Texas Instruments MSP430F6721 User manual

Errata
MSP430F6721 Microcontroller
ABSTRACT
This document describes the known exceptions to the functional specifications (advisories).
Table of Contents
1 Functional Advisories............................................................................................................................................................ 2
2 Preprogrammed Software Advisories.................................................................................................................................. 2
3 Debug Only Advisories.......................................................................................................................................................... 3
4 Fixed by Compiler Advisories............................................................................................................................................... 3
5 Nomenclature, Package Symbolization, and Revision Identification................................................................................ 4
5.1 Device Nomenclature.........................................................................................................................................................4
5.2 Package Markings..............................................................................................................................................................4
5.3 Memory-Mapped Hardware Revision (TLV Structure)....................................................................................................... 5
6 Advisory Descriptions............................................................................................................................................................6
7 Revision History................................................................................................................................................................... 25
www.ti.com Table of Contents
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 1
Copyright © 2021 Texas Instruments Incorporated

1 Functional Advisories
Advisories that affect the device's operation, function, or parametrics.
✓ The check mark indicates that the issue is present in the specified revision.
Errata Number
Rev A
ADC39 ✓
ADC42 ✓
ADC69 ✓
AUXPMM1 ✓
AUXPMM2 ✓
CPU36 ✓
CPU46 ✓
CPU47 ✓
DMA4 ✓
DMA7 ✓
DMA9 ✓
DMA10 ✓
LCDB5 ✓
LCDB6 ✓
PMM7 ✓
PMM11 ✓
PMM12 ✓
PMM14 ✓
PMM15 ✓
PMM18 ✓
PMM20 ✓
PMM26 ✓
PORT15 ✓
PORT19 ✓
SD3 ✓
UCS11 ✓
USCI36 ✓
USCI37 ✓
USCI41 ✓
USCI42 ✓
USCI47 ✓
USCI50 ✓
2 Preprogrammed Software Advisories
Advisories that affect factory-programmed software.
✓ The check mark indicates that the issue is present in the specified revision.
Errata Number
Rev A
BSL7 ✓
BSL14 ✓
Functional Advisories www.ti.com
2MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

3 Debug Only Advisories
Advisories that affect only debug operation.
✓ The check mark indicates that the issue is present in the specified revision.
Errata Number
Rev A
EEM8 ✓
EEM17 ✓
EEM19 ✓
EEM23 ✓
JTAG26 ✓
JTAG27 ✓
4 Fixed by Compiler Advisories
Advisories that are resolved by compiler workaround. Refer to each advisory for the IDE and compiler versions
with a workaround.
✓ The check mark indicates that the issue is present in the specified revision.
Errata Number
Rev A
CPU21 ✓
CPU22 ✓
CPU40 ✓
Refer to the following MSP430 compiler documentation for more details about the CPU bugs workarounds.
TI MSP430 Compiler Tools (Code Composer Studio IDE)
•MSP430 Optimizing C/C++ Compiler: Check the --silicon_errata option
•MSP430 Assembly Language Tools
MSP430 GNU Compiler (MSP430-GCC)
•MSP430 GCC Options: Check -msilicon-errata= and -msilicon-errata-warn= options
•MSP430 GCC User's Guide
IAR Embedded Workbench
•IAR workarounds for msp430 hardware issues
www.ti.com Debug Only Advisories
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 3
Copyright © 2021 Texas Instruments Incorporated

5 Nomenclature, Package Symbolization, and Revision Identification
The revision of the device can be identified by the revision letter on the Package Markings or by the HW_ID
located inside the TLV structure of the device.
5.1 Device Nomenclature
To designate the stages in the product development cycle, TI assigns prefixes to the part numbers of all MSP
MCU devices. Each MSP MCU commercial family member has one of two prefixes: MSP or XMS. These
prefixes represent evolutionary stages of product development from engineering prototypes (XMS) through fully
qualified production devices (MSP).
XMS – Experimental device that is not necessarily representative of the final device's electrical specifications
MSP – Fully qualified production device
Support tool naming prefixes:
X: Development-support product that has not yet completed Texas Instruments internal qualification testing.
null: Fully-qualified development-support product.
XMS devices and X development-support tools are shipped against the following disclaimer:
"Developmental product is intended for internal evaluation purposes."
MSP 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 (XMS) have a greater failure rate than the standard production devices.
TI 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 temperature
range, package type, and distribution format.
5.2 Package Markings
PZ100 LQFP (PZ) 100 Pin
Nomenclature, Package Symbolization, and Revision Identification www.ti.com
4MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

PN80 LQFP (PN), 80 Pin
5.3 Memory-Mapped Hardware Revision (TLV Structure)
Die Revision TLV Hardware Revision
Rev A 10h
Further guidance on how to locate the TLV structure and read out the HW_ID can be found in the device User's
Guide.
www.ti.com Nomenclature, Package Symbolization, and Revision Identification
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 5
Copyright © 2021 Texas Instruments Incorporated

