Contact Us

TPS628681ARQYR

Brand: Texas Instruments
Texas Instruments TPS628681ARQYR is a power management component designed for AI Servers & Low-Altitude Economy (Drone and UAV) and related applications. In stock at Superb Automation — request a quote for competitive pricing and fast delivery.
quote now

Product Specifications

TPS628681ARQYR — Power Management & Signal Chain | Superb Automation

TPS628681ARQYR

Texas Instruments — Power Management & Signal Chain

Product Overview

The TPS628681ARQYR 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 VoltageV
Output VoltageV
Output CurrentA
Switching FrequencykHz
Efficiency%
Quiescent CurrentµA
Topology—
Package—
Operating Temp°C
Protection—

Applications

Industrial Automation Automotive Electronics Telecom Infrastructure Server & Data Center Power Medical Devices Consumer Electronics

Request Quote — In Stock, 100% New & Original, Full Warranty

TPS628681ARQYR: 4 A step down converter with I2C voltage scaling in 5 mV steps

TPS628681ARQYR is a 4 A synchronous buck in the nine pin RQY body that runs from 2.4 V to 5.5 V, switches at 2.4 MHz, uses DCS-Control and takes its output voltage from an I2C register anywhere between 0.4 V and 1.675 V in 5 mV steps in a RQY QFN body of 9 pins. The TPS628681A is the 4 A member of the TPS62868x family, with 11 milliohm and 10.5 milliohm integrated switches and better than 90 percent efficiency at a 0.9 V output. The output voltage is set by an I2C register with a 5 mV step over the 0.4 V to 1.675 V range, and the start-up value is separately selected by a resistor on the VSET/VID pin so the rail comes up at a known level before any bus traffic arrives. Output accuracy is 1 percent over line, load and temperature, and the bus runs at up to 3.4 Mbps.

Order code and package

Order codeTPS628681ARQYR
Input voltage2.4 V to 5.5 V
Output voltage range0.4 V to 1.675 V
Voltage step5 mV
Start-up voltage range0.4 V to 1.15 V, resistor selected
Rated output current4 A
Switching frequency2.4 MHz
Integrated switch resistance11 milliohm high side, 10.5 milliohm low side
InterfaceI2C at up to 3.4 Mbps
Packagenine pin RQY QFN

Electrical parameters that decide the application

Output accuracyplus or minus 1 percent
Efficiency at 0.9 V outputgreater than 90 percent
Control topologyDCS-Control
Internal reference step2.5 mV on the 0.2 V to 0.8375 V range, 5 mV on this range, 10 mV on the 0.8 V to 3.35 V range
Enable delay, 0C variants100 us minimum, 350 us typical, 900 us maximum
SDA and SCLmust not be left floating, tie to ground if unused
Mode changeforced PWM selectable through the CONTROL register
Power goodwindow comparator with a lower threshold on most variants
Thermal padexposed pad on the nine pin body

Points to check before this part goes on a BOM

A 5 mV step across the 1.275 V span of this range is 255 codes, which is finer than the 1 percent accuracy of the rail at any point in the range. That matters because it changes what the interface is for: the step size is a convenience for a control loop that wants to walk the rail in small increments, not a way to hit an absolute voltage. At a 0.6 V output, 1 percent is 6 mV, so one code is smaller than the total error band. A production test that checks the rail against the register value has to allow for that, and a margin test that moves the rail by a few codes is testing the register, not the silicon.

Four amps through 11 milliohm is 0.176 W in the high side switch before the duty cycle is applied. At a 5 V input and a 0.9 V output the duty cycle is 18 percent, so the high side conducts for less than a fifth of the period and the conduction loss on that side is about 32 mW. The low side conducts for the rest and, at 10.5 milliohm, carries 0.16 W. The sum is close to 0.19 W, which on a nine pin body in a 2 mm square is the number that decides whether the pad has vias into a plane or only into the top copper.

