Contact Us

HXC50114

Brand: Hisiwell
Hisiwell HXC50114 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

HXC50114 — Discrete & Power Semiconductors | Superb Automation

HXC50114

Hisiwell — Discrete & Power Semiconductors

Product Overview

The HXC50114 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

HXC50114: the 60 dBm row, and why a round number is a poor reason to choose it

The HXC50114 is a Hisiwell LDMOS RF power transistor on the HF shelf at 500 W per device, two devices per pair and a band edge at 500 MHz. The vendor page calls it a 1000 W capable, highly rugged, unmatched LDMOS FET covering 200 MHz up to 500 MHz, usable in pulse and CW. A pair at 1000 W is exactly 60.00 dBm, and that is the only round number on this shelf, which is probably why this row gets asked about more than the 900 W and 1400 W rows around it.

Sixty decibels above a milliwatt is a convenient figure for a link budget and a poor figure for a purchasing decision. The nearest neighbour, the HXC50113 at 450 W per device, is 0.46 dB lower. The next one up, the HXC50115 at 700 W per device, is 1.91 dB higher. None of those three steps is large enough that a system which works on one will fail on another, so the choice between them is decided by availability, price and the cold plate, not by the power column.

The page and the row agree on class, band and rail, and they both stop short of gain, efficiency, thermal resistance, load pull, package and pulse conditions. What follows keeps the published numbers, adds arithmetic a buyer can check, and treats every assumed efficiency as an assumption rather than a specification.

The HXC50114 rows as the vendor publishes them

Part classLDMOS RF power transistor, unmatched, two devices in a pair
Frequency range, index rowHF to 0.5 GHz
Frequency range, vendor page200 MHz up to 500 MHz
Window width, calculated300 MHz
Band ratio, calculated2.5 to 1
Output per device, index row500 W
Output for the pair, calculated1000 W
Output in dBm, per device, calculated56.99 dBm
Output in dBm, pair, calculated60.00 dBm
Step from one device to the pair, calculated3.01 dB
Load voltage at 1000 W into 50 ohm, calculated223.6 V rms, 316.2 V peak
Peak against the 50 V rail, calculated6.32 times
Peak at a 2 to 1 mismatch, calculated632.5 V
Peak at a 3 to 1 mismatch, calculated547.7 V
DC input at 55 per cent, calculated1818.2 W
Drain current at 50 V and 55 per cent, calculated36.36 A
Heat to remove at 55 per cent, calculated818.2 W
DC input at 65 per cent, calculated1538.5 W
Drain current at 50 V and 65 per cent, calculated30.77 A
Heat to remove at 65 per cent, calculated538.5 W
Step against the 900 W pair, calculated0.46 dB
Step against the 1400 W pair, calculated1.91 dB
Gain, index rowNot published
Efficiency, index rowNot published
Mode, index rowCW
Modes named on the pagePulse and CW
Drain rail, both sources50 V
Thermal resistance, junction to caseNot published
Load impedance and load pullNot published
Package and drawingNot published
Orderable addendum and packingNone published
Applications listed on the pageIndustrial, scientific and medical, VHF communication, UHF TV, aerospace
Data sheetNo document linked from the page

What 1000 W at 200 to 500 MHz asks of the rest of the system

1000 W into 50 ohm is 223.6 V rms and 316.2 V peak. The peak is 6.32 times the 50 V rail, higher than the 6.00 times of the 900 W pair, and the extra is exactly the 0.46 dB of power rather than any change in topology. On a 2 to 1 mismatch the peak at the load reaches 632.5 V, which is past the point where a standard N connector should be asked to carry the signal. At this power the output connector is a 7/16 or an EIA flange, not an N.

At 55 per cent the pair draws 1818.2 W, 36.36 A at 50 V, and leaves 818.2 W to remove. At 65 per cent it draws 1538.5 W, 30.77 A, and leaves 538.5 W. The difference between those two assumptions is 280 W of heat, which is the difference between a cold plate with margin and one without. Since no efficiency is published for the row or for the page, the supply and the cooling have to be sized for the lower figure until a measurement replaces it, and the measurement should be taken at the band edge rather than at the centre, where every RF part looks its best.

Thirty six amps at 50 V is not a bench supply load. It is a busbar, a fuse rated for the DC rather than the RF, a current shunt with a low temperature coefficient, and a cable drop of a few hundred millivolts that has to be compensated at the supply rather than ignored at the device. A pair that measures 1000 W on the bench with a short cable can lose half a decibel in a rack where the supply sits two metres away, and the loss shows up as reduced output rather than as a fault.

