HXN50256
Product Specifications
HXN50256
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HXN50256 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
HXN50256: one hundred and twenty watts peak on 0.7 to 2.7 GHz for multi carrier base stations
The HXN50256 is the largest of the Hisiwell 0.7 to 2.7 GHz group, an unmatched gallium nitride HEMT on 28 V. The prose calls it a 120 W part for 700 to 2700 MHz and names MC-GSM, WCDMA and LTE as the signals it is meant for. The family index prints 120 W in the peak column and marks the mode as continuous wave, and it publishes no other electrical figure at all.
The naming of MC-GSM, WCDMA and LTE is the useful part of the page. Those are multi carrier and wideband signals with a crest factor, and a 120 W peak part driving them runs at an average well below the peak. The vendor gives no peak to average figure and no compression point, so the average at which the part is linear has to be measured. That is the single most important missing row on this page.
One hundred and twenty watts is 50.79 dBm. At 45 per cent drain efficiency the DC input is 266.7 W, the drain current is 9.52 A at 28 V and the heat is 146.7 W. At 50 per cent it is 240.0 W, 8.57 A and 120.0 W of heat. At 55 per cent it is 218.2 W, 7.79 A and 98.2 W. Even the best case leaves about 100 W to remove, which is liquid cooling or a substantial forced air sink rather than a finned block in free air.
The step down to the 60 W member is 3.01 dB and down to the 40 W member is 4.77 dB. On a level plan this part is the last stage of a base station line up, and the stage in front of it has to deliver 1.2 W to 3.8 W depending on the gain that the part actually has, which the vendor does not publish.
The parameter rows behind the HXN50256
| Part class | Unmatched gallium nitride HEMT power transistor |
| Frequency range | 0.7 to 2.7 GHz, 700 to 2700 MHz in the prose |
| Power, index row | 120 W, published in the peak column |
| Power in dBm, calculated | 50.79 dBm |
| Named signals in the prose | MC-GSM, WCDMA, LTE |
| Peak to average figure | Not published |
| Compression point | Not published |
| Saturation output | Not published |
| Power gain | Not published |
| Drain efficiency | Not published |
| DC input at 45 per cent, calculated | 266.7 W, 9.52 A, 146.7 W of heat |
| DC input at 50 per cent, calculated | 240.0 W, 8.57 A, 120.0 W of heat |
| DC input at 55 per cent, calculated | 218.2 W, 7.79 A, 98.2 W of heat |
| Drain rail | 28 V |
| Operating mode, index row | Continuous wave |
| Load voltage at 50 ohm, calculated | 77.5 V RMS, 109.5 V peak |
| Load voltage under a two to one mismatch, calculated | about 155 V peak |
| Step above the 60 W member, calculated | 3.01 dB |
| Step above the 40 W member, calculated | 4.77 dB |
| Matching | Unmatched, external network on the customer board |
| Body and package | Not published |
| Thermal resistance, junction to case | Not published |
| Input and output impedance | Not published in text |
| Gate bias point and quiescent current | Not published |
| Orderable addendum, marking and packing | Not published |
Nine and a half amps on the drain and a hundred watts in the sink
The DC side is where this part stops being easy. Nine and a half amps at 28 V through a board needs 250 mil of one ounce copper or 125 mil of two ounce, and the choke that carries it sits on the same geometry, so the drain feed is a copper pour rather than a trace. The connectors at the rail are rated well above ten amps and the bulk capacitance bank is sized for the envelope current, not the average.
The heat side is harder. Even at 55 per cent, which is the optimistic end of this class, the part leaves 98.2 W. A forced air sink with a 0.3 degree per watt interface can hold that at a 30 degree rise, and that is a large sink with a fan. A liquid plate holds it easily and costs more. A finned block in free air does not hold it at all, and a project that plans for one is planning a thermal shutdown.
The output match carries 109.5 V peak matched and about 155 V peak under a two to one mismatch. The capacitors in that network have to be 250 V parts or better, and the series element carries the full load current at that voltage. This is the row where a 120 W GaN board differs from a 40 W one far more than the datasheet power suggests.
