TPS53667RTAR
Product Specifications
TPS53667RTAR
Texas Instruments — Power Management & Signal Chain
Product Overview
The TPS53667RTAR 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
TPS53667 six phase D-CAP+ controller with NVM and PMBus for high current core rails
The TPS53667RTAR is the six phase member of the same driverless controller family as the four phase TPS53647, in the same 40 pin 6 mm by 6 mm QFN body and with the same 4.5 V to 17 V conversion range. The difference that matters on a board is headroom: six phases instead of four, each driving a CSD9549x NexFET stage, and a phase overlap mode that the four phase part does not have. In the four phase part the duty cycle of any one phase is capped at one over the phase count, which limits how low the output can go from a 12 V input at a given frequency. In the six phase part the phases are allowed to overlap in steady state and during a transient, so the same 12 V bus can reach a lower output voltage without dropping the switching frequency.
The part carries non volatile memory, so the boot voltage, the current limit, the ramp, the load line and the phase shedding thresholds can be stored in the controller and latched at start up without a host. Every one of those settings also has a pinstrap resistor route, which is what allows a board to be built and characterised before the production NVM image is signed off.
Key parameters taken from the SLUSC40B document
| Conversion input | 4.5 V to 17 V recommended, 19 V absolute, with V5 bias at 4.5 V to 5.5 V and a 3.3 V LDO on V3R3 |
| Phases | 1, 2, 3, 4, 5 or 6, with CSP2 to CSP6 tied to V3R3 to remove the phases that are not fitted |
| Boot voltage | 0.5 V to 2.5 V in 5 mV steps, 8 bit, from pinstrap or NVM |
| Power stage | CSD9549x NexFET stages, current sense gain 5 mV per amp, the CSD95490Q5MC used in the vendor reference example |
| Switching frequency | 300 kHz to 1 MHz set by the resistor from F-IMAX to ground, with table rows such as 30 k for 600 kHz nominal and 56 k for 1 MHz nominal |
| Reference output | 1.70 V typical, 1.685 V to 1.717 V on VREF |
| Output tolerance | 6 mV typical and 6.5 mV maximum in the 0.61 V to 0.995 V band at 0 to 85 degrees, 0.6 per cent up to 1.52 V, 1 per cent up to 2.5 V, 0.8 per cent over minus 40 to plus 125 degrees |
| Current report | IMON reaches 0.85 V at IMAX, with a 49.9 k resistor in the vendor's 120 A example |
| Standby current | 6.1 mA typical and 7 mA maximum on V5 with ENABLE low |
| Temperature | Minus 40 to plus 125 degrees ambient, 150 degrees junction maximum |
Phase shedding thresholds in real numbers
Dynamic phase shedding is programmed as a percentage of the per phase current limit rather than as an absolute amp figure, so the thresholds travel with the build. The vendor table gives 10 per cent and 25 per cent as the low and high thresholds for dropping phases, 60 per cent for the four to six phase step, and hysteresis of 5 per cent and 10 per cent between the two directions. On a rail that spends most of its life at 20 A and occasionally steps to 150 A, those numbers decide how many phases are switching at idle, and therefore whether the light load efficiency curve is flat or has a cliff in it.
Finding a phase that has stopped conducting
Each phase has its own current sense input, and the controller compares the per phase current against the average. When one phase drifts outside the window the part sets a warning bit, and the faulty phase number is readable in the status register. Individual phases can also be turned off deliberately over PMBus, which is the practical way to run a six phase board on four phases when a stage is unavailable, without changing the NVM image permanently. That option is worth knowing during a shortage: a rail built for six stages can often carry the same load on five, at a lower efficiency and a higher ripple, until the missing stage lands.
Configuring without a host on the bus
The pinstrap route uses resistors to VREF on the OCL-R, F-IMAX, SLEW-MODE, VBOOT and ADDR-TRISE pins. Each resistor sets a level, and the level is latched at power up. The vendor notes a specific interaction: when the SKIP-NVM pin is pulled to ground through 20 k, the resistors on O-USR, F-IMAX and the other configuration pins are read as pinstraps rather than the NVM image being used. A board that behaves differently from its sister board usually differs in that one resistor, not in the controller.
The bus side accepts 1.8 V and 3.3 V logic levels, and the address is set by the divider on ADDR-TRISE. In the vendor example the address 111 0001 is obtained with a 24 k resistor to ground and a 16.5 k resistor to VREF. Once the address is fixed, the same bus carries the fault limits, the telemetry and the on the fly phase count.
Where the six phase part is the right choice
Six phases at 600 kHz suits a core rail in the 150 A to 300 A range, a switch ASIC, or a datacentre accelerator card where the 12 V bus has a short, hot path to the load and the transient is brutal. It also suits a board that has to keep running while one stage is out of stock, because the phase count is programmable at run time. The controller itself draws no load current, so the thermal load belongs to the stages and the inductors, and the routing of the current sense returns matters more than the copper weight under the controller.
For a buyer the useful question is which stage is available and at what reel size, since the controller is common across a wide range of rails. Suppliers list the CSD9549x family on power management cross reference, and the drop in rules for a substitute stage are set out on how to verify a drop in replacement.
Ordering and incoming inspection notes
The part is a QFN with an exposed pad, so the reel label, the moisture level and the floor life are part of the specification, not paperwork. On receipt the marking should be checked against the configuration that was ordered, because a part with a different NVM image will still read as the same base part number. Where a project needs a documented second source, the six phase controller has fewer substitutes than the power stage, so the practical plan is to qualify two stages and keep one controller. Our desk keeps the stage options and the reel sizes together in the quote so the two decisions are made at the same time, and a shortage on either side is visible before the purchase order is placed. See one stop BOM sourcing for the combined view.