UCC27714DR
Product Specifications
UCC27714DR
Texas Instruments — Power Management & Signal Chain
Product Overview
The UCC27714DR 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
UCC27714DR: a 600 V half bridge gate driver with independent high side and low side inputs, 4 A peak output in both directions and a 125 ns maximum propagation delay, in a fourteen pin SOIC on a 2500 piece reel
UCC27714DR is a high side and low side driver with a floating channel that runs to 600 V on the switch node, a bootstrap supply, symmetric lockouts on both channels and 4 A of peak current to charge a large gate quickly in a SOIC body of 14 pins. The D body is a fourteen pin SOIC with 8 functional pins spread across the two rows for creepage between the high side and low side, and the part exists because a 600 V bridge needs a driver whose switch node can swing the full bus without a level shift stage that misbehaves.
Order code and package
| Catalogue string | UCC27714DR, where D is the fourteen pin SOIC and R is the 2500 piece large reel |
| Body | SOIC, D, fourteen pins, with COM and VSS separated so the low side return can carry a shunt without moving the logic reference |
| Reel | 2500 pieces on a large tape and reel, with a 50 piece tube as the prototype option |
| Finish and moisture | NIPDAU plating at level 2, 260 C, one year of floor life, on the reel row |
| Temperature | the addendum prints minus 40 to 150 C on the reel rows and 125 C on the tube row, so the grade on the order matters |
| Marking | UCC27714 on the body |
| Document | SLUSBY6B, August 2015, revised March 2017 |
| Status | Active and in production on the addendum dated 7 November 2025 |
Electrical parameters that decide the application
| Switch node rating | 600 V on the HS pin in operation, with 640 V of absolute maximum on the bootstrap supply and 20 V on the bootstrap to switch node differential |
| Output current | 4 A peak source and 4 A peak sink at 15 V of supply, measured with a 10 us pulse, which is enough for a 200 nC gate in about 50 ns |
| Bias range | 10 V to 20 V of recommended supply, with 20 V of absolute maximum and a revision that lowered the recommended ceiling from 18 V to 17 V, so a 15 V rail is the comfortable point |
| VDD lockout | rising at 8.4 V minimum, 9.1 V typical and 9.8 V maximum, falling at 7.9 V, 8.6 V and 9.3 V, with 0.4 to 0.5 V of hysteresis |
| Bootstrap lockout | rising at 7.7 V, 8.3 V and 9.0 V, falling at 6.7 V, 7.25 V and 8.05 V, with 0.5 to 1.0 V of hysteresis, and it only disables the high side |
| Quiescent current | 750 uA typical and 1050 uA maximum total on the supply, with 120 to 300 uA on the bootstrap, which sets the droop on a slow switching cycle |
| Bootstrap leakage | 20 uA maximum with the switch node at 600 V, which is the figure that decides how long a high side can stay on in a static state |
| Input thresholds | 1.7 to 2.7 V for a high and 1.2 to 2.1 V for a low, with 0.7 V of hysteresis, so a 3.3 V logic signal drives it directly and a slow edge does not chatter |
| Propagation delay | 90 ns minimum and 125 ns maximum in both directions, with 20 ns of matching between the two channels |
| Minimum pulse | 40 ns nominal and 100 ns maximum for a pulse that changes the output state, which is the number to check against a narrow pulse width modulated signal |
| Output edges | 20 ns rise and fall into 1 nF, with about 1.45 Ohm of pull up and a few Ohm of pull down, so the peak current is limited by the die rather than by a resistor |
| Thermal | 72.3 C per W junction to ambient in the fourteen pin body, with 31.8 C per W junction to board |
Family variants and how they differ
| UCC27714, the part on this page | 600 V, 4 A, 125 ns maximum, both channels independently controlled, for a half bridge or a full bridge on a rectified mains bus |
| UCC27712 | the same family with a lower current rating and a smaller body, for a bridge that drives a smaller gate |
| UCC27211 and UCC27282 | 120 V parts with 3 to 4.5 A of drive, for a low voltage bridge where 600 V of capability would be wasted |
| Against a gate drive transformer | no duty cycle limit and no volt second reset problem, at the cost of a floating supply that has to be built from a bootstrap diode and capacitor |
| Against an isolated driver | no barrier, so the high side and the controller share a ground reference and the switch node swing is limited to 600 V |
| Against a single channel driver plus a discrete level shift | one body, a defined 20 ns of matching and a lockout that behaves the same way on both channels |
Points to check before this part goes on a BOM
The 600 V switch node rating is what makes this a mains driver rather than a low voltage one, and it comes with a limit that is easy to miss. The bootstrap supply can sit at 640 V absolute maximum with respect to the switch node, but the differential between the bootstrap pin and the switch node is only 20 V, so the bootstrap capacitor and the diode have to be rated for the full bus on one side and for 20 V on the other. A diode rated at 20 V of reverse voltage is the correct choice and a 600 V diode is a waste of forward drop that will heat the bootstrap resistor.
