power aware testing strategies (slides).pdf
TRANSCRIPT
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
1/55
Power-Aware Testing and Test Strategies for Low Power
Devices
Patrick Girard, Nicola Nicolici, Xiaoqing Wen
To cite this version:
Patrick Girard, Nicola Nicolici, Xiaoqing Wen. Power-Aware Testing and Test Strategies forLow Power Devices. ATS2012: Asian Test Symposium, Nov 2012, Niigata, Japan. IEEE,.
HAL Id: lirmm-00820737
http://hal-lirmm.ccsd.cnrs.fr/lirmm-00820737
Submitted on 6 May 2013
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entific research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
Larchive ouverte pluridisciplinaire HAL, est
destinee au depot et a la diffusion de documents
scientifiques de niveau recherche, publies ou non,
emanant des etablissements denseignement et de
recherche francais ou etrangers, des laboratoires
publics ou prives.
http://hal-lirmm.ccsd.cnrs.fr/lirmm-00820737http://hal-lirmm.ccsd.cnrs.fr/lirmm-00820737https://hal.archives-ouvertes.fr/ -
8/10/2019 Power Aware Testing Strategies (Slides).pdf
2/55
1
1
Power-Aware Testing and Test Strategies
for Low Power Devices
LIRMM / CNRS
France
McMaster University
Canada
Kyushu Institute of Technology
Japan
2
Power-Aware Testing and Test
Strategies for Low Power Devices
Springer web site:
http://www.springer.com/engineering/circuits+%26+systems/book/978-
1-4419-0927-5
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
3/55
2
3
4
MARKET
Final Teston ATE
Burn-In (4-24 hrs)
Termal Cycling+ Final Test
on ATEWhole population
Sample population
Wafer sorton ATE
A
NALYSIS
Stress(up to 2000 hrs)
Test onATE
Selected products
Known Good Dies (KGDs)
Reliability measurements
(courtesy: Bernardi et al., ETS, 2009)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
4/55
3
5
6
CLK
ScanENA
CLK
ScanENA
Primary
Inputs
Primary
Outputs
D Q
ScanEna
D
Qi-1
Clk
Qi
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
5/55
4
7
CLK
ScanENA
FF1 FF2 FF3 FF4
FF1 FF2 FF3 FF4
Primary
Inputs
Primary
Outputs
8
1.35 millions of flip-flops, all are scannable !!
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
6/55
5
9
Switching (dynamic) Power
Leakage (static) Power
10
Much higher than during functional operations
determine
Excessive
Test PowerDuring
Structural
Testing
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
7/55
6
11
Much higher than during functional operations
Toggle activity under functional mode : 15%-20%
Toggle activity under test mode : 35%-40%
Power under test mode up to 3.8X power during functional mode
12
Main reasons for excessive test power
For conventional (non low-power) designs, dynamic power
is the main responsible for excessive test power !!
Leakage power is a real issue during IDDQ test (reduced sensitivity) and
during burn-in test (can result in thermal runaway condition and yield loss)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
8/55
7
13
Conventional (slow-speed) scan testing
14
At-speed scan testing with a LOC/LOS scheme
Example
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
9/55
8
15
16
one peak is not relevant
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
10/55
9
17
high average power
problems may occur
18
one peak is highly relevant
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
11/55
10
19
Elevated Average Power
Reduced ReliabilityChip Damage
Yield Loss
20
Elevated Average Power
Low Test Throughput
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
12/55
11
21
Elevated Peak Power
Manufacturing Yield Loss
As huge designs can be manufactured today, most power-related
test issues (during scan) are due to excessive peak power
22
IR Drop
I
L(di/dt)
i(t)
L
(courtesy: O. Sentieys, IRISA, France)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
13/55
12
23
(courtesy: A. Domic, Synopsys, USA)
24
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
14/55
13
25
(courtesy: K. Hatayama, STARC, Japan)
activate all modules activate only one module
26
Costly or longer test time
Straightforward Solutions
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
15/55
14
27
Main classes of dedicated solutions
Make test power dissipation comparable to functional power
Objective
28
Main classes of dedicated solutions
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
16/55
15
29
The Role of Test Data in Test Flow
Comparison
Design Data
Tester
ActualResponses
LSI Chip
Fabrication
TestPatterns
ExpectedResponses
Passing Chips Falling Chips
Probe Card
Prober
Socket Board
Handler
GO NG
Test
Data
30
Target of Scan Test Pattern Generation
PI
PPI
PO
PPO
Scan-In
Scan-Out
Scan Chain
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
17/55
16
31
Automatic Test Pattern Generation (ATPG)
X
1
X
0
1
1
X
X
CUT
32
X-Bits in Test Cubes
20
30
40
50
6070
80
90
100
0 50 100 150 200 250 300 Vec.