The three rows that bracket the HXC50114

HXC50115, 700 W per device, pair1400 W, 1.91 dB above the HXC50114
HXC50114, 500 W per device, pair1000 W, 60.00 dBm, the row this page covers
HXC50113, 450 W per device, pair900 W, 0.46 dB below the HXC50114
HXC50104, 650 W single ended650 W, 1.87 dB below the pair, one flange
HXC50119, 250 W single ended250 W, 6.02 dB below the pair
HXC50118, 170 W single ended170 W, 7.70 dB below the pair
HXC50121, 175 W per device, pair350 W, 4.56 dB below the pair
HXC50109, 130 W per device, pair260 W, 5.85 dB below the pair

A 1.91 dB step from 1000 W to 1400 W sounds like a small gain for a bigger die and a bigger supply, and on a system level it usually is. A 0.46 dB step from 900 W to 1000 W is smaller still. The practical consequence is that a buyer should pick the row that is in stock at the quantity needed, then confirm that the cold plate and the supply can absorb the difference in heat and current, because every one of these three rows needs the same kind of hardware and the same kind of fixture.

The single ended 650 W row is the exception and deserves a separate look. It removes one device, one combiner and one flange from the assembly. Where 650 W is enough, that is a better answer than 1000 W with a combiner whose reject load has to be cooled and monitored. The question to ask is not which part is strongest but which assembly is smallest while still meeting the system figure.

Combiner, reject load and the 0.46 dB nobody budgets for

A quadrature combiner is the standard way to turn two 500 W devices into 1000 W, and it comes with two loads that are easy to forget. The isolated port absorbs the amplitude and phase difference between the two legs, and at this power a one decibel imbalance sends roughly 100 W into that port. The load has to be rated for the imbalance, not for the ideal case, and it has to be cooled, because it sits inside the same enclosure as the devices.

Insertion loss in the combiner is the second figure to budget. A well built 500 MHz combiner loses 0.15 to 0.25 dB per leg, which turns 1000 W into 940 to 960 W at the output flange. That is not a fault of the transistor and cannot be recovered by driving the pair harder without eating the margin that the 6.32 times rail ratio provides. Quote the system figure, not the device figure, when the number goes into a link budget.

Phase balance is the third. The two devices sit on separate flanges with separate networks, and a five degree phase error at 500 MHz is about 3 mm of line length. That is inside the tolerance of a hand cut network, which is why the fixture should be measured for phase as well as for amplitude before the pair is run at full drive.

Incoming inspection on a 1000 W pair

The incoming plan for this part is the same shape as for the 900 W row: flatness of both flanges, a curve tracer check of the gate and drain junctions, marking and date code against the certificate, then quiescent current and gain at the band centre on the fixture that will be used for the first article. The difference at 1000 W is that the fixture itself has to be rated for the power and for the mismatch, and that the reflected port has to be instrumented before the first key up rather than after.

The incoming inspection plan page sets out the sequence and the instruments. Two additions apply here: the DC supply should be current limited at a level above the expected 36 A but below the point where a failed device destroys the fixture, and the first run should be made into a load that can take 1000 W continuously, not into a load that takes it for ten seconds.

Where the part is bought as a replacement for a part in an existing amplifier, run the incumbent and the replacement on the same fixture at the same drive and compare the drain current at equal output rather than the output at equal drive. Equal output equal current tells you the efficiency matches; equal drive with different output tells you the gain matches. A replacement that satisfies both is a drop in. One that satisfies only one needs a network adjustment, and that is a different purchase decision.

Questions buyers ask about the HXC50114

Is 1000 W the pair or one device? The pair. The row is 500 W times two and the page calls the part 1000 W capable. One device is 500 W, 56.99 dBm.

Why does the row say CW when the page says pulse and CW? The row carries the class mode and the page carries the wider statement. No duty cycle or pulse width is published, so a pulse application needs the factory to confirm the condition rather than the page to imply it.

What is the current at full power? 36.36 A at 50 V if the pair runs at 55 per cent, 30.77 A at 65 per cent. Both are arithmetic on assumed efficiencies, not published figures.

How much heat goes into the cold plate? 818.2 W at 55 per cent and 538.5 W at 65 per cent, split between two flanges. That is what sizes the plate, not the 1000 W of RF.

Can I drive it with a 10 W exciter? Only if the gain is 20 dB, which is not published. At 15 dB the drive is 31.6 W and at 18 dB it is 15.8 W. The shelf carries both classes, so the gain has to come from the factory.

Is there a document I can attach to a purchase specification? None is linked. For a part with a full document on the same shelf, look at the 200 W single ended row, which publishes maximum ratings, thermal resistance, a typical performance table and a fixture drawing.

For a quote on the HXC50114 with the HXC50113 and HXC50115 as alternates, send the quantity and the sub band to the RF component sourcing desk. Where the part is a replacement against an existing number, start from the cross reference page so the comparison is made against the incumbent rather than against the row above.