Contact Us

HXC50113

Brand: Hisiwell
Hisiwell HXC50113 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

HXC50113 — Discrete & Power Semiconductors | Superb Automation

HXC50113

Hisiwell — Discrete & Power Semiconductors

Product Overview

The HXC50113 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

HXC50113: 900 W on a 50 V rail and the 3 dB step that decides the pair

The HXC50113 is a Hisiwell LDMOS RF power transistor on the HF shelf with a 450 W row per device, two devices per pair and a band edge at 500 MHz. The vendor page calls it a 900 W capable, highly rugged, unmatched LDMOS FET covering 200 MHz up to 500 MHz, usable in pulse and CW. The index row and the page agree on the class, the band and the rail, which is not true of every row on this shelf, so the two figures can be quoted together without a footnote.

The number that separates a working pair from a damaged one is not the 900 W. It is the 3.01 dB between one device at 450 W and the pair at 900 W. Two devices combined in quadrature add 3 dB only when they are driven in phase and the combiner is lossless. Three tenths of a decibel of insertion loss in each leg brings 900 W down to about 838 W, and half a decibel of amplitude error between the legs costs a further tenth of a decibel while pushing the difference into the reject load, where it heats a resistor that no thermal budget on the drawing usually counts.

Nothing on the index row or on the vendor page publishes gain, efficiency, a pulse width, a duty cycle, a thermal resistance, a load pull table or an orderable addendum. That is the normal state of this shelf rather than a fault of this part. What follows takes the published row, adds only arithmetic that a buyer can repeat on a calculator, and labels every assumed figure as an assumption.

The HXC50113 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 row450 W
Output for the pair, calculated900 W
Output in dBm, per device, calculated56.53 dBm
Output in dBm, pair, calculated59.54 dBm
Step from one device to the pair, calculated3.01 dB
Load voltage at 900 W into 50 ohm, calculated212.1 V rms, 300.0 V peak
Peak against the 50 V rail, calculated6.00 times
Peak at a 2 to 1 mismatch, calculated600.0 V
Peak at a 3 to 1 mismatch, calculated519.6 V
DC input at 55 per cent, calculated1636.4 W
Drain current at 50 V and 55 per cent, calculated32.73 A
Heat to remove at 55 per cent, calculated736.4 W
DC input at 65 per cent, calculated1384.6 W
Drain current at 50 V and 65 per cent, calculated27.69 A
Heat to remove at 65 per cent, calculated484.6 W
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
PackageNot published on the row, no drawing on the page
Pulse width and duty cycleNot published
Orderable addendum and packingNone published
Applications listed on the pageIndustrial, scientific and medical, VHF communication, UHF TV, aerospace

What 900 W at 200 to 500 MHz asks of a 50 V rail

900 W into 50 ohm is 212.1 V rms and 300.0 V peak. That peak is six times the 50 V rail, and on an unmatched device the internal drain swing is higher again because the network transforms the 50 ohm load down to an impedance the die can drive. The six times figure is the number that decides whether the output network has margin, and it is also the number that explains why a 50 V part is used at HF when 28 V parts exist: the same power at 28 V would need a swing 10.7 times the rail, and the margin inside the package would be gone.

At a 2 to 1 mismatch the peak at the load doubles to 600 V, and at 3 to 1 it reaches 519.6 V through the vector sum rather than the scalar one. A tube amplifier at these powers is expected to survive a full mismatch into an open circuit; a solid state pair is not, and the protection has to be built into the control loop rather than assumed in the device. Fold back the drive when the reflected power crosses a threshold, and measure the reflected port rather than the forward one, because a pair with a failed combiner leg can still show a healthy forward reading.

The DC side is the part most first articles get wrong. At 55 per cent drain efficiency the pair draws 1636.4 W from the supply, which is 32.73 A at 50 V, and 736.4 W of that has to leave through the flange. At 65 per cent the same 900 W draws 1384.6 W, 27.69 A, and the heat falls to 484.6 W. Neither efficiency figure is published for this part, so the supply, the cable gauge and the cold plate all have to be sized on the pessimistic end until a measurement on the customer's own fixture replaces the assumption.

The 300 MHz window is wide for a single network. A pair that holds 900 W at 250 MHz will not hold it at 480 MHz with the same matching section, and the row states one power for the whole window. Treat the figure as a class number, cut the network for the sub band the system occupies, and measure at both edges of that sub band before signing off a build.