The input match is the second half of the problem. No gain is published, so the drive level is a measurement, and no input impedance is published, so the driver sees a load that changes across two gigahertz. A 20 dB gain assumption puts the drive at 1.2 W; a 15 dB assumption puts it at 3.8 W. Those are more than a decibel apart and they are different driver stages.
HXN50256 against the shelf
| HXN50253 | 0.7 to 2.7 GHz | 40 W peak | Not published | Not published | 4.77 dB below |
| HXN50254 | 0.7 to 2.7 GHz | 60 W peak | Not published | Not published | 3.01 dB below |
| HXN50256 | 0.7 to 2.7 GHz | 120 W peak | Not published | Not published | Reference point |
| HX1020120P2 | 1 to 2 GHz | 140 W | 20 dB | 80 per cent | 0.67 dB above, one gigahertz of band |
| HX1020240P2 | 1 to 2 GHz | 280 W | 20 dB | 80 per cent | 3.68 dB above, two rails |
| HXN50240 | 3 to 4.2 GHz | 200 W | Not published | Not published | 2.22 dB above, different band |
Against the narrow band HX1020120P2 this part gives 0.67 dB less output over nearly three times the band, and it gives up a published 20 dB of gain and 80 per cent efficiency to do it. A product that only needs 1 to 2 GHz should be built on the HX1020120P2; a product that has to cover the whole LTE group has no other 120 W member in the shelf.
Against the higher band members the choice is band, not power. The HXN50240 covers 3 to 4.2 GHz at 200 W, but it cannot serve a 700 MHz carrier. Buying two parts for two bands is normal at this power level, and the RF power transistor cross reference is where the two are put side by side before the line up is fixed.
Where the part is out of stock, backing a 200 W member down to 120 W is the usual workaround rather than switching vendors, and the component lead time solutions page covers what that does to the bias point and the match.
Board notes for a hundred and twenty watts on two gigahertz of band
The output transformer is the first item. At 109.5 V peak and a 155 V mismatch peak a lumped match is not practical, so the output is a transmission line transformer on a high frequency laminate, with the secondary referenced to a solid ground plane on both sides. The copper weight and the dielectric thickness set the impedance, and both have to be fixed before the board is ordered.
The bias is the second item. No gate voltage and no quiescent current is published. A 120 W GaN device at class AB draws a substantial quiescent current, and the drain supply has to hold the rail through the envelope of a multi carrier signal without sagging, which means the bulk bank is sized in joules rather than in microfarads alone.
The sequencer is the third item. Negative gate first, clamped, then drain, and the reverse on shutdown. On a base station this is normally a hot swap controller with a sequenced enable, and the gate clamp has to survive the failure of the negative supply rather than assume it will not fail.
The thermal interface is the fourth. With no junction to case figure published, the case temperature at rated output is the only number that closes the loop. It is measured on the first article in the final enclosure, and the acceptance criterion is written before the measurement rather than after it.
Questions buyers ask about HXN50256
What average power can it run on an LTE carrier? Not published. The peak figure is 120 W and the peak to average ratio of the signal has to be subtracted, so the average at which the part stays linear is measured on the first article and stated in the test report rather than assumed from the datasheet.
What is the gain? Not published, so a line up cannot be sized from the vendor text alone. We ask the factory for the gain at the band edges with the order, and the drive stage is sized from that answer.
Is the 120 W a continuous wave figure? The index marks the row as continuous wave while filling the peak column. For a multi carrier signal the honest reading is that the peak is 120 W and the average is set by linearity.
Do you supply a reference match? The vendor publishes no impedance data for this part, so a measured set or a reference board has to come from the factory, and it should be requested with the band edges and the board material named.
How do we start? Send the carriers, the band edges, the average output wanted at the antenna, the ambient and the rail. We will confirm the power and gain figures with the factory, quote from the current schedule and ship a sample measured at the band edges with the drive level recorded.