Designing a Connected Radio with Vybrid Automotive Solutions

TM
September 2013
•
Consumers in the mass market are driving
demand for seamless integration of their
smart-devices to bring personal content
into the vehicle
•
OEM software costs continue to rise due
to increased resource and development
needed to support higher level of
integration as well as future software
upgrade needs
•
Rearview camera systems in all
passenger vehicles driving the need for
cost-effective solution to meet upcoming
governmental safety mandates
•
The need to isolate the vehicle from
consumer device threats adds an
additional layer of complexity to the design
of driver information systems
TM
2
Gen 3 Portfolio
Power Architecture®
and ARM®
Gap and
architecture
break
Qorivva
MPC5645
ARM v7
i.MX6xx
Dual/Quad
i.MX 6Solo
Vybrid
R Series
Next Gen
Cluster / Gauge
Cluster / Gauge
Qorivva
MPC5606
8-,16- and 32bit MCUs
32-bit MCU/MPUs and
GPUs based on ARM
TM
3
Cluster / Gauge
Next Gen
Cluster / Gauge
Bridge, Scale and Unify
i.MX5xx
Gen 4 Portfolio
Today’s Connected Radio
Vybrid Automotive Solutions
Collection of discrete ICs
• Many external components:
MCU, MPU, ADC, DRAM, PMIC
• Cost-optimized solution integrates:
─ MCU + MPU
─ Video ADC
─ On-chip SRAM
─ Single 3.3V supply
• Additional assembly steps and
compromised reliability
• Higher total system cost
• Improved reliability, fewer components,
and reduced system complexity
• Simple, cost-effective package
TM
4
AUTOMOTIVE
Feature Integration
VF5xxR - Cortex A5+M4, 400MHz, 364BGA
OpenVG, TFT, MLB, USBx2, VideoADC, DDR3
VF3xxR - Cortex A5+M4, 266MHz, 176LQFP
OpenVG, TFT , MLB, USB, VideoADC
Production
Performance
TM
5
NEON optimized
CODECS/Libraries
Cortex-M4
Up to 166MHz
16KB/16KB L1
HMI Tools
Comm Stacks
Royalty-free
RTOS
Cortex-A5
Multimedia
Connectivity
Up to 400MHz
NEON/FPU
32KB/32KB L1
Power
Management
USB OTG + Phy
USB OTG + Phy
- single 3.3V supply
- low voltage reset
1.5MB SRAM
CAN
x2
System and
General Purpose
System Connectivity
Audio I/O
DMA
UART/LIN
x6
HAB 4.1 Security
Tamper Detect
ESAI x1 (i2s x6)
SPI
x3
Watchdog Timer
Other Timers (x8)
SP/DIF
Receiver/Transmitter
Media Local Bus
3-wire
Real Time Clock
Sample Rate
Convertor
10/100 Ethernet
Pulse Width
Modulator
GPIO
SAI x4 (i2s x4)
DRAM (16-bit) LP
, DDR3
2x 12-bit SAR ADC
Dual Quad-SPI
Flash (SDR and DDR)
2x 12-bit DAC
With Tone Generation
Segment Display
Controller (4x40)
Internal Temperature
Monitor
TM
•
Display I/O
2D-ACE x2
Animation &Comp
Engine
OpenVG GPU
(GC355)
External Memory
•
eMMC/SD
x2
I2C
x4
NAND/NOR Flash
8/16b
•
•
•
Camera Input, 18-bit +
Composite (4 to 1))
6
Unique dual-core architecture
with apps processor to run highlevel OS (i.e. Linux) and control
processor to run RTOS (i.e.
MQX)
Ability to segment tasks that
need predictable latencies to
execute on the M4 and execute
graphical and connectivity tasks
on A5
Secure boot and cryptographic
algorithm acceleration for
sensitive applications
Multimedia hardware IP that
offloads pixels processing from
the cores
Real time sub-system including
PWM and ADC for motor control
AUTOMOTIVE
Feature
TM
3xxR Family
5xxR Family
CPU
266MHz Cortex-A5
133MHz Cortex-M4
400MHz Cortex-A5
133MHz Cortex-M4
On-chip memory
1.5MB (512K ECC)
1.5MB (512K ECC)
OR
1MB & 512K L2 cache
Serial Flash interface
Dual QuadSPI
Dual QuadSPI
NAND
Yes (8-bit)
Up to 32-bit HW ECC
Yes (16-bit)
Up to 32-bit HW ECC
FlexBus interface
(parallel NOR)
Yes (addr / data mux’d)
Yes, (addr / data mux’d plus 8-bit
dedicated data)
DRAM interface
No
16-bit LPDDR2/DDR3
Display interface
TFT & 40x4 Segmented LCD OR
2x TFT up to WQVGA
TFT & 40x4 Segmented LCD OR
2x TFT up to WVGA
Video ADC / Camera Input
2x Composite
24-bit parallel
4x Composite
24-bit parallel
10/100 Ethernet w/ IEEE1588
1
2
10-channel 12-bit ADC
Yes
Yes
USB
1x USB OTG HS
2x USB OTG HS
Audio interface
SAI x3 (i2s x3)
ESAI x1 (2 Tx, 4 Tx or Rx)
SAI x4 (i2s x4)
ESAI x1 (2 Tx, 4 Tx or Rx)
UART, DSPI, I2C
4, 3, 4
6, 4, 4
SD/MMC interface
1
2
CAN
2x FlexCAN
2x FlexCAN
MOST
1x MLB50
1x MLB50
GPIO
Up to 115
Up to 136
Package
176LQFP
364BGA
7
AUTOMOTIVE
S V F 3 1 1 R 3 K1 C MK 2
Speed (A5 core)
2 = 266MHz
4 = 400MHz
Qualification Status
P = engineering samples
S = automotive qualified
Brand: V = Vybrid
Package
KU = 176LQFP
MK = 364BGA
Series: F = current
Family
3 = Standard
5 = Advanced
Temp Spec
C = -40 to +85C Ta
Core
1 = Cortex A5
2 = Cortex A5 + M4
3 = M4 Primary
Graphics
1 = No Open VG
2 = Open VG
Revision
K1 = Rev 1.1
Memory Option
3 = Standard (1.5MB SRAM)
2 = Optional (1MB SRAM and 512K L2 Cache)
Version
R = Auto
Not all combinations are available. Please refer to part number list
TM
8
AUTOMOTIVE
Part Number
Sample Part Number
(Superset)
Package
SVF311R3K1CKU2
PVF322R3K1CKU2
LQFP-EP 176 24*24*1.6
A5-266, 176LQFP-EP
SVF312R3K1CKU2
PVF322R3K1CKU2
LQFP-EP 176 24*24*1.6
A5-266, OpenVG GPU, 176LQFP-EP
SVF321R3K1CKU2
PVF322R3K1CKU2
LQFP-EP 176 24*24*1.6
A5-266, M4, 176LQFP-EP
SVF322R3K1CKU2
PVF322R3K1CKU2
LQFP-EP 176 24*24*1.6
A5-266, M4, OpenVG GPU, 176LQFP-EP