The two lockouts behave differently and that difference is deliberate. A VDD lockout disables both outputs, because the logic and the low side driver have lost their supply. A bootstrap lockout disables only the high side, so the low side keeps switching and recharges the bootstrap capacitor through its own diode. That behaviour is what lets a bridge start from a cold bus: the first few cycles run the low side only, the bootstrap capacitor charges, and the high side joins when the differential passes 8.3 V. A board that ties the two lockouts together, which some drivers do, cannot start without a separate charge pump.
Matching of 20 ns between the two channels is the number that sets the dead time for a bridge, and it is a maximum rather than a typical. At 20 kHz a 20 ns mismatch is 0.04 per cent of a period and can be covered by 100 ns of dead time. At 200 kHz the same mismatch is 0.4 per cent of a period, and the dead time has to be set from the switch's turn off time plus the mismatch plus the driver's own delay spread, which at 125 ns maximum is a bigger contribution than the matching. The practical rule is to take the dead time from the maximum delay, not from the typical one.
The 20 uA of bootstrap leakage at 600 V is what limits a static high side. In a bridge that holds the high side on for a long period, the bootstrap capacitor supplies the quiescent current of 120 to 300 uA plus the leakage, and a 100 nF capacitor at 270 uA droops by 2.7 V per millisecond. That is why a high side that has to stay on for tens of milliseconds needs a charge pump rather than a bootstrap, and why the datasheet's note on 100 per cent duty is conditional on the bootstrap staying above its lockout.
The input thresholds are TTL compatible with 0.7 V of hysteresis, and the pull down resistors on the inputs are 400 kOhm. A 400 kOhm pull down holds an unconnected input low, which is the safe state, but it is not a strong enough pull down to discharge a long trace quickly in a noisy environment. A controller output that goes high impedance during reset will let the input float for the length of the trace capacitance times 400 kOhm, which is microseconds, and in a bridge that is enough time for one bad pulse. A 10 kOhm resistor to ground at the driver pin is the usual fix.
The thermal figure of 72.3 C per W is measured on a board with copper, and the dissipation is dominated by the gate charge rather than by the quiescent current. Driving a 200 nC gate at 20 kHz at 15 V is 60 mW per channel, so 120 mW total, which is 9 C of rise. At 100 kHz with a 500 nC gate the same arithmetic gives 750 mW and 54 C of rise, and at that point the ambient inside an enclosure decides whether the part runs at 100 C or at 140 C. The fourteen pin body has no exposed pad, so the heat leaves through the leads and the copper around them.
Setting up the bootstrap for a real bridge
The bootstrap capacitor has to hold the gate charge of the high side switch plus the quiescent current plus the leakage for the length of the on time, and the diode has to recharge it in the low side on time. A 100 nF capacitor and a 200 nC gate leaves plenty of margin for the gate charge and none for a long on time, which is why the value is usually chosen from the on time rather than from the gate. At 270 uA of quiescent current and a 20 us on time, 100 nF droops 54 mV, which is nothing; at 20 ms the same capacitor droops 54 V and the high side is in lockout before the end of the pulse.
The bootstrap resistor is the other half of the trade. A resistor in series with the diode limits the recharge current and protects the diode, and it also delays the recharge at the start of the low side on time. A value that is too large leaves the bootstrap short at high duty, which shows up as a high side that drops out at high load and works at low load. The measurement to make is the bootstrap voltage on a scope across the full duty range rather than at one operating point.
What a 600 V driver does not protect against
The driver tolerates 600 V on the switch node and it does not tolerate a negative excursion beyond its ratings, and a bridge with a fast diode and a long loop will produce a negative spike at every commutation. The switch node can go below the low side return by the diode's forward recovery plus the loop inductance times the current slope, and the negative voltage handling is a few volts beyond the rail rather than tens of volts. A snubber or a shorter loop is the fix, and the driver's rating is the thing that tells you whether the spike is a problem.
The other event that the driver does not survive is a short circuit on the output of the bridge. The driver will deliver 4 A into the gate for as long as the controller asks, and the switch will limit the current, so the failure appears in the driver as a thermal event rather than as an electrical one. The protection has to be in the controller and in the switch, and the driver's contribution is the 125 ns maximum delay that sets how quickly the controller can react.
Where this order code fits
This is the driver for a half bridge or a full bridge on a rectified mains bus, in an offline power supply, a solar inverter, a motor drive, an induction heater or an uninterruptible supply 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 Pin to Pin Cross Reference page.
Questions buyers ask
Can the high side stay on continuously? Only while the bootstrap stays above its lockout. The quiescent current of 120 to 300 uA plus 20 uA of leakage drains the bootstrap capacitor, so a long on time needs a charge pump.
What dead time should the bridge use? Set it from the 125 ns maximum delay plus the switch turn off time plus the 20 ns of matching. A dead time taken from the typical delay will be short at the corner.
Does it need a negative bias on the low side? No. The low side is referenced to COM, and COM can be separated from the logic ground by a shunt resistor, which is why the two pins are separate.
Is 15 V enough for the supply? Yes. The recommended range is 10 V to 20 V and the lockout releases at 9.1 V typical. 15 V is the usual point for an IGBT and 10 to 12 V for a logic level FET.
What happens if the bootstrap capacitor is too small? The high side drops into lockout at high duty and the bridge loses one switch, which shows up as a loss of output at high load and normal operation at low load.
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.