i.MX 7dual Family of Applications Processors Datasheet
i.MX 7dual Family of Applications Processors Datasheet
Rev. 6, 03/2019
Data Sheet: Technical Data
MCIMX7DxDVx1nSD
MCIMX7DxEVx1nSD
Ordering Information
NXP reserves the right to change the detail specifications as may be required to permit improvements in the design of
its products.
The i.MX 7Dual family of processors is specifically useful for applications such as:
• Audio
• Connected devices
• Access control panels
• Human-machine interfaces (HMI)
• Portable medical and health care
• IP phones
• Smart appliances
• Point of Sale
• eReaders
• Wearables
• Home energy management systems
The features of the i.MX 7Dual family of processors include the following:
• Arm Cortex-A7 plus Arm Cortex-M4—Heterogeneous Multicore Processing architecture enables
the device to run an open operating system like Linux/Android on the Cortex-A7 core and an RTOS
like FreeRTOS™ on the Cortex-M4 core.
• Two Arm Cortex-A7 cores—The processor enhances the capabilities of portable, connected
applications by fulfilling the ever-increasing MIPS needs of operating systems and applications at
lowest power consumption levels per MHz.
• Multilevel memory system—The multilevel Cortex-A7 memory system is based on the L1
instruction and data caches, L2 cache, and internal and external memory. The processor supports
many types of external memory devices, including DDR3, DDR3L, LPDDR2 and LPDDR3, NOR
Flash, NAND Flash (MLC and SLC), QSPI Flash, and managed NAND, including eMMC rev.
• Power efficiency—Power management implemented throughout the IC enables features and
peripherals to consume minimum power in both active and various low-power modes.
• Multimedia—The multimedia performance is enhanced by a multilevel cache system, NEON™
MPE (Media Processor Engine) coprocessor, a programmable smart DMA (SDMA) controller.
• Up to two Gigabit Ethernet with AVB—10/100/1000 Mbps Ethernet controllers supporting IEEE
Std 1588 time synchronization.
• Electronic Paper Display Controller (EPDC)—The processor integrates an EPD controller that
supports E Ink® color and monochrome panels with up to 2048 x 1536 resolution at 106 Hz
refresh, 4096 x 4096 resolution at 20 Hz refresh, and 5-bit grayscale (32-levels per color channel).
• Human-machine interface (HMI)—i.MX 7Dual processor provides up to two separate display
interfaces (parallel display and two-lane MIPI-DSI), CMOS sensor interface (two-lane MIPI-CSI
and parallel).
• Interface flexibility—i.MX 7Dual processor supports connections to a variety of interfaces: two
high-speed USB on-the-go modules with PHY, High-Speed Inter-Chip USB, multiple expansion
card ports (high-speed MMC/SDIO host and other), two Gigabit Ethernet controllers with support
for Ethernet AVB, PCIe-II, two 12-bit ADCs with a total of 8 single-ended inputs, two CAN ports,
and a variety of other popular interfaces (such as UART, I2C, and I2S).
• Advanced security—The processors deliver hardware-enabled security features that enable secure
e-commerce, digital rights management (DRM), information encryption, secure boot, and secure
software downloads. The security features are discussed in detail in the i.MX 7Dual security
reference manual.
• Integrated power management—The processors integrate linear regulators and internally generate
voltage levels for different power domains. This significantly simplifies system power
management structure.
For a comprehensive list of the i.MX 7Dual features, see Section 1.2, “Features.”
Figure 1 describes the part number nomenclature so that the users can identify the characteristics of the
specific part number.
1 Restricted electrical specifications for parts with CPU maximum frequency of 1.2 GHz:
• Temperature range 0 to 85 degrees C (see Table 1)
• VDD_ARM requirements (see Table 9)
Figure 1. Part number nomenclature—i.MX 7Dual family of processors
1.2 Features
The i.MX 7Dual family of processors is based on Arm Cortex-A7 MPCore™ Platform, which has the
following features:
• Two Arm Cortex-A7 Cores (with TrustZone® technology)
• The core configuration is symmetric, where each core includes:
The i.MX 7Dual family of processors integrates advanced power management unit and controllers:
• PMU (power-management unit), multiple LDO supplies, for on-chip resources
• Temperature sensor for monitoring the die temperature
• Software state retention and power gating for Arm and NEON
• Support for various levels of system power modes
• Flexible clock gating control scheme
The i.MX 7Dual family of processors uses dedicated hardware accelerators to meet the targeted
multimedia performance. The use of hardware accelerators is a key factor in obtaining high performance
at low power consumption numbers, while having the CPU core relatively free for performing other tasks.
The i.MX 7Dual family of processors incorporates the following hardware accelerators:
• PXP—PiXel processing pipeline for imagine resize, rotation, overlay and CSC. Off loading key
pixel processing operations are required to support the LCD and EPDC display applications.
• EPDC—Low-power, high-performance, direct-drive, active-matrix electrophoretic display
controller, specifically designed to drive E Ink EPD panels.
Security functions are implemented by the following hardware:
• Arm TrustZone technology including separation of interrupts and memory mapping
• SJC—System JTAG Controller. Protecting JTAG from debug port attacks by regulating or
blocking the access to the system debug features.
• CAAM—Cryptographic Acceleration and Assurance Module, containing cryptographic and hash
engines, 32 KB secure RAM, and true and pseudo random number generator.
• SNVS—Secure Non-Volatile Storage, including secure real time clock
• CSU—Central Security Unit. Responsible for setting comprehensive security policy of the device.
Configured during boot and by eFuses and determines the security-level operation mode as well as
the TrustZone policy.
• HAB—High Assurance Boot—HABv4 with the new embedded enhancements: SHA-256,
2048-bit RSA key, SRK revocation mechanism, warm boot, CSU, and TrustZone initialization.
NOTE
The actual feature set depends on the part numbers as described in Table 1.
Functions, such as display and camera interfaces, connectivity interfaces,
may not be enabled for specific part numbers.
2 Architectural overview
The following subsections provide an architectural overview of the i.MX 7Dual processor system.
MMC/SD
External Memory ARM Cortex A7 Debug Clock & Reset eMMC/eSD
MPCore Platform
NAND FLASH DDR Controller DAP PLLs
CPU1
EIM CPU0 TPIU CCM
GPMI&BCH I$ 32KB D$ 32KB CTIs GPC MMC/SD
NEON FPU SDXC
NOR Flash QSPI SJC SRC
(Parallel)
SCU & Timer XTAL OSC
Internal Memory L2 Cache 512KB Timers
RC OSC Touch Panel
OCRAM 320KB WDOG(4) Control
NOR FLASH ARM Cortex M4 AP Peripherals
ROM 96KB GPT(4)
(Quad SPI) Platform
uSDHC(3)
System Counter
AXI and AHB Switch Fabric
Cortex-M4 Core
Security OCOTP
CAAM I$ 16KB D$ 16KB Flex Timer(2) Keypad
(32KB RAM) USB 2.0
Tamper MPU FPU Host (1) / OTG (2)
Detection CSU TCM 64KB Smart DMA
SDMA AVB ENET(2)
OCOTP (eFuse) 10/100/1000M
PCIe v2.1
Ethernet x2
SNVS(SRTC)
Multi-Core Unit SPBA FlexCAN(2)
RDC MU SIMv2(2)
Display Interface SEMAPHORE Shared Peripherals
I2C(4) WLAN
LCDIF eCSPI(3)
LCD Panel eCSPI(1)
MIPI DSI Image Processing SAI(3)
Pixel Processing PWM(4)
UART(3)
Camera Interface Pipeline(PXP) KPP Smart Card x2
CSI Power Management
Camera UART(4)
MIPI CSI(2 lane) Temp Monitor GPIO(7)
EPD Controller
LDOs IOMUX
ADC (2) CAN x2
USB OTG
Sensors Modem IC EPD Panel PCIe Bus
(dev/host)
3 Modules list
The i.MX 7Dual family of processors contains a variety of digital and analog modules. Table 2 describes
these modules in alphabetical order.
ADC1 Analog to Digital The ADC is a 12-bit general purpose analog to digital
ADC2 Converter converter (ADC2 is not available in the 12x12 package).
