HXN40007
Product Specifications
HXN40007
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HXN40007 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
HXN40007: a 1.2 to 1.4 GHz GaN transistor at 40 dBm from a 32 V rail
The HXN40007 is a gallium nitride power transistor with an internal matching network, published by Hisiwell in Chengdu, and the parameter list is short enough to quote in full: 1.2 to 1.4 GHz, saturation power of at least 40 dBm, gain of at least 14 dB, efficiency of at least 50 per cent, 50 ohm input and output impedance, 32 V drain, plus or minus 0.8 dB of flatness across the band, continuous wave operation. Those seven rows are the entire published characterisation, and the rest of this page is about what they do and do not tell a buyer.
The matching is the reason a part like this exists. The vendor describes an in-plane matching synthesis and a film hybrid integration process, which means the input and output matching networks are built on the same carrier as the die and brought to 50 ohm ports. A 20 W class GaN die on its own has an input impedance in the low single digits and an output impedance of a few ohms, and a board level match at 1.3 GHz on a soft substrate is a multi-section network that consumes more area than the transistor. Buying the match inside the package moves that problem to the vendor and leaves a 50 ohm port on both sides.
Forty dBm is ten watts, and 14 dB of gain means the stage needs 26 dBm, or 0.4 W, of drive to reach it. At 50 per cent efficiency the DC input at saturation is 20 W, so the drain current is 0.62 A at 32 V and the device has to shed 10 W of heat through its base. Those four numbers, 10 W out, 0.4 W in, 20 W DC and 10 W of heat, are the whole thermal and electrical case for the part, and they follow from the published rows by arithmetic rather than from any graph.
What the band covers is worth naming. From 1.2 to 1.4 GHz sits the aeronautical distance measuring equipment band up to 1215 MHz, the GNSS L2 signal at 1227.6 MHz, the 23 centimetre amateur allocation from 1240 to 1300 MHz, and the 1.3 GHz links used for unmanned aircraft video. A part with plus or minus 0.8 dB of flatness across that span is a general purpose L band building block rather than a single frequency amplifier, which is what makes it useful in a low volume, high mix environment.
The parameter rows the vendor publishes
| Part class | Gallium nitride power transistor with internal matching |
| Frequency range | 1.2 to 1.4 GHz |
| Saturation power as published | at least 40 dBm |
| Saturation power in watts, calculated | 10 W |
| Gain as published | at least 14 dB |
| Efficiency as published | at least 50 per cent |
| Output at 1 dB compression | not published |
| Input impedance | 50 ohm |
| Output impedance | 50 ohm |
| Drain voltage | 32 V |
| Flatness across the band | plus or minus 0.8 dB |
| Operating mode | continuous wave |
| DC input at saturation, calculated | 20 W |
| Dissipation at saturation, calculated | 10 W |
| Drain current at saturation, calculated | 0.62 A |
| Drive power for rated output, calculated | 26 dBm, 0.4 W |
| Matching technology | in-plane matching synthesis on a film hybrid carrier |
| Stated application | RF and microwave systems |
| Package | not published on the parameter page |
| Thermal resistance | not published |
| Small signal model | not published |
| Scattering parameters | not published |
| Storage and operating temperature | not published |
| Moisture level and floor life | not published |
| Orderable addendum | not published |
| Related higher power member | HXN40008 at 48 dBm in the same band and the same 32 V rail |
| Related higher voltage member | HXN40011 at 56 dBm from 50 V in the same band |
| Related lower band member | HXN40001 at 40 dBm from 32 V over 0.9 to 1.2 GHz |
| Related wider band member | HXN40006 at 56 dBm over 0.9 to 1.4 GHz in pulsed operation |
What 10 W of dissipation asks of the carrier
Continuous wave at 40 dBm means 10 W of heat for as long as the transmitter is keyed, and the vendor page publishes no thermal resistance at all. That absence is the single most important gap in the document. Without a junction to case figure the carrier can only be sized by measurement, and the usual route is a copper tungsten or copper molybdenum carrier under the device, a thermal interface to a heatsink, and a junction temperature measured on the bench rather than computed on paper.
The arithmetic that can be done is the failure threshold. A GaN channel at 225 C junction is at its limit, and 10 W through a carrier and interface that add up to 8 C per watt gives 80 C of rise, which sits at 105 C in a 25 C room and 165 C in a 60 C cabinet. Eight C per watt is a realistic figure for a small carrier and a thin interface, so a continuous wave build at this power level is a heatsink question before it is an electrical one.