SVF332R3K1CKU2
PVF332R3K1CKU2
LQFP-EP 176 24*24*1.6
A5-266, M4 Primary, OpenVG GPU, 176LQFPEP
SVF511R3K1CMK4
PVF522R3K1CMK4
MAP 364 17*17*1.5 P0.8
A5-400, 364BGA
SVF512R3K1CMK4
PVF522R3K1CMK4
MAP 364 17*17*1.5 P0.8
A5-400, OpenVG GPU, 364BGA
SVF521R3K1CMK4
PVF522R3K1CMK4
MAP 364 17*17*1.5 P0.8
A5-400, M4, 364BGA
SVF522R3K1CMK4
PVF522R3K1CMK4
MAP 364 17*17*1.5 P0.8
A5-400, M4, OpenVG GPU, 364BGA
SVF532R3K1CMK4
PVF532R3K1CMK4
MAP 364 17*17*1.5 P0.8
A5-400, M4 Primary, OpenVG GPU, 364BGA
SVF522R2K1CMK4
N/A
MAP 364 17*17*1.5 P0.8
A5-400, M4, L2 Cache, OpenVG GPU, 364BGA
Description
Part numbers highlighted bold are supersets
TM
9
OpenVG GPU
•
•
•
•
•
2D-ACE (DCU)
•
2D Graphics
Accelerator
Accelerates
cluster needles
at 60fps
Rendering of
true-type fonts
with 16x AA
Warping for
Head Up
Display
applications
Graphics
acceleration for
dual display
systems
•
•
•
TM
Display
controller with
2D acceleration
features
6 layer blend &
64 layer support
Enables low
memory
footprint GFx
Inline Run
Length Encode
VideoADC
•
•
•
•
•
Composite
Video (CVBS)
Decoder & AFE
PAL / NTSC
Y/C Separation
2D Comb
Filtering
Interfaces to
VIU3 for
VIU3
•
•
•
•
•
10
Digital Video
Interface
Supports ITU656, RGB, YUV
input formats
Image
Up/Downscaling
Video Deinterlacing
Brightness and
contrast
adjustment
QuadSPI
•
•
•
•
•
Serial Flash
Controller.
66MHz Dual
DDR serial flash
interface:
133MByte/sec
Flexible
buffering
Support for
multiple flash
vendors
Graphics
storage and
eXecute-inPlace (XIP)
1. Cost-optimized for Connected Radios
−
−
−
QFP-versions for lower board costs/complexity
On-chip SRAM and RTOS option for running
without DRAM
Market specific integration
2. Added value through software and enablement
−
−
“Foundation” software suite (productiongrade/terms) provided
Pre-integrated/tested software “accessories”
available from Freescale
3. 2D Animation & Composition Engine (2D-ACE)
−
−
−
Dramatically reduces VRAM requirements for
large color displays
Delivers 60fps animation w/ very low CPU load
2D GPU for line-drawing, transformations, raster
ops
4. Cortex A5 CPU & peripheral reuse enables
scalability to i.MX6 family
TM
11
Ideal as connected, displaybased radio and graphics
processor for cluster and
Multi-Function Display
Delivers eye-catching graphics
with very efficient use of
system resources
•
•
•
•
•
MCU handles CAN +
system wake-up
MPU/DSP for Media
Connectivity/Display/
UI
Radio DSP for radio
baseband & Audio
I/O
Add-on modules for
DAB / HD Radio
External ADC for
analog back-up
camera
TM
•Board Space
•PCB Complexity
•Cost of components
12
•
•
•
•
•
•
Vybrid handles all
connectivity, media
playback, vehicle
communications and
power-up/down with A5 +
M4 architecture
No ext DRAM for many
use-cases
No need for complex PMIC
No external Video ADC
Radio DSP for radio
baseband & Audio I/O
Add-on modules for DAB /
HD Radio
TM
•Reduce PCB size and Complexity
•Fewer components
•Production Grade Connectivity SW
13
TM
•
•
Impresario is a Connected Radio
software platform targeted for the
Vybrid. The platform is made up of
functional blocks to deliver
automotive connected radio
functionality
Scalable feature set from Entry
Level to Advanced solutions.
Customer Application
Impresario API
Impresario
Middleware
OS (MQX/Linux)
Drivers
Vybrid Silicon
TM
15
The Connected Radio software platform is being developed for the
Faraday silicon. The platform is made up of functional blocks to deliver
automotive connected radio functionality.
• Multimedia Framework with Radio Tuner and CD Support
− Support
•
•
•
•
•
for industry standard audio CODECs
CE Device Connectivity (iAP, USB, SD)
HMI Framework
Bluetooth Stack and Profiles with AEC/NS (HFP, A2DP, AVRCP)
Automotive Communication Stacks (CAN, LIN, MOST)
Advanced Connected Radio Functionality
− Text
to Speech
− Speech Recognition
− MirrorLink / iPod Out
TM
16
Which Operating Systems will be supported?
−
The Connected Radio platform leverages the Freescale Operating
System Abstraction layer to provide a Connected Radio Platform on
MQX and Linux*
−
Through extensions to the OSA, Connected Radio can be modified
to work with any modern operating system that supports common
OS primitives (memory management, signals, mutexes, etc)
TM
17
•
Underpinned by Freescale MQX and Linux
platform support and Commercial terms
−
Standard Warranty and Indemnification
−
Available Support and Maintenance
•
CE Device Connectivity (iPhone, USB, SD)
•
HMI Framework optimized for on-chip
graphics resources (2D-ACE, OpenVG
GPU)
•
Multimedia Framework with Radio Tuner
and CD Player
•
Bluetooth Stack and Profiles
•
Automotive Communication Stacks
•
Advanced Connected Radio Functionality
−
Text to Speech, Speech Recognition
−
iPod Out, Terminal Mode
TM
18
Which hardware will the Connected Radio platform support?
− Production
level code with full peripheral support will be available
only on the Faraday reference hardware
Which Compilers will the Connected Radio platform support?
−
Connected Radio will provide support for the following compilers

