing die area is an important design requirement,
you may need to monitor their number.

Lab 10
10. Determine the number of lockup latches DFTC inserted. DFTC names the
latches using an intuitive naming convention that is good for wildcard
searches.
get_cells LOCKUP* -hier
Question 4. How many lockup latches were added?
....................................................................................................
Note: If your design has many lockup latches counting them yourself can
become tedious. Have DFTC count them using:
sizeof_collection [get_cells LOCKUP* -hier]
Question 5. Where in the ORCA hierarchy were they placed?
....................................................................................................
Note: If the ORCA design will be placed and routed using a

top-down methodology, the placement of the lockup latches is probably
not important. If the ORCA design will be placed and routed in a
bottom-up fashion, however, you may have preferred locations for
them. For example, you may want to avoid any lockup latches at the
top-level.
11. Obtain an estimate of the design’s test coverage in DFTC.

Question 6. What is the estimated test coverage for ORCA?
....................................................................................................
12. Exit Design Compiler.

Lab 10
Task 2. Bottom-Up Scan Insertion using Test Models
Two scripts are used for bottom-up scan insertion, scan_blocks.tcl and scan_top.tcl.
These scripts currently run using full gate-level netlists.
In the subsequent steps you will first run scan_blocks.tcl to insert scan into the
blocks of the ORCA design generating test models for the blocks. Later you will run
the scan_top.tcl to complete the top-level scan chain stitching in ORCA using the
test models created for the blocks instead of the full gate-level netlists.
The ORCA top-level design contains all the I/O pad instantiations, a CLOCK_GEN
block that contains the PLLs and clock multipliers, and an ORCA_TOP block that
serves as the core of the design and contains all the blocks for which you will
perform scan insertion.
The foreach loop in the scan_blocks.tcl script iterates through a TCL list of the
blocks within ORCA_TOP. The loop first reads in the appropriate .ddc file from the
correct directory and then proceeds with block-level scan insertion. The results are
saved at the end of the loop and all design files are removed before proceeding to
the next block.
1. Locate the foreach loop within the scan_blocks.tcl script and modify it to save
the test model for each block in a file separate from the DDC file in the
test_models directory (put these commands in the script immediately before
the command that removes each block).
2. Modify the scan_top.tcl to read in the test models only from the proper
directory (these test models were saved by the scan_blocks.tcl script).
Read the test models first, and then read in the block that contains them,
ORCA_TOP. Make sure the RESET_BLOCK is not a test model.
. . .
Your commands go here
. . .
read_ddc mapped/ORCA.ddc
current_design ORCA
link
. . .

Lab 10
3. Examine the scan_blocks.tcl script and answer the following questions.

Question 7. Why specify –clock_mixing no_mix for block level
scan insertion?
....................................................................................................
....................................................................................................
Question 8. What is the result of –insert_terminal_lockup for

block level scan insertion?
....................................................................................................
....................................................................................................
Question 9. If lockup latches will be inserted at the end of all block level
scan chains, do you need to insert any lockup latches between
block level chains at the top-level?
....................................................................................................
4. Run the bottom_up.tcl script, which first runs scan_blocks.tcl on the blocks
within ORCA_TOP and then runs scan_top.tcl on the ORCA design.
UNIX% dc_shell –f scripts/bottom_up.tcl | tee

bottom_up.log
5. Examine the top-level preview_dft report for ORCA.
vi reports/ORCA_preview_dft.rpt
Question 10. Are the scan chains balanced? If not, why?
....................................................................................................

Lab 10
6. Use the approach outlined in lecture to revise the block level scan
architectures so the top-level scan chains are better balanced (stop when all
scan chain lengths are within 20 scan cells of each other).
Question 11. Which command and setting produced this better-balanced
top-level scan architecture?
....................................................................................................
Question 12. In the context of a design that contains test models, why can
you NOT use the same method for counting the number of
lockup latches in the design as used in Task 1?
....................................................................................................
....................................................................................................
Note: Since the resulting Verilog gate-level netlists of both the block level
scan insertion and the top-level scan stitching are saved to the same
UNIX directory, you can use UNIX commands to count the lockup
latches.
7. Determine the number of LOCKUP latches inserted throughout ORCA.
UNIX% fgrep LOCKUP tmax/*.v | wc -l
Note: fgrep searches for fixed string matches and wc –l outputs the
number of matched lines. In this case, that number is equal to the
number of lockup latches.
Question 13. How many lock-up latches were inserted?
....................................................................................................
Question 14. Where in the ORCA hierarchy were they placed

(Hint: do the fgrep without piping it to wc)?
....................................................................................................
....................................................................................................
....................................................................................................
Note: If the ORCA design will be placed and routed using a

top-down methodology, the placement of the lockup latches is probably
not important. If the ORCA design will be placed and routed in a
bottom-up fashion, however, you may have preferred locations for
them. For example you may want to avoid any lockup latches at the
top-level.

