TI SPNZ187

RM46x Microcontroller
Silicon Revision A
Silicon Errata
Literature Number: SPNZ187
September 2012
Contents
1
2
3
Device and Development-Support Tool Nomenclature ............................................................. 4
Device Markings ................................................................................................................. 5
Known Design Exceptions to Functional Specifications .......................................................... 6
2
Table of Contents
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
List of Figures
1
Part Marking ................................................................................................................. 5
List of Tables
....................................................................................................
..............................................................
1
Device Revision Codes
2
Known Design Exceptions to Functional Specifications
SPNZ187 – September 2012
Submit Documentation Feedback
List of Figures
Copyright © 2012, Texas Instruments Incorporated
5
6
3
Silicon Errata
SPNZ187 – September 2012
RM46x Microcontroller
This document describes the known exceptions to the functional specifications for the device.
1
Device and Development-Support Tool Nomenclature
To designate the stages in the product development cycle, TI assigns prefixes to the part numbers of all
devices and support tools. Each device has one of three prefixes: X, P, or null (no prefix).
Device development evolutionary flow:
X—
Experimental device that is not necessarily representative of the final device's electrical
specifications.
P—
Final silicon die that conforms to the device's electrical specifications but has not completed quality
and reliability verification.
null — Fully-qualified production device.
X and P devices are shipped against the following disclaimer:
"Developmental product is intended for internal evaluation purposes."
X and P devices have been characterized fully, and the quality and reliability of the device have been
demonstrated fully. TI's standard warranty applies.
Predictions show that prototype devices (X or P) have a greater failure rate than the standard production
devices. Texas Instruments recommends that these devices not be used in any production system
because their expected end-use failure rate still is undefined. Only qualified production devices are to be
used.
TI device nomenclature also includes a suffix with the device family name. This suffix indicates the silicon
revision, the package type and the temperature range (for example, "T" is -40 to 105C).
4
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Device Markings
www.ti.com
2
Device Markings
Figure 1 provides examples of the RM46Lx device markings.
RM46L
RM46L
852ZWTT
#######
852PGET
#######
Figure 1. Part Marking
Table 1 shows the applicable silicon revision.
Table 1. Device Revision Codes
DEVICE PART NUMBER
DEVICE REVISION CODE
SILICON REVISION
RM46L852
A
PART NUMBERS/COMMENTS
This silicon revision is available as A only.
RM46L852
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
5
Known Design Exceptions to Functional Specifications
3
www.ti.com
Known Design Exceptions to Functional Specifications
The following table lists the known exceptions to the functional specifications for the device.
Table 2. Known Design Exceptions to Functional Specifications
Title
......................................................................................................................................
Page
AHB_ACCES_PORT#3 (ARM ID-529470) — DAP AHB-AP Access Port Might Fail To Return Slave Error On AHB
Reset. ............................................................................................................................
BCG_IP_P_EMIF#241 (EMIF#241) — EMIF Reports Incorrect Time-out Error Data ............................................
CORTEX-R4#26 (ARM ID-577077) — Thumb STREXD Treated As NOP If Same Register Used For Both Source
Operands ........................................................................................................................
CORTEX-R4#27 (ARM ID-412027) — Debug Reset Does Not Reset DBGDSCR When In Standby Mode ...............
CORTEX-R4#33 (ARM ID-452032) — Processor Can Deadlock When Debug Mode Enables Cleared ....................
CORTEX-R4#46 (ARM ID-599517) — CP15 Auxiliary ID And Prefetch Instruction Accesses Are UNDEFINED ..........
CORTEX-R4#54 (ARM ID-639819) — Instruction Causing A Data Watchpoint Match Incorrectly Traced .................
CORTEX-R4#55 (ARM ID-722412) — CPACR.ASEDIS And CPACR.D32DIS Incorrect When Configured With
Floating Point .................................................................................................................
CORTEX-R4#57 (ARM ID-737195) — Conditional VMRS APSR_Nzcv, FPSCR May Evaluate With Incorrect Flags ....
CORTEX-R4#58 (ARM ID-726554) — DBGDSCR.Adadiscard Is Wrong When DBGDSCR.Dbgack Set ..................
CORTEX-R4#61 (ARM ID-720270) — Latched DTR-Full Flags Not Updated Correctly On DTR Access ..................
CORTEX-R4#66 (ARM ID-754269) — Register Corruption During A Load-Multiple Instruction At An Exception
Vector ...........................................................................................................................
CORTEX-R4#67 (ARM ID-758269) — Watchpoint On A Load Or Store Multiple May Be Missed. ..........................
DEVICE#B053 — CPU code execution could be halted on a device warm reset if the core power domain # 2 is
disabled by software. .........................................................................................................
DEVICE#B056 — Error flags are incorrectly set in the Error Signaling Module upon a device warm reset for modules
in core power domains that are disabled ..................................................................................
DEVICE#B058 — Device I/Os Could Toggle After Warm Reset..................................................................
DEVICE#142 — CPU Abort Not Generated on Write to Unimplemented MCRC Space ......................................
DEVICE#070 — Some multiplexing controls are not compatible with the emulation device ..................................
DMA#27 — DMA Requests Lost During Suspend Mode ..........................................................................
ANALOGIP_021_LPO#16 (F021 LPO#16) — Oscillator Fault Detection Starts too Soon ....................................
MCRC#18 — CPU Abort Generated on Write to Implemented MCRC Space After Write to Unimplemented MCRC
Space ...........................................................................................................................
MIBSPI#110 — Multibuffer Slave In 3 or 4 Pin Mode Transmits Data Incorrectly for Slow SPICLK Frequencies and for
Clock Phase = 1 ...............................................................................................................
MIBSPI#111 — Data Length Error Is Generated Repeatedly In Slave Mode when I/O Loopback is Enabled .............
NHET#52 — WCAP May not Capture the High-Resolution Offset Correctly ....................................................
NHET#53 — Enhanced Input Capture Pins May not Capture Small Pulses Correctly.........................................
SYS#46 — Clock Source Switching Not Qualified With Clock Source Enable And Clock Source Valid ....................
SYS#102 — Bit field EFUSE_Abort[4:0] in SYSTASR register is read-clear ....................................................
