HXN40017
Product Specifications
HXN40017
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HXN40017 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
HXN40017: 398 W from a 50 V pulse rail inside a 100 MHz window
The HXN40017 is a gallium nitride power transistor with the input and output matching networks built onto the device carrier, so the part presents 50 ohm at both ports and needs no external matching at the band it was tuned for. The vendor publishes five rows for it: 1.6 to 1.7 GHz, saturated output at least 56 dBm, gain at least 13 dB, drain efficiency at least 50 per cent typical, and 50 ohm ports. The index table adds a 50 V drain rail, plus or minus 0.8 dB of flatness and pulsed operation at a 1 ms pulse width and 20 per cent duty cycle.
Fifty six dBm is 398 W. That number governs everything else on this page, because at 50 per cent efficiency the DC input is 796 W, the drain current at 50 V is 15.9 A, and the heat to remove while the pulse is on is the same 398 W. On a 20 per cent duty cycle the average dissipation falls to 79.6 W, and that is the number the heatsink is sized against, not the peak.
The band is narrow on purpose. One hundred megahertz between 1.6 and 1.7 GHz covers the Iridium uplink block from 1616 to 1626.5 MHz with room either side, the lower part of the 1.7 GHz cellular uplink, and the video and telemetry links that sit between them. A plus or minus 0.8 dB flatness figure over a band that narrow describes a match tuned to one window rather than a broadband part, and a buyer planning to move the carrier more than 50 MHz should expect to re-tune or to choose a neighbour in the family instead.
The matching is described as in-plane synthesis on a film hybrid carrier, which is the construction used across this family: the die is matched on the carrier with thin film elements and the whole assembly sits in a metal flange package. That construction is why the part is 50 ohm at the ports and also why the band edges are fixed at purchase, because the match is built rather than tuned by the user.
The parameter rows Hisiwell publishes for HXN40017
| Part class | Gallium nitride power transistor, internally matched |
| Frequency range | 1.6 to 1.7 GHz |
| Saturation output power as published | at least 56 dBm |
| Saturation output power in watts, calculated | 398.1 W |
| Gain as published | at least 13 dB |
| Drain efficiency as published | at least 50 per cent, typical |
| Drain voltage | 50 V |
| Flatness across the band | plus or minus 0.8 dB |
| Operating mode | pulsed, 1 ms pulse width at 20 per cent duty |
| Input impedance | 50 ohm |
| Output impedance | 50 ohm |
| Output at 1 dB compression | not published, the index prints a slash |
| DC input at saturation, calculated | 796.2 W |
| Drain current at saturation, calculated | 15.92 A at 50 V |
| Dissipation during the pulse, calculated | 398.1 W |
| Average dissipation at 20 per cent duty, calculated | 79.6 W |
| Average drain current at 20 per cent duty, calculated | 3.18 A |
| Drive power for the rated output, calculated | 43 dBm, 19.95 W |
| Thermal resistance, junction to case | not published |
| Transient thermal impedance | not published |
| Package outline and dimensions | not published in text, one image on the product page |
| Scattering parameters | not published |
| Pinch off voltage and gate bias point | not published on this page |
| Quiescent current | not published |
| Operating and storage temperature | not published |
| Moisture level and floor life | not published |
| Orderable addendum and packing | not published |
| Neighbour at 56 dBm in the 1.2 to 1.4 GHz band | HXN40011, same rail, 13 dB gain |
| Neighbour at 49 dBm over 1.35 to 1.8 GHz | HXN40014, 28 V rail, continuous wave |
What a 20 per cent duty cycle does and does not buy
The duty cycle is the only thermal relief this part gets, and it is worth putting numbers on it. One millisecond at 796 W of DC input is 0.8 J of energy drawn from the supply in every pulse, and with a 5 ms period the supply sees an average of 159 W. A capacitor bank feeding the drain has to carry 15.9 mC per pulse, so a bank that is allowed to sag by 1 V needs 15.9 mF, and a bank allowed to sag by 3 V needs 5.3 mF. Those capacitors sit between the supply and the device and their ESR is inside the pulse, which is why a 50 V supply rated for the average current will still collapse on the leading edge of the pulse.