X-bitsRatio(%)
i.e., assigning logic values to X-bits
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
18/55
17
33
Conventional X-Filling Technique: Random-Fill
Small test pattern count due to fortuitous detection
High test (shift and LTC) power
34
Various Low-Power X-Filling Techniques
Low-Shift-Power X-Filling
0-fill1-fill
MT-filladjacent fill / repeat fill
MTR-filloutput-justification-based X-
filling
Low-LTC-Power X-Filling
PMF-fillLCP-fill
preferred fillJP-fill
CTX-fill
CCT-fillCAT-fill
PWT-fillstate-sensitive X-filling
Low-Shift-and-LTC-Power X-Filling
impact-oriented X-filling hybrid X-filling bounded adjacent fill
Low-PowerX-Filling for Compressed Scan Testing
0-fill PHS-fill CJP-fill
(source: X. Wen and S. Wang, Low-Power Test Generation, Chapter 3 in Power-Aware Testing and
Test Strategies for Low Power Devices, edited by P. Girard, N. Nicolici & X. Wen, Springer, 2009)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
19/55
18
35
Example 1: Low-Shift-Power X-filling
Goal
0-fill1-fill MT-fill adjacent-fill
0-fill
36
Example 1 (contd): Low-Shift-Power X-filling
X
X
X-
X-
X-
0 0 0 0 0
X
1 1 1 1 1
1 1 0 0 0
0 0 1 0 0
0 1
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
20/55
19
37
Example 2: Low-LTC-Power X-Filling
0
1
1
0
0 0
Stimulus Response
Combinational Circuit
Launch-to-Capture (LTC) Power
IR-Drop
Delay Increase
Malfunction
VDD
Timing
38
X0X
X
X0X1
XX
1
Example 2 (contd): Low-LTC-Power X-Filling
1
01
0
0
1
Signal Probability Calculation
Logic Value Determination
Justify
JP-Fill
0-Prob
1-Prob
No DecisionNext Pass
1
Assign
f
Test Response
(source: X. Wen et al., Proc. ITC, Paper 25.1, 2007)
too close to call
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
21/55
20
39
Example 2 (contd): Low-LTC-Power X-Filling
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0 50 100 150 200 250 300
Test Vectors.
-
OriginalJP-Fill
LTC
IR-Drop(V)
Risky
Safe
Risky Safe
JP-FillOriginalChip-Level
IR-Drop Distribution
40
Example 3: Low-Shift-&-LTC-Power X-Filling
Test Cube XXX Shift Direction
0000 0 0 0 0 0 00000
0 0 0 0 11111 0 0 0 0 0
0 0 0
0-Fill
Adjacent Fill
Bounded Adjacent Fill
000
1st0-Constraint Bit Bounding Interval = 6
X adjacent fill
The occurrence of 0 in the resulting fully-specified test vector is increased, which helps
reduce shift-out and LTC power. making use of the benefit of 0-fill
At the same time, applying adjacent fill helps reduce shift-in power.
(source: A. Chandra et al., Proc. VTS, pp. 131-138, 2008)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
22/55
21
41
Summary - (1)
Vector-Pinpoint
Area-Pinpoint
42
Summary - (2)
2012
Current
Low-Power Test
Generation
Coarse-Grained Fine-Grained
Global
20??