−
GNU
Support for porting of Connected Radio to a new compiler can be
provided by the Freescale Professional Services Team
TM
19
•
Pre-integrated key technologies from
industry-leading partners help you get to
market faster
•
Proven, tested, production-ready code
ready for evaluation and integration
•
Specifically optimized for Vybrid
performance and memory footprint
•
Partners for HMI tools, Bluetooth, and
AEC/NS
TM
20
Sybase iAnywhere
Cybercom blueGO
Alango Technologies
•
Bluetooth protocol
stack
Bluetooth software
framework
Voice Communication
Package
•
Industry leader in
Bluetooth and active
SIG leadership
•
Multi-profile Bluetooth
SW application
framework
•
Signal processing
enabling high quality
voice communications
•
Certified Bluetooth
compliance with 12
years of maturity
•
“We love making
Bluetooth easy”
•
AEC / NS
TM
21
−
Acoustic Echo Canceller
−
Noise Suppressor
TM
22
•
The 2D-ACE is an advanced graphics control module that directly
drives an external TFT LCD
− Allows
full flexibility of TFT display sizes
− Fetches bit-mapped “sprites” from on-chip or off-chip memory and
places them on graphic layers
− Blends the layers using
− Has support for a cursor separate from the “sprite” graphics
− Supports multiple graphic formats in RGB and YUV format and with and
without alpha and run length encoding