6 Advisory Descriptions
ADC39 ADC Module
Category Functional
Function Erroneous ADC10 results in extended sample mode
Description If the extended sample mode is selected (ADC10SHP = 0) and the ADC10CLK is
asynchronous to the SHI signal, the ADC10 may generate erroneous results.
Workaround 1) Use the pulse sample mode (ADC10SHP=1)
OR
2) Use a synchronous clock for ADC10 and the SHI signal.
ADC42 ADC Module
Category Functional
Function ADC stops converting when successive ADC is triggered before the previous conversion
ends
Description Subsequent ADC conversions are halted if a new ADC conversion is triggered while ADC
is busy. ADC conversions are triggered manually or by a timer. The affected ADC modes
are:
- sequence-of-channels
- repeat-single-channel
- repeat-sequence-of-channels (ADC12CTL1.ADC12CONSEQx)
In addition, the timer overflow flag cannot be used to detect an overflow
(ADC12IFGR2.ADC12TOVIFG).
Workaround 1. For manual trigger mode (ADC12CTL0.ADC12SC), ensure each ADC conversion
is completed by first checking ADC12CTL1.ADC12BUSY bit before starting a new
conversion.
2. For timer trigger mode (ADC12CTL1.ADC12SHP), ensure the timer period is greater
than the ADC sample and conversion time.
To recover the conversion halt:
1. Disable ADC module (ADC12CTL0.ADC12ENC = 0 and ADC12CTL0.ADC12ON = 0)
2. Re-enable ADC module (ADC12CTL0.ADC12ON = 1 and ADC12CTL0.ADC12ENC =
1)
3. Re-enable conversion
ADC69 ADC Module
Category Functional
Function ADC stops operating if ADC clock source is changed from SMCLK to another source
while SMCLKOFF = 1.
Advisory Descriptions www.ti.com
6MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

Description When SMCLK is used as the clock source for the ADC (ADC12CTL1.ADC12SSELx =
11) and CSCTL4.SMCLKOFF = 1, the ADC will stop operating if the ADC clock source is
changed by user software (e.g. in the ISR) from SMCLK to a different clock source. This
issue appears only for the ADC12CTL1.ADC12DIVx settings /3/5/7. The hang state can
be recovered by PUC/POR/BOR/Power cycle.
Workaround 1. Set CSCTL4.SMCLKOFF = 0 before switch ADC clock source.
OR
2. Only use ADC12CTL1.ADC12DIVx as /1, /2, /4, /6, /8
AUXPMM1 AUXPMM Module
Category Functional
Function AUXVCC1/AUXVCC2 can not be switched back to DVCC
Description When the system is running with the AUXVCC1 supply after DVCC/AVCC is lost, if the
AUXVCC1 voltage goes lower than SVSH setting for POR and above BORH level, the
system can not switch back to DVCC after DVCC ramps back up again.
Similarly, when the system is running with the AUXVCC2 supply after DVCC/AVCC is lost,
if the AUXVCC2 voltage goes lower than SVSH setting for POR and above BORH level,
the system can not switch back to DVCC after DVCC ramps back up again.
Workaround When the system is running with the AUXVCC1 supply, use SVMH to monitor AUXVCC1
voltage. When AUXVCC1 is lower than the SVMH setting, the program drives the chip
into LPMx.5. After DVCC ramps up back again, trigger one of the wake up pins. The
power supply could be switched back to DVCC again.
When the system is running with the AUXVCC2 supply, use SVMH to monitor AUXVCC2
voltage. When AUXVCC2 is lower than the SVMH setting, the program drives the chip
into LPMx.5. After DVCC ramps up back again, trigger one of the wake up pins. The
power supply could be switched back to DVCC again.
AUXPMM2 AUXPMM Module
Category Functional
Function Latch-up in AUXPMM
Description Latch-up current can appear at the AUXPMM module supply pins in the following two
scenarios:
Scenario 1: When the AUXPMM is configured for hardware- or software-controlled
switching and the module switches from DVCC to AUXVCC2, latch-up current can
appear at AUXVCC2 at the switching point defined by SVSMHCTL.SVSMHRRL (or
AUXCTL2.AUX0LVLx). The probability for this event to occur depends on:
a) Operating temperature (higher temperatures increase probability)
b) External AUXVCC2 voltage level (higher voltages increase probability)
c) SVSMHRRL level (lower levels increase probability) defining the switching level in
hardware-controlled mode
d) AUX0LVLx level (lower levels increase probability) defining the switching level in
software-controlled mode (applicable to DVCC only)
Scenario 2: When a battery is connected to DVCC, AUXVCC1 or AUXVCC2 as the first
voltage supply, due to the low internal resistance of the battery a very fast rise time is
www.ti.com Advisory Descriptions
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 7
Copyright © 2021 Texas Instruments Incorporated