SYS#111 — The Device may Generate Multiple nRST Pulses. ..................................................................
VIM#27 — Unexpected phantom interrupt ..........................................................................................
6
RM46x Microcontroller
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
AHB_ACCES_PORT#3 (ARM ID-529470) — DAP AHB-AP Access Port Might Fail To Return Slave Error On
AHB Reset.
AHB_ACCES_PORT#3 (ARM ID-529470) DAP AHB-AP Access Port Might Fail To Return Slave Error
On AHB Reset.
Severity
Medium
Expected Behavior
If an AHB reset occurs (HRESETn is asserted LOW) while the AHB-AP is performing an
access on the AHB bus, the AHB-AP should return a slave error back to the JTAG-DP.
Issue
Instead, the AHB-AP might indicate that the access completed successfully and return
unpredictable data if the access was a read.
Condition
HRESETn is asserted LOW on a specific cycle while the AHB-AP is completing an
access on the AHB bus.
Implication(s)
This will make it harder to identify cases where the AHB bus is being reset. However,
this should not affect most usage models because it is not generally possible to debug
over a bus when it is likely to be reset.
Workaround(s)
Avoid performing debug operations while the AHB bus might be reset. Resets usually
occur following powerdown, and so this can usually be achieved through the use of
features that prevent powerdown during debug.
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
7
BCG_IP_P_EMIF#241 (EMIF#241) — EMIF Reports Incorrect Time-out Error Data
www.ti.com
BCG_IP_P_EMIF#241 (EMIF#241) EMIF Reports Incorrect Time-out Error Data
Severity
Medium
Expected Behavior
Issue
Under certain conditions, the EMIF will report a time-out error (vbusp_rstatus=0x3) on
the VBUSP read status interface.The data returned for this access will be all zeros. If
this access is followed by a read to the EMIF’s MMR space, the EMIF will still report a
time-out error (vbusp_rstatus=0x3) but with the correct data for the MMR read.
Condition
This issue is only applicable if all of the following are true:
1. The EMIF is used for async memory accesses in Extended Wait mode.
2. There is a potential for a time-out error to occur. For example, the async memory will
not de-assert the pad_wait_i input.
3. If an async memory read with time-out error is followed by an MMR read.
4. The EMIF’s vbusp_rstatus output is used by the infrastructure to zero out the read
data when time-out error is reported.
Implication(s)
The EMIF will hold vbusp_rstatus=0x3 until another async access without time-out error
or an SDRAM access is performed.
Workaround(s)
If a timeout error occurs, perform a dummy read to SDRAM, or to another async chip
select that is not configured to be in Extended Wait mode, or to the same async chip
select after disabling the Extended Wait mode on that chip select.
8
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
CORTEX-R4#26 (ARM ID-577077) — Thumb STREXD Treated As NOP If Same Register Used For Both
Source Operands
CORTEX-R4#26 (ARM ID-577077) Thumb STREXD Treated As NOP If Same Register Used For Both
Source Operands
Severity
Medium
Expected Behavior
Issue
The ARM Architecture permits the Thumb STREXD instruction to be encoded with the
same register used for both transfer registers (Rt and Rt2). Because of this erratum, the
Cortex-R4 processor treats such an encoding as UNPREDICTABLE and executes it as a
NOP.
Conditions
This occurs when the processor is in Thumb state and a STREXD instruction is
executed which has Rt = Rt2. This instruction was new in ARM Architecture version 7
(ARMv7). It is not present in ARMv6T2 or other earlier architecture versions.
Implications
If this occurs the destination register, Rd, which indicates the status of the instruction, is
not updated and no memory transaction takes place. If the software is attempting to
perform an exclusive read-modify-write sequence, then it might either incorrectly
complete without memory being written, or loop forever attempting to complete the
sequence.
Workaround(s)
This can be avoided by using two different registers for the data to be transferred by a
STREXD instruction. This may involve copying the data in the transfer register to a
second, different register for use by the STREXD.
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
9
CORTEX-R4#27 (ARM ID-412027) — Debug Reset Does Not Reset DBGDSCR When In Standby Mode
www.ti.com
CORTEX-R4#27 (ARM ID-412027) Debug Reset Does Not Reset DBGDSCR When In Standby Mode
Severity
Medium
Expected Behavior
The debug reset input, PRESETDBGn, resets the processor's debug registers as
specified in the ARMv7R Architecture. The debug reset is commonly used to set the
debug registers to a known state when a debugger is attached to the target processor.
Issue
When the processor is in Standby Mode and the clock has been gated off,
PRESETDBGn fails to reset the Debug Status and Control Register (DBGDSCR).
Conditions
•
•
•
The DBGDSCR register has been written so that its contents differ from the reset
values (most fields in this register reset to zero, though a few are UNKNOWN at
reset)
The processor is in Standby Mode, and the clocks have been gated off, that is
STANDBYWFI is asserted
The debug reset, PRESETDBGn, is asserted and deasserted while the processor
clocks remain gated off
Implications
If this occurs, then after the reset the DBGDSCR register contains the values that it had
before reset rather than the reset values. If the debugger relies on the reset values, then
it may cause erroneous debug of the processor. For example, the DBGDSCR contains
the ExtDCCmode field which controls the Data Communications Channel (DCC) access
mode. If this field was previously set to Fast mode but the debugger assumes that it is in
Non-blocking mode (the reset value) then debugger accesses to the DCC will cause the
processor to execute instructions which were not expected.
Workaround(s)
This can be avoided by a workaround in the debug control software. Whenever the
debugger (or other software) generates a debug reset, follow this with a write of zero to
the DBGDSCR which forces all the fields to their reset values.
10
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
CORTEX-R4#33 (ARM ID-452032) — Processor Can Deadlock When Debug Mode Enables Cleared
CORTEX-R4#33 (ARM ID-452032) Processor Can Deadlock When Debug Mode Enables Cleared
Severity
Medium
Expected Behavior
The Cortex-R4 processor supports two different debugging modes: Halt-mode and
Monitor-mode, or both modes can be disabled. Bits [15:14] in the Debug Status and
Control Register (DBGDSCR) control which, if any, mode is enabled. Debug events can
be generated by the breakpoint unit, matching instructions executed. Deadlocks should
not occur when the debug mode is changed.
