HXN50154
Product Specifications
HXN50154
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HXN50154 from Hisiwell is a discrete power semiconductor delivering efficient switching and power conversion. Hisiwell is a recognized semiconductor supplier providing reliable components for global electronics manufacturing.
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
- Low Rds(on) for minimal conduction losses in power switching
- Fast switching speed with low gate charge for high-frequency operation
- Avalanche-rated for robust operation under inductive load conditions
- High junction temperature rating for demanding environments
- Logic-level gate drive compatible for simplified design
- RoHS and AEC-Q101 qualified options available
Key Specifications
| Vds / Vce | V |
| Id / Ic | A |
| Rds(on) / Vce(sat) | mΩ |
| Qg | nC |
| Technology | — |
| Package | — |
| Configuration | — |
| Mounting | — |
Applications
Request Quote — In Stock, 100% New & Original, Full Warranty
HXN50154: 800 W pulsed across 960 to 1200 MHz with the input match inside the package
The HXN50154 is an internally matched gallium nitride HEMT rated at 800 W across 960 to 1200 MHz for pulsed amplifier use, on a 50 V rail. The shelf row beside it prints 13 dB of gain, 60 per cent efficiency and a pulse mode. Internally matched here means the input side is matched inside the package; the output network still sits on the customer board and still has to be built for the band.
800 W is 59.03 dBm and the drive needed at 13 dB is 46.03 dBm, which is 40.1 W. That is a substantial driver, and it is the first number a system architect has to accept when a single ended 800 W stage is chosen. Into 50 ohm the load sees 200.0 V RMS and 282.8 V peak, which is 5.66 times the 50 V rail, so the output network, the connector and the circulator all have to be rated above 280 V of RF swing.
The band is 240 MHz wide, a ratio of 1.25 to 1, and the vendor states no guarantee outside it. A network that covers 960 to 1200 MHz is a reasonable ask for a single section match on a matched input part, and the fact that the vendor describes the part as matched for exactly this window is the reason the row publishes a gain figure at all.
The HXN50154 rows as the vendor publishes them
| Part class | Gallium nitride HEMT, internally matched input, pulsed amplifier use |
| Frequency range, product page | 960 to 1200 MHz |
| Frequency range, index row | 0.96 to 1.2 GHz |
| Band width, calculated | 240 MHz |
| Band ratio, calculated | 1.25 to 1 |
| Output power, both documents | 800 W |
| Output in dBm, calculated | 59.03 dBm |
| Output per megahertz of band, calculated | 3.33 W per MHz |
| Gain, index row | 13 dB |
| Drive needed at 13 dB, calculated | 46.03 dBm, 40.1 W |
| Efficiency, index row | 60 per cent |
| DC input at 60 per cent, calculated | 1333 W, 26.7 A on a 50 V rail |
| Heat to move at 60 per cent, calculated | 533 W during the pulse |
| Drain rail, index row | 50 V |
| Load voltage at 800 W into 50 ohm, calculated | 200.0 V RMS, 282.8 V peak |
| Peak swing against the rail, calculated | 5.66 times |
| Mode, index row | Pulse |
| Pulse width and duty on the page | Not published |
| Package and body | Not published |
| Thermal resistance, junction to case | Not published |
| Load impedance at the output | Not published, no load pull table |
| Applied application list | Wireless communication infrastructure, wideband amplifier, EMC testing, ISM |
| Orderable addendum and packing | None published |
An L-band avionics window, read as a procurement window
960 to 1200 MHz covers the upper part of the L band that carries secondary surveillance and several air traffic functions, and the part is described for pulse use, which fits that traffic. A programme in this window usually buys against a standard pulse condition, and the vendor publishes none here; the pulse width and the duty have to be agreed with the factory before an output figure means anything.
The second consequence of the window is that a network built for the centre at 1080 MHz will have to hold its match at both ends, and the ends are where the gain falls first. On a 1.25 to 1 band this is manageable with a single section and a tuning capacitor, but the network has to be tuned on the part it will ship with, not on a golden sample borrowed from another build.
The third is sourcing. A part on a band this specific is bought against a programme and not against a shelf, so the reel size, the packing and the lead time all become programme questions. Where the programme is an avionics build, the batch trace and the certificate of conformity are part of the specification, and they have to be requested with the first enquiry rather than at the delivery stage.
Why an internally matched input changes the driver
A die that delivers 800 W has an input impedance far below 50 ohm, and matching that on a board takes a network whose loss the driver has to make up. When the vendor puts the match inside the package, the board sees a load close to 50 ohm at the gate, and the driver can be a standard 50 ohm amplifier placed at a normal distance. That is the practical value of the matched input, and it is the reason the row can publish a gain of 13 dB.
The published gain is still a single number. There is no gain curve, no gain at the band edges and no statement of the test frequency. On a 240 MHz window the gain falls towards both ends, and the driver has to be sized for the lower end. A programme that budgets exactly 40.1 W of drive with no margin will be short at 960 MHz and at 1200 MHz.
The input match does not cover the bias. The gate has to be fed at a negative voltage through a network that does not load the RF path, and the drain has to be fed through a bias tee that presents a short at the band. Neither the gate voltage nor the quiescent current is published for this part, so the bias board has to come from a data sheet.
There is also a sequencing rule that has to be written into the bias board. A depletion mode GaN HEMT conducts with the gate at zero volts, so the drain must not be applied before the gate is held at its negative bias, and on power down the gate has to stay negative until the drain has collapsed. A bias board that sequences the two rails in the wrong order destroys the part on the first switch on, which is why the sequence belongs in the drawing and not in a habit.
Where the HXN50154 sits against the 500 W and 1500 W rows
On the same shelf the 500 W row covers 960 to 1400 MHz with a 440 MHz window, and the 1500 W pair covers 960 to 1225 MHz. The 800 W part sits between them in power and is the narrowest of the three in band. That combination, more power than 500 W on a narrower window, is the usual way a vendor trades bandwidth for watts on one die family.
Where the programme needs 400 MHz of window, the 500 W row is the honest choice and the extra decibels have to come from a second stage. Where the programme sits inside 960 to 1200 MHz and needs 800 W, this part is the shortest path, and where 1500 W is needed the pair is the only row that reaches it. Each step up in power also steps up the drive, so the driver chain grows with the output and not with the number of parts.
The three rows are compared side by side on the RF power transistor cross reference page, together with the 350 W part that reaches 400 W inside a 60 MHz window.
Questions buyers ask about the HXN50154
Is the output matched too? No. The page describes the part as internally matched, which on this shelf refers to the input. The output network is built on the customer board and no load impedance table is published, so the output match comes from a measurement or from a data sheet request.
What pulse condition applies? The row says pulse and the page says pulsed amplifier use, and neither states a width, a duty or a peak figure. A part rated at 800 W with no pulse condition is only usable once the factory confirms one, because the thermal answer depends on it entirely.
Can it be run continuous? Nothing published allows it. The mode is pulse on both documents, and a continuous carrier at 800 W would dissipate 533 W in the die, which is a different thermal problem from the one a pulse part is built for.
How is it bought and checked? Through a channel that can prove the batch, with an incoming check at 960 to 1200 MHz and a pulse test at the agreed condition. The RF component sourcing page sets out the trace and the checks, and the one stop RF supply page covers how the driver and the passives are bought against the same schedule.