16bpp RGB565, RGB1555, RGB4444, 24bpp RGB888 and 32bpp ARGB8888
Indexed colors with variable bit depths from 1 bit per pixel (bpp) to 8bpp and
APAL8
YUV format - YCbCr422
− Adjusts
the gamma of the graphics to match the TFT in use and dither
pixel colors on panels with less than 24-bit color
− Displays a test signal to allow calibration of panel and system test
TM
23
Static
objects
Any memory
eDMA
Dynamic
objects
RAM
Any
Anymemory
memory
GPU
RAM
RAM
64 obj
2D-ACE
Color conv
pre blend.
RLE exp.
32bpp
Animate.
Alpha,
Pos,
Obj …
TM
Blend
Color key
•
Pipelined
operation up to
90MHz pixel
clock
•
Memory size
optimized
•
Per obj. ani.
frame rate
Dither.
Gamma corr
CRC check
6 planes
24
•
The DCU combines layers or “sprites” to create the final content
−
−
−
−
−
There are up to 66 different sources of content possible

64 programmable layers that contain source graphics

A cursor layer

1 layer as a default color for the background
Layers are in a fixed priority to each other
For each pixel position

the DCU fetches a pixel from the topmost layer placed there AND

a pixel from the next layer in the priority

and pixels from up to four further layers (dependent on user configuration)
If indexed colors are used these are converted to 32bpp before processing
The fetched pixels are then blended to give the display content for that position.

