HX25600F
Product Specifications
HX25600F
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HX25600F 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
HX25600F: 600 W on the index, 320 W in the parameter list, tables in the sheet
The HX25600F is the largest member of the Hisiwell 2.5 GHz group on the 28 or 48 V rail: a gallium nitride power transistor in a 650F2 body, described as a 600 W device, with a parameter list that says maximum output above 320 W, best drain efficiency above 79 per cent and linear gain above 19 dB. It is also the one member of this band with a data sheet that carries real tables, which makes it the part to read first when the shelf itself has to be understood.
The index and the parameter list disagree by a factor of 1.88, or 2.75 dB, and the data sheet explains the gap without closing it. The sheet's own tables run 170 to 209 W per measurement point across the 2.5 GHz band at 48 V, and 102 to 132 W at 28 V on a single path. So the 320 W parameter is the combined figure, the 600 W index is a class label, and the per point numbers are what a board actually sees.
That is a lot of arithmetic to reconcile one part number, and it is worth doing once, because the same pattern repeats across this whole shelf and only this part documents it. A buyer who understands the HX25600F can read the neighbours correctly without a data sheet for each of them.
The HX25600F rows and the sheet behind them
| Part class | Gallium nitride power transistor, dual path, 650F2 body |
| Frequency range | HF to 2500 MHz |
| Output, index row | 600 W |
| Output, parameter list | Maximum output power above 320 W |
| Index over parameter list, calculated | Factor 1.88, 2.75 dB |
| 600 W in dBm, calculated | 57.78 dBm |
| 320 W in dBm, calculated | 55.05 dBm |
| Gain, index row | 17 dB |
| Gain, parameter list | Linear gain above 19 dB |
| Efficiency, index row | 69 per cent |
| Efficiency, parameter list | Optimal drain efficiency above 79 per cent |
| Supply | 28 or 48 V supply mode |
| Quiescent current in the 48 V table | 1000 mA |
| Quiescent current in the 28 V table | 300 mA per path |
| Thermal resistance, junction to case | 0.55 degrees per watt, both paths combined |
| Maximum junction temperature | 225 degrees, absolute 275 degrees |
| Maximum drain source voltage | 150 V |
| Maximum drain source operating voltage | 55 V |
| Maximum gate source voltage | minus 10 to plus 2 V |
| Maximum gate current, single path | 35.2 mA |
| Mismatch tolerance | 10 to 1 at 360 degrees without damage |
| Storage temperature range | minus 65 to plus 150 degrees |
| Orderable addendum and packing data | None published |
This is what a usable document looks like and it is the standard the rest of the shelf is measured against. A thermal resistance of 0.55 degrees per watt with both paths on a 85 degree base plate and 281.6 W of dissipation is a number a mechanical engineer can work with. The 10 to 1 mismatch tolerance at 360 degrees is a number a test engineer can plan a bench around. The 150 V drain source limit and the 55 V operating limit are the two that stop a supply from being built to the wrong rail.
The gain column is also instructive. The index says 17 dB and the parameter list says above 19 dB, and the data sheet's own table shows 22.6 dB at 600 MHz falling to 14.2 dB at 2.5 GHz. All three are true at different frequencies, and the fall of 8.4 dB across the band is the real constraint on a wideband stage.
The 48 V table point by point
| Frequency | Output | Output in dBm, calculated | Drain efficiency | DC input, calculated | Heat, calculated | Small signal gain |
| 600 MHz | 209 W | 53.20 dBm | 69.4 per cent | 301 W | 92 W | 22.6 dB |
| 1.0 GHz | 209 W | 53.20 dBm | 64.6 per cent | 324 W | 115 W | 21.8 dB |
| 1.3 GHz | 204 W | 53.10 dBm | 61.1 per cent | 334 W | 130 W | 20.4 dB |
| 1.5 GHz | 195 W | 52.90 dBm | 61.9 per cent | 315 W | 120 W | 19.9 dB |
| 2.0 GHz | 186 W | 52.70 dBm | 60.7 per cent | 306 W | 120 W | 18.5 dB |
| 2.5 GHz | 170 W | 52.30 dBm | 62.2 per cent | 273 W | 103 W | 14.2 dB |
The measured conditions are 48 V, 1000 mA of quiescent current, a 100 microsecond pulse at 10 per cent duty, and the part bolted to a heat sink. That duty cycle matters: these are pulse numbers, not continuous ones, so the heat column is an instantaneous figure and not a continuous rating. A customer who runs this part continuous has to re rate the whole table, and the sheet says so by stating the condition on every page.