Lab 10
Question 15. How does this compare to the number of lock-up latches
inserted in Task 1 by the top-down script?
....................................................................................................
8. Examine the estimated coverage report for ORCA and answer the following
questions.
UNIX% cat reports/ORCA_coverage.rpt
Question 16. How does this coverage compare to the estimated coverage
obtained when inserting scan top-down into ORCA?
....................................................................................................
....................................................................................................
Question 17. Why is the estimated coverage obtained using test models not
meaningful?
....................................................................................................
....................................................................................................
Note: In order to obtain a more meaningful estimate for the test coverage, you
must read the entire design into TetraMAX.
9. Run a script to obtain the actual test coverage in TetraMAX and then answer
the following questions.
UNIX% cd tmax
UNIX% tmax –shell run.tmax
Question 18. Why is the total number of faults in TetraMAX much greater
than the number of faults in DFTC?
....................................................................................................
....................................................................................................
Question 19. What test coverage does TetraMAX obtain for ORCA?
....................................................................................................
10. Exit Design Compiler.

Lab 10
Task 3. Optional: Bottom-up Scan Insertion using ILMs

with Test Models Attached
1. Repeat Task 2 but instead use ILMs with test models attached.
• Modify scripts/ilm_scan_blocks.tcl to generate and save the ILMs.
• Modify scripts/ilm_scan_top.tcl to read in the ILMs.
If you get stuck, check the .solutions directory.

Answers / Solutions
Task 1. Top-Down Scan Insertion using RSS

3. Modify this script to use the RSS features you learned about in lecture.
Compared to the original script in the lab directory, your revised script should:
a. Modify the original design as little as possible
b. Run scan insertion faster
c. Consume less memory
set_dft_insertion_configuration \
–synthesis none \
–preserve_design_name true
Question 1. Are the scan chains well balanced? If not, why?
No. The scan chains are very unbalanced because clock

mixing is not enabled.
7. Open the top_down.tcl file again and change it so that you can achieve
balanced top-level scan chains.
set_scan_configuration –clock_mixing mix_clocks
Question 2. Are the scan chains more balanced now?
Yes. All 6 chains are either 488 or 487 elements long.

Question 3. What DFT structure did DFTC have to add to achieve these
balanced scan chains?
Lockup latches.
Question 4. How many lockup latches were added?
2
Question 5. Where in the ORCA hierarchy were they placed?
dc_shell-xg-t> get_cells LOCKUP* -hier

{I_ORCA_TOP/I_RESET_BLOCK/LOCKUP I_ORCA_TOP/I_SDRAM_WRITE_FIFO/LOCKUP}

Question 6. What is the estimated test coverage for ORCA?
97.67%
Task 2. Bottom-Up Scan Insertion using Test Models
1. Locate the foreach loop within the scan_blocks.tcl script and modify it to save
the test models only for each block:
foreach design $designs {

. . .
# Save the Test Model HERE
write_test_model -format ddc \
-output test_models/${design}.ddc
. . .
}
2. Modify the scan_top.tcl to read in the test models only from the proper
directory (these test models were saved by the scan_blocks.tcl script.)
Read the test models first, and then read in the block that contains them,
ORCA_TOP.
read_ddc [glob test_models/*.ddc]

use_test_model -true [get_designs *]
remove_design RESET_BLOCK
read_ddc mapped_scan/RESET_BLOCK.ddc
use_test_model -false RESET_BLOCK
read_ddc mapped/ORCA_TOP.ddc
change_names –rule verilog
read_ddc mapped/ORCA.ddc
current_design ORCA
link
. . .
Question 7. Why specify –clock_mixing no_mix for block level

scan insertion?
Mixing clocks in block level scan chains would restrict

top-level scan chain balancing. It is therefore better to
postpone scan chain clock mixing until you perform the
final top-level scan stitching.
Question 8. For block-level scan insertion, what is the result of using
–insert_terminal_lockup?
Each block level scan chain ends with a lockup latch,

eliminating the need for lockup latches at the top-level.

