HXN50126
Product Specifications
HXN50126
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HXN50126 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
HXN50126: 240 W continuous from DC to 2 GHz on a 50 V rail
The HXN50126 is a single ended gallium nitride HEMT from Hisiwell, filed on the 50 V unmatched shelf with a DC to 2 GHz window and a 240 W saturation row. The vendor calls this one internally prematched, which matters at 2 GHz: the input match sits inside the package, so the network on the customer board starts at the output port and at the bias feed rather than at both ports.
240 W is 53.80 dBm by calculation. On the shelf around it that puts the part in the middle of the high power group: 0.67 dB below the HXN50127 on the same DC to 2 GHz window, 2.66 dB above the 130 W HXN50122 beside it, and 1.40 dB above the 100 W HXN50121 on the 3 GHz window. The spacing is worth reading before a part is chosen, because 0.67 dB is inside the tolerance of most production test benches and outside the tolerance of a link budget.
What the index row does not carry is gain and efficiency. Both columns are blank, which is normal across this family. Blank is not the same as unavailable: the product page links a data sheet that carries the sweep. What it does mean is that a quotation issued off the index row alone cannot promise a drive level, a DC current or a heat load, and none of the three should appear in a proposal until the sheet has been read.
The HXN50126 rows as the vendor publishes them
| Part class | Gallium nitride HEMT power transistor, single ended, internally prematched |
| Frequency range | DC to 2 GHz |
| Band width, calculated | 2,000 MHz |
| Saturation output, index row | 240 W |
| Saturation in dBm, calculated | 53.80 dBm |
| Output at 2 GHz, calculated | 120 W per gigahertz of band |
| Drain rail, index row | 50 V |
| Mode, index row | Continuous wave |
| Gain | Not published on the index row, carried on the data sheet |
| Drain efficiency | Not published on the index row, carried on the data sheet |
| Peak output | Not published, the row is a saturation figure only |
| Average output | Not published |
| Package and body | Not published on the index row |
| Thermal resistance, junction to case | Not published |
| Gate bias point and quiescent current | Not published |
| Load impedance at the output | Not published, no load pull table on the page |
| Applied application list | Wireless communication infrastructure, wideband amplifier, EMC testing, industrial scientific and medical |
| Datasheet link on the product page | Present, target address not resolvable from the public page |
| Orderable addendum | None published |
| Packing quantity and moisture level | None published |
The row set is short for a 240 W part and that is the first thing to notice. Twenty published rows on a device that dissipates a hundred watts or more is thin, and the three figures that decide whether a board closes are all in the not published set: gain, efficiency and load impedance. A buyer who needs those numbers gets them from the data sheet or from a measured report on the exact part, not from the index.
The application list is the vendor's, not a claim about this site. Wireless infrastructure, wideband amplifier, EMC testing and industrial scientific and medical cover a broad range, and a 240 W continuous part at DC to 2 GHz fits all four for the obvious reason: the band starts at DC, so the same device serves an HF exciter chain, a 900 MHz cellular stage, a 1.8 GHz lab amplifier and a 2 GHz driver without a second part number.
Fifty volts changes the current, not the heat
| Drain efficiency assumed | DC input, calculated | Drain current at 50 V, calculated | Heat at 240 W output, calculated |
| 50 per cent | 480 W | 9.60 A | 240 W |
| 55 per cent | 436 W | 8.73 A | 196 W |
| 60 per cent | 400 W | 8.00 A | 160 W |
| 65 per cent | 369 W | 7.38 A | 129 W |
| 70 per cent | 343 W | 6.86 A | 103 W |
This table exists because the efficiency column is blank. It is not a prediction, it is the envelope the rail and the output imply. The spread from 50 to 70 per cent moves the DC input by 137 W and the current by 2.7 A, so the power supply has to be sized on the low end unless the vendor sheet says otherwise, and the heat sink has to be sized on the low end for the same reason.
What the 50 V rail changes is the current, not the dissipation. The same 240 W at 60 per cent draws 8.00 A here and would draw 14.29 A on a 28 V rail. Lower current means a thinner drain trace, a smaller choke and less copper loss in the bias network, and that is the whole case for the 50 V shelf on a continuous part. The heat load is identical in both cases, so the thermal stack does not get easier with the higher rail.
Ten amps on a continuous part is a connection question before it is a copper question. A screw terminal or a heavy bus on the drain, a choke rated well above the saturation current, and a bulk bank sized for the modulation envelope rather than the average. The vendor publishes no pulse or burst rating on this row, so the supply has to hold the continuous current without droop, and any envelope behaviour is on the system to manage.
