Lab Instructions 2012 dft

TDDC33 Design for Test of Digital Systems

Design for Test of Digital Systems

TDDC33

Lab Instructions

Date of last revision

24/08/2012

2012

Dimitar Nikolov, IDA/SaS ESLAB

TDDC33 Design for Test of Digital Systems 1

Table of Contents

1. Introduction ......................................................................................................................................... 3

2. Initial preparations .............................................................................................................................. 4

3. Synthesis ............................................................................................................................................. 5

3.1. Input ............................................................................................................................................. 5

3.2. Output .......................................................................................................................................... 5

3.3. Synthesis procedure ..................................................................................................................... 5

4. Design for test using DFTAdvisor .................................................................................................. 7

4.1. Input ............................................................................................................................................. 7

4.2. Output .......................................................................................................................................... 7

4.3. Starting DFTAdvisor ................................................................................................................... 7

4.4. Circuit Setup ................................................................................................................................ 7

4.5. Test Synthesis .............................................................................................................................. 7

4.5.1 Full Scan Insertion ..................................................................................................................... 8

4.5.2 Partial Scan Insertion ................................................................................................................. 8

5. Fault Coverage Analysis and Test Pattern Generation using FastScan ............................................ 10

5.1. Input ........................................................................................................................................... 10

5.2. Output ........................................................................................................................................ 10

5.3. Writing test patterns ................................................................................................................... 10

5.4. Starting FastScan ....................................................................................................................... 11

5.5. Circuit setup ............................................................................................................................... 11

5.6. Fault simulation ......................................................................................................................... 12

5.7. Test pattern generation ............................................................................................................... 12

5.8. Results and analysis ................................................................................................................... 13

6. Test point insertion in VHDL ........................................................................................................... 14

6.1. Input ........................................................................................................................................... 14

6.2. Output ........................................................................................................................................ 14

6.3. Inserting test points .................................................................................................................... 14

7. Board test using Boundary Scan (IEEE 1149.1) ............................................................................... 15

7.1. Input ........................................................................................................................................... 15

7.2. Output ........................................................................................................................................ 15


7.3. Starting Trainer1149 .................................................................................................................. 15

7.4. How to use Trainer1149 ............................................................................................................. 15

7.5. Introducing a fault ...................................................................................................................... 16

7.6. Writing and verifying a test program using the Test Constructor .............................................. 17

7.7. Writing and verifying a test program using the TAP Controller................................................ 19

Appendix A S27_TP VHDL Description ............................................................................................. 20

References ............................................................................................................................................. 25


1. Introduction

This document describes the environmental setup and the tools needed to complete the lab

assignments in the course TDDC33 Design for Test of Digital Systems. The tools that are used in this

course are listed in Table 1. For the synthesis, automatic test pattern generation (ATPG), and design

for test (DFT), the core cells from AMS 0.35m [1] standard cell library will be used together with

the test library c35_CORELIB.atpg. A design named s27 will be used as an example throughout these

instructions. The s27 design is described in VHDL (s27.vhdl) and stored in a directory named

s27_test. The name of the design and the directory will later be changed as you solve the lab

assignments.

The rest of this document is organized as follows. Chapter 1 contains information about how to setup

the system in order to start the tools. The following chapters, Chapter 2 to Chapter 7, contain

instructions for the synthesis, DFT, test pattern generation, test point insertion, and board testing,

respectively.

Task Tool Tool vendor

Synthesis Leonardo Spectrum (Mentor Graphics)

Design for test DFTAdvisor (Mentor Graphics)

Test pattern generation FastScan (Mentor Graphics)

Boundary Scan Trainer1149 (Testonica)

Boundary Scan TSTAP-Studio (SAAB Aerotech)

Table 1 Covered tools


2. Initial preparations

The following commands are all executed in a terminal window.

Add the modules /mentor/tessent and mentor/fpgadv if they are missing.