Question 9. If lockup latches will be inserted at the end of all block-level

scan chains, do you need to insert any lockup latches
between block-level chains at the top-level?
No, doing so would create back-to-back lock-up latches in

the scan chains.
Question 10. Are the scan chains balanced? If not, why?
No. Some block level scan chains are much longer than
others. This affects the balance of the top-level scan chains;
they are very unbalanced.
Question 11. Which command and setting produced this better-balanced
top-level scan architecture?
In scan_blocks.tcl:
set_scan_configuration –max_length 100
Question 12. In the context of a design that contains test models, why can
you NOT use the same method for counting the number of
lockup latches in the design as used in Task 1?
You cannot count the latches within a test model because

the test model is an empty design in DFTC.
Question 13. How many lock-up latches were inserted?
34
The number of lock-up latches seen will be dependent on
the setting for –max_length. The number provided is for a
setting of “-max_length 100”.

Question 14. Where in the ORCA hierarchy were they placed

(Hint: do the fgrep without piping it to wc)?
unix% fgrep LOCKUP tmax/*.v | awk '{printf ("%s %s %s\n", $1, $2, $3)}'
BLENDER_gates.v: lanlq1 LOCKUP

BLENDER_gates.v: lanlq1 LOCKUP1
BLENDER_gates.v: lanlq1
BLENDER_gates.v: lanlq1
CONTEXT_MEM_gates.v: lanlq1 LOCKUP
ORCA_scan.v: lanlq1 LOCKUP
ORCA_scan.v: lanlq1 LOCKUP1
PARSER_gates.v: lanlq1 LOCKUP
PARSER_gates.v: lanlq1 LOCKUP1
PCI_CORE_gates.v: lanlq1 LOCKUP
PCI_CORE_gates.v: lanlq1 LOCKUP1
PCI_RFIFO_gates.v: lanlq1 LOCKUP
PCI_RFIFO_gates.v: lanlq1 LOCKUP1
PCI_WFIFO_gates.v: lanlq1 LOCKUP
PCI_WFIFO_gates.v: lanlq1 LOCKUP1
RESET_BLOCK_gates.v: lanlq1 LOCKUP3
RISC_CORE_gates.v: lanlq1 LOCKUP
RISC_CORE_gates.v: lanlq1 LOCKUP1
RISC_CORE_gates.v: lanlq1 LOCKUP2
SDRAM_RFIFO_gates.v: lanlq1 LOCKUP
SDRAM_RFIFO_gates.v: lanlq1 LOCKUP1
SDRAM_WFIFO_gates.v: lanlq1 LOCKUP
SDRAM_WFIFO_gates.v: lanlq1 LOCKUP1
Question 15. How does this compare to the number of lock-up latches
inserted in Task 1 by the top-down script?
There are many more lockup latches, one for each block
level chain.
Question 16. How does this coverage compare to the estimate coverage
obtained when inserting scan top-down into ORCA?
The ~45% estimated test coverage is much lower than that

obtained for the full gate-level top-down scan inserted
design.

Question 17. Why is the estimated coverage obtained using test models
not meaningful?
A test model in DFTC is an empty design. When obtaining

the estimated test coverage for a design in DFTC that uses
test models, these empty designs will also be passed to
TetraMAX.
Question 18. Why is the total number of faults in TetraMAX much
greater than the number of faults in DFTC?
TetraMAX has the full gate-level netlist for the entire design
hierarchy. DFTC had the gate-level netlist for ORCA_TOP
and the designs above it (CLOCK_GEN and ORCA); but all
the blocks within ORCA_TOP except for RESET_BLOCK
were test models (empty designs) in DFTC.
Question 19. What test coverage does TetraMAX obtain for ORCA?
98.50%

12 DFT MAX flow
Learning Objectives

• Insert Adaptive Scan using top down flow
• Understand needed components for Adaptive Scan
• Run TetraMAX flow on Adaptive Scan desgin
Lab Duration:
45 minutes
DFT MAX Flow Lab 12-1

Lab 12
Overview
During this lab you will:

1. Perform the Mapped DFT flow for a test-ready design through scan insertion while
enabling Adaptive Scan insertion.
2. Identify needed files to export to TetraMAX.
Mapped Flow:
2. Specify scan constraints and preview the scan architecture that will result from
applying them.
3. Add needed option to enable Adaptive Scan Flow.
4. Insert the scan chains and Adaptive Scan logic.
5. Understand post scan insertion dft_drc checks.
Lab 12-2 DFT MAX Flow

Lab 12
File Locations
lab12_dftmax.
Directory Structure
lab12_dftmax Current working directory
../ref/rtl/ORCA/vhdl Design files
.solutions Solution scripts

