HX446050F2
Product Specifications
HX446050F2
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HX446050F2 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
HX446050F2: fifty watts at 4.4 to 6 GHz and forty one watts of heat
The HX446050F2 is the larger member of the 4.4 to 6 GHz pair from Hisiwell, on the same 28 V rail and the same 55 per cent drain efficiency as the 30 W part below it. The index row reads 50 W of saturation and 15 dB of power gain. Fifty watts is 46.99 dBm by calculation, and with 15 dB of gain the drive at saturation is 1.58 W, or 31.99 dBm.
One decibel of extra gain against the 30 W member is the detail that makes this part easier to drive than its size suggests. Doubling the output would normally cost 3 dB of drive if the gain were unchanged. Here the gain rises by 1 dB, so the drive only rises from 1.19 W to 1.58 W, and a driver line up that was sized for the small part still covers the large one without a change.
The heat is where the size shows. At 55 per cent the drain takes 90.9 W and the part leaves 40.9 W, against 24.5 W for the 30 W member. Forty one watts is a flange and a forced air sink, and it is the number that decides the enclosure depth rather than the 50 W output figure.
The band is 1.6 GHz wide, 30.8 per cent at a 5.2 GHz centre. Combined with a 1.58 W drive this makes the part usable across the whole 4.4 to 6 GHz window with one network, which is the reason it turns up in test equipment and in wideband satcom up converters rather than in a single channel radio.
The drain current is 3.25 A at saturation on 28 V. That is a feed trace of 200 mil at one ounce copper, a choke above 5 A and a bulk bank sized for the crest of the envelope rather than for the average.
The parameter rows behind the HX446050F2
| Part class | Gallium nitride HEMT power transistor, unmatched |
| Frequency range | 4.4 to 6 GHz |
| Band width, calculated | 1,600 MHz |
| Fractional bandwidth at 5.2 GHz centre, calculated | 30.8 per cent |
| Saturation output, index row | 50 W |
| Saturation in dBm, calculated | 46.99 dBm |
| Peak output column | Not published, saturation only |
| Average output column | Not published |
| Power gain | 15 dB |
| Gain as a ratio, calculated | 31.6 times |
| Drive at saturation, calculated | 1.58 W, 31.99 dBm |
| Drain efficiency | 55 per cent |
| DC input at 55 per cent, calculated | 90.9 W |
| Drain current, calculated | 3.25 A at 28 V |
| Heat at saturation, calculated | 40.9 W |
| Gain against the 30 W member, calculated | 1 dB more for 2.22 dB more output |
| Drain rail | 28 V |
| Mode | Continuous wave |
| Load voltage at 50 ohm, calculated | 50.0 V RMS, 70.7 V peak |
| Load voltage under a two to one mismatch, calculated | 100.0 V peak |
| Output per gigahertz of band, calculated | 31.3 W per GHz |
| Matching | Unmatched, network on the customer board |
| Package and body | Not published on the index row |
| Thermal resistance, junction to case | Not published |
| Gain flatness over the band | Not published |
| Pulse width and duty cycle limits | Not published |
Ninety one watts in and fifty watts out
The supply has to carry 90.9 W at 3.25 A, and a 100 W module with 20 per cent headroom is the normal choice. The crest of a modulated carrier is what sets the bulk capacitance, not the average current, and at six times crest factor the drain lead sees nearly 20 A for a few microseconds.
Forty one watts of heat is the point where the flange joint stops being a detail. A copper coin with a via field under the die, a thin interface and a flat sink face are all required. A 0.1 mm void under the flange at 40 W is a hot spot that shows up as gain drift on the second hour of a burn in test, not on the first.
The output voltage is the other number to watch. Fifty watts at 50 ohm is 70.7 V peak before any mismatch and 100 V peak under two to one. That leaves no room for 63 V capacitors and only just enough for 100 V parts with a derating policy, so the match and the protection are sized together.
Thirty one watts per gigahertz of band is what separates this part from the narrow band members of the 5 GHz shelf, where the same output lands in a 1 GHz window. Wideband and efficient at the same time is the reason the 55 per cent matters here more than it would on a single channel part.
HX446050F2 against the other members of the shelf
| HX446050F2 | 4.4 to 6 GHz | 50 W saturation | 15 dB | 55 per cent | Reference point |
| HX446030F2 | 4.4 to 6 GHz | 30 W saturation | 14 dB | 55 per cent | 2.22 dB less output, 1 dB less gain, 16.4 W less heat |
| HX5060140P2 | 5 to 6 GHz | 140 W saturation | 11 dB | 50 per cent | 4.47 dB more output, 4 dB less gain, 11.1 W of drive |
| HXN50214 | 5 to 6 GHz | 50 W peak | Not published | Not published | Same output, narrower window, no efficiency row |
| HXN50217 | 5 to 6 GHz | 100 W peak | Not published | Not published | 3.01 dB more peak, no efficiency row |
| HXN50209 | 4.4 to 5.9 GHz | 15 W saturation, 25 W peak, 2 W average | 15 dB | 26 per cent | 5.23 dB less output, same gain, much less efficient |
Against the 140 W part the arithmetic is blunt: 4.47 dB more output costs 4 dB of gain and 11.1 W of drive, so the line up in front of a 140 W stage is a project of its own. Where the antenna needs 47 dBm rather than 51.5 dBm, this part is the end of the chain and not a driver.
Against the 5 to 6 GHz peak parts the window is the difference. Those are filed on a 1 GHz band with a peak column and no efficiency row, which suits a pulsed radar. This part is filed on a 1.6 GHz band with a saturation row and an efficiency, which suits a continuous wideband amplifier.
Both 4.4 to 6 GHz parts sit on the RF power transistor cross reference, and the driver and passives that go with them are collected on RF procurement.
Board notes for a fifty watt wideband stage
The first item is the match across 30.8 per cent. A two section ladder on a low loss laminate holds it, and the network is tuned at 4.4, 5.2 and 6 GHz rather than at the centre alone.
The second item is the capacitor rating at the output. At 70.7 V peak a 100 V part is the minimum and a 250 V part is the safer choice when a mismatch is possible. The match inductors carry higher circulating current than the load current, which is the second half of the same decision.
The third item is the bias network. Fifteen decibels of gain with no published quiescent current means the class is set on the first article against the linearity the system needs, and the gate network is temperature compensated so the point does not walk with the case.
The fourth item is the flange. At 40.9 W the joint is the thermal path of the whole amplifier, and a torque controlled fastener with a specified interface is the only way to keep it repeatable.
Questions buyers ask about HX446050F2
Is 50 W continuous? The row is a saturation figure for continuous wave operation. A backed off continuous point is lower, and no derating curve is published.
How much drive does it need? 1.58 W, which is 31.99 dBm. That is only 0.4 dB more than the 30 W member of the same band.
Is the efficiency 55 per cent at saturation? Yes, the single efficiency figure belongs to the single power figure. At a backed off point it is lower and unmeasured on this page.
Can it be pulsed? The mode column reads continuous wave, so a pulsed application needs a written confirmation of pulse width and duty cycle. Neither is published.
How do we start? Send the band, the output at the connector, the signal, the ambient and the rail. We will quote from the schedule and ship a sample measured at 4.4, 5.2 and 6 GHz with gain, efficiency and the flange temperature on the curve.