Issue
If there are active breakpoints at the time when the debugging mode is changed, this
can cause the processor to deadlock.
Conditions
•
•
•
•
At least one breakpoint is programmed and active
Either halt-mode debugging or monitor mode debugging is enabled and the debug
enable input, DBGEN is HIGH
An instruction which matches a breakpoint is fetched and executed
After the instruction is fetched, but before it has completed execution, both halt-mode
and monitor-mode debugging are disabled by means of a write to DBGDSCR
Implications
Changing the debugging mode while breakpoints or watchpoints are active is
UNPREDICTABLE. Therefore the above conditions cannot be met in normal
(recommended) operation of the processor. However, if the conditions do occur, then the
processor will deadlock, which is outside the bounds of permitted UNPREDICTABLE
behavior.
Workaround(s)
None
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
11
CORTEX-R4#46 (ARM ID-599517) — CP15 Auxiliary ID And Prefetch Instruction Accesses Are UNDEFINED
www.ti.com
CORTEX-R4#46 (ARM ID-599517) CP15 Auxiliary ID And Prefetch Instruction Accesses Are
UNDEFINED
Severity
Medium
Expected Behavior
The ARMv7-R architecture requires implementation of the following two features in CP15
:
1. An Auxiliary ID Register (AIDR), which can be read in privileged modes, and the
contents and format of which are IMPLEMENTATION DEFINED.
2. The operation to prefetch an instruction by MVA, as defined in the ARMv6
architecture, to be executed as a NOP.
Issue
CP15 accesses to Auxiliary ID Register (AIDR) or an operation to prefetch an instruction
by MVA will generate an UNDEFINED exception on Cortex-R4.
Conditions
Either of the following instructions is executed in a privileged mode:
• - MRC p15,1,<Rt>,c0,c0,7 ; Read IMPLEMENTATION DEFINED Auxiliary ID
Register
• - MCR p15,0,<Rt>,c7,c13,1 ; NOP, was Prefetch instruction by MVA in ARMv6
Implications
If software attempts to read the AIDR as part of its process of identifying the processor
on which it is running, it will encounter an unexpected UNDEFINED exception. If
software attempts to execute an instruction prefetch using the ARMv6 CP15 prefetch
operation, it will encounter an unexpected UNDEFINED exception.
Workaround(s)
In the first case, because the contents of the AIDR are IMPLEMENTATION DEFINED, it
must always be read in conjunction with the Main ID Register (MIDR). A simple way of
avoiding this would be to read and analyze the MIDR first and, if this indicates that the
processor is Cortex-R4 skip the read of the AIDR. In the second case, any instruction to
prefetch an instruction by MVA should be removed or replaced by a true NOP
instruction.
12
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
CORTEX-R4#54 (ARM ID-639819) — Instruction Causing A Data Watchpoint Match Incorrectly Traced
CORTEX-R4#54 (ARM ID-639819) Instruction Causing A Data Watchpoint Match Incorrectly Traced
Severity
Medium
Expected Behavior
When tracing a program execution using the ETM, an extra instruction should not be
traced.
Issue
When tracing a program execution using the ETM, an extra instruction is traced if a data
watchpoint matches and causes a debug exception. The extra instruction that is traced is
the instruction which caused the data watchpoint to match.
Conditions
The extra instruction is traced if the following occurs:
• A hardware watchpoint matches
• DBGEN is asserted
• Monitor debug-mode is enabled
Implications
Trace analysis tools will incorrectly consider the instruction causing a data watchpoint
match to have executed. If any of the ETM address comparators are configured to match
on address of the instruction and the exact match bit is set, the comparator will
incorrectly fire. This might cause an unexpected trigger or change in any ETM resources
which are configured to be sensitive to the address comparator.
Workaround(s)
If a data abort exception is taken and the cause was a data watchpoint, the instruction
traced immediately before the entry to the exception handler was not executed and must
be discarded.
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
13
CORTEX-R4#55 (ARM ID-722412) — CPACR.ASEDIS And CPACR.D32DIS Incorrect When Configured With Floating Point
www.ti.com
CORTEX-R4#55 (ARM ID-722412) CPACR.ASEDIS And CPACR.D32DIS Incorrect When Configured
With Floating Point
Severity
Medium
Expected Behavior
Issue
In implementations of the VFPv3-D16 architecture that do not also implement the
Advanced SIMD functionality, the ASEDIS and D32DIS bits, (bits [31] and [30]
respectively of the Coprocessor Access Control Register (CPACR)) are read-asone/writes ignored (RAO/WI). This indicates that Advanced SIMD functionality and
registers D16-D31 are disabled. Because of this erratum, these bits read zero in
implementations of Cortex- R4F which include the floating-point unit. When the floating
point unit is not included, all bits of the CPACR correctly read-as-zero (RAZ).
Conditions
In an implementation of Cortex-R4F with the floating-point unit included, the CPACR is
read.
Implications
If software uses the CPACR to determine if Advanced SIMD functionality and registers
D16-D31 are available, it will incorrectly determine that they are. Any attempt to use
these features will lead to an unexpected UNDEFINED exception.
Workaround(s)
Because the Cortex-R4F processor never includes Advanced SIMD functionality or
registers D16-D31, this can be correctly determined by reading the part number from
one of the ID registers rather than examining ASEDIS and D32DIS in the CPACR.
14
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
CORTEX-R4#57 (ARM ID-737195) — Conditional VMRS APSR_Nzcv, FPSCR May Evaluate With Incorrect
Flags
CORTEX-R4#57 (ARM ID-737195) Conditional VMRS APSR_Nzcv, FPSCR May Evaluate With Incorrect
Flags
Severity
Medium
Expected Behavior
A conditional VMRS APSR_nzcv, FPSCR instruction should evaluate its condition codes
using the correct flags.
Issue
Under certain circumstances, a conditional VMRS APSR_nzcv, FPSCR instruction may
evaluate its condition codes using the wrong flags and incorrectly execute or not
execute.
Conditions
The erratum requires the following sequence of instructions in ARMstate: - VMRS<c>
APSR_nzcv, FPSCR (formerly FMSTAT<c>), where the condition on the instruction is
not always.