−
−
The blending attributes are determined per layer and the lowest priority pixel’s blending
attributes are ignore
Each resulting pixel can be gamma corrected
The output format is 8-bits per channel(24bpp)
TM
25
•
A layer is the mechanism by which
graphics are displayed on the
panel
• The DCU has a set of 9 registers
to configure each layer
• The layer registers configure
− Height
& width of layer (pixels)
− Signed position on panel (x,y)
− Pointer to graphic (32-bit)
− Graphic coding (bpp) & CLUT,
blending, type, tile & safety
− Chroma limits (max & min)
− Tile size
− Transparency mode colors
TM
x
26
y
TM
27
Layer 20:
16x480
(0,0)
YCbCr
Alpha:100%
Serial flash
TM
28
Layer 20:
800x480
(0,0)
YCbCr
Alpha:100%
Tile size: 16x480
Serial flash
TM
29
Layer 12:
800x430
(0,50)
8BPP
Alpha:100%
Simple chroma
DRAM
TM
30
Layer 10:
52x34
(350,70)
4BPP
Alpha:100%
SRAM
TM
31
Layer 9:
60x60
(80,348)
8BPP
Alpha:100%
Simple Chroma
SRAM
TM
Layer 8:
80x80
(100,350)
8BPP
Alpha:100%
Simple Chroma
Serial flash
32
Layer 6:
40x40
(400,30)
4BPP
Alpha:100%
Simple Chroma
DRAM
Layer 6:
40x40
(5,120)
4BPP
Alpha:100%
Simple Chroma
Serial flash
TM
33
TM
34
TM
35
•
Full fixed function hardware vector
graphics GPU
•
Hardware Tessellation
Minimum CPU involvement
•
•
•
•
•
•
•
16X FSAA
Photorealistic quality
No performance degradation
Multi-format rendering
sRGB color transformation
Video image conversion
•
High-quality vector font rendering
support
•
Dedicated GPU for QoS
requirements
TM
36
•
•
•
Vector graphics are drawn and stored as mathematical vector
formulae
Each vector and fill is assigned a
color value, instead of assigning color
to each separate pixel
A black circle can be represented as:
= r cos θ
y = r sin θ
or:
− r2 = x2+ y2
− With color value 0000 for black
−x
•
Benefits
–
–
–
Infinitely scalable
Independent of screen resolution
Saves data memory
TM
37
GC355 GPU Core
•
Fully HW accelerated OpenVG
pipeline
−
−
−
−
−
−
−
−
•
−
−
Memory Controller
Graphics
Pipeline
Front End
Vector Graphics Pipeline
Vector Graphics Engine
Imaging Engine
VG Pixel
Engine
Entire OpenVG pipeline hardware accelerated
Continuous high frame rates;
Outstanding 16X MSAA quality,
Very low OpenVG CPU driver load
Native rendering of true-type fonts,
w/16x Anti-Aliasing
Additional graphics acceleration for dual
display systems
Also used in i.MX6D/Q
TM
AXI
Host Interface
Benefits:
−
•
Stage 1: Path, Transformation, Stroke,
&Paint
Stage 2: Stroked Path Generation
Stage 3: Transformation
Stage 4: Rasterization
Stage 5: Clipping and Masking
Stage 6: Paint Generation
Stage 7: Image Interpolation
Stage 8: Blending and Antialiasing
AHB
38
The “Video Subsystem”
Digital out:
YUV888
Video
Analogue Front
End (AFE)
Decoder
(VDEC)
Mux
Analogue
In: CVBS
Digital Path
YUV888,RGBnnn,
or ITU656
TM
39
Graphics
VIU
Memory
TFT
DCU
LCD
•
Integrated LCD driver with support up to 40 front-planes and 8
back-planes
− 160
LCD segments
− Adjustable for 3V3 only operation
− Operation can continue of stop mode if required
•
Contrast adjustable in two ways
− VLCD
control
− Contrast adjustment phases for electronic control of contrast
•
Configurable for a wide range of LCDs
− Individual
enables for front- and back-plane signals
− Bias voltage adjustable
− Ability to reconfigure BP and FP signals on the interface to adjust
ratio of FP and BP
TM
40
TM
41
Simple power supply – No PMIC needed
• Only two levels of power sequencing
•
− Controlled
−
•
by Vybrid via NPN transistor
In Low Power modes external NPN is turned off
Need regulators for:
− 3.3V
for main power
− 1.5V/.75V for DDR3; 1.2V/0.6V for LPDDR2
− 5V for USB Host
TM
42
•
Run Mode
− High
•
frequency using PLLs, all peripherals operational
Low Power Modes
Power Run (LPRUN) Mode – PLLs OFF, Sys clock is 24 MHz
− Ultra Low Power Run (ULPRUN) Mode – PLLs OFF, Sys clock is 32
kHz
− WAIT Mode – Both the CA5 and CM4 cores are halted
− Low

