INA128UA-2K5
Product Specifications
INA128UA-2K5
Texas Instruments — Power Management & Signal Chain
Product Overview
The INA128UA-2K5 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
INA128UA-2K5: the hyphenated spelling of the A grade reel code
This entry is the hyphenated spelling of the A grade reel code, INA128UA-2K5, and the part it names is the A grade SOIC body on a large reel. The catalogue carries all three spellings of this family's reel codes, and the reason to keep them on separate pages is search rather than engineering: a buyer who has a purchase history in one form should not have to translate it before reading a specification.
The specification is the family one with the tighter bin: 50 µV maximum offset on the family row and tighter on this grade, 0.5 µV per degree drift, 5 nA bias current, 8 nV per root hertz noise, 120 dB common mode rejection, 700 µA quiescent current, gain from 1 to 10,000 with G = 1 + 50 kΩ divided by RG, and a supply window from plus or minus 2.25 V to plus or minus 18 V.
The INA128UA-2K5 rows as the addendum publishes them
| Part class | Precision instrumentation amplifier, three op amp, current feedback |
| Vendor status | Active, production |
| Data sheet | SBOS051G, revised January 2026 |
| Vendor string | Printed as INA128UA/2K5 with a slash |
| Catalogue string | Also indexed with a hyphen and without a separator |
| Grade | A suffix, tighter offset and gain error window |
| Gain equation | G = 1 + 50 kΩ divided by RG |
| Input offset voltage | Family row 50 µV maximum, tighter on the A bin |
| Offset drift | 0.5 µV per degree maximum |
| Input bias current | 5 nA maximum |
| Bandwidth | 1.3 MHz at a gain of 1, 200 kHz at a gain of 100 |
| Input protection | Plus or minus 40 V without damage |
| Supply range | Plus or minus 2.25 V to plus or minus 18 V |
| Quiescent current | 700 µA typical |
| Body | SOIC, D outline, 8 pins |
| Carrier | Large reel, quantity per the addendum row |
| Moisture level | Published per row, level 2 and level 3 both appear in the addendum |
| Plain sibling | INA128U/2K5 and INA128U-2K5 |
Why the INA129 uses 49.4 kΩ and the INA128 uses 50 kΩ
The two parts in this document are the same silicon with one difference in the internal gain network: the INA128 uses 50 kΩ so that G = 1 + 50 kΩ divided by RG, and the INA129 uses 49.4 kΩ so that G = 1 + 49.4 kΩ divided by RG. The 49.4 kΩ is there so that a resistor value chosen for a competing part gives the same gain in the INA129, which is what makes it a drop-in replacement on an existing board without a resistor change. On a gain of 100 the difference between the two equations is 1.2 per cent, which is larger than the resistor tolerance and large enough to fail a calibrated channel.
Because of that, INA128UA-2K5 must be bought as the part the board was drawn for, and the substitution question is not symmetrical. An INA129 in an INA128 socket runs 1.2 per cent low at a gain of 100 and 0.12 per cent low at a gain of 10, so it passes a functional test and fails a calibration. An INA128 in an INA129 socket runs high by the same amount. Neither failure is visible without a gain measurement, which is why the two codes should never be treated as interchangeable.
Gain error rows, the A grade and the 0.024 per cent window
Gain error on this family is quoted per gain setting and per grade, and the tightest row is 0.024 per cent at a gain of 1 on the A grade parts. That figure is smaller than the tolerance of the resistor that sets the gain, so the resistor, not the amplifier, is what a calibration adjusts. At a gain of 100 the family rows widen and the A grade still holds a window about half as wide as the plain grade, which is the practical difference between the two bins on a channel that is calibrated once rather than per unit.
Nonlinearity is 0.001 per cent of full scale on the plain grade and 0.0005 per cent on the A grade, which is well below any converter this part is likely to drive. The number that limits a real channel is the gain drift, quoted in ppm per degree, because it cannot be trimmed at the bench and it moves with the cabinet temperature. On INA128UA-2K5 the drift row is the one to put in a specification that has to hold over temperature, and the room temperature gain error row is the one to put in a production test limit.
700 µA of quiescent current across a plus or minus 2.25 V to 18 V window
INA128UA-2K5 draws 700 µA typical, which is four times the current of the micropower INA126 and a twentieth of what a fully featured instrumentation amplifier usually costs. The supply window of plus or minus 2.25 V to plus or minus 18 V covers a plus or minus 15 V industrial rack and a plus or minus 2.5 V single cell analogue front end in the same part number, and the common mode range tracks the rails rather than sitting at a fixed offset from ground.