• A flag-setting integer multiply or multiply and accumulate instruction (e.g. MULS)
• A single-precision floating-point multiply-accumulate (FP-MAC) instruction (e.g.
VMLA), timed such that the accumulate operation is inserted into the pipeline in the
cycle in which the VMRS instruction is first attempted to be issued.
To meet the above timing requirements, the VMRS instruction must be three pipeline
stages behind the FPMAC. Depending on the rate in which the instructions are
fetched, interlocks within this sequence and dual-issuing, this can be up to three
other instructions between this pair, plus the multiply. Out-of-order completion of FPMAC instructions must be enabled.
Implications
If this erratum occurs, the VMRS instruction will pass or fail its condition codes
incorrectly, and this will appear in any trace produced by the ETM. This can corrupt the
N, Z, C, V flag values in the CPSR which will typically affect the program flow.
Workaround(s)
This erratum can be avoided by disabling out-of-order single-precision floating point
multiply-accumulate (SPMAC) instruction completion. Set DOOFMACS, bit [16] in the
Secondary Auxiliary Control Register. This will have the side-effect of reducing the
performance of SP-MAC operations, though the impact will depend on how these
instructions are used in your code.
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
15
CORTEX-R4#58 (ARM ID-726554) — DBGDSCR.Adadiscard Is Wrong When DBGDSCR.Dbgack Set
www.ti.com
CORTEX-R4#58 (ARM ID-726554) DBGDSCR.Adadiscard Is Wrong When DBGDSCR.Dbgack Set
Severity
Medium
Expected Behavior
When the DBGDSCR.ADAdiscard bit is set, asynchronous data aborts are discarded,
except for setting the DBGDSCR.ADAbort sticky flag. The Cortex-R4 processor ensures
that all possible outstanding asynchronous data aborts have been recognised before it
enters debug halt state. The flag is immediately on entry to debug halt state to indicate
that the debugger does not need to take any further action to determine whether all
possible outstanding asynchronous aborts have been recognized.
Issue
Because of this erratum, the Cortex-R4 processor also sets the DBGDSCR.ADAdiscard
bit when the DBGDSCR.DBGack bit is set. This can cause the DBGDSCR.ADAbort bit
to become set when the processor is not in debug halt state, and it is not cleared when
the processor enters debug halt state. However, the processor does not discard the
abort. It is pending or generates an exception as normal.
Conditions
•
•
•
The processor is not in debug halt state
The DBGDSCR.DBGack bit is set
An asynchronous data abort (for example, SLVERR response to a store to Normaltype memory) is recognized
NOTE:
it is not expected that DBGDSCR.DBGack will be set in any Cortex-R4
system
If this erratum occurs, and the processor subsequently enters debug halt state, the
DBGDSCR.ADAbort bit will be set, when in fact no asynchronous data abort has
occurred in debug state. Before exiting debug state, the debugger will check this bit and
will typically treat it as an error. If no other asynchronous data abort has occurred in
debug state, this is a false error.
Implications
Workaround(s)
16
None
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
CORTEX-R4#61 (ARM ID-720270) — Latched DTR-Full Flags Not Updated Correctly On DTR Access
CORTEX-R4#61 (ARM ID-720270) Latched DTR-Full Flags Not Updated Correctly On DTR Access
Severity
Medium
Expected Behavior
When the debug Data Transfer Register (DTR) is in non-blocking mode, the latched
DTR-full flags (RXfull_l and TXfull_l) record the state of the DTR registers as observed
by the debugger and control the flow of data to and from the debugger to prevent race
hazards. For example, when the target reads data from DBGDTRRXint, the associated
flag RXfull is cleared to indicate that the register has been drained, but the latched value
Rxfull_l remains set. Subsequent debugger writes to DBGDTRRXext are ignored
because RXfull_l is set. RXfull_l is updated from RXfull when the debugger reads
DBGDSCRext such that a debugger write to DBGDTRRXext will only succeed after the
debugger has observed that the register is empty. The ARMv7 debug architecture
requires that RXfull_l be updated when the debugger reads DBGDSCRext and when it
writes DBGDTRRXext. Similarly, TXfull_l must be updated when the debugger reads
DBGDSCRext and when it reads DBGDTRTXext.
Issue
Because of this erratum, RXfull_l and TXfull_l are only updated when the debugger
reads DBGDSCRext.
Conditions
The DTR is in non-blocking mode, that is, DBGDSCR.ExtDCCmode is set to 0b00 and
EITHER:
• The debugger reads DBGDSCRext which shows that RXfull is zero, that is,
DBGDTRRX is empty.
• The debugger writes data to DBGDTRRXext.
• Without first reading the DBGDSCRext, and before the processor has read from
DBGDTRRXint, the debugger performs another write to DBGDTRRXext.
OR
• The debugger reads DBGDSCRext which shows that TXfull is one, that is,
DBGDTRTX is full.
• The debugger reads data from DBGDTRTXext,
• The processor writes new data into DBGDTRTXint,
• Without first reading the DBGDSCRext, the debugger performs another read from
DBGDTRTXext.
Implications
The ARMv7 debug architecture requires the debugger to read the DBGDSCRext before
attempting to transfer data via the DTR when in non-blocking mode. This erratum only
has implications for debuggers that violate this requirement. If the erratum occurs via
data transfer, data loss may occur. The architecture requires that data transfer never
occur.
Texas Instruments has verified that TI's Code Composer Studios IDE is not affected by
this issue.
Workaround(s)
None
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
17
CORTEX-R4#66 (ARM ID-754269) — Register Corruption During A Load-Multiple Instruction At An Exception Vector
www.ti.com
CORTEX-R4#66 (ARM ID-754269) Register Corruption During A Load-Multiple Instruction At An
Exception Vector
Severity
Medium
Expected Behavior
Issue
Under certain circumstances, a load multiple instruction can cause corruption of a
general purpose register.