Arm Arm Platform Arm The Arm Core Platform includes two Cortex-A7
coresand 1x Cortex-M4. It also includes associated
sub-blocks, such as the Level 2 Cache Controller, SCU
(Snoop Control Unit), GIC (General Interrupt
Controller), private timers, watchdog, and CoreSight
debug modules.
BCH Binary-BCH ECC System control The BCH module provides up to 62-bit ECC
Processor peripherals encryption/decryption for NAND Flash controller
(GPMI)
CCM Clock Control Module, Clocks, resets, and These modules are responsible for clock and reset
GPC General Power power control distribution in the system, and also for the system
SRC Controller, System Reset power management.
Controller
CSI Parallel CSI Multimedia The CSI IP provides parallel CSI standard camera
peripherals interface port. The CSI parallel data ports are up to 24
bits. It is designed to support 24-bit RGB888/YUV444,
CCIR656 video interface, 8-bit YCbCr, YUV or RGB,
and 8-bit/10-bit/16-bit Bayer data input.
CSU Central Security Unit security The Central Security Unit (CSU) is responsible for
setting comprehensive security policy within the i.MX
7Dual platform.
DAP Debug Access Port System control The DAP provides real-time access for the debugger
peripherals without halting the core to access:
• System memory and peripheral registers
• All debug configuration registers
The DAP also provides debugger access to JTAG scan
chains.
ENET1 Ethernet Controller Connectivity The Ethernet Media Access Controller (MAC) is
ENET2 peripherals designed to support 10/100/1000 Mbps Ethernet/IEEE
802.3 networks. An external transceiver interface and
transceiver function are required to complete the
interface to the media. The module has dedicated
hardware to support the IEEE 1588 standard. See the
ENET chapter of the i.MX 7Dual Application Processor
Reference Manual (IMX7DRM) for details.
EPDC Electrophoretic Connectivity The EPDC is a feature-rich, low power, and
Display peripherals high-performance direct-drive, active matrix EPD
Controller controller. It is specifically designed to drive E Ink EPD
panels, supporting a wide variety of TFT backplanes.
Various levels of flexibility and programmability have
been introduced, as well as hardware support for
different E Ink image enhancing algorithms, such as
Regal D waveform support.
FLEXCAN1 Flexible Controller Area Connectivity The CAN protocol was primarily, but not only, designed
FLEXCAN2 Network peripherals to be used as a vehicle serial data bus, meeting the
specific requirements of this field: real-time processing,
reliable operation in the Electromagnetic interference
(EMI) environment of a vehicle, cost-effectiveness and
required bandwidth. The FlexCAN module is a full
implementation of the CAN protocol specification,
Version 2.0 B, which supports both standard and
extended message frames.
FLEXTIMER1 Flexible Timer Module Timer Peripherals Provide input signal capture and PWM support
FLEXTIMER2
GPIO1 General Purpose I/O System control Used for general purpose input/output to external ICs.
GPIO2 Modules peripherals Each GPIO module supports up to 32 bits of I/O.
GPIO3
GPIO4
GPIO5
GPIO6
GPIO7
GPMI General Purpose Memory Connectivity The GPMI module supports up to 8x NAND devices and
Interface peripherals 62-bit ECC encryption/decryption for NAND Flash
Controller (GPMI2). GPMI supports separate DMA
channels for each NAND device.
GPT General Purpose Timer Timer peripherals Each GPT is a 32-bit “free-running” or “set and forget”
mode timer with programmable prescaler and compare
and capture register. A timer counter value can be
captured using an external event and can be configured
to trigger a capture event on either the leading or trailing
edges of an input pulse. When the timer is configured to
operate in “set and forget” mode, it is capable of
providing precise interrupts at regular intervals with
minimal processor intervention. The counter has output
compare logic to provide the status and interrupt at
comparison. This timer can be configured to run either
on an external clock or on an internal clock.
I2C1 I2C Interface Connectivity I2C provide serial interface for external devices. Data
I2C2 peripherals rates of up to 320 kbps are supported.
I2C3
I2C4
IOMUXC IOMUX Control System control This module enables flexible IO multiplexing. Each IO
peripherals pad has default and several alternate functions. The
alternate functions are software configurable.
KPP Key Pad Port Connectivity KPP Supports 8x8 external key pad matrix. KPP
peripherals features are:
• Open drain design
• Glitch suppression circuit design
• Multiple keys detection
• Standby key press detection
LCDIF LCD interface Multimedia The LCDIF is a general purpose display controller used
peripherals to drive a wide range of display devices varying in size
and capability. The LCDIF is designed to support dumb
(synchronous 24-bit Parallel RGB interface).
MIPI-CSI MIPI Camera Interface Multimedia This module provides a two-lane MIPI camera interface
(two-lane) peripherals operating up to a maximum bit rate of 1.5 Gbps.
MIPI DSI MIPI Display Interface Connectivity This module provides a two-lane MIPI display interface
(two-lane) peripherals operating up to a maximum bit rate of 1.5 Gbps.
DDRC DDR Controller Connectivity The DDR Controller has the following features:
peripherals • Supports 16/32-bit DDR3/DDR3L, LPDDR3, and
LPDDR2-1066
• Supports up to 2 Gbyte DDR memory space
OCOTP_CTRL OTP Controller Security The On-Chip OTP controller (OCOTP_CTRL) provides
an interface for reading, programming, and/or
overriding identification and control information stored
in on-chip fuse elements. The module supports
electrically-programmable poly fuses (eFUSEs). The
OCOTP_CTRL also provides a set of volatile
software-accessible signals that can be used for
software control of hardware elements, not requiring
non-volatility. The OCOTP_CTRL provides the primary
user-visible mechanism for interfacing with on-chip fuse
elements. Among the uses for the fuses are unique chip
identifiers, mask revision numbers, cryptographic keys,
JTAG secure mode, boot characteristics, and various
control signals, requiring permanent non-volatility.
OCRAM On-Chip Memory Data path The On-Chip Memory controller (OCRAM) module is
controller designed as an interface between system’s AXI bus
and internal (on-chip) SRAM memory module.
In i.MX 7Dual processors, the OCRAM is used for
controlling the 128 KB multimedia RAM through a 64-bit
AXI bus.
PCIe PCI Express 2.0 Connectivity The PCIe IP provides PCI Express Gen 2.0
peripherals functionality.
PMU Power Management Unit Data path Integrated power management unit. Used to provide
power to various SoC domains.
PWM1 Pulse Width Modulation Connectivity The pulse-width modulator (PWM) has a 16-bit counter
PWM2 peripherals and is optimized to generate sound from stored sample
PWM3 audio images and it can also generate tones. It uses
PWM4 16-bit resolution and a 4x16 data FIFO to generate
sound.
PXP PiXel Processing Pipeline Display peripherals A high-performance pixel processor capable of 1
pixel/clock performance for combined operations, such
as color-space conversion, alpha blending,
gamma-mapping, and rotation. The PXP is enhanced
with features specifically for gray scale applications. In
addition, the PXP supports traditional pixel/frame
processing paths for still-image and video processing
applications, allowing it to interface with the integrated
EPD.
QSPI Quad SPI Connectivity Quad SPI module act as an interface to external serial
peripherals flash devices. This module contains the following
features:
• Flexible sequence engine to support various flash
vendor devices
• Single pad/Dual pad/Quad pad mode of operation
• Single Data Rate/Double Data Rate mode of
operation
• Parallel Flash mode
• DMA support
• Memory mapped read access to connected flash
devices
• Multi-master access with priority and flexible and
configurable buffer for each master
SAI1 Synchronous Audio Connectivity The SAI module provides a synchronous audio
SAI2 Interface peripherals interface (SAI) that supports full duplex serial interfaces
SAI3 with frame synchronization, such as I2S, AC97, TDM,
and codec/DSP interfaces.
