Contact Us

TMS320F28335PGFA

Brand: Texas Instruments
Texas Instruments TMS320F28335PGFA 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

TMS320F28335PGFA — Power Management & Signal Chain | Superb Automation

TMS320F28335PGFA

Texas Instruments — Power Management & Signal Chain

Product Overview

The TMS320F28335PGFA 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

TMS320F28335PGFA: 150 MHz Delfino with a floating point unit in a 176-pin LQFP

TMS320F28335PGFA is a 32-bit floating point control microcontroller in a LQFP body of 176 pins. It is the F28335: 150 MHz, a single precision floating point unit, 256K by 16 of flash and 34K by 16 of SARAM, in the 176-pin LQFP. The PGFA order code is the A grade at minus 40 to plus 85 degrees, and the floating point unit is what removes the scaling work that fixed point control code carries.

Order code and package

Order codeTMS320F28335PGFA
Package176-pin LQFP, PGF body
Packingtray
Core32-bit C28x at 150 MHz with a single precision floating point unit
Memory256K by 16 flash, 34K by 16 SARAM, 1K by 16 OTP
External busXINTF present
Temperatureminus 40 to plus 85 degrees Celsius on the A grade

Electrical parameters that decide the application

ADC12-bit, 16 channels, 80 ns conversion, two sample and hold
PWMsix ePWM modules with high resolution mode
DMAsix channel direct memory access
Serialthree SCI, two SPI, one I2C, one eCAN, two McBSP
Timersthree 32-bit CPU timers and a watchdog
Supply3.3 V input and output, 1.8 V core
Code security128-bit key over flash, SARAM and OTP

Family variants and how they differ

F28335256K flash and 34K SARAM with the floating point unit, the part on this page
F28334128K flash and 34K SARAM, same core
F2833264K flash and 26K SARAM, reduced peripherals
F28235the fixed point version of the same peripheral set

Points to check before this part goes on a BOM

The floating point unit is the practical difference from the fixed point members of the family: a current loop written in floating point runs without the per unit scaling, the saturation code and the lookup tables that a fixed point implementation needs. That saves engineering time and program space, and it costs a small amount of cycle time on the conversions the compiler inserts.

256K by 16 of flash is 512 kilobytes, and 34K by 16 of SARAM is 68 kilobytes. That is enough for a servo drive with a field bus, a motion profile table and a logging ring, which is why this device number outlived the rest of the family in industrial products. The SARAM is the tighter of the two limits once a real time operating system is added.

The 176-pin LQFP shares pins between the external interface, the ePWM outputs and the capture inputs. A product that uses the external bus has to give up some peripheral pins, and the pin map is fixed by the silicon revision. Planning that map from the pin multiplexing tables before the first board avoids a second spin.

Six ePWM modules with high resolution mode give the fine duty resolution that a digital power supply needs at high switching frequency. At 150 MHz the base PWM resolution is coarse for a 500 kHz converter, and the high resolution extension is what makes the difference between a usable duty step and a limit cycle oscillation.

Two rails are required, 1.8 V for the core and 3.3 V for the input and output rings, and the data sheet specifies a power up order. Many products use a single sequencing regulator and never notice; the ones that fail usually bring up 3.3 V first with a slow ramp and let the reset release early. The reset supervisor is part of the power tree, not an afterthought.

The A grade in this order code stops at 85 degrees ambient. The S grade of the same device number covers 125 degrees and is a different line item. In a sealed enclosure the choice should come from the measured internal ambient plus the calculated junction rise, not from the room temperature on the specification sheet.

The external interface is often used to attach an FPGA for a high channel count acquisition front end. That works, but the interface has wait states and no cache, so the achievable throughput is a fraction of the core clock. A product that needs streaming data should measure the bus throughput early rather than size the FPGA from the clock rate.

This part is now bought for repair and legacy builds as much as for new products. Supply comes from allocation and from the aftermarket, and the two are not the same thing: a lot with no traceable date code cannot be assumed to meet the same revision of the data sheet, which matters where the errata for an older silicon revision affect a peripheral.

Floating point changes the arithmetic but not the timing discipline. A current loop in floating point still has to finish inside the PWM period, and the compiler's conversion instructions and the library calls add cycles that a fixed point implementation does not have. Measuring the loop with a spare PWM output is the only way to know where the time goes.

The high resolution PWM extension calibrates itself against the system clock at start up, and the calibration assumes the clock is stable. A product that changes the clock after that point loses the fine resolution without any error flag, and the symptom is a limit cycle at high duty resolution that appears only after a mode change.

The SARAM is 34K words and it is the tighter limit once a real time operating system and a communication stack are added. Code runs from flash, and buffers and stacks live in RAM, so the map file at link time is the number that matters, not the flash size on the front page.

The two rail power up order in the data sheet is not a suggestion. A product that brings 3.3 V up first and releases reset before the 1.8 V rail is valid can start executing with a core that is not yet at voltage, and the failure looks like a dead device on the bench and a working one after a power cycle.

The device is supported by a mature tool chain and a large body of application notes, which is part of why it is still bought. A new programme that starts on it should treat the supply as an allocation question from the first prototype, and should qualify a second device number early rather than at the first shortage. For a programme that is still in production, the sensible sourcing plan is to hold a buffer of the exact order code and to qualify the neighbouring device number in parallel, because the family members share the peripheral set and a board change is smaller than a port to another core generation.

Where this order code fits

It is used in servo drives, solar and wind converters, UPS systems and instrumentation where a floating point control loop runs beside a field bus. Teams comparing it with a neighbouring part start from our MCU Cross Reference page, and teams replacing a part that has gone end of life start from the Component Obsolescence Management page.

Questions buyers ask

Does the F28335 have a floating point unit? Yes, single precision IEEE 754. The F28235 in the same peripheral family does not, and code written for one does not compile unchanged for the other at the same performance.

How much flash and RAM? 256K by 16 bit words of flash, which is 512 kilobytes, and 34K by 16 bit words of SARAM, 68 kilobytes.

Is the 176-pin body the only option? No. The family also lists BGA bodies and the temperature suffix changes with the body, so the order code has to be read as a whole.

What temperature does PGFA cover? Minus 40 to plus 85 degrees ambient. The S grade of the same device number covers minus 40 to plus 125 degrees.

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 Component Lead Time Solutions page sets out the checks that keep a substitute from failing on the bench.