Conditions
All the following conditions are required for this erratum to occur:
• A UDIV or SDIV instruction is executed with out-of-order completion of divides
enabled
• A multi-cycle instruction is partially executed before being interrupted by either an
IRQ, FIQ or imprecise abort. In this case, a multi-cycle instruction can be any of the
following:
– LDM/STM that transfers 3 or more registers
– LDM/STM that transfers 2 registers to an unaligned address without write back
– LDM/STM that transfers 2 registers to an aligned address with write back
– TBB/TBH
• A load multiple instruction is executed as the first instruction of the exception handler
• The load multiple instruction itself is interrupted either by an IRQ, FIQ, imprecise
abort or external debug halt request. This erratum is very timing sensitive and
requires the UDIV or SDIV to complete when the load multiple is in the Issue stage of
the CPU pipeline. The register that is corrupted is not necessarily related to the loadmultiple instruction and will depend on the state in the CPU store pipeline when the
UDIV or SDIV completes.
Implications
For practical systems, it is not expected that an interruptible LDM will be executed as the
first instruction of an exception handler, because the handler is usually required to save
the registers of the interrupted context. Therefore, it is not expected that this erratum has
any implications for practical systems. If the situation of the erratum occurs it will result in
the corruption of the register bank state and could cause a fatal failure if the corrupted
register is subsequently read before being written.
Workaround(s)
To work around this erratum, set bit [7] of the Auxiliary Control Register to disable out-oforder completion for divide instructions. Code performance may be reduced depending
on how often divide operations are used.
18
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
CORTEX-R4#67 (ARM ID-758269) — Watchpoint On A Load Or Store Multiple May Be Missed.
CORTEX-R4#67 (ARM ID-758269) Watchpoint On A Load Or Store Multiple May Be Missed.
Severity
Medium
Expected Behavior
The Cortex-R4 supports synchronous watchpoints. This implies that for load and store
multiples, a watchpoint on any memory access will generate a debug event on the
instruction itself.
Issue
Due to this erratum, certain watchpoint hits on multiples will not generate a debug event.
Conditions
All the following conditions are required for this erratum to occur:
• A load or store multiple instruction is executed with at least 5 registers in the register
list
• The address range accessed corresponds to Strongly-Ordered or Device memory
• A watchpoint match is generated for an access that does not correspond to either the
first two or the last two registers in the list.
Under these conditions the processor will lose the watchpoint. Note that for a "store
multiple" instruction, the conditions are also affected by pipeline state making them
timing sensitive.
Implications
Due to this erratum, a debugger may not be able to correctly watch accesses made to
Device or Strongly-ordered memory. The ARM architecture recommends that
watchpoints should not be set on individual Device or Strongly-ordered addresses that
can be accessed as part of a load or store multiple. Instead, it recommends the use of
the address range masking functionality provided to set watchpoints on an entire region,
ensuring that the watchpoint event will be seen on the first access of a load or store
multiple to this region. If this recommendation is followed, this erratum will not occur.
Workaround(s)
None
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
19
DEVICE#B053 — CPU code execution could be halted on a device warm reset if the core power domain # 2 is disabled by
software.
www.ti.com
DEVICE#B053
CPU code execution could be halted on a device warm reset if the core power
domain # 2 is disabled by software.
Severity
Medium
Expected Behavior
The CPU code execution must start from the reset vector (address 0x00000000) upon a
device warm reset and is not affected by the state of any switchable device power
domain.
Issue
CPU code execution could be halted upon a warm reset if the core power domain # 2
has been disabled by software prior to the device warm reset.
Conditions
The behavior is not dependent on any particular operating condition.
Implications
CPU code execution is halted so that a system hang occurs. An external monitor must
be present to prevent the system from entering an unsafe state when this happens.
Workaround(s)
The application must not disable the core power domain # 2 in software via the Power
Management Module (PMM) registers, even if the modules inside this core power
domain are not used in the application.
20
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
DEVICE#B056 — Error flags are incorrectly set in the Error Signaling Module upon a device warm reset for
modules in core power domains that are disabled
DEVICE#B056
Error flags are incorrectly set in the Error Signaling Module upon a device warm
reset for modules in core power domains that are disabled
Severity
High
Expected Behavior
Once a core power domain is disabled (turned off) no error must be indicated to the
ESM from a module in this power domain.
Issue
When a core power domain is disabled, a device warm reset causes error signals to be
sent to the ESM that appear to be from the power domain that has been disabled. This
is caused due to the incorrect power domain isolation implemented on these error
signals. All such error signals are connected to ESM group1 channels.
Conditions
The behavior is not dependent on any particular operating condition.
Implications
If the application has enabled the generation of interrupts or the assertion of the
nERROR output when any flag gets set in the ESM module's group1 status registers,
then the CPU will receive an interrupt and/or the nERROR pin will be driven for an error
condition ascribed to a module that is in a core power domain that has been disabled.
Workaround(s)
When a core power domain is disabled, the application must ensure that any error signal
related to a module within this power domains is not configured to generate an interrupt
to the CPU or assert the device nERROR output. Alternately, the application could clear
the ESM group1 status flags upon each device warm reset condition before it enables
interrupts and nERROR drive features for the ESM group1 channels.
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
21
DEVICE#B058 — Device I/Os Could Toggle After Warm Reset
www.ti.com
DEVICE#B058
Device I/Os Could Toggle After Warm Reset
Severity
Medium
Expected Behavior
All device I/Os are tri-stated (inputs) after a device warm reset. No toggle is expected on
any of the device I/Os without software configuration.
Issue
All core power domains get enabled on a system reset. After the core domains are
turned ON, the domain outputs are unstable for a few cycles after the system reset is
released. This could cause the I/Os to toggle.
Conditions
The behavior is not dependent on any particular operating condition.
Implications
Any external circuit connected to the device I/Os could see activity on the connections to
the device after a devcie warm reset.
Workaround(s)
There are no software or hardware workarounds for this issue.
22
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
DEVICE#142 — CPU Abort Not Generated on Write to Unimplemented MCRC Space
www.ti.com
DEVICE#142
CPU Abort Not Generated on Write to Unimplemented MCRC Space
Severity
Low
Expected Behavior
A write to the illegal address region of the MCRC module will generate an abort
Issue
A CPU Abort does not get generated per the following condition:
Conditions
When a normal mode write to an illegal address region of MCRC register space is
followed by a debug mode access.