seen by the AUXPMM and latch-up current can appear at the connected supply if:
a) Rise times are in the range of 140 kV/s (faster rise times increase probability)
b) Device operates at temperatures of 75 deg C and above (higher temperatures increase
probability)
The latch-up current disappears after complete power cycles of all supply sources.
Workaround For scenario 1:
- Increase SVSMRRL to a level above maximum external voltage expected on AUXVCC2.
SVSMRRL = 6 or 7 (requires VCORE level of 3) is applicable for AUXVCC2 of up to
maximum voltage, 3.58V, while a lower SVSMRRL setting can be selected if a lower
voltage (e.g. 3.3V) is expected on AUXVCC2.
Or
- Connect all 3 supplies via 3 external diodes to DVCC and realize the switching externally
without using the internal AUXPMM switches. See application report "Implementation of a
Three-Phase Electronic Watt-Hour Meter Using the MSP430F471xx" for details.
Or
- Use AUXVCC1 instead of AUXVCC2 for backup supply
For scenario 2:
Limit the supply voltage ramp up time through a series resistor (e.g. 10 Ohm) in the critical
supply path. Side effects such as voltage dips due to high current consumption of the
device need to be considered.
BSL7 BSL Module
Category Software in ROM
Function BSL does not start after waking up from LPMx.5
Description When waking up from LPMx.5 mode, the BSL does not start as it does not clear the Lock
I/O bit (LOCKLPM5 bit in PM5CTL0 register) on start-up.
Workaround 1. Upgrade the device BSL to the latest version (see Creating a Custom Flash-Based
Bootstrap Loader (BSL) Application Note - SLAA450 for more details)
OR
2. Do not use LOCKLPM5 bit (LPMx.5) if the BSL is used but cannot be upgraded.
BSL14 BSL Module
Category Software in ROM
Function BSL request to unlock the JTAG
Description The feature in the BSL to keep the JTAG unlocked by setting the bit
BSL_REQ_JTAG_OPEN in the return value has been disabled in this device.
Workaround None
CPU21 CPU Module
Category Compiler-Fixed
Advisory Descriptions www.ti.com
8MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

Function Using POPM instruction on Status register may result in device hang up
Description When an active interrupt service request is pending and the POPM instruction is used to
set the Status Register (SR) and initiate entry into a low power mode , the device may
hang up.
Workaround None. It is recommended not to use POPM instruction on the Status Register.
Refer to the table below for compiler-specific fix implementation information.
IDE/Compiler Version Number Notes
IAR Embedded Workbench Not affected
TI MSP430 Compiler Tools (Code
Composer Studio) v4.0.x or later
User is required to add the compiler
or assembler flag option below. --
silicon_errata=CPU21
MSP430 GNU Compiler (MSP430-
GCC) MSP430-GCC 4.9 build 167 or later
CPU22 CPU Module
Category Compiler-Fixed
Function Indirect addressing mode with the Program Counter as the source register may produce
unexpected results
Description When using the indirect addressing mode in an instruction with the Program Counter (PC)
as the source operand, the instruction that follows immediately does not get executed.
For example in the code below, the ADD instruction does not get executed.
mov @PC, R7
add #1h, R4
Workaround Refer to the table below for compiler-specific fix implementation information.
IDE/Compiler Version Number Notes
IAR Embedded Workbench Not affected
TI MSP430 Compiler Tools (Code
Composer Studio) v4.0.x or later
User is required to add the compiler
or assembler flag option below. --
silicon_errata=CPU22
MSP430 GNU Compiler (MSP430-
GCC) MSP430-GCC 4.9 build 167 or later
CPU36 CPU Module
Category Functional
Function PC corruption when single-stepping through flash erase
Description When single-stepping over code that initiates an INFOD Flash memory erase, the
program counter is corrupted.
www.ti.com Advisory Descriptions
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 9
Copyright © 2021 Texas Instruments Incorporated