What the duty cycle does not buy is a thermal answer, because the vendor publishes no transient thermal impedance curve. The average of 79.6 W tells the person sizing the heatsink what the steady state load is, but the junction temperature inside a 1 ms pulse depends on the die area, the solder interface and the carrier, and none of those are on the page. A buyer working with 79.6 W of average load on a 0.5 C per watt carrier is looking at 40 C of average rise, which is comfortable; the same buyer cannot say from the published data whether the junction climbs another 30 C or another 130 C during the pulse.
Efficiency is printed as a typical with no minimum, and the arithmetic is sensitive to it. At 45 per cent the dissipation at 398 W output rises to 486.6 W and the drain current to 17.7 A; at 55 per cent it falls to 325.7 W and 14.5 A. A delivered part at the low end of the efficiency spread therefore asks for 22 per cent more heat removal and 22 per cent more average current than a part at the high end, on the same order code.
Where the published page stops short
The first gap is compression. The index table has a column for output at 1 dB compression and every row in this family prints a slash in it, so 56 dBm is a saturated figure and nothing on the page describes behaviour backed off from saturation. A radar or datalink waveform with any peak to average ratio has to be planned against a saturation point well above the average power it needs, and the margin has to be measured rather than read.
The second gap is the small signal behaviour. The 50 ohm ports and the plus or minus 0.8 dB flatness are saturated power figures; input return loss, gain slope and isolation between the ports at a backed off drive are not published. A cascade of two of these parts, or one of these parts behind a driver, depends on the input match of the second stage and on how much of the drive is reflected at the band edges.
The third gap is bias. No pinch off voltage, no recommended gate voltage and no quiescent current is printed for this part. The family's one full data sheet, discussed on the HXN42001 page, prints a pinch off window of minus 1.5 to minus 3.5 V and a 5 V gate ceiling, which is a starting point for a first article and not a substitute for the data sheet of this part. A 50 V drain rail with 15.9 A of pulse current also needs the gate supply established before the drain is applied, every time, including after a brownout.
The fourth gap is mechanical. The product page carries one outline image and no dimension table, and at 398 W in a pulse the flange flatness, the mounting torque and the carrier material all decide whether the solder interface survives thermal cycling. Those are questions for the vendor with the quotation, not for the first article.
Where a 1.6 to 1.7 GHz stage at this power fits
The fit is a fixed frequency transmitter with a pulse budget: a satellite handset or messaging uplink in the Iridium block, a ground station driver, a telemetry transmitter, or a laboratory amplifier used to test receivers at 1.65 GHz. In each case the 100 MHz band is a feature rather than a limit, because the system owns the window and the internal match removes the tuning that a discrete die would need.
It is a poor fit for continuous wave at this power, because 398 W of steady dissipation would need a cooling stack that no flange this size can feed, and the vendor's own 28 V continuous wave neighbour at 49 dBm exists for that job. It is also a poor fit for a wideband requirement, and a poor fit as a first stage: 43 dBm of drive, or 19.95 W, has to come from a power amplifier of its own.
Our RF power transistor cross reference work keeps the band, the rail, the duty cycle and the drive level side by side for this whole family, because the parts that look interchangeable on output power alone differ by 20 V of rail and by a factor of four in drive, and both of those decide the supply and the stage count.
Questions we answer about HXN40017 before a lot is booked
The first is the pulse width. This part is rated at a 1 ms pulse, and a system that keys it for 5 ms at the same duty cycle is outside the published condition even though the average dissipation looks the same. Ask what the longest pulse is, not just the duty cycle.
The second is the drain supply. 796 W of peak DC input at 50 V means a supply with current limit set above 16 A, a capacitor bank sized for the pulse, and a drain feed that does not ring at 15.9 A of edge current. Our RF procurement process raises the rail, the pulse shape and the droop budget with the vendor before the order, because a part that meets 56 dBm on a test bench with a short pulse can miss it on a customer supply with a long one.
The third is what arrives with the lot. With no moisture level, no marking code and no orderable addendum published, the purchase order has to name the band, the pulse condition and the flange, and the incoming inspection plan then records the flange, the marking and the DC behaviour at the gate and drain that can be measured without an RF bench. A power sweep at 1.65 GHz is done on the customer bench or at a partner RF laboratory, and the quotation says so.