To check which modules are already loaded issue the following command:

module list

If the modules /mentor/tessent and mentor/fpgadv are missing from the list, use the following

command to add the modules.

module add prog/mentor/fpgadv module add prog/mentor/tessent

Note! If you have added the modules by yourself, then each time you close the terminal window you

will need to add them again. To avoid adding the modules each time you close the terminal window,

use the following command:

module initadd prog/mentor/fpgadv module initadd prog/mentor/tessent

Observe that these commands do not add the modules instantly, instead the modules /mentor/tessent

and mentor/fpgadv will be loaded by default from the next time you login to your account.

Make a directory s27_test

mkdir s27_test

Download and extract the required files. (Described in the labs)

Download the labX.tar.gz file where X is the number of the lab gunzip labX.tar.gz tar xvf labX.tar Copy the following files to the s27_test directory:

c35_CORELIB.atpg

fflop.vhd

gates.vhd

s27.vhdl

Set the environment variable MODEL_TECH

setenv MODEL_TECH /sw/mentor/fpgadv/6.2/Modeltech/bin

Generate the work directory

$MODEL_TECH/vlib work

Compile the vhdl files

$MODEL_TECH/vcom -93 fflop.vhd $MODEL_TECH/vcom -93 gates.vhd $MODEL_TECH/vcom s27.vhdl


3. Synthesis

This chapter describes the synthesis procedure using Leonardo Spectrum from Mentor Graphics. It is

assumed that the initial preparations, described in the Chapter 2, have been made.

3.1. Input

A compiled VHDL-file

3.2. Output

A synthesized Verilog netlist

3.3. Synthesis procedure

Start the synthesis program from a command prompt

leonardo & Click OK

In the Quick Setup tab

Select library file:

technology->ASIC->AMS->c35_CORELIB

Select input file

Input->Open files: Select the vhdl file (s27.vhd)

Select the name of the output file

s27.v

In the Advanced tab

Output -> Format->Verilog

In the Quick Setup tab

Check that the Optimize Effort is set to Fastest Runtime Check that the output filename is correct

Start synthesis

Run Flow Check the results from the synthesis in the Exemplar.log file Check that the Verilog file (s27.v) was generated


Figure 1 Synthesis Setup (Quick Tab)

Figure 2 Synthesis setup (Advanced Tab)


4. Design for test using DFTAdvisor

This chapter describes how to perform scan chain insertion using DFTAdvisor tool from Mentor

Graphics.

4.1. Input

A design netlist (Verilog netlist)

Library file

4.2. Output

A new design netlist

ATPG setup files (dofile and procedure file)

4.3. Starting DFTAdvisor

Invoke DFTAdvisor from the command prompt

dftadvisor netlist_file lib c35_CORELIB.ATPG

Note! The command is issued from a working directory that contains both the netlist_file and the

library file (c35_CORELIB.ATPG)

By default, the tool enters the SETUP mode. In this mode, the user can give further inputs to the tool

regarding the design that is provided, i.e. specify the primary inputs and outputs, specify clocks etc.

The tool is able to directly identify the primary inputs and outputs, but further inputs are required to

specify the clocks. To check the primary inputs and outputs that the tool has identified try the

following commands:

report primary inputs

report primary outputs

Next, we present how to specify clocks.

4.4. Circuit Setup

To specify the clocks, issue the following command:

add clocks off_state pin_pathname

where off_state represents the pin value (0 or 1) that results in a scan elements clock input being at its

inactive state (for latches) or state prior to a capturing transition (for edge-triggered devices), and

pin_pathname represents a primary input pin that is to be assigned as a clock pin.

Example:

add clocks 0 CLK

Once all clocks are added, the circuit setup is completed, and we can continue working with the tool

and insert scan chains. To do that, we need to switch the system mode to DFT. We set the new system

mode by issuing the following command:

set system mode dft

4.5. Test Synthesis

In this section, we provide the commands that are required to insert scan chains into your design using

DFTAdvisor. Regarding scan chain insertion, we will describe the steps to perform a full scan

insertion, where all scannable elements are connected in a single scan chain, and partial scan

insertion where a number of scannable elements are stitched in a single chain.