Choosing between the HXC50113, the HXC50114 and the rest of the shelf

HXC50114, 500 W per device, pair1000 W, 0.46 dB more than the HXC50113
HXC50113, 450 W per device, pair900 W, the row this page covers
HXC50115, 700 W per device, pair1400 W, 1.91 dB more than the HXC50113
HXC50121, 175 W per device, pair350 W, 4.10 dB below the HXC50113
HXC50119, 250 W single ended250 W, 5.56 dB below the HXC50113 pair
HXC50118, 170 W single ended170 W, 7.24 dB below the HXC50113 pair
HXC50104, 650 W single ended650 W, 1.41 dB below the pair, one flange instead of two
HXC50109, 130 W per device, pair260 W, 5.39 dB below the pair

The step from the HXC50113 to the HXC50114 is 0.46 dB. On a bench with a calibrated power meter and a good attenuator that difference is measurable, but on a first article built with hand cut lines and a coaxial combiner it is inside the spread between two builds of the same part. A buyer comparing only on the power column is therefore comparing noise. The real differences are availability, price at the quantity on the order, and whether the pair has to be matched to a specific flange and cold plate arrangement.

The step down to a single ended 650 W part is more interesting than the step up. One flange means one set of mounting hardware, one thermal interface and no combiner to lose 0.3 dB and to fail in the field. Where the system can accept 650 W instead of 900 W, the single ended row removes a whole assembly from the bill of materials, and 1.41 dB is usually cheaper to recover in antenna gain than in a second device.

Heat, clamping and the hardware around a 900 W pair

Between 484 W and 736 W of heat leaves through two flanges, so each flange carries between 242 W and 368 W at the assumed efficiency. That number, not the RF power, is what sizes the cold plate. A copper block with a machined channel and a flow of two litres per minute will move it; an extruded heat sink with a fan will not, and no amount of thermal compound compensates for a flange that is not flat to better than 25 micrometres across its length.

The vendor publishes no junction to case resistance for this row. Without it there is no way to turn the heat figure into a junction temperature, and without a junction temperature there is no way to argue a lifetime. The two things to ask the factory before committing a build are the thermal resistance for the package actually shipped and the maximum junction temperature the part is qualified to, because the two numbers decide the cold plate specification and cannot be inferred from the 900 W row.

Clamping matters more at this power than at 100 W. A flanged package held by two screws at the ends will bow in the middle, and the bow shows up as a hot spot under the die that a case temperature reading at the flange edge will never reveal. Use a clamp bar across the full width, torque to the figure the factory states rather than to the figure that feels tight, and check the interface after the first thermal cycle rather than before it.

Incoming inspection and first article on an unmatched pair

An unmatched pair cannot be tested by substitution into a working amplifier, because the network in that amplifier is cut for a different part. The first article needs its own fixture, built for the sub band, with the drive level set so that the pair reaches the rated power and no more, and with the reflected power monitored for the whole run. The fixture should be kept, because the second article will be compared against it and the comparison is only meaningful if the fixture is the same one.

A workable incoming plan for this class of part checks four things before any RF is applied: the flange flatness, the gate to source and drain to source resistance with a curve tracer, the marking against the order, and the date code against the certificate. After the first power up it adds the quiescent current at the stated gate voltage, the gain at the band centre, and the drain current at rated output. The full sequence is on the incoming inspection plan page, and it is written to be run on a bench that has a power meter, a directional coupler and a current shunt rather than a full load pull system.

Where the part is being bought as a second source against an existing amplifier, the first article should also be compared against the incumbent device on the same fixture, at the same drive, with the same network. That comparison is the only one that answers the question the buyer actually has, which is whether the replacement will make the same power in the box that already exists.

Questions buyers ask about the HXC50113

Is 900 W the output of one device or of the pair? Of the pair. The index row reads 450 W times two, and the page calls the part 900 W capable. One device is 450 W, which is 56.53 dBm.

What supply does it need? 50 V on the row and on the page. That is higher than the 28 V rows on the same shelf, which is why the current at 900 W is 32.73 A rather than the 58 A a 28 V part would need at the same efficiency.

Can it run CW at 900 W? The row says CW and the page says pulse and CW, so the class answer is yes. The honest answer is that no duty cycle, no pulse width and no thermal resistance is published, so the CW rating can only be confirmed by the factory for the flange and the cold plate in the actual build.

How much drive does it take? Not published. At 15 dB it would be 28.5 W and at 18 dB 14.3 W for 900 W, and the shelf carries rows at both gains, so the number has to come from the factory rather than from a neighbouring row.

Is there a data sheet? No document is linked from the page, and no orderable addendum or packing column is published. A buyer who needs a controlled document for an incoming specification should ask for one and hold the order until it arrives.

What is the closest part with a published document? The single ended 200 W row on the same shelf carries a data sheet with maximum ratings, thermal resistance, a typical performance table over 410 to 600 MHz and a fixture drawing. It is a different power class, but it shows what the factory can publish when it chooses to, and it is the template to ask for on this part.

For pricing and lead time on the HXC50113 against the neighbouring rows, send the quantity, the sub band and the rail to the RF component sourcing desk. For a replacement against an existing part number, the cross reference page is the faster route, because it starts from the incumbent and works back to what is on the shelf.