Contact Us

HXC50108

Brand: Hisiwell
Hisiwell HXC50108 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

HXC50108 — Discrete & Power Semiconductors | Superb Automation

HXC50108

Hisiwell — Discrete & Power Semiconductors

Product Overview

The HXC50108 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

HXC50108: 400 W on a 50 V rail, and the exact 200 V that makes it the top of the line

The HXC50108 is the largest pair on the Hisiwell HF to 1.0 GHz line: 200 W per device, two devices, 400 W total, on a 50 V rail, in CW with the vendor page adding that it supports both pulse and CW. There is no gain and no efficiency on the index row, and no document was found for it in the copy read this round, so everything below is either the published row, the page text, or arithmetic that a buyer can repeat.

The arithmetic is worth doing because it lands on a round number that tells the whole story. 400 W into 50 ohm is 141.4 V rms and exactly 200.0 V peak. Against a 50 V rail that is 4.00 times the supply. This is a push pull pair with each device driving half the load line, and each drain therefore swings to twice the rail, which is where a 50 V part reaches 100 V on the drain and where the 110 V breakdown of the neighbouring 260 W document starts to look less generous than it did at first reading.

Put differently, this row is the one where the rail, the breakdown voltage and the load line stop having spare room between them. A 400 W pair on 50 V is a legitimate product and the vendor sells it as such, but the buyer who takes it is buying a part whose margin lives in the matching network and the heat sink rather than in the die. That is the difference between this row and the 260 W row below it, and it is the reason both exist on the shelf.

The HXC50108 rows as the vendor publishes them

Part classLDMOS RF power transistor, thermally enhanced, unmatched, two devices in a pair
Frequency range, index rowHF to 1 GHz
Output, index row200 W per device, 400 W for the pair
Output in dBm, calculated53.01 dBm per device, 56.02 dBm for the pair
Step up from the 130 W pair, calculated1.87 dB
Step down from the 500 W pair at HF to 500 MHz, calculated0.97 dB, on a different band
Drain rail, index row50 V
Mode, index row and pageCW, page adds pulse and CW
Load voltage at 400 W into 50 ohm, calculated141.4 V rms, 200.0 V peak
Peak against the 50 V rail, calculated4.00 times
Load voltage at a 2 to 1 mismatch, calculated400 V peak
Load voltage at a 3 to 1 mismatch, calculated600 V peak
DC input at 60 per cent efficiency, calculated667 W, 13.33 A at 50 V
Heat to remove at 60 per cent, calculated267 W
DC input at 55 per cent efficiency, calculated727 W, 14.55 A at 50 V
Heat to remove at 55 per cent, calculated327 W
DC input at 70 per cent efficiency, calculated571 W, 11.43 A at 50 V
Heat to remove at 70 per cent, calculated171 W
GainNot published on the row or the page
EfficiencyNot published on the row or the page
Thermal resistance, junction to caseNot published for this row
Breakdown voltage, maximum ratingsNot published for this row
PackageDescribed as thermally enhanced, body style not published for this row
Orderable addendumNone published
Moisture level and packing quantityNot published
Applications listed on the pageIndustrial, scientific and medical equipment, FM radio, VHF television

What 400 W on a 50 V rail asks of the supply and the output network

The current is the first constraint. At 60 per cent efficiency the pair draws 13.33 A from a 50 V bus and at 55 per cent it draws 14.55 A. Those are not numbers to carry on a two layer board with a thin copper pour: a 50 V bus at 14 A with a few milliohms in the wrong place loses a volt and a half, which is 3 per cent of the rail and shows up as compression at the top of the drive range. The bus bar or the heavy copper that carries this current is part of the amplifier, not an accessory to it.

The voltage on the load is the second constraint. 200 V peak at 400 W means the matching section, the combining transformer and the connector all have to be rated above that with margin, and the mismatch case pushes 400 V and 600 V across the same parts. On a pair at this power the output combiner is usually a transmission line transformer wound on a ferrite that saturates well before the capacitors fail, so the ferrite material and its flux swing belong on the specification, next to the capacitor voltage rating.

The third constraint is the one buyers forget: the pair is driven in quadrature, so the two halves have to be matched in amplitude and phase. A tenth of a decibel of amplitude error between the halves puts a fraction of the power into the reject load and shows up as heat in a resistor that no thermal calculation on the drawing has accounted for. At 400 W the reject load is not a diagnostic aid, it is a component with a power rating.

Thermal: what the missing figure costs the buyer

This row publishes no thermal resistance, so the junction temperature cannot be computed from the vendor data. What can be done is to bound it with the figure the neighbouring 260 W row publishes, 0.5 degrees per watt, and label the result as an estimate. At 60 per cent efficiency the heat is 267 W; at 0.5 degrees per watt that is a 133 degree rise, and at an 85 degree case the junction would be at 218 degrees against a 225 degree limit. Seven degrees of margin on an estimate is not margin, it is a question for the vendor.