SDMA Smart Direct Memory System control The SDMA is a multichannel flexible DMA engine. It
Access peripherals helps in maximizing system performance by offloading
the various cores in dynamic data routing. It has the
following features:
• Powered by a 16-bit Instruction-Set micro-RISC
engine
• Multi-channel DMA supporting up to 32 time-division
multiplexed DMA channels
• 48 events with total flexibility to trigger any
combination of channels
• Memory accesses including linear, FIFO, and 2D
addressing
• Shared peripherals between Arm and SDMA
• Very fast Context-Switching with 2-level priority
based preemptive multi-tasking
• DMA units with auto-flush and prefetch capability
• Flexible address management for DMA transfers
(increment, decrement, and no address changes on
source and destination address)
• DMA ports can handle unidirectional and
bidirectional flows (Copy mode)
• Up to 8-word buffer for configurable burst transfers
for EMIv2.5
• Support of byte-swapping and CRC calculations
• Library of Scripts and API is available
SIMv2-1 Smart Card Connectivity Smart card interface designed to be compatible with
SIMv2-2 peripherals ISO7816.
SJC System JTAG Controller System control The SJC provides JTAG interface (designed to be
peripherals compatible with JTAG TAP standards) to internal logic.
The i.MX 7Dual family of processors uses JTAG port for
production, testing, and system debugging.
Additionally, the SJC provides BSR (Boundary Scan
Register) standard support, designed to be compatible
with IEEE 1149.1 and IEEE1149.6 standards.
The JTAG port must be accessible during platform
initial laboratory bring-up, for manufacturing tests and
troubleshooting, as well as for software debugging by
authorized entities. The i.MX 7Dual SJC incorporates
three security modes for protecting against
unauthorized accesses. Modes are selected through
eFUSE configuration.
SNVS Secure Non-Volatile Security Secure Non-Volatile Storage, including Secure Real
Storage Time Clock, Security State Machine, Master Key
Control, and Violation/Tamper Detection and reporting.
TZASC Trust-Zone Address Security The TZASC (TZC-380 by Arm) provides security
Space Controller address region control functions required for intended
application. It is used on the path to the DRAM
controller.
UART1 UART Interface Connectivity Each of the UARTv2 modules support the following
UART2 peripherals serial data transmit/receive protocols and
UART3 configurations:
UART4 • 7- or 8-bit data words, 1 or 2 stop bits, programmable
UART5 parity (even, odd or none)
UART6 • Programmable baud rates up to 4 Mbps. This is a
UART7 higher max baud rate relative to the 1.875 MHz,
which is stated by the TIA/EIA-232-F standard.
• 32-byte FIFO on Tx and 32 half-word FIFO on Rx
supporting auto-baud
WDOG1 Watchdog Timer peripherals The Watch dog timer supports two comparison points
WDOG3 during each counting period. Each of the comparison
WDOG4 points is configurable to evoke an interrupt to the Arm
core, and a second point evokes an external event on
the WDOG line.
WDOG2 Watchdog (TrustZone Timer peripherals The TrustZone Watchdog (TZ WDOG) timer module
(TrustZone) technology) protects against TrustZone starvation by providing a
method of escaping Normal mode and forcing a switch
to the TZ mode. TZ starvation is a situation where the
normal OS prevents switching to the TZ mode. Such
situation is undesirable as it can compromise the
system’s security. Once the TZ WDOG module is
activated, it must be serviced by TZ software on a
periodic basis. If servicing does not take place, the
timer times out. Upon a time-out, the TZ WDOG asserts
a TZ mapped interrupt that forces switching to the TZ
mode. If it is still not served, the TZ WDOG asserts a
security violation signal to the CSU. The TZ WDOG
module cannot be programmed or deactivated by a
normal mode SW.
The package contact assignments can be found in Section 6, “Package information and contact
assignments.” Signal descriptions are provided in the i.MX 7Dual Application Processor Reference
Manual (IMX7DRM).
Table 3. Special signal considerations
CCM_CLK1_P/ One general purpose differential high speed clock input/output and one single-ended clock input
CCM_CLK1_N are provided.
CCM_CLK2 Either or both of them can be used:
• To feed an external reference clock to the PLLs and to the modules inside the SoC, for
example, as an alternate reference clock for PCIe, Video/Audio interfaces and so forth.
• To output the internal SoC clock to be used outside the SoC as either a reference clock or as
a functional clock for peripherals; for example, it can be used as an output of the PCIe master
clock (root complex use)
See the i.MX 7Dual Application Processor Reference Manual (IMX7DRM) for details on the
respective clock trees.
The CCM_CLK1_* inputs/outputs are an LVDS differential pair.
Alternatively, a single-ended signal may be used to drive CCM_CLK1_P input. In this case
corresponding CCM_CLK1_N input should be tied to the constant voltage level equal to 1/2 of the
input signal swing.
Termination should be provided in case of high frequency signals.
See the LVDS pad electrical specification for further details. CCM_CLK2 is a single-ended input
referenced to ground.
After initialization:
• The CCM_CLK1_* inputs/outputs can be disabled if not used. Any of the unused CCM_CLK1_*
pins may be left floating.
• The CCM_CLK2 input should be grounded if not used.
RTC_XTALI/RTC_XTALO If the user wishes to configure RTC_XTALI and RTC_XTALO as an RTC oscillator, a 32.768 kHz
crystal, (100 k ESR, 10 pF load) should be connected between RTC_XTALI and RTC_XTALO. It
is recommended to use the configurable load capacitors provided in the IP instead of adding them
externally. To hit the exact oscillation frequency, the configurable capacitors need to be reduced
to account for board and chip parasitics.
The integrated oscillation amplifier is self biasing, but relatively weak. Care must be taken to limit
parasitic leakage from RTC_XTALI and RTC_XTALO to either power or ground (>100 M). This will
debias the amplifier and cause a reduction of startup margin. Typically RTC_XTALI and
RTC_XTALO should bias to approximately 0.5 V.
If it is desired to feed an external low frequency clock into RTC_XTALI, the RTC_XTALO pin
should be left floating or driven with a complimentary signal. The logic level of this forcing clock
should not exceed VDD_SNVS_CAP level.
In the case when a high-accuracy realtime clock is not required, the system may use internal low
frequency oscillator. It is recommended to connect RTC_XTALI to ground and keep RTC_XTALO
floating. This will however result in increased power consumption, because the internal oscillator
uses higher power than the RTC oscillator. Thus for lowest power configuration it is recommended
to always install a crystal.
XTALI/XTALO A 24.0 MHz crystal should be connected between XTALI and XTALO.
DRAM_VREF When using DDR_VREF with DDR I/O, the nominal reference voltage must be half of the
NVCC_DRAM supply. The user must tie DDR_VREF to a precision external resistor divider. Use
a 1 kΩ 0.5% resistor to GND and a 1 kΩ 0.5% resistor to NVCC_DRAM. Shunt each resistor with
a closely-mounted 0.1 µF capacitor.
To reduce supply current, a pair of 1.5 kΩ 0.1% resistors can be used. Using resistors with
recommended tolerances ensures the ± 2% DDR_VREF tolerance (per the DDR3 specification)
is maintained when four DDR3 ICs plus the i.MX 7Dual are drawing current on the resistor divider.
It is recommended to use regulated power supply for “big” memory configurations (more than
eight devices)
ZQPAD DRAM calibration resistor 240 Ω 1% used as reference during DRAM output buffer driver
calibration should be connected between this pad and GND.
PCIE_VPH/PCIE_VPH_TX/ Short these pins to VDDA_PHY1P8 if using PCIe. User can tie these pins to ground if not using
PCIE_VPH_RX PCIe.
PCIE_VP/PCIE_VP_TX/PC Short these pins to VDDD_1P0CAP if using PCIe. User can tie these pins to ground with a 10 KΩ
IE_VP_RX resistor if not using PCIe.
VDDA_MIPI_1P8 Short these pins to VDDA_PHY_1P8 if using MIPI. User can leave these pins floating or grounded
if not using MIPI.
VDD_MIPI_1P0 Short these pins to VDDD_1P0_CAP if using MIPI. User can leave these pins floating or grounded
if not using MIPI.
GPANAIO This signal is reserved for manufacturing use only. User must leave this connection floating.
JTAG_nnnn The JTAG interface is summarized in Table 4. Use of external resistors is unnecessary. However,
if external resistors are used, the user must ensure that the on-chip pull-up/down configuration is
followed. For example, do not use an external pull down on an input that has on-chip pull-up.
JTAG_TDO is configured with a keeper circuit such that the floating condition is eliminated if an
external pull resistor is not present. An external pull resistor on JTAG_TDO is detrimental and
should be avoided.