Workaround None.
NOTE: This erratum applies to debug mode only.
CPU40 CPU Module
Category Compiler-Fixed
Function PC is corrupted when executing jump/conditional jump instruction that is followed by
instruction with PC as destination register or a data section
Description If the value at the memory location immediately following a jump/conditional jump
instruction is 0X40h or 0X50h (where X = don't care), which could either be an instruction
opcode (for instructions like RRCM, RRAM, RLAM, RRUM) with PC as destination
register or a data section (const data in flash memory or data variable in
RAM), then the PC value is auto-incremented by 2 after the jump instruction is executed;
therefore, branching to a wrong address location in code and leading to wrong program
execution.
For example, a conditional jump instruction followed by data section (0140h).
@0x8012 Loop DEC.W R6
@0x8014 DEC.W R7
@0x8016 JNZ Loop
@0x8018 Value1 DW 0140h
Workaround In assembly, insert a NOP between the jump/conditional jump instruction and program
code with instruction that contains PC as destination register or the data section.
Refer to the table below for compiler-specific fix implementation information.
IDE/Compiler Version Number Notes
IAR Embedded Workbench IAR EW430 v5.51 or later
For the command line version add
the following information Compiler:
--hw_workaround=CPU40
Assembler:-v1
TI MSP430 Compiler Tools (Code
Composer Studio) v4.0.x or later
User is required to add the compiler
or assembler flag option below. --
silicon_errata=CPU40
MSP430 GNU Compiler (MSP430-
GCC) Not affected
CPU46 CPU Module
Category Functional
Function POPM peforms unexpected memory access and can cause VMAIFG to be set
Description When the POPM assembly instruction is executed, the last Stack Pointer increment is
followed by an unintended read access to the memory. If this read access is performed
on vacant memory, the VMAIFG will be set and can trigger the corresponding interrupt
(SFRIE1.VMAIE) if it is enabled. This issue occurs if the POPM assembly instruction is
performed up to the top of the STACK.
Advisory Descriptions www.ti.com
10 MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

Workaround If the user is utilizing C, they will not be impacted by this issue. All TI/IAR/GCC pre-built
libraries are not impacted by this bug. To ensure that POPM is never executed up to the
memory border of the STACK when using assembly it is recommended to either
1. Initialize the SP to
a. TOP of STACK - 4 bytes if POPM.A is used
b. TOP of STACK - 2 bytes if POPM.W is used
OR
2. Use the POPM instruction for all but the last restore operation. For the the last restore
operation use the POP assembly instruction instead.
For instance, instead of using:
POPM.W #5,R13
Use:
POPM.W #4,R12
POP.W R13
Refer to the table below for compiler-specific fix implementation information.
IDE/Compiler Version Number Notes
IAR Embedded Workbench Not affected
C code is not impacted by this bug.
User using POPM instruction in
assembler is required to implement
the above workaround manually.
TI MSP430 Compiler Tools (Code
Composer Studio) Not affected
C code is not impacted by this bug.
User using POPM instruction in
assembler is required to implement
the above workaround manually.
MSP430 GNU Compiler (MSP430-
GCC) Not affected
C code is not impacted by this bug.
User using POPM instruction in
assembler is required to implement
the above workaround manually.
CPU47 CPU Module
Category Functional
Function An unexpected Vacant Memory Access Flag (VMAIFG) can be triggered
Description An unexpected Vacant Memory Access Flag (VMAIFG) can be triggered, if a PC-
modifying instruction (e.g. - ret, push, call, pop, jmp, br) is fetched from the last addresses
(last 4 or 8 byte) of a memory (e.g.- FLASH, RAM, FRAM) that is not contiguous to a
higher, valid section on the memory map.
www.ti.com Advisory Descriptions
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 11
Copyright © 2021 Texas Instruments Incorporated

In debug mode using breakpoints the last 8 bytes are affected.
In free running mode the last 4 bytes are affected.
Workaround Edit the linker command file to make the last 4 or 8 bytes of affected memory sections
unavailable, to avoid PC-modifying instructions on these locations.
Remaining instructions or data can still be stored on these locations.
DMA4 DMA Module
Category Functional
Function Corrupted write access to 20-bit DMA registers
Description When a 20-bit wide write to a DMA address register (DMAxSA or DMAxDA) is interrupted
by a DMA transfer, the register contents may be unpredictable.
Workaround 1. Design the application to guarantee that no DMA access interrupts 20-bit wide
accesses to the DMA address registers.
OR
2. When accessing the DMA address registers, enable the Read Modify Write disable bit
(DMARMWDIS = 1) or temporarily disable all active DMA channels (DMAEN = 0).
OR
3. Use word access for accessing the DMA address registers. Note that this limits the
values that can be written to the address registers to 16-bit values (lower 64K of Flash).
DMA7 DMA Module
Category Functional
Function DMA request may cause the loss of interrupts
Description If a DMA request starts executing during the time when a module register containing an
interrupt flags is accessed with a read-modify-write instruction, a newly arriving interrupt
from the same module can get lost. An interrupt flag set prior to DMA execution would not
be affected and remain set.
Workaround 1. Use a read of Interrupt Vector registers to clear interrupt flags and do not use read-
modify-write instruction.
OR
2. Disable all DMA channels during read-modify-write instruction of specific module
registers containing interrupts flags while these interrupts are activated.
DMA9 DMA Module
Category Functional
Function DMA stops transferring bytes unexpectedly
Description When the DMA is configured to transfer bytes from the eUSCI_A or eUSCI_B transmit
or receive buffers, the transmit or receive triggers (TXIFG and RXIFG) may not be seen
by the DMA module and the transfer of the bytes is missed. Once the first byte in a
Advisory Descriptions www.ti.com
12 MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