4.5.1 Full Scan Insertion

To setup full scan insertion run the following commands:

setup scan identification full_scan

run

After running these commands, in the command window you will get a brief report stating the number

of sequential elements that were identified in your design and how many of them have been converted

into scan cells. Next, we insert the test logic in the design by issuing the following command:

insert test logic

This command adds the new primary inputs and outputs to the circuit, replaces the sequential

elements into scan cells, and connects the scan cells in a single chain. To save the modified design,

you use the command:

write netlist filename

The new design has the scan chains included. To be able to run an ATPG for the scannable design,

there are some further inputs that are to be given to the ATPG tool (FastScan in this case) so that it

can identify the presence of the scan chains, and get to know how the scan operations are to be

performed. This information is included in the ATPG setup files (the procedure file and the dofile).

To save these files, use the following command:

write atpg setup basename

After running this command, two files are generated: basename.testproc (a file that contains the

information of how to perform the scan operations) and basename.dofile (a file that informs the

ATPG tool for the existence of scan circuitry in the design)

4.5.2 Partial Scan Insertion

To add partial scan, you need to first write a configuration file, where you explicitly select how many

and which memory elements to be converted into scan cells and connected in a scan chain. Note that

this you do manually, due to the lack of the tool to automatically run partial scan insertion. In the ideal

case, when a tool supports automatic partial scan insertion, the tool can select the specified number of

memory elements, which is given by the user as a number or maybe a percentage, by applying some

algorithm and performing some analysis so that it selects only the most important memory elements

to be converted into scan cells. For example, SCOAP can be used to identify which memory elements

are the best to be converted into scan cells and connected in a scan chain, so that the fault coverage is

maximized. Next, we describe how to write this configuration file to perform partial scan insertion.

The configuration file is a text file that contains lines, each of them with three parameters:

instance_pathname cell_id chain_id

instance_pathname - A string that specifies the name of the non-scan cell that you want DFTAdvisor

to put in the scan chain. To find instance_pathname open the netlist file and read its contents.

cell_id - An integer that specifies the placement of the instance_pathname in relation to other

instance_pathnames. DFTAdvisor places the instance having the smallest cell_id closest to the scan

chain output. All instances in the same chain must have unique cell_ids.

chain_id - An integer that specifies the scan chain in which you want DFTAdvisor to place the

instance_pathname. DFTAdvisor places instances with the same chain_id in the same chain.


Once you have prepared the configuration file you only need to insert the test logic that is specified

there. To do this, use the following command:

insert test logic config_file

where config_file stands for the configuration file that you have written. Next you need to do, is to

write the netlist file and the atpg setup files, as described in the previous section.


5. Fault Coverage Analysis and Test Pattern Generation using FastScan

This chapter describes the fault coverage analysis and test pattern generation process using FastScan

from Mentor Graphics. It is assumed that the initial preparations, described in Chapter 2, have been

made.

5.1. Input

A design netlist (Verilog netlist)

Library file

External test patterns (optional)

Command files: dofile and testproc file (optional)

5.2. Output

Test patterns

Fault coverage

5.3. Writing test patterns

Manually crafted test patterns should be written in a text file using the following format:

Combinational design (c17 design)

ASCII_PATTERN_FILE_VERSION = 2;

SETUP =

declare input bus "PI" = "/INP[0]", "/INP[1]", "/INP[2]", "/INP[3]",

"/INP[4]";

declare output bus "PO" = "/OUTP[0]", "/OUTP[1]";

end;

SCAN_TEST =

pattern = 0;

force "PI" "10010" 0;

measure "PO" "00" 1;

pattern = 1;

force "PI" "10010" 0;

measure "PO" "00" 1;

end;


Sequential design (could be the s27 design)

ASCII_PATTERN_FILE_VERSION = 2;

SETUP =

declare input bus "PI" = "/INP[0]", "/INP[1]", "/INP[2]", "/INP[3]",

"/H";

declare output bus "PO" = "/OUTP[0]";

clock "/H" =

off_state = 0;

pulse_width = 1;

end;

end;

SCAN_TEST =

pattern = 0 clock_sequential ;

force "PI" "01110" 0;

pulse "/H" 1;

force "PI" "11010" 2;

measure "PO" "1" 3;

pattern = 1 clock_sequential ;

force "PI" "01010" 0;

pulse "/H" 1;

force "PI" "10110" 2;

measure "PO" "1" 3;

end;

5.4. Starting FastScan

To invoke FastScan use the following command:

fastscan netlist_file lib c35_CORELIB.ATPG

By default, the tool enters the SETUP system mode, where you can provide more inputs to the tool

regarding the design for which you intend to perform fault simulation or want to run automatic test

pattern generation.

