HXN40064
Product Specifications
HXN40064
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HXN40064 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
HXN40064: 47 dBm continuous wave across 5.3 to 5.9 GHz
HXN40064 is the 47 dBm row of the Hisiwell GaN internally matched family for 5.3 GHz to 5.9 GHz. The vendor table lists 11 dB of gain, 45 percent drain efficiency, a 28 V drain rail, plus or minus 0.8 dB of flatness and a continuous wave rating. The part is matched to 50 ohm at both ports inside the package, so a 50 W stage is a flange, a launch, a bias feed and a heatsink rather than a tuning exercise.
47 dBm is 50.12 W, and this is the row where a continuous wave product starts to be a thermal project. At 45 percent efficiency the device draws 111.4 W of DC, which is 3.98 A from a 28 V rail, and 61.3 W of that leaves as heat in the flange. Those three figures, not the output power, are what set the heatsink, the supply and the enclosure.
The numbers on this page are the vendor table rows for this part, and the derived figures are the direct arithmetic from them. They are quoted here because the summary table on a vendor site gives the output power and the gain and leaves the reader to work out the rest, and the rest is what a quotation and a thermal budget depend on.
Fifty watts continuous: the arithmetic behind the flange
61.3 W in the flange is a different problem from the 24.4 W of a 20 W part. A flange to heatsink interface with a thermal pad is about 0.5 degrees per watt, so the interface alone drops 31 degrees. A heatsink at 0.6 degrees per watt, which is a finned extrusion with forced air, adds another 37 degrees. That is 68 degrees of rise on top of ambient before the junction to case number is counted, and a 50 degree enclosure puts the flange at 118 degrees.
The same product with a 28 V rail at 3.98 A also has to carry that current through the bias feed. A 15 mm microstrip at 0.5 mm width handles it thermally but not without loss: 3.98 A through 20 milliohm of copper and connector is 0.32 W, which is small against 61 W but grows with every connector and every screw terminal in the path. At this current the drain feed is a busbar question, not a trace question.
| Function | Internally matched GaN power transistor, 50 ohm in and out |
| Frequency band | 5.3 GHz to 5.9 GHz |
| Saturated output power | 47 dBm, which is 50.12 W |
| Gain | 11 dB at saturation |
| Drain efficiency | 45 percent at saturation |
| Drain voltage | 28 V nominal |
| DC input at saturation | 111.4 W |
| DC current at saturation | 3.98 A from a 28 V rail |
| Heat in the flange at saturation | 61.3 W continuous |
| Flatness in band | plus or minus 0.8 dB, which is 20.2 percent of power |
| Operating mode | Continuous wave |
| Gain step to the next row up | 1 dB less gain at 50 dBm |
| Input return loss expectation | Set by the internal match, confirm on the data sheet curve |
| Bias feed | Quarter wave at band centre, rated for 4 A |
| Bias sequencing | Gate negative before drain, drain removed before gate |
| Package | Flange mount, screw fixed, thermal pad at the interface |
| Interface | 50 ohm microstrip launch at both ports |
| Cooling | Heatsink and airflow sized on 61.3 W continuous |
| Enclosure | Air path through the fins, not across them |
Where HXN40064 sits between 43 and 50 dBm
The ladder in this band runs 37, 40, 43, 47, 50, 51 and 53 dBm, so HXN40064 is the fourth step and the first one where the heatsink decides the enclosure. The 3 dB step from 43 dBm to 47 dBm doubles the power, and it also nearly doubles the DC input, from 44.3 W to 111.4 W, because the efficiency stays at 45 percent across the band. A product that moves up this step inherits a 2.5 times larger thermal problem, not a 2 times larger one, because the gain also drops a dB and the driver has to make up for it.