transfer sequence is missed, all the following bytes are missed as well. All eUSCI_A
modes (UART, SPI, and IrDA) and all eUSCI_B modes (SPI and I2C) are affected.
Workaround 1) Use Interrupt Service Routines to transfer data to and from the eUSCI_A or eUSCI_B.
OR
2) When using DMA channel 0 for transferring data to and from the eUSCI_A or
eUSCI_B, use DMA channel 2 (lower priority than DMA channel 0) to read the same
register of the eUSCI_A or eUSCI_B that DMA channel 0 is working with. Use the same
USCI IFG (e.g. UCA0RXIFG) as trigger source for these both DMA channels.
DMA10 DMA Module
Category Functional
Function DMA access may cause invalid module operation
Description The peripheral modules MPY, CRC, USB, RF1A and FRAM controller in manual mode
can stall the CPU by issuing wait states while in operation. If a DMA access to the
module occurs while that module is issuing a wait state, the module may exhibit undefined
behavior.
Workaround Ensure that DMA accesses to the affected modules occur only when the modules are
not in operation. For example with the MPY module, ensure that the MPY operation is
completed before triggering a DMA access to the MPY module.
EEM8 EEM Module
Category Debug
Function Debugger stops responding when using the DMA
Description In repeated transfer mode, the DMA automatically reloads the size counter (DMAxSZ)
once a transfer is complete and immediately continues to execute the next transfer unless
the DMA Enable bit (DMAEN) has been previously cleared. In burst-block transfer mode,
DMA block transfers are interleaved with CPU activity 80/20% - of ten CPU cycles, eight
are allocated to a block transfer and two are allocated for the CPU.
Because the JTAG system must wait for the CPU bus to be clear to halt the device, it can
only do so when two conditions are met:
- Three clock cycles after any DMA transfer, the DMA is no longer requesting the bus.
and
- The CPU is not requesting the bus.
Therefore, if the DMA is configured to operate in the repeat burst-block transfer mode,
and a breakpoint is set between the line of code that triggers the DMA transfers and the
line that clears the DMAEN bit, the DMA always requests the bus and the JTAG system
never gains control of the device.
Workaround When operating the DMA in repeat burst-block transfer mode, set breakpoint(s) only when
the DMA transfers are not active (before the start or after the end of the DMA transfers).
EEM17 EEM Module
Category Debug
Function Wrong Breakpoint halt after executing Flash Erase/Write instructions
www.ti.com Advisory Descriptions
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 13
Copyright © 2021 Texas Instruments Incorporated

Description Hardware breakpoints or Conditional Address triggered breakpoints on instructions that
follow Flash Erase/Write instructions, stops the debugger at the actual Flash Erase/Write
instruction even though the flash erase/write operation has already been executed. The
hardware/conditional address triggered breakpoints that are placed on either the next two
single opcode instructions OR the next double opcode instruction that follows the Flash
Erase/Write instruction are affected by this erratum.
Workaround None. Use other conditional/advanced triggered breakpoints to halt the debugger right
after Flash erase/write instructions.
Note
This erratum affects debug mode only.
EEM19 EEM Module
Category Debug
Function DMA may corrupt data in debug mode
Description When the DMA is enabled and the device is in debug mode, the data written by the DMA
may be corrupted when a breakpoint is hit or when the debug session is halted.
Workaround This erratum has been addressed in MSPDebugStack version 3.5.0.1. It is also available
in released IDE EW430 IAR version 6.30.3 and CCS version 6.1.1 or newer.
If using an earlier version of either IDE or MSPDebugStack, do not halt or use breakpoints
during a DMA transfer.
Note
This erratum applies to debug mode only.
EEM23 EEM Module
Category Debug
Function EEM triggers incorrectly when modules using wait states are enabled
Description When modules using wait states (USB, MPY, CRC and FRAM controller in manual mode)
are enabled, the EEM may trigger incorrectly. This can lead to an incorrect profile counter
value or cause issues with the EEMs data watch point, state storage, and breakpoint
functionality.
Workaround None.
Note
This erratum affects debug mode only.
JTAG26 JTAG Module
Category Debug
Function LPMx.5 Debug Support Limitations
Advisory Descriptions www.ti.com
14 MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