NC Do not connect. These signals are reserved and should be floated by the user.
POR_B This cold reset negative logic input resets all modules and logic in the IC.
May be used in addition to internally generated power on reset signal (logical AND, both internal
and external signals are considered active low).
ONOFF In Normal mode, may be connected to ON/OFF button (De-bouncing provided at this input).
Internally this pad is pulled up. Short connection to GND in OFF mode causes internal power
management state machine to change state to ON. In ON mode short connection to GND
generates interrupt (intended to SW controllable power down). Long above ~5s connection to
GND causes “forced” OFF.
TEST_MODE TEST_MODE is for factory use. This signal is internally connected to an on-chip pull-down device.
The user must tie this signal to GND.
PCIE_REXT The impedance calibration process requires connection of reference resistor 4.7 KΩ 1% precision
resistor on PCIE_REXT pad to ground.
USB_OTG1_REXT/USB_O The bias generation and impedance calibration process for the USB OTG PHYs requires
TG2_REXT connection of 200 Ω (1% precision) reference resistors on each of the USB_OTG1_REXT and
USB_OTG2_REXT pads to ground.
USB_OTG1_CHD_B An external pullup resistor with value in range from 10 kΩ to 100 kΩ should be connected
between open-drain output USB_OTG1_CHD_B and supply VDD_USB_OTG1_3P3_IN for 3.3 V
signaling. Optionally, a similarly valued pullup resistor could be connected instead between
USB_OTG1_CHD_B and an unrelated supply up to 1.8 V, but in that case the output is only valid
when both that supply and VDD_USB_OTG1_3P3_IN are powered.
Recommendation
Module Package Net Name
if Unused
Recommendation
Module Package Net Name
if Unused
TEMPSENSOR_RESERVE Floating
VDD_TEMPSENSOR_1P8 1.8 V
4 Electrical characteristics
This section provides the device and module-level electrical characteristics for the i.MX 7Dual family of
processors.
FPBGA case “X” and case “Y” package thermal resistance on page 22
4.1.2.1 FPBGA case “X” and case “Y” package thermal resistance
Table 8 displays the thermal resistance data.
Per JEDEC JESD51-2, the intent of thermal resistance measurements is solely for a thermal performance
comparison of one package to another in a standardized environment. This methodology is not meant to
and does not predict the performance of a package in an application-specific environment.
12x12 19x19
Rating Test conditions Symbol Unit
pkg value pkg value
Junction to Ambient1 Single-layer board (1s); natural convection2 RθJA 55.4 44.4 oC/W
Junction to Ambient1 Single-layer board (1s); airflow 200 ft/min2,3 RθJA 41.8 34.3 oC/W
Four-layer board (2s2p); airflow 200 ft/min2,3 RθJA 28.0 25.8 oC/W
1
Junction temperature is a function of die size, on-chip power dissipation, package thermal resistance, mounting site (board)
temperature, ambient temperature, air flow, power dissipation of other components on the board, and board thermal
resistance.
2 Per JEDEC JESD51-2 with the single layer board horizontal. Thermal test board meets JEDEC specification for the specified
package.
3 Per JEDEC JESD51-6 with the board horizontal.
4 Thermal resistance between the die and the printed circuit board per JEDEC JESD51-8. Board temperature is measured on
5
Thermal resistance between the die and the case top surface as measured by the cold plate method (MIL SPEC-883 Method
1012.1).
6
Thermal characterization parameter indicating the temperature difference between package top and the junction temperature
per JEDEC JESD51-2. When Greek letters are not available, the thermal characterization parameter is written as Psi-JT.
Parameter
Symbol Min Typ Max1 Unit Comment
Description
Run Mode VDD_ARM 0.95 1.0 1.25 V Operation at 800 MHz and
below
1.045 1.1 1.25 V Operation between 800 MHz
and 1 GHz
Power VDDA_1P8 1.71 1.8 1.89 V Power for analog LDO and
Supply internal analog blocks. Must
Analog match the range of voltages
Domain and that the rechargeable backup
LDOs battery supports.
LDO for VDD_LPSR 1.71 1.8 1.89 V Power rail for Low Power State
Low-Power Retention mode
State
Retention
mode
USB supply VDD_USB_OTG1_3 3.0 3.3 3.6 V This rail is for USB
voltages P3_IN
Parameter
Symbol Min Typ Max1 Unit Comment
Description
Fuse power FUSE_FSOURCE 1.710 1.8 1.890 V Power supply for internal use
T o
Junction J -20 — 105 C See Table 1 for complete list of
temperature, junction temperature
industrial capabilities.
1
Applying the maximum voltage results in maximum power consumption and heat generation. A voltage set point = (Vmin + the
supply tolerance) is recommended. This results in an optimized power/speed ratio. Operating a voltage of 1.2V and above will
reduce the overall lifetime of the part. For details, see i.MX 7Dual/Solo Product Lifetime Usage (AN5334).
Table 10 shows on-chip LDO regulators that can supply on-chip loads.
PCIE_VP
PCIE_VP_RX
PCIE_VP_TX
PCIE_VPH
PCIE_VPH_RX
PCIE_VPH_TX
1
On-chip LDOs are designed to supply i.MX 7Dual loads and must not be used to supply external loads.
The typical values shown in Table 11 are required for use with NXP BSPs to ensure precise time keeping
and USB operation. For RTC_XTALI operation, two clock sources are available. If there is not an
externally applied oscillator to RTC_XTALI, the internal oscillator takes over.
• On-chip 32 kHz RC oscillator—this clock source has the following characteristics:
— Approximately 25 µA more IDD than crystal oscillator
— Approximately ±10% tolerance
— No external component required
— Starts up faster than 32 kHz crystal oscillator
— Three configurations for this input:
– External oscillator
– External crystal coupled to RTC_XTALI and RTC_XTALO
– Internal oscillator
External crystal oscillator with on-chip support circuit:
— At power up, RC oscillator is utilized. After crystal oscillator is stable, the clock circuit
switches over to the crystal oscillator automatically.
— Higher accuracy than RC oscillator
— If no external crystal is present, then the RC oscillator is utilized
The decision of choosing a clock source should be taken based on real-time clock use and precision
timeout.
Imax = N × C × V × (0.5 × F)
where:
are typically available from the memory vendors. They take into account factors such as signal termination. See the application
note, i.MX 7DS Power Consumption Measurement (AN5383) for examples of DRAM power consumption during specific use
case scenarios.
7
1 Low
OFF RUN
8
Power
3 5
2
4 6
SNVS LPSR
The following table summarizes the external power supply state in all the power modes.
VDD_SNVS_IN OFF ON ON ON ON
The NVCC_DRAM_CKE can be still ON during SNVS/LPSR mode to keep the CKE/RESET pad in
correct state to hold DRAM device in self-refresh mode.
The NVCC_XXX can be off in RUN mode / Low Power mode if all the pads in that IO bank is not used
in the application, the NVCC_XXX supply could be tied to GND.
The VDD_USB_OTG1_3P3_IN and VDD_USB_OTG2_3P3_IN are fully asynchronous to other power
rails, so it can be either ON/OFF in any of the power modes.
When LPSR mode is not needed for the application, the VDD_LPSR can be connected to VDDA_1P8 and
NVCC_GPIO1/2 can be connected to the same power supply as NVCC_XXX for other GPIO banks.
In LPSR mode, the supported wakeup source are RTC alarm, ONOFF event, security/tamper and also the
16 GPIO pads.
power saving. For example, the PMIC can change the DCDC rails to PFM mode to reduce the power
consumption.
The power consumption in low power modes is defined in Table 15.
Power rail Voltage Current Power Voltage Current Power Voltage Current Power Voltage Current Power
(V) (mA) (mW) (V) (mA) (mW) (V) (mA) (mW) (V) (mA) (mW)
VDD_ARM 1.0 2.7 2.70 1.0 0.428 0.43 1.0 0.3 0.30 0.0 — 0.00
VDD_SOC 1.0 19.38 19.38 1.0 1.423 1.42 1.0 0.6 0.60 0.0 — 0.00
VDDA_1P8_IN 1.8 3.46 6.23 1.8 0.206 0.37 1.8 0.4 0.72 0.0 — 0.00
VDD_SNVS_IN 3.0 0.006 0.018 3.0 0.005 0.015 3.0 0.006 0.018 3.0 0.003 0.009
VDD_LPSR_IN 1.8 0.04 0.07 1.8 0.041 0.07 1.8 0.039 0.0702 1.8 0.04 0.07
NVCC_GPIO1/2 1.8 0.072 0.13 1.8 0.073 0.13 1.8 0.072 0.13 1.8 0.072 0.13
All the power numbers defined in Table 15 are based on typical silicon at 25°C.