Optionally can keep core clocked to debug in WAIT mode
Mode – Cores halted, Peripheral clocks are gated off, Optional
Regulator Power down, Deep Sleep option
− STOP
•
Power Gated Modes
− LPSTOP[3-1]
- Peripheral clocks are gated off, power domains are deenergized, and some (or all) SysRam is lost
TM
43
Power Mode
Current Consumption
Run Mode (@ A5=396 MHz, M4=132 MHz)
- Full Tower using TWR-LCD-RGB and
TWR-SER w/ Ethernet, DDR
600 mA
(Measured on 3.3V supply for
system)
Run Mode (@ A5=396 MHz, M4=132 MHz)
- Linux (150 Threads, 24FPS, 63% CPU)
using TWR-LCD-RGB + TWR-SER2, DDR
775 mA
(Measured on 3.3V supply for
system)
LPRUN (FIRC; with no PLLs)
16 mA
ULPRUN (Slow XOSC)
8 mA
STOP (with Well Bias)
4 - 5 mA
LPS3 (48K memory retained)
270 uA
LPS2 (16K memory retained)
250 uA
LPS1 (FIRC disabled, no memory retained)
40 uA
TM
44
TM
45
•
Vybrid has no onboard flash. In order to store important nonvolatile, constant information we use fuses.
•
For example: Serial No., MAC Address, Boot information
•
Various Boot Interfaces Supported
− Quad
SPI (QSPI[0:1])
− NOR
Flash (FlexBus)
− SD/eSD/MMC/eMMC
− NAND
Flash (NFC)
− FlexCAN[0:1]
− Serial
(ESDHC[0:1])
(using Serial Download Protocol)
Flash (I2C[0:3]/DSPI[0:3]) (Recovery Boot)
− USB
− UART[0:3]
TM
46
•
Various Boot Interfaces Supported
− Quad
SPI (QSPI[0:1])
− NOR
Flash (FlexBus)
− SD/eSD/MMC/eMMC
− NAND
Flash (NFC)
− FlexCAN[0:1]
− Serial
•
(ESDHC[0:1])
(using Serial Download Protocol)
Flash (I2C[0:3]/DSPI[0:3]) (Recovery Boot)
Serial Download Interfaces
− USB
− UART[0:3]
TM
47
BootROM
SPI Flash
SDCard
NAND
NOR
U-Boot
with boot
header –
u-boot.imx
BootROM changes Program
Counter to NOR/QuadSPI region. UBoot starts executing there.
BootROM retrieves code to SRAM,
then change program counter to
SRAM for U-Boot to start executing
TM
QuadSPI Flash
48
•
Serial Flash:
−
−
−
−
•
Communications interface between MCU and external flash memory
Interface similar to standard SPI but optionally utilises 2 or 4 data lines to
transfer Can optionally support DDR to further increase throughput
Command driven interface
Supports both 24- and 32-bit addressing.
Quad Mode SDR Read Command Sequence:
TM
49
•
Dual QuadSPI architecture supports:
−
−
−
−
•
2 external Serial Flashes per QuadSPI module
SDR and DDR Serial Flash
Programmable Sequence Engine for
compatibility to any Serial flash
Supports XIP (Execute-In-Place)
QuadSPI can control 2 x 4-bit serial flashes :
−
−
Accessed separately or….
Parallel mode enabling ‘octal flash’ with data
recombination internally in QuadSPI
•
66MHz clock => 132MByte/sec peak
bandwidth
• Flexible Buffering Scheme:
−
−
−
−
Sub-buffers allocated to specific masters.
Master prioritization
Pre-fetch capability
Suspend & resume for lower priority masters
TM
50
TM