Lab 12
Instructions
Your goal is to complete and run the four steps of the Mapped Flow portion of the
DFT Insertion Flow. The scripts you will modify are for previewing the scan
architecture, inserting scan chains and finally, handing off the needed files.
The four needed scripts:
Mapped Flow
4read_gates_and_protocol.tcl
5preview_dft.tcl
6insert_dft.tcl
7handoff.tcl
Task 1. Read gate-level Design and its Test Protocol
1. Invoke DFTC.
cd lab12_dftmax
dc_shell | tee logs/mylab12.log
2. Examine and execute the script that performs step 4 of the DFT flow: reading
in the gate-level design and creating the protocol.
exec cat scripts/4read_gate_and_protocol.tcl

Lab 12
1. Investigate side effects of specifying long chains.
set_scan_configuration \
-clock_mixing mix_clocks -chain_count 2
Question 1. How many scan chains will be inserted, and how will this
affect the number of tester cycles on the tester?
....................................................................................................
Question 2. What is one way to reduce Test Data Volume and Test Time?
....................................................................................................
2. Change the scripts/settings_insert_dft.tcl script to enable compression

with a compression ratio of 5.
set_dft_configuration -scan_compression enable

set_scan_compression_configuration –minimum_compression 5
set_scan_configuration –chain_count 5
Note: Since the design is so small for lab purposes, setting a compression ratio
of 5 is enough to see the benefits.
Question 3. Do you need to specify the compression ratio?
....................................................................................................
3. Do another preview_dft run and compare what was architected now.
Note: For the class, a Tcl script is provided to display content in a separate
window. Try “v preview_dft -show scan_summary”
Question 4. What do you see now in the preview report?
....................................................................................................
4. Investigate the use of the Test_mode pin.

Lab 12
Question 5. Why was there a test_modea pin added to the design?
....................................................................................................
Question 6. Can you share the existing test_mode pin that might be
selecting Test Clocks?
....................................................................................................
....................................................................................................
5. Edit the file scripts/settings_insert_dft.tcl again and specify your own top
level Test Mode pin for compression. Then rerun the script.
create_port –direction in TM_COMP

set_dft_signal –view spec –type TestMode –port TM_COMP
6. Since a new port was added you need to recreate the test protocol file and
rerun dft_drc.
create_test_protocol –capture_procedure multi_clock
dft_drc
preview_dft –show scan_summary
Question 8. Which port is now being used to select between modes?
....................................................................................................
7. Now insert_dft to insert the chains and compression logic.
insert_dft
Question 9. What new messages do you see with insert_dft?
....................................................................................................

Lab 12
Task 3. Hand-off files for ATPG for two modes
Question 10. How many STIL Test Protocol files do you need to write out
with an Adaptive Scan flow?
....................................................................................................
Question 11. How many netlists are needed?
....................................................................................................
1. Check drc rule violations on Internal_scan mode:
current_test_mode Internal_scan
dft_drc
2. Check drc rule violations for ScanCompression_mode:
current_test_mode ScanCompression_mode
dft_drc
Question 12. What are the differences that you see?
....................................................................................................
Note: Internal_scan mode is also known as reconfigured scan mode.
3. Edit the file scripts/7handoff.tcl and add the command to write out the
ScanCompression_mode protocol.
The ScanCompression_mode protocol file should be named
“scancompress.spf” and written to the “tmax” directory.
4. Run the handoff script:
s scripts/7handoff.tcl
Question 13. Look at both test protocol files, do you see a difference?
....................................................................................................
Task 4. Run ATPG in Compression mode
1. Run the TetraMAX command file ./tmax/compression.cmd to run ATPG in

Adaptive Scan mode:

Lab 12
cd tmax
tmax compression.cmd &
Question 14. What is the test coverage reported by TetraMAX ?
....................................................................................................
Congratulations! This completes the DFTMAX lab.

Answers / Solution
For help with the scripts/commands, see the .solutions directory.
Question 1. How many scan chains will be inserted, and how will this
affect the number of tester cycles on the tester?
2 scan chains. This will make the scan chains longer, which
will affect the number of tester cycles needed per pattern
hence increasing the Test Data Volume and Test Time.
Question 2. What is one way to reduce Test Data Volume?
Use Adaptive Scan technology to insert compression.