The start-up voltage comes from a resistor on the VSET/VID pin and is independent of the bus. That is a deliberate ordering: a system whose firmware fails to start still gets a valid rail at the resistor value, so the processor can be debugged instead of being held in reset by a supply that never came up. The corollary is that the resistor value is a safety parameter for the load. A core rail that comes up at 1.15 V when the device is only rated for 0.95 V is a dead processor on the bench, so the code is worth checking against the load rating before the board is powered the first time.

At 2.4 MHz the period is 417 ns, and a 5 V input to a 0.4 V output is a duty cycle of 8 percent, which is 33 ns of on time. That is the corner of this device rather than its centre. Any converter has a minimum controllable on time, and below it the loop either skips pulses or lets the output rise, so the highest input paired with the lowest output is the case to measure rather than assume. A rail that spends its life at 0.4 V should be run from a lower input if the option exists, which is what the 2.4 V minimum input allows.

SDA and SCL are real pins with real edges, and the document asks for them to be tied to ground if unused. A floating bus line on a switching board is an antenna, and the receiver in this device can interpret noise as a write to the voltage register. The failure looks like a rail that moves on its own at a particular load, which is difficult to reproduce and easy to prevent with two resistors.

The 6 A sibling in the same family, the TPS62869x, shares the body and the pinout with this 4 A part. A board built for one accepts the other, which makes a substitution look free on the drawing. It is not free on the passive side: the inductor and the output capacitor were sized for the ripple of the 4 A part, and the 6 A part pushes more current through them at the same load. The swap is safe when the passives were sized for the larger current and not otherwise, and the reel label is the only place where the difference appears once the parts are on the board.

The I2C interface runs at up to 3.4 Mbps and the register map is small, so a voltage change is one transaction. The ramp between two settings is set by the output capacitor and the load current rather than by the interface: a 22 uF output at 2 A of load moves 100 mV in about a microsecond if the converter can source the current, which is faster than most loads accept. The step size helps here, because a sequence of small steps with a delay between them produces a controlled ramp out of a converter that has no internal ramp generator for voltage changes.

Where this order code fits

This order code belongs on processor core rails, FPGA auxiliary rails, DDR termination and sensor supplies in handheld instruments, portable medical devices and battery powered industrial terminals, where the rail has to move while the system runs. It is also the part chosen where the rail must come up to a safe value with no firmware running, because the resistor on the VSET pin defines the start-up level independently of the bus. 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 BOM Shortage Recovery page.

Questions buyers ask

What sets the voltage before the bus is active? A resistor on the VSET/VID pin selects the start-up voltage from a table between 0.4 V and 1.15 V. The rail reaches that value with no bus traffic, so a processor can boot and then move the rail through the I2C register. If the bus never arrives the rail stays at the resistor value, which is a known state rather than an undefined one.

Can this part replace the 6 A version? It shares the body and the pinout with the 6 A member of the same family, so the footprint is the same. The passives are not automatically the same, because the ripple current and the transient current of the larger part are higher. A substitution in the direction of the smaller part is safe on the passives side and a substitution in the other direction needs the inductor and the capacitor checked.

How fine is the voltage resolution in practice? The register step is 5 mV on this 0.4 V to 1.675 V range, and the rail accuracy is 1 percent, which at 0.6 V is 6 mV. The step is therefore finer than the error, so moving the rail by one code does not reliably move the delivered voltage by 5 mV. The resolution is useful for walking a rail in a controlled way, not for trimming an absolute value.

What happens at a 5 V input and a 0.4 V output? The duty cycle is 8 percent, which at 2.4 MHz is 33 ns of on time. That is close to the minimum controllable pulse width of the regulator, so the loop has to skip or stretch pulses to hold the rail, and the output ripple rises. A board that needs both the highest input and the lowest output should measure that corner rather than take the datasheet efficiency curve as the answer.

SUPERB holds and sources this converter line against the exact order code, with the reel, date code and country of origin stated on the offer. When a rail has to move to a second part, the Component Lead Time Solutions page sets out the checks that separate a working substitute from one that fails on the bench.