Two things follow. The first is that a buyer who needs this power level should ask the factory for the thermal resistance and the maximum junction temperature in writing before the heat sink is specified, because the difference between 0.5 and 0.4 degrees per watt moves the junction by 26 degrees at this heat load. The second is that the case temperature becomes a measured quantity rather than an assumed one: a thermocouple on the flange, read at thermal equilibrium at full power, is the only honest input to the calculation.

If the enclosure cannot hold the flange at 85 degrees, the answer is not a bigger fan at the same power. It is the 260 W row below this one, which publishes 0.5 degrees per watt and 260 W, or it is two 200 W pairs combined, which spreads the same heat over twice the flange area and turns one difficult thermal problem into two easier ones.

Gain, drive and the consequence of an unpublished figure

No gain is published for this row. On the same shelf, the pair at 260 W publishes 19 dB and the pair at 60 W publishes 23 dB, so the reasonable expectation for a 400 W pair is between 15 and 19 dB, but an expectation is not a specification. Until the vendor confirms it, the driver has to be sized on the pessimistic end: at 15 dB the drive for 400 W is 12.7 W, and at 19 dB it is 5.0 W. A driver sized on 5 W and installed behind a part that needs 12.7 W will not reach rated output, and the failure will look like a device problem rather than a drive problem.

The same caution applies to the input match. An unmatched LDMOS pair at this power has an input impedance that falls with frequency and a return loss that is typically between minus 6 and minus 10 dB across the band. The driver sees a load that changes with frequency, so the gain flatness of the complete amplifier is the sum of the driver flatness and the pair flatness, and the pair flatness is unknown for this row. Measure the assembled amplifier across the band before the gain budget is frozen.

The practical route is to buy a small quantity, measure the pair on the intended fixture, and only then release the driver specification. That sequence costs a week and avoids a rebuild of the driver board after the first build, which costs a month.

Where the HXC50108 sits, and what to compare it with

The ladder above this row on the same 50 V rail is the 500 W pair at HF to 500 MHz and the 700 W pair at the same band edge, both of which give more output but stop at 500 MHz. Below it are the 260 W pair at HF to 1 GHz and the single ended rows at 200, 170, 140, 100 and 80 W. So this row is the top of the 1 GHz ladder and the point where a system that needs more than 300 W across HF to 1 GHz has to decide between this pair and two smaller pairs combined.

The comparison that matters most is with the 260 W pair, because the two are 1.87 dB apart and share a band, a rail and an application list. The 260 W part publishes gain, efficiency, thermal resistance, breakdown voltage, capacitances and a full broadband table. The 400 W part publishes a row and a paragraph. A buyer who needs the last 1.87 dB takes the larger part and pays for measurement; a buyer who needs a documented thermal path takes the smaller part and combines two if the power is short. Both are legitimate, and the difference is documentation rather than silicon.

Against the single ended 200 W row the choice is simpler: two 200 W singles in a pair give the same 400 W with two devices to heat sink and two gate biases to set, which is more work at assembly and more parts to match. The pair in one package is the better route when the flange can take the heat. The ladder with drive, current and heat at each rung is laid out on the RF power transistor cross reference page, and the supply side of the same decision, including the second source question that follows an unpublished thermal figure, sits on the one stop RF supply page.

At goods inwards, this row needs three checks beyond the label: the flange flatness, because a pair at 400 W lives or dies on the case to heat sink interface; the gate to source resistance, because a part with a damaged gate oxide reads a short and looks like a handling failure; and the packing quantity against the purchase order, because no addendum is published and the reel or tray count is whatever the vendor ships. The inspection sequence for those three checks is on the incoming inspection plan page.

Questions buyers ask about the HXC50108

Is 400 W one device or two?

Two. The index row is 200 W per device, the page calls the part a 400 W transistor, and the arithmetic is 200 W plus 200 W, which is 3.01 dB. Each half needs its own gate bias, and the pair has to be driven in quadrature through a combiner, so the part is a subsystem rather than a single ended stage.

What heat sink does 400 W need?

At 60 per cent efficiency the heat is 267 W, and if the flange has to stay at or below 85 degrees with a 45 degree ambient inside the box, the sink to ambient resistance has to be about 0.15 degrees per watt. That is a forced air tunnel with a copper base or a liquid cooled plate, not an extruded sink in still air. At 55 per cent efficiency the heat is 327 W and the requirement tightens further.

Can this pair be run at 40 V?

No figure is published for it. The 260 W row on the same shelf is documented at 40 V with more than 150 W, and if this pair scales the same way its output at 40 V would be about 256 W, which is the square of the voltage ratio applied to 400 W. That is arithmetic and not a vendor rating, so it belongs in a feasibility note rather than in a specification.