Efficiency helps less than it appears to. At exactly 50 per cent, half the DC input becomes heat, so a 3 dB improvement in efficiency would halve the dissipation at the same output; the vendor prints 50 per cent as a floor with no typical, which means the delivered part may be better and cannot be assumed to be. Buyers who plan a thermal solution on a hoped for 60 per cent should ask for the measured efficiency at their frequency before committing the carrier.
Where the datasheet stops and the bench starts
The published list has no output power at 1 dB compression, and that is the row a communication system needs. Saturation power is a saturated, compressed figure: it says nothing about the linearity a modulated signal will see, and a part quoted only at saturation can be several dB worse at the compression point a multicarrier signal requires. Continuous wave operation plus a saturation only rating is a fair description of a part built for a fixed frequency, saturated transmitter, and a poor description of one intended for a linear application.
There are also no scattering parameters. At 1.2 to 1.4 GHz the internal match brings the ports to 50 ohm, which makes the part easy to cascade, but the return loss across the band and the phase behaviour at the band edges are exactly what a multi stage chain needs and exactly what the seven row list does not give. A buyer planning a two stage amplifier has to ask for the measured small signal data at their two band edges, or measure it.
The third gap is temperature. No operating range and no derating curve is published, so the only honest statement about a build in a 70 C cabinet is that the vendor has not said. For a part that sheds 10 W in continuous wave, that is a question to put in writing before an order rather than after a first article.
The fourth is package and handling. The parameter page names no body, no moisture level and no floor life, so a production line cannot set a reflow profile or a bake rule from it. Ask for the outline drawing and the moisture classification with the quotation, since both are normally available and neither is on the page.
Where a 10 W internally matched part still fits
The clear fit is the output stage of a short range L band transmitter: a 1.3 GHz video or telemetry link, a 1.2 to 1.3 GHz amateur or industrial link, or a driver for a larger pulsed stage. Ten watts at 14 dB of gain from a 32 V rail is a stage that can be built with one transistor, one bias network and a 50 ohm line on a PTFE or ceramic filled board, with no external match at all.
It also fits a driver position ahead of a 48 or 56 dBm device in the same family. Driving an HXN40008 at 48 dBm needs 35 dBm, which is 3.2 W, and this part at 40 dBm has 5 dB of headroom for the splitter loss and the cable. A two stage chain built that way uses the same 32 V rail for both devices, which is worth more in a compact enclosure than the last dB of gain.
It is a poor fit for a linear application with a high peak to average ratio, for the compression reason above, and a poor fit for a pulsed radar front end where the pulse droop over a long pulse is not characterised. For radar the family has members with published pulse widths and duty cycles, 1 ms at 10 per cent on the 53 and 56 dBm parts, and the continuous wave part is the wrong row of the same table.
Our RF power transistor cross reference work holds the band, the saturation power, the gain and the drain voltage per approved alternate, because two L band GaN parts at the same output power can differ by 3 dB of gain and by 10 V of drain rail, and the rail decides the whole bias supply. The RF component sourcing page describes how we handle the band specific documentation that a part like this needs.
Questions we answer before an order is placed
The first is the drive level the system can supply. Twenty six dBm is not a small signal, and a driver stage that delivers 0.4 W at 1.3 GHz is a second amplifier problem behind the first. Ask what the preceding stage delivers at the top of the band, because 14 dB of gain is a floor and the loss of a filter or a coupler between the two stages is subtracted directly from the output.
The second is the thermal stack. Ten watts in continuous wave with no published junction to case figure means the carrier has to be chosen from experience and then verified. We ask for the carrier material, the interface and the heatsink, and we do not quote a continuous wave build as a drop in for a pulsed one, since the same part in 1 ms pulses at 10 per cent duty sheds a tenth of the heat.
The third is the measurement route. RF figures on this device come from the customer bench or from a partner RF laboratory, since our own production test covers DC parameters, functionality and customer fixtures rather than power sweeps at 1.3 GHz. That is a division of work worth stating at the quotation stage rather than after the first article.
The fourth is the documentation set. The vendor page carries seven parameter rows and no full document, so the request for quotation should carry a request for the outline drawing, the measured efficiency at the customer frequency, the small signal data and the moisture classification. Our incoming inspection plan records what arrives against that request, and for parts of this class that record is the only traceable evidence of what the delivered lot actually is.