At the output, 240 W into 50 ohm is 109.5 V RMS and 154.9 V peak by calculation. That is above the 100 V capacitor family and inside the 200 V family, and a two to one mismatch takes the peak to 219 V. Output capacitors, the coupler and any relay in the path have to be chosen against the mismatched figure and not the matched one.
Where the HXN50126 sits on the 50 V shelf
| HXN50122 | DC to 2 GHz | 130 W | Gain and efficiency not published | Unmatched, not prematched |
| HXN50126 | DC to 2 GHz | 240 W | Gain and efficiency not published | Internally prematched |
| HXN50127 | DC to 2 GHz | 280 W | Gain and efficiency not published | Internally prematched |
| HX25400F | DC to 2.5 GHz | 400 W index, 240 W maximum output on the parameter list | 19.5 dB linear gain, 82 per cent best drain efficiency | 28 or 48 V rail, flanged package |
| HX25600F | DC to 2.5 GHz | 600 W index, 320 W maximum output on the parameter list | 19 dB linear gain, 79 per cent best drain efficiency | 28 or 48 V rail, 650F2 body |
| HXN50121 | DC to 3 GHz | 100 W | Gain and efficiency not published | Continuous wave and pulsed continuous wave |
| HXN50141 | DC to 3 GHz | 280 W | Gain and efficiency not published | Single ended, thermally optimised for continuous wave |
Two groups live on this shelf and they are not interchangeable. The HXN members publish a power row and little else. The HX members publish a gain, an efficiency and a package, and the two HX parts on the 2.5 GHz window carry a 1.67 and a 1.88 factor between the index power and the maximum output on the parameter list, which is the single biggest thing to settle before a quotation goes out on either.
Against the HX25400F the HXN50126 gives up 2.22 dB of index power and gains nothing measurable in data. Against the HXN50127 it gives up 0.67 dB and nothing else. The honest reading is that the 240 W row is the volume point of the family and the 280 W row is the graded bin, and a system that needs 53.8 dBm and no more should not pay for the bin above it.
The wider comparison across vendors sits on the RF power transistor cross reference, which is the page to use when two shelves have to be lined up on one axis. When the question is how a measured sweep is obtained before a purchase order, RF component sourcing covers the route and the documents that come back with it.
What an order has to state, and what it cannot yet
The vendor publishes no orderable addendum, no packing column, no marking code and no moisture sensitivity level for this part. Nothing of that kind is quoted on this page, and no order code is offered, because an invented suffix on a 240 W transistor is a returns problem waiting for a purchase order. What the order has to carry instead is the band, the 50 V rail, the continuous wave condition, the output point that matters to the system, the drive level expected at the input and the quantity.
Two documents should be requested with the first order and both are on the buyer to ask for: the data sheet for the exact part, and a measured report at the frequency of use. The data sheet is linked from the vendor product page, but the link target is not resolvable from the public page as it stands, so the copy that arrives has to be checked against the part number on the label rather than trusted from the browse path.
Lead time is not published for this family either. On a 50 V shelf part the two variables that move it are the package and the test, and neither is visible from the index row, so the honest answer before a factory reply is that the schedule depends on a question that has not been asked yet. That is the state of the record and it is written here rather than filled with a number.
For a programme that needs a second source, the useful move is a cross to a part with published curves at the same band and rail, not a cross to a part with the same watts. One stop RF supply sets out how the substitute is qualified and what has to be re measured on the board before it is released.
Questions buyers ask about the HXN50126
Is 240 W the continuous rating or the peak? It is the saturation row, and the vendor publishes no peak and no average column for this part. On a continuous wave carrier the drain sees the saturation point; on a modulated carrier the average power has to come down until the junction temperature and the load line are both inside the sheet, and the sheet is where that number has to be read.
Can it run at 48 V instead of 50 V? Nothing published says no, and nothing published confirms it either. The index row states 50 V and the part family beside it carries dual 28 and 48 V ratings on some members. A rail change is a factory question on this part, not a catalogue question, and it should be asked before the supply is specified.
Is the input match inside the package? The vendor calls this part internally prematched, which is a statement about the input side. The output still needs a network on the board, and the vendor publishes no load impedance table for this row, so the output match has to come from a measured set or from the data sheet rather than from an estimate.
How much drive does it need? With no gain published on the row, no figure is quoted here. On GaN parts of this class a single driver stage is normally enough and two stages are normally unnecessary, but the number that decides it is on the sheet and the sheet has to be read first.
For a part in this band where the question is which of two similar shelves to qualify, the comparison is on the cross reference page and the next step is a short list with bands, rails and powers on one line each.