VDD_USB_OTG1_3P3_IN VDD_USB_OTG2_3P3_IN
Table 17. USB PHY current consumption with VBUS Valid Comparators disabled1
VDD_USB_OTG1_3P3_IN VDD_USB_OTG2_3P3_IN
In Power-Down mode, everything is powered down, including the USB_VBUS valid comparators and
their associated bandgap circuity in typical condition. Table 18 shows the USB interface current
consumption in Power-Down mode.
VDD_USB_OTG1_3P3_IN VDD_USB_OTG2_3P3_IN
VDD Low Power Supply Voltage for PHY Core 1V 0.95 1.05 V
High Power Supply Voltage for PHY Core 1.8 V 1.71 1.89
Note: VDD should have no more than 40 mVpp AC power supply noise superimposed on the high power supply voltage for the
PHY core (1.8 V nominal DC value). At the same time, VDD should have no more than 20 mVpp AC power supply noise
superimposed on the low power supply voltage for the PHY core.
The power supply voltage variation for the PHY core should have less than +/-5% including the board-level power supply variation
and on-chip power supply variation due to the finite impedances in the package.
Slumber Mode — 7 — mW
TX Serial output rise time (20% to 80%) TTXRISE 1.5 Gbps 50 — 273 ps
2.5 Gbps 50 — —
5.0 Gbps 30 — —
6.0 Gbps 33 — 80
TX Serial output fall time (80% to 20%) TTXFALL 1.5 Gbps 50 — 273 ps
2.5 Gbps 50 — —
5.0 Gbps 30 — —
6.0 Gbps 33 — 80
TX Serial data output voltage (Differential, pk–pk) ΔVTX 1.5 Gbps 400 — 600 mVp–p
PCIe Tx deterministic jitter < 1.5 MHz TRJ 2.5 Gbps/ — — 3 ps ps, rms
5.0 Gbps
PCIe Tx deterministic jitter > 1.5 MHz TDJ 5.0 Gbps/ — — 30 ps/ ps, pk–pk
2.5 Gbps 60 ps
RX Serial data input voltage (Differential pk–pk) ΔVRX 1.5 Gbps 325 — 600 mVp–p
Duty Cycle DC 40 — 60 %
Period Jitter — — 25 ps
3.0 Gbps — 75 —
TXD[0:30] setup time to the rising edge of TBC TSETUP.TX 20-bit 1.5 Gbps 2.0 — — ns
3.0 Gbps
3.0 Gbps
6.0 Gbps
TXD[0:30] hold time to the rising edge of TBC THOLD.TX 20-bit 1.5 Gbps 2.0 — — ns
3.0 Gbps
3.0 Gbps
6.0 Gbps
Latency from the rising edge of TBC to the leading edge TLAT.TX — 1.5 Gbps — 70 — bits
of the corresponding first transmitted serial output bit
TXP/TXN 2.5 Gbps — 100 —
3.0 Gbps — 95 —
3.0 Gbps — 75 —
RXD[0:30] delay time from the falling edge of RBC TDLY,RX — — — — 1.33 ns
Latency from the leading edge of the corresponding first TLAT.RX 20-bit 1.5 Gbps — 100 — bits
received serial input bit, RXP/RXN, to the rising edge of RBC
2.5 Gbps — 230 —
PCIe — — 250 —
3.0 Gbps — 75 —
The system design must comply with power-up sequence, power-down sequence, and steady state
guidelines as described in this section to guarantee the reliable operation of the device. Any deviation
from these sequences may result in the following situations:
• Excessive current during power-up phase
• Prevention of the device from booting
4.2.1.1 LDO_1P2
The LDO_1P2 regulator implements a programmable linear-regulator function from VDDA_1P8_IN (see
Table 9 for minimum and maximum input requirements). The typical output of the LDO, VDD_1P2_CAP,
is 1.2 V. It is intended for use with the USB HSIC PHY, which uses this voltage level for its output driver.
For additional information, see the “Power Management Unit (PMU)” chapter of the i.MX 7Dual
Application Processor Reference Manual (IMX7DRM).
4.2.1.2 LDO_1P0D
The LDO_1P0D regulator implements a programmable linear-regulator function from VDDA_1P8_IN
(see Table 9 for minimum and maximum input requirements). The typical output of the LDO,
VDD_1P0D_CAP, is 1.0 V. It is intended for use with the internal physical interfaces, including MIPI and
PCIe PHY. For additional information, see the i.MX 7Dual Application Processor Reference Manual
(IMX7DRM).
4.2.1.3 LDO_1P0A
The LDO_1P0A regulator implements a programmable linear-regulator function from VDDA_1P8_IN
(see Table 9 for minimum and maximum input requirements). The typical output of the LDO,
VDD_1P0A_CAP, is 1.0 V. It is intended for use with the internal analog modules, including the XTAL,
ADC, PLL, and Temperature Sensor. For additional information, see the i.MX 7Dual Application
Processor Reference Manual (IMX7DRM).
4.2.1.4 LDO_USB1_1PO/LDO_USB2_1P0
The LDO_USB1_1P0/LDO_USB2_1P0 regulators implement a fixed linear-regulator function from
VDD_USB_OTG1_3P3_IN and VDD_USB_OTG2_3P3_IN power inputs respectively (see Table 9 for
minimum and maximum input requirements). The typical output voltage is 1.0 V. It is intended for use
with the internal USB physical interfaces (USB PHY1 and USB PHY2). For additional information, see
the i.MX 7Dual Application Processor Reference Manual (IMX7DRM).
4.2.1.5 LDO_SVNS_1P8
1.8 V LDO from coin cell to generate 1.8 V power for SNVS and 32 K RTC. The LDO_SNVS_1P8
regulator implements a fixed linear-regulator function from VDD_SNVS_IN (see Table 9 for minimum
and maximum input requirements). The typical output is 1.7 V. It is intended for use with the internal
SNVS circuitry and 32 K RTC. For additional information, see the i.MX 7Dual Application Processor
Reference Manual (IMX7DRM).
ENET_PLL Clock output range 650 MHz–1.3 GHz, set to 1.0 GHz
4.4.1 OSC24M
Power for the oscillator is supplied from a clean source of VDDA_1P8. This block implements an
amplifier that when combined with a suitable quartz crystal and external load capacitors implements an
oscillator. The oscillator is powered from VDDA_1P8.
The system crystal oscillator consists of a Pierce-type structure running off the digital supply. A straight
forward biased-inverter implementation is used.
4.4.2 OSC32K
This block implements an internal amplifier, trimable load capacitors and a resistor that when combined
with a suitable quartz crystal implements a low power oscillator.
In addition, if the clock monitor determines that the OSC32K is not present then the source of the 32 kHz
clock will automatically switch to the internal relaxation oscillator of lesser frequency accuracy.
CAUTION
The internal RTC oscillator does not provide an accurate frequency and is
affected by process, voltage and temperature variations. NXP strongly
recommends using an external crystal as the RTC_XTALI reference. If the
internal oscillator is used instead, careful consideration must be given to the
timing implications on all of the SoC modules dependent on this clock.
The OSC32k runs from VDD_SNVS_1p8_CAP, which is regulated from
VDD_SNVS. The target battery is an ~3 V coin cell for VDD_SNVS and
the regulated output is ~1.75V.
Fosc — 32.768 KHz — This frequency is nominal and determined by the crystal selected. 32.0 K
would work as well.
Current consumption — 350 nA — The typical value shown is only for the oscillator, driven by an external
crystal. If the interrelaxation oscillator is used instead of an external crystal
then approximately 250 nA should be added to this value.