Description The JTAG connection to the device might fail at device-dependent low or high supply
voltage levels if the LPMx.5 debug support feature is enabled. To avoid a potentially
unreliable debug session or general issues with JTAG device connectivity and the
resulting bad customer experience Texas Instruments has chosen to remove the LPMx.5
debug support feature from common MSP430 IDEs including TIs Code Composer Studio
6.1.0 with msp430.emu updated to version 6.1.0.7 and IARs Embedded Workbench
6.30.2, which are based on the MSP430 debug stack MSP430.DLL 3.5.0.1 http://
www.ti.com/tool/MSPDS
TI plans to re-introduce this feature in limited capacity in a future release of the debug
stack by providing an IDE override option for customers to selectively re-activate LPMx.5
debug support if needed. Note that the limitations and supply voltage dependencies
outlined in this erratum will continue to apply.
For additional information on how the LPMx.5 debug support is handled within the
MSP430 IDEs including possible workarounds on how to debug applications using
LPMx.5 without toolchain support refer to Code Composer Studio User's Guide for
MSP430 chapter F.4 and IAR Embedded Workbench User's Guide for MSP430 chapter
2.2.5.
Workaround 1. If LPMx.5 debug support is deemed functional and required in a given scenario:
a) Do not update the IDE to continue using a previous version of the debug stack such as
MSP430.DLL v3.4.3.4.
OR
b) Roll back the debug stack by either performing a clean re-installation of a previous
version of the IDE or by manually replacing the debug stack with a prior version such as
MSP430.DLL v3.4.3.4 that can be obtained from http://www.ti.com/tool/MSPDS.
2. In case JTAG connectivity fails during the LPMx.5 debug mode, the device supply
voltage level needs to be raised or lowered until the connection is working.
Do not enable the LPMx.5 debug support feature during production programming.
JTAG27 JTAG Module
Category Debug
Function Unintentional code execution after programming via JTAG/SBW
Description The device can unintentionally start executing code from uninitialized RAM addresses
0x0006 or 0x0008 after being programming via the JTAG or SBW interface. This can
result in unpredictable behavior depending on the contents of the address location.
Workaround 1. If using programming tools purchased from TI (MSP-FET, LaunchPad), update to CCS
version 6.1.3 later or IAR version 6.30 or later to resolve the issue.
2. If using the MSP-GANG Production Programmer, use v1.2.3.0 or later.
3. For custom programming solutions refer to the specification on MSP430 Programming
Via the JTAG Interface User's Guide (SLAU320) revision V or newer and use
MSPDebugStack v3.7.0.12 or later.
For MSPDebugStack (MSP430.DLL) in CCS or IAR, download the latest version of the
development environment or the latest version of the MSPDebugStack
www.ti.com Advisory Descriptions
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 15
Copyright © 2021 Texas Instruments Incorporated

NOTE: This only affects debug mode.'
LCDB5 LCDB Module
Category Functional
Function Static DC charge can built up on dedicated COMx pins.
Description If the device is set into LPMx.5, its dedicated COMx pins (not shared with GPIO function)
are floating. External leakage paths to these pins can result in dedicated COMx pins being
charged. This can lead to static DC voltages being applied to the external LCD display.
This might cause long term over-stress to the LCD display and/or cause certain LCD
segments to flare up when device wakes up from LPMx.5 mode.
Workaround Connect a high-resistance resistor between the dedicated COM pins and Vss to
permanently discharge the affected pins.
LCDB6 LCDB Module
Category Functional
Function LCD outputs may be corrupted by modifying register fields VLCDx and/or LCDCPEN of
LCDCVCTL register while LCDON (LCDCCTL0) is set
Description Writing to VLCDx and/or LCDCPEN register bits in LCDCVCTL register while LCDC
is enabled (LCDON = '1' in LCDCCTL0 register) may corrupt the LCD output due to
incorrect start-up of LCD-controller and internal voltage generation.
Workaround Do not modify VLCDx and/or LCDCPEN bits in LCDCVCTL register while LCDON = '1'
PMM7 PMM Module
Category Functional
Function PMMRIE default conditions different than user guide
Description The user guide specifies that, after a BOR reset condition, the SVS will not be configured
to trigger a POR signal in the condition that the monitored voltages fall below the SVS
level(s). This is not true for this device. The SVS Low and SVS High Side POR Enable
bits (SVSLPE/SVSHPE) in the Power Management System Reset Enable and Interrupt
Enable register are set by default (PMMRIE = 0x1100).
Workaround If this behavior is not desired, reset the SVSLPE/SVSHPE bits in the PMMRIE register at
the beggining of the application.
PMM11 PMM Module
Category Functional
Function MCLK comes up fast on exit from LPM3 and LPM4
Description The DCO exceeds the programmed frequency of operation on exit from LPM3 and
LPM4 for up to 6 us. This behavior is masked from affecting code execution by default:
SVSL and SVML run in normal-performance mode and mask CPU execution for 150
us on wakeup from LPM3 and LPM4. However ,when the low-side SVS and the SVM
are disabled or are operating in full-performance mode (SVMLE= 0 and SVSLE= 0, or
SVMLFP= 1 and SVSLFP= 1) AND MCLK is sourced from the internal DCO running over
Advisory Descriptions www.ti.com
16 MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

