topic: far-field emi analysis methodology and verification o … · 2019-03-13 · speakers...
TRANSCRIPT
Image
Topic:
o Nam elementum commodo mattis. Pellentesque
malesuada blandit euismod.
Topic:
o Nam elementum commodo mattis. Pellentesque
malesuada blandit euismod.
o Nam elementum commodo mattis. Pellentesque
malesuada blandit euismod.
Topic:
o Nam elementum commodo mattis. Pellentesque
malesuada blandit euismod.
TITLE
Far-Field EMI Analysis Methodology and Verification on SSD Boards
Seong-Jin Mun, (Samsung Electronics Inc.)
Jin-Sung Youn (Samsung Electronics Inc.), Kyung-Nam Park (Samsung Electronics Inc.)
Jieun Park (Samsung Electronics Inc.), Daehee Lee (Samsung Electronics Inc.)
Chan-Seok Hwang (Samsung Electronics Inc.), Jong-Bae Lee (Samsung Electronics Inc.)
SPEAKERS
Seong-Jin Mun
Engineer, Samsung Electronics Inc.
Received the B.S. degree from Department of Electronic Engineering, Sogang
University, Seoul, Korea, in 2010. Since 2010, He has been an engineer with the
Semiconductor Business, Samsung Electronics, Korea. His research interests are
in the area of signal integrity (SI), power integrity (PI), electromagnetic interference
(EMI), thermal integrity (TI) and package/board layout design verification.
Agenda
Background
o Trends of Data Storage
o Necessity of EMI Simulation Methodology
Far-Field EMI Simulation Methodology
o Proposed EMI Simulation Flow
o PCB-Level EMI Solution
Correlation with Far-Field Measurement Results
Relationship Between Board Design and Far-Field EMI
Conclusion
Hard Disk Drive (HDD) Solid-State Drive (SSD)
o Higher Read/Write Rate, Faster Access Time, Lower Power Consumption
Trends of Data Storage
SSD HDD Difference
Media NAND FLASH Magnetic Platters
Read/Write Speed Sequential [MB/s] 540 / 330 60 / 160 ⅹ9 / 2
Random [IOPS*] 98000 / 70000 450/400 ⅹ217 / 175
Data Access Time [ms] 0.1 10~12 ⅹ100~120
Power Consumption Active(Idle) [W] 0.127(0.046) 1.75(0.8) ⅹ13 ↓(ⅹ17 ↓)
HDD SSD
Source: www.samsung.com Source: www.google.com
EMI Problem in SSD Products
EMI(Electromagnetic Interference) becomes a critical issue!
SSD’s speed and density is continuously increasing
Em
issio
n(d
B)
Digital Signal
Spectrum Far Field Emission
Fknee
Increasing Speed Far-Field EMI
Measurement Setup Increasing Density
Overall Emission Spectrum (Speed x1)
Overall Emission Spectrum (Speed x2)
Source: www.google.com Source: high-speed digital system design
Measurement-based EMI verification requires additional cost and time to debug
EMI mechanisms and root causes of the radiated field need to analyze
Necessity of EMI Simulation Methodology
PCB Design-Stage
Improvement
Product
Design Fabrication
EMI
Measurement
Mass
Production Fail
Pass
Product
Re-Design
EMI
Simulation
Product
Design Fabrication
EMI
Measurement
Mass
Production Fail
Pass
Product
Re-Design
∙ F/W solution (Slew Rate Control)
∙ Operating Frequency Change
∙ SSC (Spread Spectrum Clock)
∙ Shielding Solution
∙ Design Revision …
Proposed
Process
Existing
Process
EMI Noises in SSD Products
1
DRAM
1st SATA/
PCIe
>3
NAND
2nd
NAND
3rd DRAM
2nd
NAND
4th
NAND
1st
Radiated
Energy
Frequency
[GHz]
Regulation
Composed of various devices
o NAND, DRAM, CTRL, PMIC ...
o Operated with different speeds / voltages
Data Path
o NAND interface, DRAM interface, Host interface
Higher supply voltage, longer board routing
NAND
Flash
Memory
NAND
Flash
Memory Controller
DRAM
PMIC
Source: www.samsung.com
Far-field EMI simulation methodology [Ref. APEMC2015, Benson Wei et. al.]
