Contact Us

HXN40066

Brand: Hisiwell
Hisiwell HXN40066 is a discrete semiconductors component designed for DC-DC Converters & Power Supplies and related applications. In stock at Superb Automation — request a quote for competitive pricing and fast delivery.
quote now

Product Specifications

HXN40066 — Discrete & Power Semiconductors | Superb Automation

HXN40066

Hisiwell — Discrete & Power Semiconductors

Product Overview

The HXN40066 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 / VceV
Id / IcA
Rds(on) / Vce(sat)mΩ
QgnC
Technology—
Package—
Configuration—
Mounting—

Applications

Switch-Mode Power Supplies Motor Drives & Inverters Battery Management Systems DC-DC Converters Load Switching Automotive Power Distribution

Request Quote — In Stock, 100% New & Original, Full Warranty

HXN40066: 51 dBm on a 32 V drain rail at 5.3 to 5.9 GHz

HXN40066 is the 32 V member of the Hisiwell GaN internally matched family for 5.3 GHz to 5.9 GHz. The vendor table gives 51 dBm of saturated output, 10 dB of gain, 45 percent drain efficiency, plus or minus 0.8 dB of flatness and a continuous wave rating, with every other row in the band on 28 V. That single difference is the reason this page exists: a part that makes 125.9 W on 32 V draws 8.74 A instead of the 9.98 A the same power would need on 28 V.

51 dBm is 125.9 W, the second highest continuous wave row in the band, and at 45 percent efficiency the DC input is 279.8 W with 153.9 W of heat in the flange. The rail is the interesting part. A 32 V rail is not a standard telecom or automotive rail, so a product that uses this part either has a boost converter in front of it or a dedicated supply, and the arithmetic of that choice is on this page.

The numbers below are the vendor table rows for this part. Anything derived is arithmetic from them, and the arithmetic for the alternative rail is shown so that a product on 28 V can see exactly what it gives up by staying at 28 V.

Why a 32 V rail is worth a converter

At 51 dBm the load needs 125.9 W. On 28 V and 45 percent efficiency that is 9.98 A; on 32 V it is 8.74 A. The 1.24 A difference is 12 percent less current in the bias feed, the connectors and the supply, and at these currents it is also 12 percent less loss in every milliohm of that path. A bias feed with 20 milliohm of copper and connector loses 1.99 W at 9.98 A and 1.53 W at 8.74 A, which is small against 154 W of device heat but real in a sealed enclosure.

The converter that makes 32 V from a 28 V source has to deliver 8.74 A, which is 280 W, and a boost stage at 95 percent efficiency adds 14.7 W of its own heat and takes 9.3 A from the 28 V rail. The net effect is that the converter moves heat from the flange to the supply and adds a switching node to the enclosure. For a product that already has a 48 V bus, a buck to 32 V is simpler than a buck to 28 V plus nothing, and the 32 V part is then the natural choice.

FunctionInternally matched GaN power transistor, 50 ohm in and out
Frequency band5.3 GHz to 5.9 GHz
Saturated output power51 dBm, which is 125.9 W
Gain10 dB at saturation
Drain efficiency45 percent at saturation
Drain voltage32 V nominal, the only 32 V row in this band
DC input at saturation279.8 W
DC current at saturation8.74 A from a 32 V rail
DC current if the same power were made on 28 V9.98 A, which is 1.24 A more
Heat in the flange at saturation153.9 W continuous
Flatness in bandplus or minus 0.8 dB
Operating modeContinuous wave
Supply arrangement32 V at 9 A, or a boost converter from 28 V
Boost converter loss at 95 percent14.7 W, drawing 9.3 A from a 28 V source
Bias feedQuarter wave at band centre, rated for 9 A
Bias sequencingGate negative before drain, drain removed before gate
PackageFlange mount, screw fixed, cold plate interface
CoolingCold plate sized on 153.9 W continuous
Interface50 ohm microstrip launch at both ports

One hundred and fifty four watts in a flange

153.9 W is the highest continuous heat load in the band short of the pulsed 53 dBm row, and it puts this part in the same mechanical class as a 300 W LDMOS pallet. A 0.1 degrees per watt cold plate at 40 degree coolant holds the flange at 55 degrees, which leaves the junction to case budget intact. A 0.8 degrees per watt forced air heatsink puts the flange at 163 degrees above ambient, which no vendor rating survives.

The die area is not published in the summary table, and the junction to case figure comes from the data sheet. At this heat the case temperature is the number to measure on the first article: a flange at 85 degrees with a 0.3 degrees per watt junction to case path sits at 131 degrees junction, which is close to the usual GaN limit. A flange at 105 degrees is a failed thermal plan even though the part still makes its power on a cold bench.

The band ladder and the 3 dB steps

PartPsat (dBm)WattsDC input (W)Heat (W)Vd (V)
HXN40061375.0111.16.128
HXN400624010.0022.212.228
HXN400634319.9544.324.428
HXN400644750.12111.461.328
HXN4006550100.00222.2122.228
HXN4006651125.89279.8153.932
HXN4006753199.53443.4 pulsed243.9 pulsed28

The 1 dB step from 100 W to 125.9 W is unusual in a family that otherwise moves in 3 dB steps, and it exists because the 32 V rail allows a larger die to be driven at the same current density. The practical reading is that HXN40066 is not a general purpose step up from HXN40065; it is the row for a product whose supply is already above 28 V, and a 28 V product that buys it will spend more on a converter than it gains in power.

Gate bias, current and the temperature coefficient

A 154 W flange runs hot, and the gate threshold of a GaN device falls as the junction heats. With a fixed gate voltage the drain current rises with temperature, which is the wrong direction, so the bias has to be set with a current loop or with a gate supply that tracks the case temperature. The vendor curve gives the coefficient; the board gives the tracking.

The bias feed carries 8.74 A and is a quarter wave at the band centre, about 15 mm on 0.25 mm laminate with a dielectric constant near 3. At this current the line width is set by electromigration and heating rather than by impedance, and a 1.5 mm wide line on 35 micron copper at 8.74 A runs about 25 degrees above the board, which is 25 degrees of drift in the bias point if the loop is not closed.

Enclosure, airflow and the supply in the same box

The device dissipates 153.9 W and the converter that feeds it dissipates 14.7 W, so a single enclosure has 168.6 W to remove before the control electronics are counted. A 40 degree inlet with a 10 degree rise through the box allows 168.6 W at about 28 cubic metres per hour of air, which is a 60 mm fan at low speed or a passive chimney if the box is tall enough. The arithmetic is the same as any power supply; the mistake is to size the airflow on the device alone.

On the RF side, the launch and the flange have to stay flat and the lid has to be grounded every 20 mm across the band. The lid resonance at 5.3 GHz to 5.9 GHz is the usual cause of a gain dip at the top of the band and it is a mechanical fix, not an electrical one.

What we test and what the customer measures

Our line tests DC parameters and function, which covers the gate threshold window, the drain to source insulation, the bias network on the carrier and the flange flatness. Gain and power at 5.3 GHz to 5.9 GHz are measured on the customer fixture or at a partner RF laboratory, and we do not publish RF figures of our own, because a DC screen cannot see a 154 W thermal path.

The traceability record is what we control. The part marking, the date code and the carrier are read at incoming against our incoming inspection plan, and the lot history that follows the part to a unit serial is on our component traceability report. When the question is whether a 28 V part can replace this one, the answer is no without a converter, and the comparison that shows why sits on our RF power transistor cross reference.