Implications
When debugging, either a breakpoint on the instruction after the illegal write, or single
stepping through the illegal write will not generate an abort.
Workaround(s)
None
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
23
DEVICE#070 — Some multiplexing controls are not compatible with the emulation device
www.ti.com
DEVICE#070
Some multiplexing controls are not compatible with the emulation device
Severity
Medium
Expected Behavior
For the common functions, it is expected that the derivative parts will maintain
compatibility to the superset emulation part.
Issue
There are multiple issues with the multiplexing controls:
1. For GIOB[2], to take input from the alternate path,the PINMMR29[16] has to cleared.
This is incompatible with the emulation device, where the PINMMR29[16] needs to
be set to select the alternate path as input for GIOB[2].
2. PINMMR29[24] is cleared by default to select MII interface for the EMAC. This is
incompatible with the emulation device, where the PINMMR29[24] was set by default
and selects the RMII interface.
Conditions
The behavior is not dependent on any particular operating condition.
Implications
The application must manage the configuration of the PINMMR registers based on the
actual part being used. The code running on the emulation device cannot be run as-is on
this device.
Workaround(s)
There are no software workarounds.
24
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
DMA#27 — DMA Requests Lost During Suspend Mode
www.ti.com
DMA#27
DMA Requests Lost During Suspend Mode
Severity
Medium
Expected Behavior
Issue
DMA requests from a peripheral can be lost.
Condition
This only happens when:
• The device is suspended by a debugger
• A peripheral continues to generate requests while the device is suspended
• The DMA is setup to continue the current block transfer during suspend mode with
DMA DEBUG MODE set to ‘01’
• And the request trigger type TTYPE is set to frame trigger
Implication(s)
When the DMA comes out of the suspend mode to resume the transfer, the data
transfers corresponding to the third and subsequent requests will be lost.
Workaround(s)
Either use TTYPE = Block transfer when DMA DEBUG MODE is '01' (Finish Current
Block Transfer) or use DMA DEBUG MODE = '00' (Ignore suspend) when using TTYPE
= Frame transfer to complete block transfer even after suspend/halt is asserted.
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
25
ANALOGIP_021_LPO#16 (F021 LPO#16) — Oscillator Fault Detection Starts too Soon
www.ti.com
ANALOGIP_021_LPO#16 (F021 LPO#16) Oscillator Fault Detection Starts too Soon
Severity
Low
Expected Behavior
The oscillator fault detection circuit allows approximately 200ms after the release of
nPORRST for the oscillator to become valid before monitoring for oscillator faults.
Issue
Sometimes the oscillator fault detection starts monitoring the oscillator shortly
(approximately 100us) after the release of nPORRST.
Condition
This occurs only after power on.
Implication(s)
If the oscillator is slow to start, an oscillator fault condition may be detected and the
device will switch to using the HFLPO as a clock source.
Workaround(s)
This condition is avoided if nPORRST is held low long enough for the oscillator to
stablize.
This condition can be corrected by software. The software should check for oscillator
fault after detecting a power-on reset. If an oscillator fault has occured, the software can
attempt to restart the oscillator. Refer to the device Technical Reference Manual for
information on how to recover from an oscillator failure.
26
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
MCRC#18 — CPU Abort Generated on Write to Implemented MCRC Space After Write to Unimplemented
MCRC Space
MCRC#18
CPU Abort Generated on Write to Implemented MCRC Space After Write to
Unimplemented MCRC Space
Severity
Low
Expected Behavior
A write to the legal address region of the MCRC module should not generate an abort
Issue
An Unexpected CPU Data Abort is generated in the following condition:
Conditions
When a normal mode write to an illegal address region of MCRC register space is
followed by a write to a legal address region of the MCRC register space.
Implications
A write to an illegal address region of the MCRC register space generates a data abort
as expected. The next write to a legal address region of the MCRC register space
generates an unexpected second data abort.
Workaround(s)
None
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
27
MIBSPI#110 — Multibuffer Slave In 3 or 4 Pin Mode Transmits Data Incorrectly for Slow SPICLK Frequencies and for Clock
Phase = 1
www.ti.com
MIBSPI#110
Multibuffer Slave In 3 or 4 Pin Mode Transmits Data Incorrectly for Slow SPICLK
Frequencies and for Clock Phase = 1
Severity
Medium
Expected Behavior
The SPI must be able to transmit and receive data correctly in slave mode as long as
the SPICLK is slower than the maximum frequency specified in the device datasheet.
Issue
The MibSPI module, when configured in multi-buffered slave mode with 3 functional pins
(CLK, SIMO, SOMI) or 4 functional pins (CLK, SIMO, SOMI, nENA), could cause
incorrect data to be transmitted.
Conditions
This issue can occur under the following condition:
• Module is configured to be in mult-buffered mode, AND
• Module is configured to be a slave in the SPI communication, AND
• SPI communication is configured to be in 3-pin mode or 4-pin mode with nENA, AND
• Clock phase for SPICLK is 1, AND
• SPICLK frequency is VCLK frequency / 12 or slower
Implications
Under the above described condition, the slave MibSPI module can transmit incorrect
data.
Workaround(s)
The issue can be avoided by setting the CSHOLD bit in the control field of the TX RAM.
The nCS is not used as a functional signal in this communication, hence setting the
CSHOLD bit does not cause any other effect on the SPI communication.
28
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
www.ti.com
MIBSPI#111 — Data Length Error Is Generated Repeatedly In Slave Mode when I/O Loopback is Enabled
MIBSPI#111
Data Length Error Is Generated Repeatedly In Slave Mode when I/O Loopback is
Enabled
Severity
Medium
Expected Behavior
After a data length (DLEN) error is generated and the interrupt is serviced the SPI
should abort the ongoing transfer and stop.
Issue
When a DLEN error is created in Slave mode of the SPI using nSCS pins in IO
LoopBack Test mode, the SPI module re-transmits the data with the DLEN error instead
of aborting the ongoing transfer and stopping.
Conditions
This is only an issue for a IOLPBK mode Slave in Analog Loopback configuration, when
the intentional error generation feature is triggered using
CTRL_DLENERR(IOLPBKTSTCR.16).
Implications
The SPI will repeatedly transmit the data with the DLEN error when configured in the
above configuration.