Propose 3 items to enhance simulation accuracy and efficiency
o EMI source extraction
o Package and reference plane modeling
o Huygens’ box optimization for N/F to F/F transform
Proposed Far-Field EMI Simulation Flow
EMI Source
Extraction
Structure
Modeling
EM Solving
Based on
Huygens’ Principle
3m
SSD
Near-Field EMI
Result
Far-Field EMI
Huygens’
box
Stage 1: EMI Source Extraction
Read operation at 460 Mbps
(NAND to Controller Interface)
Block diagram
o NAND Flash I/O buffer
o Controller I/O buffer
o Board model (S-parameter)
Input Stimulus
o I/O Buffer : PRBS 27-1
o Strobe and clock signals
– Periodic pattern
– 5% duty cycle distortion
: To include power noise
ControllerInput Stimulus
NAND
1 …0 0 11 00…
0 …1 1 00 11…
1 …0 0 01 10… …
1 …0 0 11 00…
0 …1 1 00 11…
1 …1 0 10 10…
DQSB
IO0
IO7
DQS
REB
RE
Input Stimulus
SSD
Board
Stage 2: Structure Modeling (1)
Frequency [MHz]
Real PKG Model
Virtual PKG Model
Far-
Fie
ld
Ra
dia
ted
En
erg
y [
dB
]
100 200 300 400 500 600 700 800 900 1000
EM simulation with package and board together
o Impractical solution due to simulation time and hardware resources
Propose virtual package model with metal plane
Both results shows similar simulation results
<PKG>
<Virtual PKG model>
Stage 2: Structure Modeling (2)
Far-
Fie
ld
Ra
dia
ted
En
erg
y
[dB
]
2nd
harmonic
3rd
harmonic 1st
harmonic
Measurement
GND reference
GND+Power reference
Propose the merged reference plane
o Power and ground plane are modeled in connection to the one plane
Represents return path through the capacitor of the die
o Simulation result is similar to the measurement
GND
PWR
GND reference GND+Power reference
Stage 3: EM Solving Based on Huygens’ Principle
Proposed near- to far-field transform method based on Huygens’ principle
o Radiated energy simulation at 3-m distance from micro-unit SSD board
Attach I/O current as EMI source Maximum electric field calculation at 3-m sphere surface
How to optimize Huygens’ box size?
3m
Huygens’ Box
N/F Simulation – 2.5D EM solver F/F Simulation – 3D EM solver
Stage 3: EM Solving Based on Huygens’ Principle
Huygens’ box size optimization
o Proper box size is necessary to minimize error for near- to-far field transform
o The radiated field is saturated from 10mm box size
Correlation with Far-Field Measurement Results
NAND read operation (460Mbps)
Good agreement up to 1GHz between simulation and measurement
Far-
Fie
ld E
MI
at
3-m
Dis
tan
ce
[d
Bu
(V/m
)]
30
Frequency [MHz]
300
Measurement (Vertical)
Measurement (Horizontal)
Simulation
1000
Measurement (Vertical)
Measurement (Horizontal)
230
460 690
920
Case Analysis (1)
Investigate on relationship between board design and far-field EMI
o Routing scheme, Signaling scheme, Number of layer, Number of channel F
ar-
Fie
ld E
MI
at
3-m
Dis
tan
ce
[dB
u(V
/m)]
2nd
harmonic
3rd
harmonic
4th
harmonic
Outer routing
Inner routing
Far-
Fie
ld E
MI
at
3-m
Dis
tan
ce
[dB
u(V
/m)]
1st
harmonic
2nd
harmonic
3rd
harmonic
Pseudo differential
Fully differential
vs. Outer
Layer Inner
Layer vs.
Pseudo
Diff.
Fully
Diff.
Investigate on relationship between board design and far-field EMI
o Routing scheme, Signaling scheme, Number of layer, Number of channel
Case Analysis (2)
Far-
Fie
ld E
MI
at
3-m
Dis
tan
ce
[dB
u(V
/m)]
1st
harmonic
2nd
harmonic
3rd
harmonic
10 Layer
12 Layer
Far-
Fie
ld E
MI
at
3-m
Dis
tan
ce
[dB
u(V
/m)]
2nd
harmonic
3rd
harmonic
8 Channel
4 Channel
vs. Add
Layer
2-channel 1-channel vs.
Conclusion
Far-field EMI analysis methodology for commercial SSD products
o EMI source extraction
o PCB structure modeling
o EM solving method based on huygens’ principle
Good correlation between simulation and measurement
Investigation of the relationship between board designs and the radiated energy
o Routing scheme, signaling scheme, number of layer and number of channel
EMI analysis in the design stage prior to the manufacturing process