Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Definition
Design for Testability (DFT) - 1 Ad-hoc methods
Scan design
Design rules
Scan register
Scan flip-flops
Scan test sequences
Mohammad Tehranipoor Overhead
Electrical and Computer Engineering
Scan design system
Summary
University of Connecticut
Design for testability (DFT) refers to those design Good design practices learned through
techniques that make test generation and test
experience are used as guidelines:
application cost-effective.
Don’t-s and Do-s
DFT methods for digital circuits:
Avoid asynchronous (unclocked) feedback.
Ad-hoc methods
Avoid delay dependant logic.
Structured methods:
Avoid parallel drivers.
Scan
Avoid monostables and self-resetting logic.
Partial Scan
1
Ad-Hoc DFT Methods Scan Design
Objectives
Disadvantages of ad-hoc DFT methods: Simple read/write access to all or subset of storage
Experts and tools not always available elements in a design.
Test generation is often manual with no Direct control of storage elements to an arbitrary value
guarantee of high fault coverage (0 or 1).
Functional patterns Direct observation of the state of storage elements and
Design iterations may be necessary hence the internal state of the circuit.
Very time consuming
Gates
Inputs
Outputs Make input/output of each scan shift register
controllable/observable from PI/PO.
Use combinational ATPG to obtain tests for all testable faults in
Scan-in (SI) the combinational logic.
Add shift register tests and convert ATPG tests into scan
Scan Flip-Flop sequences for use in manufacturing test.
Comb.
logic D1 Q
Use only clocked D-type flip-flops for all state
Comb.
variables. D2
FF
logic
CK
At least one PI pin must be available for test; more
pins, if available, can be used.
All clocks must be controlled from PIs. Comb.
logic
Clocks must not feed data inputs of flip-flops. Q
D1
D2 FF Comb.
logic
CK
2
Correcting a Rule Violation Scan Flip-Flop (Master-Slave)
SE CK D flip-flop
SCK D flip-flop
SFF
SD
Normal
MCK
mode
Logic TCK
overhead
TCK MCK
mode
Scan
SCANIN S1 S2
PI I1 I2 O1 O2 PO TC 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0000000
Combinational
SCANIN
SCANOUT PO O1 O2
TC
logic
Next
Present S2 N1 N2
S1 state SCANOUT N1 N2
state
Sequence length = (nsff + 1) ncomb + nsff clock periods
ncomb = number of combinational vectors
nsff = number of scan flip-flops
3
Testing Scan Register Multiple Scan Registers
Scan flip-flops can be distributed among any number of
shift registers, each having a separate scanin and scanout
Scan register must be tested prior to application of scan test pin.
sequences. Test sequence length is determined by the longest scan
A shift sequence 00110011 . . . of length nsff+4 in scan shift register.
mode (TC=0) produces 00, 01, 11 and 10 transitions in all Just one test control (TC) pin is essential.
flip-flops and observes the result at SCANOUT output.
Total scan test length: PI/SCANIN
Combinational
PO/
M SCANOUT
((nsff + 1) ncomb + nsff ) + (nsff + 4) clock periods. logic U
SFF X
(ncomb + 2) nsff + ncomb + 4 clock periods.
SFF
Example: 2,000 scan flip-flops, 500 comb. vectors, total scan SFF
test length ~ 106 clocks.
TC
Multiple scan registers reduce test length.
CK
4
Scan Area Overhead Example: Scan Layout
2,000-gate CMOS chip
Linear dimensions of active area:
X = (C + S) / r Fractional area under flip-flop cells, s = 0.478
X’ = (C + S + αS) / r y = track dimension, wire Scan flip-flop (SFF) cell width increase, α = 0.25
width+separation
β) / T
Y’ = Y + ry = Y + Y(1--β C = total comb. cell width Routing area fraction, β = 0.471
S = total non-scan FF cell Cell height in routing tracks, T = 10
Area overhead width
Calculated overhead = 17.24%
X’Y’--XY s = fractional FF cell area
= S/(C+S) Actual measured data and performance:
= -------------- x 100% α = SFF cell width fractional
XY increase
β
1--β r = number of cell rows Scan implementation Area overhead Normalized clock rate
or routing channels
= [(1+α αs)(1+ -------) – 1] x 100% β = routing fraction in active
______________________________________________________________________
T area None 0.0 1.00
T = cell height in track
1--ββ dimension y Hierarchical 16.93% 0.87
= (αs + ------- ) x 100% Optimum layout 11.90% 0.91
T
27 October 2009 25 27 October 2009 26