5.5. Circuit setup

If you are about to work with a sequential design, then you are supposed to specify which of the

primary inputs represent the clock pins. To specify clocks, use the following command:

add clocks off_state pin_pathname

where off_state represents the pin value (0 or 1) that results in a scan elements clock input being at its

inactive state (for latches) or state prior to a capturing transition (for edge-triggered devices), and

pin_pathname represents a primary input pin that is to be assigned as a clock pin.

Example:

add clocks 0 CLK


If you are about to work with a design that contains scan chains, and you want the ATPG tool to use

these test feature while generating the patterns, then you need to provide the dofile to the tool by

issuing the following command:

dofile filename.dofile

Once you are done with the circuit setup, you can continue further with either running fault

simulation, in which case you should switch to the FAULT system mode, or running the ATPG

process, in which case you should switch to the ATPG system mode. Switching the system mode is

done by using the following command:

set system mode $SM

where $SM denotes one of the following system modes: SETUP, GOOD, FAULT and ATPG.

5.6. Fault simulation

To perform fault simulation, you need to make sure that you have set the system mode to FAULT.

Before you run fault simulation, you need to specify the type of faults that you would like to be

simulated and you should add those faults in the fault list which will later be used by the tool to

identify which of them have been detected using the set of test patterns that you supply to the tool. To

add faults in the fault list, use the following command:

add faults -all -stuck_at 01

The given command adds all stuck_at faults for the provided netlist, meaning that at each fault site

(each net in the netlist) it considers both the stack_at 1 and stuck_at 0 faults. Given this fault list

during the fault simulation the tool will try to identify which of those faults are covered by the

supplied test pattern set. To observe which are the faults that have been added in the faults list, what is

their type and where are the fault sites in your netlist, you can use the following command:

report faults

To supply your own test pattern set, you should use the following command:

set pattern source external filename

Note! We assume that you have already created a test pattern file, using the instructions presented

earlier in this tutorial (see 5.3. Writing test patterns).

Once you have added the required faults in the fault list and have supplied a set of test patterns, you

perform the fault simulation by using the following command:

run

To get a detailed report about the fault coverage, use the following command

report statistics

5.7. Test pattern generation

To perform test pattern generation, you need to make sure that you have set the system mode to

ATPG. The next step is to add all the faults in the fault list, so that the tool can later generate test

patterns for detecting the faults that are present in the fault list. Here is the set of commands that you

should issue in order to generate test patterns.

add faults -all -stuck_at 01

create patterns

In the reported statistics you will be able to see what is the fault coverage for the generated test

patterns.


5.8. Results and analysis

The following metrics are reported after running test pattern generation or fault simulation. Here is a

brief description for each of them.

Test Coverage - percentage of all testable faults that are detected by the patterns.

Fault Coverage - percentage of all faults both testable and untestable that are detected by the

patterns.

ATPG effectiveness percentage - a measure of the ability of the ATPG tool to either provide a test

to detect a fault, or prove that a test cannot be created.


6. Test point insertion in VHDL

This chapter describes how the testability of a design can be improved by inserting test points. It is

assumed that the VHDL description of a design is available.

6.1. Input

A design described in VHDL

6.2. Output

A new design with test points.

6.3. Inserting test points

Add the new input and output ports in the ENTITY block

Add new signals in the ARCHITECTURE block (if needed)

Modify the design (introduction of new gates may be required) such that the new ports are

used to control and observe the hard-to-test parts of the design.

Please refer to Appendix A for an example of VHDL code with inserted test points.


7. Board test using Boundary Scan (IEEE 1149.1)

This chapter describes the boundary scan board test methodology. For the board test programming

and test, a program called Trainer1149 from Testonica Lab is used.