| Part | Psat (dBm) | Watts | DC input (W) | Heat (W) | Gain (dB) |
| HXN40061 | 37 | 5.01 | 11.1 | 6.1 | 12 |
| HXN40062 | 40 | 10.00 | 22.2 | 12.2 | 12 |
| HXN40063 | 43 | 19.95 | 44.3 | 24.4 | 12 |
| HXN40064 | 47 | 50.12 | 111.4 | 61.3 | 11 |
| HXN40065 | 50 | 100.00 | 222.2 | 122.2 | 10 |
| HXN40066 | 51 | 125.89 | 279.8 | 153.9 | 10 |
| HXN40067 | 53 | 199.53 | 443.4 pulsed | 243.9 pulsed | 10 |
That table is the reason a 47 dBm part is often the practical top of an air cooled product and 50 dBm is where a liquid loop or a heat pipe starts to be discussed. It is also the reason two 47 dBm parts combined with 0.3 dB of loss are frequently a better answer than one 50 dBm part: the pair spreads 122 W across two flanges and two heatsinks, and the enclosure gets the same air through two smaller paths.
Bias, sequencing and the 4 A rail
The gate has to be negative before the drain comes up and the drain has to be removed before the gate. On a 4 A rail the failure is quick: a device that is allowed to conduct with the gate floating will pull whatever the supply can deliver, and the flange is destroyed before a fuse sees it. The clamp belongs on the board, with a resistor across gate and source and a diode to the negative rail, and the sequence belongs in the power supply rather than in the controller code.
The bias feed is a quarter wave at the centre of the band, which on 0.25 mm laminate with a dielectric constant near 3 is about 15 mm, and at 4 A it has to be wide enough to carry the current without heating. A 40 degree rise across a 15 mm line at 4 A is normal and it moves the bias point as the part heats, which is why the current is set at thermal equilibrium rather than on a cold board. Constant current bias with a temperature coefficient that follows the vendor curve is the usual answer on a 50 W stage.
Combining, isolation and the antenna side
Two HXN40064 parts combined with a 90 degree hybrid give 50.1 W plus 50.1 W minus the combining loss, which is 97 W at 0.3 dB, and the hybrid load has to take the imbalance when one part fails. A Wilkinson combiner has no load and therefore no failure tolerance but less loss. The choice is made by what has to happen when one device dies, and a radar or a broadcast transmitter usually wants the hybrid because the system can keep running at reduced power.
The antenna side is where the match matters most. A 50.12 W stage into an antenna with 10 dB return loss sees 5 W reflected, which at 61 W of heat is 8 percent more and at the flange is a hot spot. Circulators at this band and power are available and we stock them with the parts, which is covered on our one stop RF supply page.
What the driver stage has to deliver
At 11 dB of gain this part needs 36 dBm, which is 3.98 W, at its input to reach saturation. A driver that delivers 4 W with 1 dB of margin is 5 W, and the two candidates in this band are a GaAs amplifier or a smaller member of this GaN family. The GaN driver brings the same bias polarity and the same sequencing requirement, which keeps the bias supply simple, while the GaAs driver needs a negative rail of its own and a different bias point.
The driver also sets the linearity of the pair. A 47 dBm part run at 6 dB of backoff needs the driver to stay clean at 30 dBm, and a driver that is already compressing at that level puts its own intermodulation into the output. The practical test is a two tone measurement at the backoff power with the driver in place, not a single tone sweep of the final alone.
What we check and what the customer measures
Our line tests DC parameters and function: gate threshold window, insulation, the bias network on the carrier and the flange workmanship. Gain and power at 5.3 GHz to 5.9 GHz are measured on the customer fixture or at a partner RF laboratory and are not figures we publish, because a DC screen cannot see a 61 W thermal path and only a calibrated bench can say what the part does at the top of the band.
What we do control is the record. 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 serial number is on our component traceability report. When a customer is comparing this row against another vendor for a drop in, the two figures that break the comparison are the gain, 11 dB here against 10 dB on the rows above, and the drain voltage, which is 28 V here and 32 V on HXN40066. That comparison is on our RF power transistor cross reference.