Bias resistor — 200 MΩ This is the integrated bias resistor that sets the amplifier into a high gain
state. Any leakage through the ESD network, external board leakage, or
even a scope probe that is significant relative to this value will debias the
amp. The debiasing will result in low gain and will impact the circuit's ability
to start up and maintain oscillations.
Cload — 10 pF — Usually, crystals can be purchased tuned for different Cload. This Cload
value is typically 1/2 of the capacitances realized on the PCB on either side
of the quartz. A higher Cload will decrease oscillation margin but increases
current oscillating through the crystal. The Cload is programmable in 2 pF
steps.
ESR — 50 KΩ — Equivalent series resistance of the crystal. Choosing a crystal with a higher
value will decrease oscillating margin.
High-level output voltage VOH IOH= –1.8mA, –3.6mA, –7.2mA, –10.8mA 0.8 × OVDD OVDD V
Low-level output voltage VOL IOL=1.8mA, 3.6mA, 7.2mA, 10.8mA 0 0.2 × OVDD V
DDRMC operation with the standards stated above is contingent upon the board DDR design adherence
to the DDR design and layout requirements stated in the hardware development guide for the i.MX 7
application processor.
Test
Parameters Symbol Min Max Unit
Conditions
Test
Parameters Symbol Min Max Unit
Conditions
Vocm Output common mode voltage 0.9 1 1.1 Core supply is used
Icc Power supply current (ovdd) Rload=100 Ω between padp 4.7 mA This is not including
and padn current through
external
Rload=100 Ω
1
VOH_max = Vos_max + Vod_max/2 = 1.1+0.225 = 1.325 V. VOH_min = Vos_min + Vod_min/2 = 0.9+0.125 = 1.025 V.
2
VOL_max = Vos_max - Vod_min/2 = 1.1-0.125 = 0.975 V. VOL_min = Vos_min - Vod_max/2 = 0.9 - 0.225 = 0.675 V
OVDD
80% 80%
20% 20%
Output (at pad) 0V
tr tf
0 0 1 1× Slow Slew 4.9 6.0 12.5 4.8 6.1 11.9 5.4 6.7 14.6
0 0 0 1× Fast Slew 3.8 4.7 11.2 3.8 5.1 12.8 4.2 5.3 13.5
0 1 1 2× Slow Slew 4.1 4.8 8.2 4.2 4.9 8.8 4.5 5.3 9.1
0 1 0 2× Fast Slew 2.8 3.3 6.4 2.9 3.4 7.2 3.1 3.7 7.2
1 0 1 4× Slow Slew 3.6 4.1 6.0 3.7 4.1 6.4 3.9 4.4 6.6
1 0 0 4× Fast Slew 2.2 2.5 4.1 2.3 2.6 4.6 2.4 2.8 4.8
1 1 1 6× Slow Slew 3.6 4.0 5.5 3.6 4.0 5.9 3.8 4.3 6.2
1 1 0 6× Fast Slew 2.0 2.3 3.4 2.1 2.3 3.8 2.2 2.5 3.9
1 1 0 6× Fast Slew 250 225 140 240 215 120 235 205 120
1
Maximum frequency value is obtained with lumped capacitor load. If you consider transmission line or SSN noise
effect, it could be worse than suggested value.
Tphld Output Differential propagation delay high to low Rload=100 Ω between padp and 1
— — 0.61 ns
padn,
Tplhd Output Differential propagation delay low to high Cload = 2pF — — 0.61
Tphlr Input Differential propagation delay high to low Rload=100 Ω between padp and — — 0.33 ns 3
OVDD
PMOS (Rpu)
Ztl Ω, L = 20 inches
ipp_do pad
predriver
Cload = 1p
NMOS (Rpd)
OVSS
U,(V)
Vin (do)
VDD
t,(ns)
0
U,(V)
Vout (pad)
OVDD
Vref1 Vref2
Vref
t,(ns)
0
Vovdd - Vref1
Rpu = × Ztl
Vref1
Vref2
Rpd = × Ztl
Vovdd - Vref2
Table 40 shows DDR I/O output buffer impedance of i.MX 7Dual family of processors.
Table 40. DDR I/O output buffer impedance
Typical
Test Conditions DSE
Parameter Symbol NVCC_DRAM=1.5 V NVCC_DRAM=1.2 V Unit
(Drive Strength)
(DDR3) (LPDDR2)
DDR_SEL=11 DDR_SEL=10
Note:
1. Output driver impedance is controlled across PVTs using ZQ calibration procedure.
2. Calibration is done against 240 Ω external reference resistor.
3. Output driver impedance deviation (calibration accuracy) is ±5% (max/min impedance) across PVTs.
POR_B
(Input)
CC1
WDOGx_B
(Output)
CC3
NOTE
RTC_XTALI is approximately 32 kHz. RTC_XTALI cycle is one period or
approximately 30 μs.
NOTE
WDOGx_B output signals (for each one of the Watchdog modules) do not
have dedicated pins, but are muxed out through the IOMUX. See the
IOMUXC chapter of the i.MX 7Dual Application Processor Reference
Manual (IMX7DRM) for detailed information.
WE2
... WE3
EIM_BCLK
WE8 WE9
EIM_WE_B
WE10 WE11
EIM_OE_B
WE12 WE13
EIM_EBx_B
WE14 WE15
EIM_LBA_B
WE16 WE17
Output Data
EIM_BCLK
WE18
Input Data
WE19
WE20
EIM_WAIT_B
WE21
1
t is the maximum EIM logic (axi_clk) cycle time. The maximum allowed axi_clk frequency depends on the fixed/non-fixed
latency configuration, whereas the maximum allowed EIM_BCLK frequency is:
—Fixed latency for both read and write is 132 MHz.
—Variable latency for read only is 132 MHz.
—Variable latency for write only is 52 MHz.
In variable latency configuration for write, if BCD = 0 & WBCDD = 1 or BCD = 1, axi_clk must be 104 MHz. Write BCD = 1 and
104 MHz axi_clk, will result in a EIM_BCLK of 52 MHz. When the clock branch to EIM is decreased to 104 MHz, other buses
are impacted which are clocked from this source. See the CCM chapter of the i.MX 7Dual Application Processor Reference
Manual (IMX7DRM) for a detailed clock tree description.
2
EIM_BCLK parameters are being measured from the 50% point, that is, high is defined as 50% of signal value and low is
defined as 50% as signal value.
3
For signal measurements, “High” is defined as 80% of signal value and “Low” is defined as 20% of signal value.
Figure 14 to Figure 17 provide few examples of basic EIM accesses to external memory devices with the
timing parameters mentioned previously for specific control parameters settings.
EIM_WE_B
WE14
EIM_LBA_B WE15
WE10 WE11
EIM_OE_B
WE12 WE13
EIM_EBx_B
WE18
EIM_DATAxx D(v1)
WE19
Figure 14. Synchronous memory read access, WSC=1
EIM_BCLK
WE4 WE5
EIM_ADDRxx Last Valid Address Address V1
WE6 WE7
EIM_CSx_B
WE8 WE9
EIM_WE_B
WE14
EIM_LBA_B
WE15
EIM_OE_B
WE13
WE12
EIM_EBx_B
WE16 WE17
EIM_DATAxx D(V1)
Figure 15. Synchronous memory, write access, WSC=1, WBEA=0 and WADVN=0
EIM_BCLK
WE5 WE16 WE17
EIM_ADDRxx/ WE4
EIM_ADxx Last Valid Address Address V1 Write Data
WE6 WE7
EIM_CSx_B
WE8 WE9
EIM_WE_B
WE14 WE15
EIM_LBA_B
EIM_OE_B
WE10 WE11
EIM_EBx_B
Figure 16. Muxed Address/Data (A/D) mode, synchronous write access, WSC=6, ADVA=0, ADVN=1, and
ADH=1
NOTE
In 32-bit Muxed Address/Data (A/D) mode the 16 MSBs are driven on the
data bus.