7.1. Input

A board design, which consists of one bsdl-file for each chip and another file that contains a list of

connections between the chips.

7.2. Output

A test program.

7.3. Starting Trainer1149

Verify that the current installed version of java is 1.6 or higher

java version

module add prog/jdk/1.6

Start the Trainer1149 program

java jar trainer1149.jar &

Select an existing project or make a new project

7.4. How to use Trainer1149

The Trainer1149 program has three different modes, Project Mode, Debug Mode, and Board Edit

Mode. The Trainer1149 program will start in the project mode, which allows you to view and modify

the board layout. In Figure 3, the twochips.nl netlist has been chosen in the Project Explorer.


Figure 3 Trainer1149 in Project mode

Select appropriate item in the Mode menu to change the working mode. Another way is to use the

toolbar buttons illustrated in Figure 4.

Figure 4 Mode selection buttons

In Debug Mode it is possible to perform boundary scan operations on the current design. In the Board

Edit mode the user can create and/or modify the design. In this tutorial we focus on boundary scan

operations in Debug Mode.

7.5. Introducing a fault

The Trainer1149 has a trainer function where faults can be inserted in the design. The trainer function

is used to verify whether the test program detects the faults as intended. The following instructions

show how to insert a stuck-at 0 in net2.

Insert a fault in the design.

Training->Inject Fault

The Injection Fault window, illustrated in Figure 5, will appear.


Figure 5 Inject Fault window

Insert the stuck-at 0 fault.

Select Stuck-at 0 in Open Fault panel.

Select net2 in Select net panel.

Press Inject fault button.

7.6. Writing and verifying a test program using the Test Constructor

The following procedure describes how to write a test program that detects if there is a stuck-at 0 fault

present at the net2 for the design presented in Figure 3. (to identify the net, just place the cursor over

the wire and the label of the net will appear)

Select the Debug Mode

Mode->Debug

The Test Constructor panel will appear as illustrated in Figure 6.

Figure 6 Test Constructor panel


The buttons TLR, IR, and DR are used for test logic reset, perform scan on instruction register, and

perform scan on data register, respectively.

Specify the instruction and test vector.

Select the EXTEST for chip1 and chip2.

Specify the following vectors to detect the stuck-at 0 fault:

Chip1 : 1111111111111111

Chip2 : 1111111111111111

Press the RUN button, to run the test

After applying the test vectors, the produced test responses can be compared with the expected test

responses and thus detect if there is a fault. An example is illustrated in Figure 7 where the stuck-at 0

fault on net2 is detected (expected responses are the values you observe in the white boxes, while the

produced test responses are the values you observe in the green boxes).

If you have inserted a fault as described in Section 6.3, you can check if you have successfully

detected it.

Check if you have detected the fault.

Training->Check Fault

Select the net that you suspect has a fault

Figure 7 Stuck-at 0 fault detected

net 2


7.7. Writing and verifying a test program using the TAP Controller

The following procedure describes how to write a test program that detects if there is a stuck-at 0 fault

present at the net2 in Figure 20.

Open the TAP Controller state machine.

Press TAP State Diagram button

The TAP State Diagram window will appear as illustrated in Figure 8

Figure 8 The TAP state machine

Specify the input signals TDI and TMS by pressing the buttons TDI(0) and TMS(0). The test clock is

toggled by pressing TCK(0).

Load the EXTEST instruction and apply the test vector.

The stuck-at 0 fault is detected by applying the test procedure presented in Table 2.

TDI(Value) TMS(Value) TCK (No. of clicks) Comment

0 0 1 Run-Test/Idle

0 1 2 Select IR-Scan

0 0 2 Shift-IR

0 0 15 Shift in EXTEST for both chips

0 1 3 Select DR-Scan

0 0 2 Shift-DR

1 0 35 Shift in test vector (all 1s)

0 1 3 Select DR-Scan

0 0 1 Capture DR (responses captured)

0 0 15 Shift-DR (shift out results)

0 1 2 Update-DR

0 0 1 Run-Test/Idle

Table 2 Test procedure


Appendix A S27_TP VHDL Description

---------------------------------------------------------------------------------------------

--This file is modified by Anders Larsson

--Increased testability by introducing a new control point.

