LMK00804BPWR
Product Specifications
LMK00804BPWR
Texas Instruments — Power Management & Signal Chain
Product Overview
The LMK00804BPWR 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
LMK00804BPWR: One Clock In, Four Single Ended Clocks Out, 35 ps of Skew
LMK00804BPWR is a one to four LVCMOS clock fanout buffer in a TSSOP (PW) body of 16 pins. It copies one input clock to four LVCMOS or LVTTL outputs at up to 350 MHz with 35 ps of output skew and 0.04 ps RMS of additive jitter at 125 MHz, and it takes either a differential input pair or a single ended clock. The clock enable is synchronised internally, so starting and stopping the outputs does not produce a narrow pulse.
Order code and package
| Ordering string | LMK00804BPWR |
| Family | LMK00804B fanout buffer, document SNAS642A |
| Package | TSSOP, PW body, 16 pins, 5.00 mm by 4.40 mm nominal |
| Reel and packing | 2500 pieces on a large reel, tin lead finish |
| Tube row | LMK00804BPW at 92 pieces in a tube, for prototypes and repair |
| Moisture level | Level 1, 260 C peak, unlimited floor life |
| Temperature range | Minus 40 to plus 85 degrees, industrial grade |
| Core supply | 3.3 V plus or minus 5 percent |
| Output supply combinations | 3.3 V, 2.5 V, 1.8 V or 1.5 V, paired with the 3.3 V core |
| Output impedance | 7 ohm nominal per output, which sets the series resistor choice |
Electrical parameters that decide the application
| Output count | Four LVCMOS or LVTTL outputs |
| Maximum output frequency | 350 MHz |
| Output skew | 35 ps maximum between outputs of the same device |
| Part to part skew | 700 ps maximum, the figure that decides whether two packages can be paralleled |
| Additive jitter | 0.04 ps RMS typical at 125 MHz |
| Phase noise floor | Minus 166 dBc per hertz typical at 125 MHz |
| Input selection | Two inputs, a CLK and nCLK pair or a dedicated LVCMOS_CLK pin |
| Differential input formats | LVPECL, LVDS, HCSL, SSTL, LVHSTL and LVCMOS or LVTTL |
| Clock enable | Synchronised internally, outputs held low while disabled |
| Output enable | A separate pin that puts the outputs into high impedance |
| Recommended input slew rate | 3 V per nanosecond or faster for the published jitter |
| Slow input behaviour | Below 0.05 V per nanosecond the output can chatter from switching noise feeding back |
Family variants and how they differ
| LMK00804BPW | The 92 piece tube row of the same body |
| CDCLVC1310 | A ten output LVCMOS buffer from the same vendor when four outputs are not enough |
| LMK00725 | A differential input fanout buffer when the load wants LVPECL or LVDS rather than LVCMOS |
| LMK00308 | An eight output differential buffer for a clock tree that has to change format per bank |
| LMK1C1104 | A four output LVCMOS buffer for a lower cost clock tree |
| LMK6C series | Oscillators that can feed this buffer directly when the reference is not on the board |
Points to check before this part goes on a BOM
The input slew rate is a requirement, not a recommendation. The document states that 3 V per nanosecond or faster is needed to reach the quoted additive jitter, and that a sine wave input slower than 0.05 V per nanosecond can make the output chatter because switching noise feeds back into the input stage. A 32.768 kHz tuning fork crystal fed straight in is exactly that case and needs a comparator in front.
The 7 ohm output impedance is part of the termination. Driving a long trace with no series resistor produces ringing at the far end, and the usual fix is a resistor of 22 to 33 ohm close to the pin, chosen so that the sum lands near the trace impedance. The impedance is quoted at a nominal value, so the resistor value has to be checked against the actual load and the slew rate that the receiver accepts.
Two enables do two different jobs. The clock enable is synchronised and leaves the outputs driven low, while the output enable leaves them high impedance. A bus that is shared between two clock sources needs the high impedance state, and a receiver that must never see a floating input needs the driven low state. Wiring both pins to the same signal gives the wrong one of the two behaviours.
The core supply stays at 3.3 V whatever the output supply does. Dropping the whole part to 1.8 V to save power is outside the recommended conditions, and the four permitted combinations all pair a 3.3 V core with a lower output rail.
350 MHz is the ceiling for a clean LVCMOS edge, and the practical limit on a board is lower. Above roughly 200 MHz the trace has to be treated as a transmission line, and the 35 ps skew figure only holds when all four outputs drive similar loads.
The two inputs are not interchangeable in use. The CLK and nCLK pair needs a bias network when the source is AC coupled, while the LVCMOS_CLK pin is single ended and referenced to the local ground. A board that wires an LVDS oscillator to the single ended pin loses the common mode rejection that the differential route provides, and the added jitter rises accordingly.
Where four single ended clocks go in a converter front end
An analog front end with four converters, or with two converters and two digital blocks, needs four copies of one clock and cannot afford four separate oscillators. This buffer takes one reference and drives all four loads from a single die, so the four clocks stay within 35 ps of each other and the sampling instants of the converters line up without a per channel calibration.
The differential input is what makes that arrangement quiet. A single ended clock routed across a converter board picks up the switching noise of the supplies, while a differential pair from the oscillator cancels most of it before the buffer. The part accepts both, so the same board can start with a single ended oscillator and move to a differential one without a board change.
What the enable pins are for in a real product
Test equipment and medical front ends often need the clock to stop while a calibration or a self test runs, and they need it to stop without a runt pulse that a converter would treat as a valid sample clock. The synchronised clock enable does exactly that, because the internal logic waits for a clean edge boundary before it stops driving.
A shared bus is the other case. Two buffers driving one clock line need a high impedance state, and that is the output enable rather than the clock enable. The distinction is easy to miss because both pins are active low on this part, and a board that ties them together gets a buffer that never releases the line.
Where this order code fits
The part is used where one reference clock has to reach several single ended loads on the same board and the sampling instants have to stay aligned. Teams comparing it with a neighbouring part start from our ADC DAC Cross Reference page, and teams replacing a part that has gone end of life start from the Component Lead Time Solutions page.
Questions buyers ask
Can the four outputs drive different loads? Yes, and the skew figure assumes similar loads. An output driving a short trace into a high impedance receiver behaves differently from one driving a 50 ohm line, and the difference shows up as added skew between the two. Loads that differ a lot should be moved to separate banks or compensated at the receiver.
Does the part work with a 1.8 V output and a 3.3 V core? Yes, that is one of the four published combinations, along with 3.3 V, 2.5 V and 1.5 V outputs. The core is always 3.3 V, and the output rail sets the swing that the receiver sees.
What happens if the input clock disappears? The outputs stop at whatever level the internal stage holds and the part does not report a loss of clock. Systems that must detect a missing reference need a separate supervisor or a clock monitor, because this buffer has no loss of signal output.
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 BOM Shortage Recovery page sets out the checks that keep a substitute from failing on the bench.