Future
Power-Safe Test
Generation
Regional
(source: X. Wen, ETS, Invited Talk, 2012)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
23/55
22
43
Main classes of dedicated solutions
Make test power dissipation comparable to functional power
Objective
44
Shift Power Reduction LTC Power Reduction
Partial Capture
circuit modification- scan chain disable
- one-hot clocking
- capture-clock staggering Scan Chain Modification
scan cell reordering
- scan chain segmentation
scan chain disable
Shift Impact Blocking
blocking gate special scan cell first-level power supply gating
Scan Clock Manipulation
splitting staggering
- multi-duty clocking
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
24/55
23
45
Example 1: Low power scan cell
(source: A. Hertwig and H.-J. Wunderlich, Proc. ETW, pp. 49-53, 1998)
46
ENAB
SISO
Example 2: Scan chain segmentation
ENAA
ScanENA
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
25/55
24
47
CK/2CK/2
CK
SISO
Example 3: Staggered clocking
48
Example 4: Scan cell reordering
(courtesy: H.-J. Wunderlich and C. Zoellin, Univ. Stuttgart)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
26/55
25
49
Example 5: Inserting logic into scan chains
(courtesy: H.-J. Wunderlich and C. Zoellin, Univ. Stuttgart)
50
Example 6: Scan segment inversion
(courtesy: H.-J. Wunderlich and C. Zoellin, Univ. Stuttgart)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
27/55
26
51
Main classes of dedicated solutions
52(source: P. Girard et. al., Proc. VTS, pp. 407-412, 1999)
Example 1: Masking logic insertion during BIST
ViVj
Vk
Vl
V0
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
28/55
27
53
Example 2: Adaptation to Scan-Based BIST
(source: F. Corno et. al., Proc. DFT, pp. 219-226, 1999)
54
(source: S. Wang and S. Gupta, Proc. ITC, pp. 848-857, 1997)
Example 3: Dual-speed LFSR for BIST
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
29/55
28
55(source: S. Wang and S. Gupta, Proc. ITC, pp. 848-857, 1997)
Example 3 (contd): Dual-speed LFSR for BIST
56
Example 4: Coding for compression and test power
(courtesy: K. Chakrabarty and S.K. Goel, Duke Univ.)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
30/55
29
57
Example 5: Linear Finite State Machines
(source: F. Czyusz et al., Proc. VTS, pp. 75-83, 2007)
58
Main classes of dedicated solutions
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
31/55
30
59
Improve Test Throughput by Exploiting Design Modularity
(source: Y. Zorian, Proc. VTS, pp. 4-9, 1993)
60
Example 1: Resource Allocation and Incompatibility Graphs
(source: R. Chou et al., IEEE Trans. on VLSI, Vol. 5, No. 2, pp. 175-185, 1997)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
32/55
31
61
Example 1 (contd): Power Model and Test Schedule
(source: R. Chou et al., IEEE Trans. on VLSI, Vol. 5, No. 2, pp. 175-185, 1997)
62
Example 2: Power Profile Manipulation
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
33/55
32
63
Example 3: Thermal Considerations
64
Test Power Reduction Strategies
High
High
Low
Low
Low
Low
Low
Low
Minimum
Minimum
(source: S. Ravi, TI, ITC07)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
34/55
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
35/55
34
67
The new power-performance paradigm:
Adoption of low-power design and power management techniques
Density 2X
Performance 1.4X
Power 1X
Cost 0.5X
Applications
Units
Users
Revenue
(courtesy: M. Hirech, Synopsys, USA)
68
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
36/55
35
69
Main LPD techniquesPower reduction
These techniques are often combined together to
achieve the maximum power optimization value
70
Even more critical for Low-Power Design !!
PM structures
often disabled
during test
application
(courtesy: X. Wen, KIT)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
37/55
36
71
And still target:
72
Main classes of dedicated solutions
Make test power dissipation comparable to functional power
Objective (again)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
38/55
37
73
Main classes of dedicated solutions
74
Multi-Voltage Design Styles
0.9V0.9V0.7V0.7V
0.9V0.9V
OFF0.7 0.9V0.7 0.9V
PWR
CTRL
PWR
CTRL
0.7V0.7V0.9
V
OFF0.9V0.9V0.7V0.7V
0.9V0.9V
Dynamic Voltage
Frequency Scaling
(DVFS)
Multi-Voltage
with power gating
Multi-Voltage
Multi-supply
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
39/55
38
75
Example 1: Multi-Voltage Aware Scan Cell Ordering
L ogical Physical M ulti-
Voltage
1
2
3
4
5
6
7
8
9
Design (1.2V)
B1
B3
B2
C2
C1
C3
A1
A2
A3
Block A
0.9V
Block B
1.2V
Block C
0.9V
Position
Ordering
76
EncounterTM
Example 2: Power-Aware Scan Chain Assembly
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
40/55
39
77
Example 3: Voltage scaling in scan mode
78
Example 3 (contd): Voltage scaling in scan mode
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
41/55
40
79
S
OS
I
PD4
SC1 SC4
PD1
Test
Controller
Power
Controll
Power
Controller
Test
Controller
PD2
SC2
PD3
SC3
Example 4: Power Domain Test Planning
Objective:
(Source: M. Hirech, Synopsys, DATE, 2008)
80
Main classes of dedicated solutions
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
42/55
41
81
not
only in logic portions but also
in clock trees (> 50%)
Basic Clock Gating Design
GClk
Clock Gator
82
Impact of Clock Gating on Test
Impact on Shift ModeNegative
Impact on Capture ModePositive
Shift Operation Guarantee Needed
Good for LTC Power Reduction
Dynamic In-ATPG Techniques
Static In-ATPG Techniques
Post-ATPG X-Filling Techniques
DfT for Clock Gating Logic
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
43/55
42
83
Example 1: DfT for Clock Gating Logic
Shift ModeSE
SE
Capture ModeSE
Clock Gator
SE
84
Example 1 (contd): DfT for Clock Gating Logic
shift mode
SE
capture modeSE
Shift Register Operation
LTC Power Reduction
1
SE
11
SE
0
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
44/55
43
85
Example 2: LTC Power Reduction by Clock Gating
(source: R. Illman et al., Proc. LPonTR, pp. 45-46, 2008)
default value
A preset value is to be used to fill an X-bit in a test cube
Identify all clock gators
For each clock gator, calculate a set of input settings that set the clock off
The values in each input setting are default values..