--More control and/or observable points should be added to further increase the testability.

---------------------------------------------------------------------------------------------

library IEEE;

use IEEE.std_logic_1164.all;

use work.all;

ENTITY s27_bench IS

PORT (

--Add a new input

INP: in std_ulogic_vector(0 to 4);

OUTP : out std_ulogic_vector(0 to 0);

H : in std_ulogic

);

END s27_bench ;

ARCHITECTURE structural OF s27_bench IS

component andg

generic (tpd_hl : time;

tpd_lh : time);

port (in1, in2 : std_logic;

out1 : out std_logic);

end component;

component org


tpd_lh : time);



end component;

component xorg


tpd_lh : time);



end component;

component xnorg


tpd_lh : time);



end component;


component nandg


tpd_lh : time);



end component;

component norg


tpd_lh : time);



end component;

component invg


tpd_lh : time);

port (in1 : std_logic;


end component;

component buffg


tpd_lh : time);

port (in1 : std_logic;


end component;

-- ******* Portes generiques sur le nombre d'entr

component andg_n

generic (n : integer ;

tpd_hl : time ;

tpd_lh : time);

port (inp : std_logic_vector(0 to n-1);

out1 : out std_logic) ;

end component;

component nandg_n


tpd_hl : time ;

tpd_lh : time );



end component;

component org_n


tpd_hl : time ;

tpd_lh : time) ;



end component;


component norg_n


tpd_hl : time ;

tpd_lh : time) ;



end component;

component xorg_n


tpd_hl : time ;

tpd_lh : time) ;



end component;

component xnorg_n


tpd_hl : time ;

tpd_lh : time) ;



end component;

component DFFC


tpd_lh : time);

port (DFFC,H,C : std_logic;

Q : out std_logic);

end component;

component DFF


tpd_lh : time);

port (D,H : std_logic;

Q : out std_logic);

end component;

component TFFC


tpd_lh : time);

port (T,H,C : std_logic;

Q : out std_logic);

end component;

signal INTERP : std_ulogic_vector(0 to 11):=(others=>'0') ;

signal OUTPI : std_ulogic_vector(OUTP'range):=(others=>'0') ;


BEGIN

DFF0 : DFF generic map (1 ns,1 ns)

port map (

D => INTERP(1),

H => H,

Q => INTERP(0));


port map (

D => INTERP(3),

H => H,

Q => INTERP(2));


port map (

D => INTERP(5),

H => H,

Q => INTERP(4));

INV0 : INVG generic map (1 ns,1 ns)

port map (

in1 => INP(0),

out1 => INTERP(6));

INV1 : INVG generic map (1 ns,1 ns)

port map (

in1 => INTERP(3),

out1 => OUTPI(0));

AND0 : ANDG_N generic map (2,1 ns,1 ns)

port map (

inp(0) => INTERP(6),


out1 => INTERP(7));

OR0 : ORG_N generic map (2,1 ns,1 ns)

port map (



out1 => INTERP(8));

OR1 : ORG_N generic map (2,1 ns,1 ns)

port map (

inp(0) => INP(3),


out1 => INTERP(10));

NAND0 : NANDG_N generic map (2,1 ns,1 ns)

port map (



out1 => INTERP(11));


NOR0 : NORG_N generic map (2,1 ns,1 ns)

port map (



out1 => INTERP(1));


port map (



out1 => INTERP(3));


port map (

inp(0) => INP(1),


out1 => INTERP(9));


port map (

inp(0) => INP(2),


out1 => INTERP(5));

BUFFER_OUT : OUTP '0') ;

signal OUTPI : std_ulogic_vector(OUTP'range):=(others=>'0') ;

BEGIN

REGVECT : BLOCK (H='1' AND NOT H'STABLE)

BEGIN

DFF3 : INTERP(0)


References

[1] http://asic.austriamicrosystems.com/databooks/c35/databook_c35_33/, April 2006.

Documents

Lab Instructions 2012 dft