The current is flat with load and nearly flat with temperature, so the supply budget can be set from the data sheet row rather than measured. What is not in the row is the reference divider and the load current. A plus or minus 15 V part driving a 10 kΩ load at 10 V adds a milliamp, which is more than the amplifier, and a single supply channel with an unbuffered mid rail divider can add a quarter of a milliamp. The amplifier is therefore rarely the largest item in the budget, and the budget is still small enough that a battery instrument can afford two channels.
The D body at 4.9 by 6 mm against the P body at 9.8 by 9.4 mm
The family ships in two bodies, an eight pin SOIC measuring 4.9 mm by 6 mm and an eight pin plastic DIP measuring 9.81 mm by 9.43 mm, and the difference is a process decision rather than an electrical one. The D body runs on the reflow line with the rest of the board and is the default for a new build; the P body needs plated through holes, a separate insertion step and either wave or selective solder, and on a board with a handful of DIP parts that second process usually costs more than the part.
The reason to keep the P body is mechanical. A socketed DIP can be replaced in the field or swapped during calibration without a rework station, and the longer leads absorb a mismatch between the board and a connector. A probe amplifier that is trimmed on the bench is a genuine case for the socket. On INA128UA-2K5 the die and the pinout are the same in both bodies, so a board can move from one to the other without an electrical change, and the footprint is the only thing that has to be revisited.
Settling time against gain on a multiplexed front end
The settling time curve on this family is the one to read before a channel is put behind a multiplexer. At a gain of 100 the part settles to 0.01 per cent in a few microseconds, which allows a scan rate in the tens of kilohertz; at a gain of 1000 the same figure stretches out and the scan rate falls with it. The stage before the converter has to settle from the previous channel's value, not from zero, so the worst case is a full scale step between two adjacent channels, and that is the step the vendor curve uses.
The other half of the settling budget is the multiplexer and the filter between it and the amplifier. A 1 kΩ series resistor with a 100 nF capacitor has a 100 µs time constant, which dominates the amplifier completely and is the usual reason a scan runs slower than the data sheet suggests. On INA128UA-2K5 the settling figure is only useful when the input network in front of it is faster than the amplifier, and a channel that ignores that rule will show an inter-channel error that moves with the source impedance.
The REF pin, the 8 Ω rule and single supply references
The REF pin sets the output level and is part of the signal path in a way that is easy to forget. Any resistance in series with it appears as a common mode error at the input and as a gain error at the output at the same time, and the vendor states that 8 Ω is enough to degrade common mode rejection on a typical part. A divider made of two 10 kΩ resistors has a source impedance of 5 kΩ, which is three orders of magnitude above that, so an unbuffered divider is not a reference for INA128UA-2K5 at any gain.
On a bipolar supply the REF pin goes to ground, and ground has to be the same ground the converter uses or the difference appears as an offset. On a single supply the pin goes to a buffered mid rail, and the buffer has to be able to sink and source the current the REF pin draws, which is small but not zero. A shunt reference with a 1 µF capacitor is the cheap answer; an op amp buffer is the answer when the reference also has to drive other channels. Both are one more part than a divider and both are the reason a single supply channel measures what it claims.
Buying INA128UA-2K5
Three spellings and two grades on one family means the order line has to carry both the separator form and the grade letter, and the confirmation has to come back in the same form. A receiving check that compares the label against the order catches the grade error and the quantity error at the same time, and both are more expensive than a spelling error.
The rest of the purchase is the storage rule. A large reel is a production carrier, and a level 3 floor life means a clock that runs whenever the reel is out of its bag. Recording the open date on the reel label is the simplest control and it costs nothing. Our incoming inspection plan covers the checks at the door, and the traceability report page covers the record kept against each reel.
Questions that arrive with this part number
Whether the separator matters to the vendor is the first question; it does not, and the confirmation should still come back in the vendor form. The second is whether the A grade is worth its premium at a gain of 100, and it is on a small full scale channel because the input referred offset falls with gain while the resistor error does not.
The third is the reference pin. On a single supply the mid rail has to be buffered, and on a bipolar supply the pin goes to analogue ground. Either way the source impedance has to be far below the 8 Ω the vendor names, which rules out a bare divider and explains why so many channels measure worse than their data sheet.