4 MHz, 7 MHz,11 MHz,or 14 MHz at core voltage levels 0, 1, 2, and 3, respectively, the
mask lasts only 2 us. MCLK is, therefore, susceptible to run out of spec for 4 us.
Workaround Set the MCLK divide bits in the Unified Clock System Control 5 Register (UCSCTL5) to
divide MCLK by two prior to entering LPM3 or LPM4 (set DIVMx= 001). This prevents
MCLK from running out of spec when the CPU wakes from the low-power mode.
Following the wakeup fromthe low-power mode, wait 32, 48, 80, or 100 cycles for core
voltage levels 0, 1, 2, and 3, respectively, before resetting DIVM xto zero and running
MCLK at full speed [for example, __delay_cycles(100)]
PMM12 PMM Module
Category Functional
Function SMCLK comesup fast on exit from LPM3 and LPM4
Description The DCO exceeds the programmed frequency of operationon exit from LPM3 and LPM4
for up to 6 us. When SMCLK is sourced by the DCO, it is not masked on exit from LPM3
or LPM4. Therefore, SMCLK exceeds the programmed frequency of operation on exit
from LPM3 and LPM4 for up to 6 us. The increased frequency has the potential to change
the expected timing behavior of peripherals that select SMCLK as the clock source.
Workaround - Use XT2 as the SMCLK oscillator source instead of the DCO
or
- Do not disable the clock request bit for SMCLKREQEN in the Unified Clock System
Control 8 Register (UCSCTL8). This means that all modules that depend on SMCLK to
operate successfully should be halted or disabled before entering LPM3 or LPM4. If the
increased frequency prevents the proper function of an affected module, wait 32, 48, 80
or 100 cycles for core voltage levels 0, 1, 2, or 3, respectively, before re-enabling the
module. (for example, __delay_cycles(100)
PMM14 PMM Module
Category Functional
Function Increasing the core level when SVS/SVM low side is configured in full-performance mode
causes device reset
Description When the SVS/SVM low side is configured in full performance mode
(SVSMLCTL.SVSLFP = 1), the setting time delay for the SVS comparators is ~2us. When
increasing the core level in full-performance mode; the core voltage does not settle to the
new level before the settling time delay of the SVS/SVM comparator expires. This results
in a device reset.
Workaround When increasing the core level; enable the SVS/SVM low side in normal mode
(SVSMLCTL.SVSLFP=0). This provides a settling time delay of approximately 150us
allowing the core sufficient time to increase to the expected voltage before the delay
expires.
PMM15 PMM Module
Category Functional
Function Device may not wake up from LPM2, LPM3, or LPM4
www.ti.com Advisory Descriptions
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 17
Copyright © 2021 Texas Instruments Incorporated

Description Device may not wake up from LPM2, LPM3 or LMP4 if an interrupt occurs within 1
us after the entry to the specified LPMx; entry can be caused either by user code or
automatically (for example, after a previous ISR is completed). Device can be recovered
with an external reset or a power cycle. Additionally, a PUC can also be used to reset
the failing condition and bring the device back to normal operation (for example, a PUC
caused by the WDT).
This effect is seen when:
- A write to the SVSMHCTL and SVSMLCTL registers is immediately followed
by an LPM2, LPM3, LPM4 entry without waiting the requisite settling time
((PMMIFG.SVSMLDLYIFG = 0 and PMMIFG.SVSMHDLYIFG = 0)).
or
The following two conditions are met:
- The SVSL module is configured for a fast wake-up or when the SVSL/SVML module is
turned off. The affected SVSMLCTL register settings are shaded in the following table.
and
-The SVSH/SVMH module is configured to transition from Normal mode to an OFF
state when moving from Active/LPM0/LPM1 into LPM2/LPM3/LPM4 modes. The affected
SVSMHCTL register settings are shaded in the following table.
Workaround Any write to the SVSMxCTL register must be followed by a settling delay
(PMMIFG.SVSMLDLYIFG = 0 and PMMIFG.SVSMHDLYIFG = 0) before entering LPM2,
LPM3, LPM4.
Advisory Descriptions www.ti.com
18 MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated

and
1. Ensure the SVSx, SVMx are configured to prevent the issue from occurring by the
following:
- Configure the SVSL module for slow wake up (SVSLFP = 0). Note that this will increase
the wakeup time from LPM2/3/4 to twakeupslow (~150 us).
or
- Do not configure the SVSH/SVMH such that the modules transition from Normal mode
to an OFF state on LPM entry and ensure SVSH/SVMH is in manual mode. Instead
force the modules to remain ON even in LPMx. Note that this will cause increased power
consumption when in LPMx.
Refer to the MSP430 Driver Library(MSPDRIVERLIB) for proper PMM configuration
functions.
Use the following function, PMM15Check (void), to determine whether or not the existing
PMM configuration is affected by the erratum. The return value of the function is 1 if the
configuration is affected, and 0 if the configuration is not affected.
unsigned char PMM15Check (void)
{
// First check if SVSL/SVML is configured for fast wake-up
if ( (!(SVSMLCTL & SVSLE)) || ((SVSMLCTL & SVSLE) && (SVSMLCTL & SVSLFP)) ||
(!(SVSMLCTL & SVMLE)) || ((SVSMLCTL & SVMLE) && (SVSMLCTL & SVMLFP)) )
{ // Next Check SVSH/SVMH settings to see if settings are affected by PMM15
if ((SVSMHCTL & SVSHE) && (!(SVSMHCTL & SVSHFP)))
{
if ( (!(SVSMHCTL & SVSHMD)) || ((SVSMHCTL & SVSHMD) &&
(SVSMHCTL & SVSMHACE)) )
return 1; // SVSH affected configurations
}
if ((SVSMHCTL & SVMHE) && (!(SVSMHCTL & SVMHFP)) && (SVSMHCTL &
SVSMHACE))
return 1; // SVMH affected configurations
}
return 0; // SVS/M settings not affected by PMM15
}
}
2. If fast servicing of interrupts is required, add a 150us delay either in the interrupt
service routine or before entry into LPM3/LPM4.
PMM18 PMM Module
Category Functional
Function PMM supply overvoltage protection falsely triggers POR
Description The PMM Supply Voltage Monitor (SVM) high side can be configured as overvoltage
protection (OVP) using the SVMHOVPE bit of SVSMHCTL register. In this mode a POR
should typically be triggered when DVCC reaches ~3.75V.
If the OVP feature of SVM high side is enabled going into LPM234, the SVM might trigger
at DVCC voltages below 3.6V (~3.5V) within a few ns after wake-up. This can falsely
www.ti.com Advisory Descriptions
SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
MSP430F6721 Microcontroller 19
Copyright © 2021 Texas Instruments Incorporated

cause an OVP-triggered POR. The OVP level is temperature sensitive during fail scenario
and decreases with higher temperature (85 degC ~3.2V).
Workaround Use automatic control mode for high-side SVS & SVM (SVSMHCTL.SVSMHACE=1). The
SVM high side is inactive in LPM2, LPM3, and LPM4.
PMM20 PMM Module
Category Functional
Function Unexpected SVSL/SVML event during wakeup from LPM2/3/4 in fast wakeup mode
Description If PMM low side is configured to operate in fast wakeup mode, during wakeup
from LPM2/3/4 the internal VCORE voltage can experience voltage drop below the
corresponding SVSL and SVML threshold (recommendation according to User's Guide)
leading to an unexpected SVSL/SVML event. Depending on PMM configuration, this
event triggers a POR or an interrupt.
Note
As soon the SVSL or the SVML is enabled in Normal performance mode the
device is in slow wakeup mode and this erratum does not apply. In addition, this
erratum has sporadic characteristic due to an internal asynchronous circuit. The
drop of Vcore does not have an impact on specified device performance.
Workaround If SVSL or SVML is required for application (to observe external disruptive events at
Vcore pin) the slow wakeup mode has to be used to avoid unexpected SVSL/SVML
events. This is achieved if the SVSL or the SVML is configured in "Normal" performance
mode (not disabled and not in "Full" Performance Mode).
PMM26 PMM Module
Category Functional
Function Device lock-up if RST pin pulled low during write to SVSMHCTL or SVSMLCTL
Description Device results in lock-up condition under one of the two scenarios below:
1) If RST pin is pulled low during write access to SVSMHCTL, with the RST/NMI pin
is configured to reset function and is pulled low (reset event) the device will stop code
execution and is continuously held in reset state. RST pin is no longer functional. The only
way to come out of the lock-up situation is a power cycle.
OR
2) If RST pin is pulled low during write access to SVSMLCTL and only if the code that
checks for SVSMLDLYIFG==1 is implemented without a timeout. The device will be stuck
in the polling loop polling since SVSMLDLYIFG will never be cleared.
Workaround Follow the sequence below to prevent the lock-up for both use cases:
1) Disable RST pin reset function and switch to NMI before access SVSMHCTL or
SVSMLCTL.
then
2) Activate NMI interrupt and handle reset events in this time by SW (optional if reset
functionality required during access SVSMHCTL or SVSMLCTL)
then
3) Enable RST pin reset function after access to SVSMHCTL or SVSMLCTL
Advisory Descriptions www.ti.com
20 MSP430F6721 Microcontroller SLAZ339AF – OCTOBER 2012 – REVISED MAY 2021
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated