TPS65218D0RSLR
Product Specifications
TPS65218D0RSLR
Texas Instruments — Power Management & Signal Chain
Product Overview
The TPS65218D0RSLR from Texas Instruments is a power management / signal chain IC delivering high efficiency, precision regulation, and robust protection. Texas Instruments (TI) is a global leader in analog ICs, power management solutions, and embedded processors.
Superb Automation partners with leading brands including Qorvo, TI, ADI, XILINX, NVIDIA, Intel, ST, MACOM, Renesas, and Coilcraft to provide complete BOM and supply-chain solutions.
Key Features
- High-efficiency power conversion with low quiescent current
- Wide input voltage range with precision output regulation
- Integrated over-current, over-temperature, and short-circuit protection
- Small footprint packaging for space-constrained designs
- AEC-Q100 qualified options available for automotive applications
- Flexible sequencing and soft-start control
Key Specifications
| Input Voltage | V |
| Output Voltage | V |
| Output Current | A |
| Switching Frequency | kHz |
| Efficiency | % |
| Quiescent Current | µA |
| Topology | — |
| Package | — |
| Operating Temp | °C |
| Protection | — |
Applications
Request Quote — In Stock, 100% New & Original, Full Warranty
TPS65218D0RSLR: a four rail PMIC for Cortex-A8 and A9 systems in the 6 mm VQFN
TPS65218D0RSLR is a power management IC with three bucks, a buck boost, two backup rails and a load switch set in a VQFN (RSL) body of 48 pins. The RSL body is 6.00 mm by 6.00 mm with forty eight pins, and the D0 in the string is the silicon revision rather than a package code: the revision register at subaddress 5h reads 4.0 for D0. The document is SLDS206E, revised February 2021, and its first page still describes the family by rail rather than by revision.
Order code and package
| TPS65218D0RSLR | VQFN (RSL), 48 pins, 6.00 mm by 6.00 mm; the revision register 5h reads 4.0 for the D0 revision |
| Addendum rows | the published rows in SLDS206E are the B1 strings, TPS65218B1RSLR at 2500 pieces on a large reel and TPS65218B1RSLT at 250 on a small reel, every one of them marked NRND |
| HTQFP alternative | the same die in the 48 pin PHP body at 7.00 mm by 7.00 mm, 1000 pieces on a large reel for the B101 and B1 strings |
| Finish and moisture | NIPDAU lead finish, MSL level 3 at 260 C with a 168 hour floor life, -40 to 105 C operating range |
| Supervisor | built in monitors on DCDC1 and DCDC2 at plus or minus 4 percent and on DCDC3, DCDC4 and LDO1 at plus or minus 5 percent |
| Interface | I2C slave at address 0x24, data rates to 400 kbps, with a separate power good output for the backup domain |
Electrical parameters that decide the application
| DCDC1 and DCDC2 | 1.8 A each from a 2.7 V to 5.5 V input, 1.1 V default, adjustable from 0.85 V to 1.675 V, 100 percent duty for lowest dropout |
| DCDC3 | 1.8 A from the same input, 1.2 V default, adjustable from 0.9 V to 3.4 V |
| DCDC4 | buck boost, 1.6 A, 3.3 V default, adjustable from 1.175 V to 3.4 V, which is the rail that holds up when the input crosses the output |
| DCDC5 and DCDC6 | 1 V and 1.8 V fixed, fed from the system supply or from a coin cell on the CC pin, with an IN_nCC flag that reports which source is in use |
| LDO1 | 400 mA from 1.8 V to 5.5 V in, 1.8 V default, adjustable 0.9 V to 3.4 V, dropout 400 mV maximum at 2.7 V in |
| Load switches | LS1 350 mA from 1.2 V to 3.6 V with 110 mOhm at 1.35 V; LS2 100 mA or 500 mA selectable from 4 V to 5.5 V with 500 mOhm; LS3 the same current options at the higher voltage |
| Switching frequency | 2400 kHz on the integrated bucks, which is what keeps the inductors small in a 6 mm package footprint |
| Sequencing | rail order and timing are set through I2C registers rather than by external components, and the settings live in OTP |
| Protection | undervoltage lockout per rail, overtemperature warning ahead of shutdown, and active output discharge when a rail is disabled |
| Always on | a push button monitor that stays alive on the backup domain, which is why the coin cell path exists at all |
Family variants and how they differ
| TPS65218B1 | the earlier revision, published in the addendum as NRND, so a board that was qualified on B1 and needs D0 has to confirm the register map differences rather than assume they are identical |
| TPS6521815 and TPS6521835 | OTP images of the same die with different default rail voltages and sequencing, selected by the first register |
| RSL against PHP | 6 mm VQFN against 7 mm HTQFP, same 48 pins of function, different thermal path and different reel sizes |
| Against a discrete set | four bucks, an LDO, three load switches and a supervisor in one body against the same functions as separate parts |
| No boost for the main rails | DCDC4 is the only rail that can run above its input, and it is limited to 1.6 A |
Points to check before this part goes on a BOM
The order code needs a check before anything else, because SLDS206E publishes B1 rows only and marks every one of them NRND. The D0 revision appears in the document as a value in the revision register, not as an orderable row, so a buyer holding a TPS65218D0RSLR requirement is asking for a revision whose reel quantity and lifecycle status are not printed in this revision of the document. That is worth settling with the vendor or with a date coded stock check rather than assuming 2500 pieces from the B1 row.
The rail budget is fixed by the silicon, so the useful question is which rails a given system actually loads. DCDC1 and DCDC2 at 1.8 A each normally feed the core and the MPU domain, DCDC3 at 1.8 A feeds DDR or a 1.2 V peripheral domain, and DCDC4 at 1.6 A is the one that has to survive an input that crosses its output, which is why it is a buck boost and not a fourth buck.
DCDC5 and DCDC6 are the reason this part shows up in battery products. They run from either the system rail or a coin cell on the CC pin, and the IN_nCC flag tells the processor which one is live so the software can decide whether to keep the real time clock running or start a controlled shutdown. That behaviour is inside the part, so it does not depend on the processor staying awake.
Everything programmable is programmed over I2C at 0x24 and then stored, which means an unprogrammed part comes up on its OTP defaults. A board that expects 1.8 V on LDO1 and gets the 1.8 V default is fine, but a board that reorders the rails has to have the sequencing written before the processor is released from reset, otherwise the first boot and every later boot behave differently.
What the load switch set is for
LS1 is a 350 mA switch with 110 mOhm at 1.35 V, and LS2 and LS3 offer 100 mA or 500 mA at 5 V with 500 mOhm. In practice LS1 powers a low voltage peripheral domain such as an SD card or a sensor hub, and LS2 and LS3 switch USB or HDMI 5 V branches. All three are controlled through registers, so a rail can be dropped in software without a GPIO.
The current limit is selectable on LS2 and LS3 and fixed on LS1, which is the detail that decides whether a 5 V branch with a 400 mA inrush can be switched at all: the 100 mA setting will fold back on the inrush, and the 500 mA setting is the only one that lets the branch start.
Why 2.4 MHz matters here
At 2400 kHz the buck inductors are small enough to sit beside a 6 mm package, which is the whole point of an integrated PMIC. The cost is switching loss and a tighter minimum on time, so a rail that converts 5 V to 0.85 V runs at a low duty and has to be checked against the minimum controllable on time rather than assumed to regulate.
The high frequency also sets what the output capacitors have to be. A 1.8 A rail at 2.4 MHz with ceramic output capacitors needs far less capacitance than the same rail at 500 kHz, but the ESR and the loop stability still have to be checked against the document's recommended values rather than scaled by intuition.
Where this order code fits
It is the power tree for a Cortex-A8 or A9 board: industrial HMI panels, point of sale terminals, test equipment and personal navigation units running from a 5 V adapter or a single cell pack. Teams comparing it with a neighbouring part start from our Power Management Cross Reference page, and teams replacing a part that has gone end of life start from the Component Obsolescence Management page.
Questions buyers ask
Is this part still recommended for new builds? The addendum rows in SLDS206E are marked NRND, so a new build should confirm lifecycle status with the vendor before committing. Existing builds are unaffected.
What does the D0 in the order code mean? It is the silicon revision, not a package. The revision register at 5h reads 4.0 for D0, and the document publishes B1 orderable rows only.
Can the backup rails run from a coin cell alone? Yes. DCDC5 and DCDC6 take their input from the CC pin when the system rail is absent, and the IN_nCC flag reports which source is in use.
How many pieces are on a large reel? 2500 for the RSL VQFN in the published B1 row and 1000 for the PHP HTQFP. The D0 string row is not printed, so confirm the quantity on the offer.
SUPERB holds and sources this line for production and repair demand, quoted against the exact ordering string with date codes, packing and country of origin stated on the offer. When a part is moving, the Last-Time-Buy and LTB Planning page sets out the checks that keep a substitute from failing on the bench.