Question 3. Do you need to specify the compression ratio?
Yes. The compression ratio defaults to 10. So you only need

to specify it if you are reducing or increasing this number.
Question 4. What do you see now in the preview report?
You see two modes specified. One for Internal_scan with 5

chains and one for ScanCompresion_mode with 30 much
shorter chains. You also see a new Test Mode Controller
Information section.
Question 5. Why was there a test_modea pin added to the design?
A control signal is need to select between the two modes

created by Adaptive Scan, Internal_scan and
ScanCompression_mode.
Question 6. Can you share the existing test_mode pin that might be
selecting Test Clocks?
No. The test_mode port was specified as a Constant type

signal. It could be a signal used to fixed resets or clocks. We
need a dedicated Test Mode pin for Adaptive Scan.
Same pins as specified for Normal scan

ScanDataIn Ports:
pad[0], pad[1], pad[2], pad[3], pad[4]
ScanDataOut Ports:
sd_A[0], sd_A[1], sd_A[2], sd_A[3], sd_A[4]
ScanEnable Port:
scan_en
Question 8. Which port is now being used to select between modes?
TM_COMP.
Question 9. What new messages do you see with insert_dft?
Architecting Scan Compression structures

Architecting Scan Chains
Architecting Load Compressor (version 1.3)
Number of inputs/chains/internal modes = 5/30/4
Architecting Load Decompressor (version 1.3)
Number of outputs/chains = 5/30
Combinational initialization successfully completed.
Question 10. How many STIL Test Protocol files do you need to write
out with an Adaptive Scan flow?
Two Test Protocol files, one for normal scan, and one for
adaptive scan.
Question 11. How many netlists are needed?
Only one.
Question 12. What are the differences that you see?
There are new scan compression violations:

517 SCAN COMPRESSION VIOLATIONS
150 X on chain affects observe ability of other chains violations (R11)
367 X source gate is propagatable to other scancells violations (R14)
Question 13. Look at both test protocol files, do you see a difference?
Visually inspect the files to see added sections for Adaptive

Scan.
Question 14. What is the test coverage reported by TetraMAX?
99.56%


DFTC1 2007.12 LabGuide

Uploaded by

Copyright:

Available Formats

DFTC1 2007.12 LabGuide

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

DFTC1 2007.12 LabGuide

Uploaded by

Copyright:

Available Formats

CUSTOMER EDUCATION SERVICES

Synopsys Customer Education Services

Workshop Registration: 1-800-793-3448

Destination Control Statement

Registered Trademarks (®)

Service Marks (SM)

Document Order Number: 30-I-011-SLG-012

Synopsys Customer Education Services

After completing this lab, you should be able to:

Creating Test Protocols Lab 4-1

If you need help…

dc_shell> help prev*

Answers & Solutions

Lab 4-2 Creating Test Protocols

Task 1. Set Up Your Environment For Reading

Add the following lines to the .synopsys_dc.setup file.

# for reading and linking

The pound (#) at the beginning introduces a TCL comment.

Creating Test Protocols Lab 4-3

3. Create shortcuts by editing the .synopsys_dc.setup file.

a. Create the following commonly-used alias for the frequently-used

Question 3. To use the history command in the Design Compiler

4. Further examine the .synopsys_dc.setup file. Note the two variables

Lab 4-4 Creating Test Protocols

Test Protocol Definition Flow

Create a test protocol

Save the test protocol file

1. Open the script called scripts/unmapped.tcl.

Creating Test Protocols Lab 4-5

2. Execute your script by running the unmapped.tcl script in batch mode.

unix% dc_shell -f scripts/unmapped.tcl | \

3. Determine if there were any problems in the unmapped flow by looking at

Question 5. a. What value is the TEST_MODE port set to?

Question 6. Look at the results of running the dft_drc command, are

Lab 4-6 Creating Test Protocols

Task 3. Verify that the Design was Compiled Properly

2. Determine if the design has been properly scan replaced.

Task 4. Part A Review Questions

Question 10. What is the purpose of the .synopsys_dc.setup file?

Question 13. What is the function of the target_library variable?

3. Exit from DFTC.

Creating Test Protocols Lab 4-7

Lab 4-8 Creating Test Protocols

Creating Test Protocols Lab 4-9

Task 1. Internally Generated test_mode Signal Using

Lab 4-10 Creating Test Protocols

Task 2. Read a Mapped Design and Create a Test

1. Change directories to the project directory lab4b_init.

3. Create a shortcut to make running the scripts easier.

alias s "source -echo -verbose"

exec cat scripts/1read_design.tcl

5. Add the appropriate command to the 2create_test_protocol.tcl script.

Task 3. Create an Initialization Sequence

Developing the initialization test protocol from scratch is not

1. Save out the current test protocol for the design.

2. Add the required Vector statements (“V {}”) to the orca_mapped.spf.

Note: The “test_setup” or initialization section is only part of

Lab 4-12 Creating Test Protocols

create_test_protocol –capture_procedure multi_clock

Congratulations! You have completed the lab.