Generate a test cube for fault detection
Assign default values to the X-bits in the test cube
If there are multiple choices of default values, use the one that can turn-off more FFs.
86
Example 2 (contd): LTC Power Reduction by Clock Gating
Test Cubes after ATPG Test Cubes after Assigning Default Values
Default Values
S1 = 0 / S2 = 0
S1 = 1 / S2 = 1
(source: R. Illman et al., Proc. LPonTR, pp. 45-46, 2008)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
45/55
44
87
Example 2 (contd): LTC Power Reduction by Clock Gating
Circuit Statistics
(source: R. Illman et al., Proc. LPonTR, pp. 45-46, 2008)
88
Summary
shift mode
capture mode
Assign pre-determined clock-gator disabling values to dont-care bits
(X-bits) in a test cube initially generated for fault detection.
Find and assign proper logic values to dont-care (X-bits) in a test
cube so as to disable as many clock gators as possible.
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
46/55
45
89
Main classes of dedicated solutions
90
Typical Power Management (PM) Structures
Unified Power Format Common Power Format
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
47/55
46
91
Segmented Switch
Segmented Switch
Header Switch Footer Switch Symmetric Switch
Header Switch/ Footer Switch/ Symmetric Switch
Example 1: Test for Power Switches
92
Example 1 (contd): Test for Power Switches
Test for Short TE standby_t
Out Out
Test for Open TE standby_t
Out Out
from powercontrol logic
(source: Goal (NXP) et al., Proc. ETS, pp. 145-150, 2006)
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
48/55
47
93
Example 2: Parametric Test of Micro Switches
94
Example 2 (contd): Parametric Test of Micro Switches
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
49/55
48
95
Example 2 (contd): Parametric Test of Micro Switches
96
Example 3: Test for State Retention Registers (SRRs)
Master
Latch
Retention
Latch
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
50/55
49
97
Example 3 (contd): Test for State Retention Registers (SRRs)
SRR
V
(2) Save
Restore
V
Repeat for V = 0 and V = 1
Retention Capability Test
98
Example 4: Test for Isolation Cells - (1)
SLEEP_1
SLEEP_2
ISO_1 OUT
OUT
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
51/55
50
99
Example 4: Test for Isolation Cells - (2)
OR
Power Domain OFF
Power Domain ON
Check for 1
AND
Power Domain OFF
Power Domain ON
Check for 0
FF
Power Domain OFF
Power Domain ON
Check for 0 and 1
Power Domain ON
Power Domain ON
Power Domain ON
100
Example 5: Test for Level Shifters
Power Domain 1 Power Domain 2
Level Shifter
Test as a whole under all power supply voltage combinations
of Power Domain 1 and Power Domain 2.
(VDD-1, VDD-2, ) (VDD-a, VDD-b, )
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
52/55
51
101
Example 6: Test for Power Mode Control Logic in SRAMs
Low-Power SRAM
corecellarray
IOIO control
wordlinedriver
corecellarray
102
Summary
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
53/55
52
103
(source: CADENCE, 2007)
104
Solutions for high quality at-speed fault coverage are needed !
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
54/55
53
105
Synopsys:
http://www.synopsys.com/Tools/Implementation/RTLSynthesis/Test/Pages/def
ault.aspx
106
Mentor:
http://www.mentor.com/products/silicon-yield/logic_test/
Cadence:
http://www.cadence.com/products/ld/test_architect/pages/default.aspx
-
8/10/2019 Power Aware Testing Strategies (Slides).pdf
55/55
107
Thank You !