EIM_BCLK
WE4 WE5 WE19
EIM_ADDRxx/ Last Valid Address Address V1 Data
EIM_ADxx WE6 WE18
EIM_CSx_B
WE7
EIM_WE_B
WE14 WE15
EIM_LBA_B WE10
WE11
EIM_OE_B
WE12 WE13
EIM_EBx_B
Figure 17. 16-Bit Muxed A/D Mode, Synchronous Read Access, WSC=7, RADVN=1, ADH=1, OEA=0
INT_CLK
EIM_CSx_B MAXCSO
EIM_WE_B
WE39 WE40
EIM_LBA_B
WE35 WE36
EIM_OE_B
WE37 WE38
EIM_EBx_B
WE44
EIM_DATAxx[7:0] MAXCO
D(V1)
WE43 MAXDI
Figure 18. Asynchronous memory read access (RWSC = 5)
start of end of
access access
INT_CLK
MAXCSO
EIM_CSx_B
EIM_ADDRxx/ WE31 MAXDI
EIM_ADxx Addr. V1 D(V1)
WE32A
WE44
EIM_WE_B
WE40A
WE39
EIM_LBA_B
WE35A WE36
EIM_OE_B
WE37 WE38
EIM_EBx_B
MAXCO
Figure 19. Asynchronous A/D muxed read access (RWSC = 5)
EIM_CSx_B
WE31 WE32
EIM_ADDRxx Last Valid Address Address V1 Next Address
WE33 WE34
EIM_WE_B
WE39 WE40
EIM_LBA_B
EIM_OE_B
WE45 WE46
EIM_EBx_B
WE42
EIM_DATAxx D(V1)
WE41
Figure 20. Asynchronous memory write access
EIM_CSx_B
WE41
EIM_ADDRxx/ WE31
Addr. V1 D(V1)
EIM_DATAxx WE32A
WE42
WE33 WE34
EIM_WE_B
WE40A
WE39
EIM_LBA_B
EIM_OE_B
WE45 WE46
EIM_EBx_B
WE42
EIM_CSx_B
EIM_ADDRxx WE31 WE32
Last Valid Address Address V1 Next Address
EIM_WE_B
WE39 WE40
EIM_LBA_B
WE35 WE36
EIM_OE_B
WE37 WE38
EIM_EBx_B WE44
D(V1)
EIM_DATAxx[7:0] WE43
WE48
EIM_DTACK_B
WE47
EIM_CSx_B
WE31 WE32
EIM_ADDRxx Last Valid Address Address V1 Next Address
WE33 WE34
EIM_WE_B
WE39 WE40
EIM_LBA_B
EIM_OE_B
WE45 WE46
EIM_EBx_B
WE42
EIM_DATAxx D(V1)
WE41 WE48
EIM_DTACK_B
WE47
Figure 23. DTACK Mode write access (DAP=0)
Table 46. EIM asynchronous timing parameters table relative chip to select
Determination by
Ref No. Parameter Synchronous measured Min Max Unit
parameters1
Table 46. EIM asynchronous timing parameters table relative chip to select(continued)
Determination by
Ref No. Parameter Synchronous measured Min Max Unit
parameters1
WE40A EIM_CSx_B Valid to WE14 – WE6 + (ADVN + ADVA –3 + (ADVN + 3 + (ADVN + ADVA + ns
(muxed EIM_LBA_B Invalid + 1 – CSA) ADVA + 1 – CSA) 1 – CSA)
A/D)
WE41 EIM_CSx_B Valid to Output WE16 – WE6 – WCSA — 3 – WCSA ns
Data Valid
Table 46. EIM asynchronous timing parameters table relative chip to select(continued)
Determination by
Ref No. Parameter Synchronous measured Min Max Unit
parameters1
DDR1
DRAM_SDCLKx_N
DRAM_SDCLKx_P DDR2
DDR4
DRAM_CSx_B
DDR5
DRAM_RAS_B
DDR5
DDR4
DRAM_CAS_B
DDR4
DDR5
DDR5
DRAM_SDWE_B
DRAM_ODTx /
DRAM_SDCKEx
DDR6 DDR4
DDR7
DRAM_ADDRxx ROW/BA COL/BA
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
2
All measurements are in reference to Vref level.
3
Measurements were done using balanced load and 25 Ω resistor from outputs to VDD_REF.
Figure 25 shows the DDR3 write timing diagram. The timing parameters for this diagram appear in
Table 48.
DRAM_SDCLKx_P
DRAM_SDCLKx_N
DDR21 DDR22
DDR23
DRAM_SDQSx_P
(output) DDR18
DDR17 DDR17 DDR18
DRAM_DQMx DM DM DM DM DM
DM DM DM
(output)
DDR17 DDR17
DDR18 DDR18
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
DDR21 DRAM_SDQSx_P latching rising transitions to associated clock edges tDQSS -0.25 +0.25 tCK
1
To receive the reported setup and hold values, write calibration should be performed in order to locate the DRAM_SDQSx_P in
the middle of DRAM_DATAxx window.
2 All measurements are in reference to Vref level.
3 Measurements were taken using balanced load and 25 Ω resistor from outputs to DDR_VREF.
Figure 26 shows the DDR3 read timing diagram. The timing parameters for this diagram appear in
Table 49.
DRAM_SDCLKx_P
DRAM_SDCLKx_N
DRAM_SDQSx_P
(input)
DRAM_DATAxx
(input) DATA DATA DATA DATA DATA DATA DATA DATA
DDR26
Figure 26. DDR3 read cycle
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
1
To receive the reported setup and hold values, read calibration should be performed in order to locate the DRAM_SDQSx_P
in the middle of DRAM_DATAxx window.
2 All measurements are in reference to Vref level.
3 Measurements were done using balanced load and 25 Ω resistor from outputs to VDD_REF.
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
Figure 28 shows the LPDDR3 write timing diagram. The timing parameters for this diagram appear in
Table 51.
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
Figure 29 shows the LPDDR3 read timing diagram. The timing parameters for this diagram appear in
Table 52.
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
LP26 Minimum required DRAM_DATAxx valid window width for LPDDR3 — 460 — ps
1 To receive the reported setup and hold values, read calibration should be performed in order to locate the DRAM_SDQSx_P
in the middle of DRAM_DATA_xx window.
2 All measurements are in reference to V level.
ref
3 Measurements were done using balanced load and 25 Ω resistor from outputs to DDR_VREF.
DRAM_SDCLKx_P
LP1
DRAM_CSx_B LP4 LP2
LP3
DRAM_SDCKEx LP3
LP3
LP4
DRAM_CAS_B
LP3 LP4
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
Figure 31 shows the LPDDR2 write timing diagram. The timing parameters for this diagram appear in
Table 54.
DRAM_SDCLKx_P
DRAM_SDCLKx_N
LP21
LP23
DRAM_SDCLKx_P LP22
(output) LP18 LP17
LP17 LP18
DRAM_DATAxx Data Data Data Data Data Data Data Data
(output)
DRAM_DQMx DM DM DM DM DM DM DM DM
(output)
LP17 LP17
LP18 LP18
Figure 31. LPDDR2 write cycle
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
LP21 DRAM_SDQSx_P latching rising transitions to associated clock edges tDQSS -0.25 +0.25 tCK
1 To receive the reported setup and hold values, write calibration should be performed in order to locate the DRAM_SDQS in
the middle of DRAM_DATAxx window.
2 All measurements are in reference to Vref level.
3 Measurements were done using balanced load and 25 Ω resistor from outputs to DDR_VREF.
Figure 32 shows the LPDDR2 read timing diagram. The timing parameters for this diagram appear in
Table 55.
DRAM_SDCLKx_P
DRAM_SDCLKx_N
DRAM_SDQSx_P
(input)
LP26
DRAM_DATAxx
DATA DATA DATA DATA DATA DATA DATA DATA
(input)
CK = 533 MHz
ID Parameter Symbol Unit
Min Max
LP26 Minimum required DRAM_DATAxx valid window width for LPDDR2 — 230 — ps
1
To receive the reported setup and hold values, read calibration should be performed in order to locate the DRAM_SDQSx_P
in the middle of DRAM_DATA_xx window.
2 All measurements are in reference to Vref level.
3 Measurements were done using balanced load and 25 Ω resistor from outputs to DDR_VREF.
E&ϭ
.!.$?#,%
E&ϯ
.!.$?#%?"
E&ϭϬ
.!.$?7%?" E&ϱ E&ϭϭ
.!.$?!,% E&ϲ E&ϳ
E&ϴ E&ϵ
EEͺddždž ĚĚƌĞƐƐ
.!.$?#,% E&ϭ
.!.$?#%?" E&ϯ
E&ϭϬ
.!.$?7%?" E&ϱ E&ϭϭ
.!.$?!,% E&ϲ E&ϳ
E&ϴ E&ϵ
.!.$?$!4!XX ĂƚĂƚŽE&
.!.$?#,%
.!.$?#%?"
E&ϭϰ
.!.$?2%?" E&ϭϯ E&ϭϱ
.!.$?2%!$9?" E&ϭϮ
E&ϭϲ E&ϭϳ
.!.$?$!4!XX ĂƚĂĨƌŽŵE&
Figure 36. Read Data Latch cycle timing diagram (Non-EDO Mode)
.!.$?#,%
.!.$?#%?"
E&ϭϰ
.!.$?2%?" E&ϭϯ E&ϭϱ
.!.$?2%!$9?" E&ϭϮ E&ϭϳ
E&ϭϲ
EEͺddždž ĂƚĂĨƌŽŵE&
Figure 37. Read Data Latch cycle timing diagram (EDO mode)
Timing
ID Parameter Symbol T = GPMI Clock Cycle Unit
Min. Max.
NF1 NAND_CLE setup time tCLS (AS + DS) × T - 0.12 [see notes2,3] ns
NF2 NAND_CLE hold time tCLH DH × T - 0.72 [see note2] ns
NF3 NAND_CE0_B setup time tCS (AS + DS + 1) × T [see notes3,2] ns
NF4 NAND_CE0_B hold time tCH (DH+1) × T - 1 [see note2] ns
NF5 NAND_WE_B pulse width tWP DS × T [see note2] ns
NF6 NAND_ALE setup time tALS (AS + DS) × T - 0.49 [see notes3,2] ns
NF7 NAND_ALE hold time tALH (DH × T - 0.42 [see note2] ns
NF8 Data setup time tDS DS × T - 0.26 [see note2] ns
NF9 Data hold time tDH DH × T - 1.37 [see note2] ns
NF10 Write cycle time tWC (DS + DH) × T [see note2] ns
NF11 NAND_WE_B hold time tWH DH × T [see note2] ns
NF12 Ready to NAND_RE_B low tRR4 (AS + 2) × T [see 3,2] — ns
NF13 NAND_RE_B pulse width tRP DS × T [see note2] ns
NF14 READ cycle time tRC (DS + DH) × T [see note2] ns
NF15 NAND_RE_B high hold time tREH DH × T [see note2] ns
NF16 Data setup on read tDSR — (DS × T -0.67)/18.38 [see ns
notes5,6]
NF17 Data hold on read tDHR 0.82/11.83 [see notes5,6] — ns
1
GPMI’s Asynchronous mode output timing can be controlled by the module’s internal registers
HW_GPMI_TIMING0_ADDRESS_SETUP, HW_GPMI_TIMING0_DATA_SETUP, and HW_GPMI_TIMING0_DATA_HOLD.
This AC timing depends on these registers settings. In the table, AS/DS/DH represents each of these settings.
2 AS minimum value can be 0, while DS/DH minimum value is 1.
3
T = GPMI clock period -0.075ns (half of maximum p-p jitter).
4 NF12 is guaranteed by the design.
5
Non-EDO mode.
6
EDO mode, GPMI clock ≈ 100 MHz
(AS=DS=DH=1, GPMI_CTL1 [RDN_DELAY] = 8, GPMI_CTL1 [HALF_PERIOD] = 0).
In EDO mode (Figure 36), NF16/NF17 are different from the definition in non-EDO mode (Figure 35).
They are called tREA/tRHOH (RE# access time/RE# HIGH to output hold). The typical value for them
are 16 ns (max for tREA)/15 ns (min for tRHOH) at 50 MB/s EDO mode. In EDO mode, GPMI will
sample NAND_DATAxx at rising edge of delayed NAND_RE_B provided by an internal DPLL. The
delay value can be controlled by GPMI_CTRL1.RDN_DELAY (see the GPMI chapter of the i.MX 7Dual
Application Processor Reference Manual [IMX7DRM]). The typical value of this control register is 0x8
at 50 MT/s EDO mode. But if the board delay is big enough and cannot be ignored, the delay value should
be made larger to compensate the board delay.
1)
1$1'B&/(
1) 1)
1)
1$1'B$/(
1) 1)
1$1'B:(5(B%
1)
1$1'B&/.
1$1'B'46
1$1'B'46
2XWSXWHQDEOH
1) 1)
1) 1)
1$1'B'$7$>@
2XWSXWHQDEOH
Figure 38. Source Synchronous mode command and address timing diagram
1)
1)
.!.$?#%?"
1) 1)
.!.$?#,% 1) 1)
1) 1)
.!.$?!,% 1) 1)
1$1'B:(5(B%
1)
.!.$?#,+
1)
.!.$?$13 1)
.!.$?$13
2XWSXWHQDEOH
1) 1)
.!.$?$1;=
1) 1)
.!.$?$1;=
2XWSXWHQDEOH
1)
.!.$?#%?" 1)
1) 1)
.!.$?#,% 1) 1)
1) 1)
1$1'B$/( 1) 1)
.!.$?7%2% 1)
1)
1)
1)
.!.$?#,+
.!.$?$13
.!.$?$13
/UTPUT ENABLE
.!.$?$!4!;=
.!.$?$!4!;=
/UTPUT ENABLE
.!.$?$13
E&ϯϬ
.!.$?$!4!;= Ϭ ϭ Ϯ ϯ
E&ϯϬ E&ϯϭ E&ϯϭ
Timing
T = GPMI Clock Cycle
ID Parameter Symbol Unit
Min. Max.
NF18 NAND_CE0_B access time tCE CE_DELAY × T - 0.79 [see note 2] ns
NF19 NAND_CE0_B hold time tCH 0.5 × tCK - 0.63 [see note2] ns
NF20 Command/address NAND_DATAxx setup time tCAS 0.5 × tCK - 0.05 ns
NF21 Command/address NAND_DATAxx hold time tCAH 0.5 × tCK - 1.23 ns
NF22 clock period tCK — ns
NF23 preamble delay tPRE PRE_DELAY × T - 0.29 [see note2] ns
NF24 postamble delay tPOST POST_DELAY × T - 0.78 [see note2] ns
NF25 NAND_CLE and NAND_ALE setup time tCALS 0.5 × tCK - 0.86 ns
NF26 NAND_CLE and NAND_ALE hold time tCALH 0.5 × tCK - 0.37 ns
NF27 NAND_CLK to first NAND_DQS latching transition tDQSS T - 0.41 [see note2] ns
NF28 Data write setup 0.25 × tCK - 0.35
NF29 Data write hold 0.25 × tCK - 0.85
NF30 NAND_DQS/NAND_DQ read setup skew — 2.06
NF31 NAND_DQS/NAND_DQ read hold skew — 1.95
1
GPMI’s Source Synchronous mode output timing can be controlled by the module’s internal registers
GPMI_TIMING2_CE_DELAY, GPMI_TIMING_PREAMBLE_DELAY, GPMI_TIMING2_POST_DELAY. This AC timing
depends on these registers settings. In the table, CE_DELAY/PRE_DELAY/POST_DELAY represents each of these settings.
2 T = tCK(GPMI clock period) –0.075 ns (half of maximum p-p jitter).
For DDR Source Synchronous mode, Figure 41 shows the timing diagram of
NAND_DQS/NAND_DATAxx read valid window. The typical value of tDQSQ is 0.85 ns (max) and 1 ns
(max) for tQHS at 200 MB/s. GPMI will sample NAND_DATA[7:0] at both rising and falling edge of an
delayed NAND_DQS signal, which can be provided by an internal DPLL. The delay value can be
controlled by GPMI register GPMI_READ_DDR_DLL_CTRL.SLV_DLY_TARGET (see the GPMI
chapter of the i.MX 7Dual Application Processor Reference Manual [IMX7DRM]). Generally, the typical
delay value of this register is equal to 0x7 which means 1/4 clock cycle delay expected. But if the board
delay is big enough and cannot be ignored, the delay value should be made larger to compensate the board
delay.
DEV?CLK
.!.$?#%X?"
.&