Workaround(s)
After the DLEN_ERR interrupt is detected in IOLPBK mode, disable the transfers by
clearing the SPIEN bit of SPIGCR1 register (bit 24) and then re-enable the transfers by
setting SPIEN.
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
29
NHET#52 — WCAP May not Capture the High-Resolution Offset Correctly
www.ti.com
NHET#52
WCAP May not Capture the High-Resolution Offset Correctly
Severity
High
Expected Behavior
The N2HET is a complex timer that operates by executing a set of timing instructions
each loop resolution period. Normally counters and capture registers implemented by
such a scheme would be limited in resolution to the timer's loop period; but the N2HET
also includes dedicated hardware timer extensions that enable operation with resolution
as high as 1/128 of the timer loop period. The WCAP instruction is designed to support
time-stamping with high resolution. When an event occurs, the WCAP instruction is
supposed to capture both the loop-resolution counter value and the high-resolution offset
measuring the interval between the start of the loop and actual event occurrence, and
return the combined result. This result can be interpreted as the number of timer loops
represented as fixed point number with 25 integral bits and 7 fractional bits.
Issue
WCAP may not capture the high-resolution offset correctly.
Conditions
If the edge event that the WCAP instruction is time-stamping happens to align with the
start of the internal loop resolution period, then the WCAP does not properly capture the
7-bit fractional offset correctly
Implications
Under the above described condition, the fractional offset of the previous event captured
by WCAP is what is stored in the WCAP data field. Because of this boundary condition,
the WCAP instruction should not be used in high-resolution mode
Workaround(s)
A similar instruction, PCNT, is known to operate correctly at high resolution. While
WCAP returns the absolute time-stamp of an event, PCNT returns the interval between
an event and the previous event on the same pin.
Instead of WCAP, if an absolute time-stamp is required it is recommended to use a
PCNT instruction followed by an ADD instruction that accumulates the time between
successive edge measurements to produce a result that is a WCAP equivalent.
An example for using PCNT as a work around for WCAP:
WCAP {next = ERRPIN, cond_addr = JMP, hr_lr=high, reg=T, event=RISE, pin=16,
data=0}
work around:
SAVE: MOV32 {remote=TRISE, type=REMTOREG, reg=R}
TRISE: PCNT {next=TIMESTAMP, hr_lr=high, type=RISE2RISE, pin=16, data=0}
TIMESTAMP: ADD {src1=REM, src2=R, rdest=REM, remote=TRISE, dest=NONE,
next=CHK, data=0, hr_data=0}
CHK: BR {cond_addr=JMP, event=RISE, pin=20,next=ERRPIN}
30
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
NHET#53 — Enhanced Input Capture Pins May not Capture Small Pulses Correctly
www.ti.com
NHET#53
Enhanced Input Capture Pins May not Capture Small Pulses Correctly
Severity
High
Expected Behavior
The PCNT and WCAP instructions can capture pulse length or time stamp of small
pulses that have two edges within a single loop resolution.
Issue
The high resolution value may be captured incorrectly.
Conditions
If the second edge event that the PCNT or WCAP instruction is using happens to align
with the start of the internal loop resolution period, then the instruction does not properly
capture the high resolution value correctly.
Implications
Because of this boundary condition, the PCNT and WCAP instructions should not be
used on small pulses.
Workaround(s)
None
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
31
SYS#46 — Clock Source Switching Not Qualified With Clock Source Enable And Clock Source Valid
www.ti.com
SYS#46
Clock Source Switching Not Qualified With Clock Source Enable And Clock
Source Valid
Severity
Low
Expected Behavior
An attempt to switch to a clock source which is not valid yet should be discarded.
Issue
Switching a clock source by simply writing to the GHVSRC bits of the GHVSRC register
may cause unexpected behavior. The clock will switch to a source even if the clock
source was not ready.
Conditions
The clock source is enabled (CSDIS register) but the clock source is not yet valid
(CSVSTAT register).
Implications
Unexpected behavior stated above.
Workaround(s)
Always check the CSDIS register to make sure the clock source is turned on and check
the CSVSTAT register to make sure the clock source is valid. Then write to GHVSRC to
switch the clock.
32
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
SYS#102 — Bit field EFUSE_Abort[4:0] in SYSTASR register is read-clear
www.ti.com
SYS#102
Bit field EFUSE_Abort[4:0] in SYSTASR register is read-clear
Severity
Medium
Expected Behavior
The documentation states that EFUSE_Abort[4:0] of the SYSTASR register should be
write-clear in privilege mode.
Issue
However, these bits are implemented as read-clear.
Conditions
Always.
Implications
Software implementation for error handling needs to take care of this.
Workaround(s)
None
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
33
SYS#111 — The Device may Generate Multiple nRST Pulses.
www.ti.com
SYS#111
The Device may Generate Multiple nRST Pulses.
Severity
Medium
Expected Behavior
When an event on the device causes a system reset to be asserted, the device drives
the system reset (nRST) terminal low for 8 VCLK periods. Then the nRST terminal is
released and gets pulled High, allowing the device to start its warm boot sequence.
Issue
It is possible that the device may generate multiple system resets instead of releasing
the system reset (nRST) at the appropriate time and starting the warm boot sequence.
Conditions
This issue occurs when either the external reset pulse or the 8 VCLK internally
generated reset pulse is the same length as the delay caused by the glitch filter on the
nRST signal. The delay time of the glitch filter is a function of device process, the
temperature and the core voltage. This has a range from 500ns to 2μs. The length of the
8 VCLK reset extension is a function of the crystal speed, the I/O voltage, and the
resistance and capacitance on the nRST pin. There is no problem when reset is properly
asserted with the PORRST pin for 1ms or more as the nRST pin will be held low much
longer than 2us.
Implications
Usually only one or two extra reset pulses are generated. However, it is possible, in rare
cases, that a long string of reset pulses could be generated. This would prevent the
device from booting up in a timely manner.