The output falls 0.9 dB from 600 MHz to 2.5 GHz while the efficiency falls 7.2 points and the gain falls 8.4 dB. Those three curves do not move together, which is what a wideband part looks like and why a single wattage on an index row is such a poor description of it. The heat load peaks at 1.3 GHz, not at the top of the band, because the efficiency dips there.
At 2.5 GHz the load impedance printed in the table is 1.3 minus j3.9 ohm and the source impedance is 1.6 minus j3.7 ohm. Those are the numbers that let a customer start a board without a load pull. On this shelf they exist for one part only, and on the rest of the family the equivalent request has to be made to the factory.
The 28 V table and what it says about the rail claim
| Frequency | Output, single path | Drain efficiency | Heat, calculated | Gain |
| 600 MHz | 110 W | 83.4 per cent | 22 W | 26.8 dB |
| 1.0 GHz | 115 W | 82.1 per cent | 25 W | 26.4 dB |
| 1.3 GHz | 126 W | 81.0 per cent | 30 W | 23.9 dB |
| 1.5 GHz | 123 W | 80.8 per cent | 29 W | 24.2 dB |
| 2.0 GHz | 132 W | 78.3 per cent | 37 W | 20.2 dB |
| 2.5 GHz | 102 W | 75.0 per cent | 34 W | 15.2 dB |
Two things stand out. First, the 28 V efficiency is higher than the 48 V efficiency at every point, 83.4 against 69.4 at 600 MHz, which is normal when a lower rail runs closer to the load line and is the reason a 28 V amplifier is often the cooler of the two at the same output. Second, the 28 V table is labelled single path and the 48 V table is not, so the two sets cannot be compared watt for watt without knowing how many paths each describes.
The 48 V table is described in the sheet with a thermal resistance given for both paths combined and a dissipation figure of 281.6 W across them. That reads as a two path device with the table reporting one path, which would make the combined 48 V output around 340 to 418 W depending on frequency, and that is the number that lands near the 320 W parameter and far from the 600 W index.
For a 28 V build the single path numbers put the combined output near 204 to 264 W, again below the parameter list. That is the arithmetic a buyer should do before a rail is chosen, and it is only possible because this part publishes a table.
Ordering and qualification
No orderable addendum, marking code, packing quantity or moisture level is published even for this part, so nothing of that kind is quoted here. The purchase order should carry the part number, the 650F2 body, the rail, the duty cycle, the frequency of use and the quantity. The gain, the efficiency and the load impedance at the frequency of use should be attached as a measured point rather than quoted from the index.
The data sheet is the one document on this shelf that a buyer can defend a build against, and it is worth sending it with the request for quotation so that the factory prices the same operating point that the customer specified. RF component sourcing covers how those documents travel with an order and what is checked on arrival.
Where two shelves have to be compared on one line, the RF power transistor cross reference is the page to use, and where a second source is being qualified at this band and rail, one stop RF supply sets out what has to be re measured.
Questions buyers ask about the HX25600F
Is it 600 W or 320 W? The index says 600 W, the parameter list says above 320 W, and the tables run 170 to 209 W per point at 48 V. For a specification, 320 W is the defensible combined figure and the tables are what the board sees.
Is the 48 V table continuous? No. It is a 100 microsecond pulse at 10 per cent duty with a bolted heat sink. Continuous operation has to be re rated on the thermal resistance and the junction limit.
Can it be run at 28 V? Yes, there is a full 28 V table on the sheet, single path, 102 to 132 W with 75 to 83 per cent efficiency and a 300 mA quiescent point per path.
What package is the 650F2? The sheet states the body code and the flange, and the mechanical drawing is a legitimate request with the first order. The thermal resistance of 0.55 degrees per watt is given for both paths combined at a 85 degree base plate.
Where this part is being used as the reference for a shelf comparison, the cross reference is the page that keeps the rails and the bands on one axis.