Workaround(s)
In the case of an externally generated system reset, this issue can be avoided by
asserting the system reset signal for more than 2μs. If the issue is seen when there is an
internal system reset condition then the issue can be avoided by using an 8MHz or
slower crystal. Another workaround is to avoid the use of internal system resets such as
software reset, oscillator fault reset, watchdog reset, or debug reset. This erratum has
been fixed on newer devices by extending the time of the internally generated
nRSTpulse to 32 VCLK cycles. This extends the reset pulse to a minimum time of 3.2us
(with a 20MHz crystal), which exceeds the maximum delay time of the glitch filter.
34
RM46x Microcontroller
SPNZ187 – September 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
VIM#27 — Unexpected phantom interrupt
www.ti.com
VIM#27
Unexpected phantom interrupt
Severity
High
Expected Behavior
When responding to an interrupt and a subsequent interrupt is received, the
corresponding VIM request should be flagged as pending in the VIM status registers.
When the CPU is ready to service the subsequent interrupt, the correct service routine
address should be fetched by the CPU.
Issue
On rare occasions the VIM may return the phantom interrupt vector instead of the real
interrupt vector.
Condition
This condition is specific to software and hardware vectored modes. This is not
applicable for legacy interrupt servicing mode. The ratio of GCLK to VCLK must be 3:1
or greater for hardware vectored mode. The ratio of GCLK to VCLK must be 5:1 or
greater for software vectored mode. A subsequent interrupt request must occur when the
VIM is finishing acknowledging a previous interrupt.
Implication(s)
The subsequent interrupt request vectors to the phantom interrupt routine instead of the
correct service routine.
Workaround(s)
This issue can be avoided if the clock ratio GCLK:VCLK is 1:1 or 2:1. Otherwise the VIM
status register can be checked within the phantom interrupt routine to determine the
source of the subsequent interrupt.
SPNZ187 – September 2012
Submit Documentation Feedback
RM46x Microcontroller
Copyright © 2012, Texas Instruments Incorporated
35
IMPORTANT NOTICE
Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, enhancements, improvements and other
changes to its semiconductor products and services per JESD46, latest issue, and to discontinue any product or service per JESD48, latest
issue. Buyers should obtain the latest relevant information before placing orders and should verify that such information is current and
complete. All semiconductor products (also referred to herein as “components”) are sold subject to TI’s terms and conditions of sale
supplied at the time of order acknowledgment.
TI warrants performance of its components to the specifications applicable at the time of sale, in accordance with the warranty in TI’s terms
and conditions of sale of semiconductor products. Testing and other quality control techniques are used to the extent TI deems necessary
to support this warranty. Except where mandated by applicable law, testing of all parameters of each component is not necessarily
performed.
TI assumes no liability for applications assistance or the design of Buyers’ products. Buyers are responsible for their products and
applications using TI components. To minimize the risks associated with Buyers’ products and applications, Buyers should provide
adequate design and operating safeguards.
TI does not warrant or represent that any license, either express or implied, is granted under any patent right, copyright, mask work right, or
other intellectual property right relating to any combination, machine, or process in which TI components or services are used. Information
published by TI regarding third-party products or services does not constitute a license to use such products or services or a warranty or
endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property of the
third party, or a license from TI under the patents or other intellectual property of TI.
Reproduction of significant portions of TI information in TI data books or data sheets is permissible only if reproduction is without alteration
and is accompanied by all associated warranties, conditions, limitations, and notices. TI is not responsible or liable for such altered
documentation. Information of third parties may be subject to additional restrictions.
Resale of TI components or services with statements different from or beyond the parameters stated by TI for that component or service
voids all express and any implied warranties for the associated TI component or service and is an unfair and deceptive business practice.
TI is not responsible or liable for any such statements.
Buyer acknowledges and agrees that it is solely responsible for compliance with all legal, regulatory and safety-related requirements
concerning its products, and any use of TI components in its applications, notwithstanding any applications-related information or support
that may be provided by TI. Buyer represents and agrees that it has all the necessary expertise to create and implement safeguards which
anticipate dangerous consequences of failures, monitor failures and their consequences, lessen the likelihood of failures that might cause
harm and take appropriate remedial actions. Buyer will fully indemnify TI and its representatives against any damages arising out of the use
of any TI components in safety-critical applications.
In some cases, TI components may be promoted specifically to facilitate safety-related applications. With such components, TI’s goal is to
help enable customers to design and create their own end-product solutions that meet applicable functional safety standards and
requirements. Nonetheless, such components are subject to these terms.
No TI components are authorized for use in FDA Class III (or similar life-critical medical equipment) unless authorized officers of the parties
have executed a special agreement specifically governing such use.
Only those TI components which TI has specifically designated as military grade or “enhanced plastic” are designed and intended for use in
military/aerospace applications or environments. Buyer acknowledges and agrees that any military or aerospace use of TI components
which have not been so designated is solely at the Buyer's risk, and that Buyer is solely responsible for compliance with all legal and
regulatory requirements in connection with such use.
TI has specifically designated certain components which meet ISO/TS16949 requirements, mainly for automotive use. Components which
have not been so designated are neither designed nor intended for automotive use; and TI will not be responsible for any failure of such
components to meet such requirements.
Products
Applications
Audio
www.ti.com/audio
Automotive and Transportation
www.ti.com/automotive
Amplifiers
amplifier.ti.com
Communications and Telecom
www.ti.com/communications
Data Converters
dataconverter.ti.com
Computers and Peripherals
www.ti.com/computers
DLP® Products
www.dlp.com
Consumer Electronics
www.ti.com/consumer-apps
DSP
dsp.ti.com
Energy and Lighting
www.ti.com/energy
Clocks and Timers
www.ti.com/clocks
Industrial
www.ti.com/industrial
Interface
interface.ti.com
Medical
www.ti.com/medical
Logic
logic.ti.com
Security
www.ti.com/security
Power Mgmt
power.ti.com
Space, Avionics and Defense
www.ti.com/space-avionics-defense
Microcontrollers
microcontroller.ti.com
Video and Imaging
www.ti.com/video
RFID
www.ti-rfid.com
OMAP Applications Processors
www.ti.com/omap
TI E2E Community
e2e.ti.com
Wireless Connectivity
www.ti.com/wirelessconnectivity
Mailing Address: Texas Instruments, Post Office Box 655303, Dallas, Texas 75265
Copyright © 2012, Texas Instruments Incorporated