HXNM40004
Product Specifications
HXNM40004
Hisiwell — Discrete & Power Semiconductors
Product Overview
The HXNM40004 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
HXNM40004: ten watts and twenty eight decibels of gain on 2.7 to 3.5 GHz
The HXNM40004 is the widest and highest gain member of the Hisiwell high gain family that covers S band. The vendor index row reads 2.7 to 3.5 GHz, saturated output at least 40 dBm, gain at least 28 dB, drain efficiency at least 45 per cent, 28 V drain rail, continuous wave. That is 10.00 W across 800 MHz with a gain floor that no standard internally matched part in this catalogue reaches.
Eight hundred megahertz at a 3.1 GHz centre is 25.81 per cent fractional bandwidth. That is not a channelised part, it is a band part, and it covers the S band radar allocation, the 3.3 to 3.4 GHz broadband wireless access window and the 2.7 to 3.5 GHz point to point links in one amplifier. For a system that has to serve several of those with one board, the window is the reason to choose it.
The power budget is ordinary. At 45 per cent efficiency 10.00 W of output draws 22.2 W from the rail, which is 0.794 A at 28 V, and leaves 12.2 W of heat. Those are calculated figures from the published rows, and they sit inside what a board with a copper paddle and a via field can carry.
The drive is what makes it different. With 28 dB of gain the input for rated output is 12 dBm, which is 16 mW. A modem output, a synthesizer or a single MMIC amplifier reaches that directly, so the transmit chain is one stage rather than three.
The parameter rows Hisiwell publishes for HXNM40004
| Part class | Gallium nitride power transistor, internally matched, high gain family |
| Frequency range | 2.7 to 3.5 GHz |
| Bandwidth, calculated | 800 MHz, 25.81 per cent fractional at a 3.10 GHz centre |
| Saturation output power as published | at least 40 dBm |
| Saturation output power in watts, calculated | 10.00 W |
| Gain as published | at least 28 dB |
| Drain efficiency as published | at least 45 per cent |
| Drain voltage | 28 V |
| Flatness across the band | not published, the index prints a slash in that column |
| Operating mode | continuous wave |
| Output at 1 dB compression | not published, the index prints a slash in that column |
| DC input at saturation, calculated | 22.2 W |
| Drain current at saturation, calculated | 0.794 A at 28 V |
| Dissipation at saturation, calculated | 12.2 W |
| Dissipation if efficiency falls to 40 per cent, calculated | 15.0 W |
| Drive power for the rated output, calculated | 12 dBm, 15.8 mW |
| Output per 100 MHz of the band, calculated | 1.25 W |
| Load voltage at saturation into 50 ohm, calculated | 22.4 V RMS, 31.6 V peak |
| Thermal resistance, junction to case | not published |
| Package outline and dimensions | not published in text |
| Scattering parameters | not published |
| Stability data and load mismatch tolerance | not published |
| Pinch off voltage and gate bias point | not published |
| Operating and storage temperature | not published |
| Orderable addendum and packing | not published |
| Family neighbour at 36 dBm on 2.7 to 3.1 GHz | HXNM40003, same rail, 25 dB gain |
Twenty five per cent bandwidth is the hard part
A 25.81 per cent window from an internally matched part means the matching network on the carrier is doing most of the work across a very wide span. The published gain floor of 28 dB is a minimum across that window, and it is unusually high for a part this wide, which suggests the match is synthesised rather than tuned in a narrow cell.
What the vendor does not publish is the flatness. The standard family carries plus or minus 0.8 dB in that column and this family prints a slash, so there is no figure for the peak to peak variation across the 800 MHz. For a system with a fixed drive and a link budget built on the 28 dB floor, that missing number is the first thing to establish on a bench.
The wide window also changes the external filter. A single bandpass covering 2.7 to 3.5 GHz is a different component from three narrow filters and a switch, and it usually has more insertion loss at the edges. The system budget should count that loss against the 40 dBm output rather than against the amplifier alone.
Where the application is a radar that uses only part of the window, the extra bandwidth costs nothing and buys margin against a channel change. Where the application is a communications link, the filter at the output decides what reaches the antenna and the amplifier is simply wide enough to stay out of the way.
One stage instead of three
A conventional chain for 10 W at 3 GHz is a 20 dBm source, a 1 W driver and a 10 W output stage, each with its own bias and match. With 28 dB of gain the HXNM40004 takes the 12 dBm source directly, so the chain is one device, one bias network and one thermal path.
The saving in board area is significant but the saving in risk is larger. Two interstage matches disappear, and each of those is a place where the flatness and the stability of the chain can go wrong. What remains is one input match to the source and one output match to the filter.
The risk that arrives with high gain is instability. A 28 dB part with an output filter that presents a reactive load out of band can oscillate, and with no stability data published the cure has to be built in: a small resistive pad at the input, an isolator at the output, or both. On a 10 W stage the isolator is the usual answer.
Gain also falls with temperature and no coefficient is published. A part with 28 dB at 25 C may have 26 dB at 85 C, which moves the required drive by 2 dB. A source with a fixed output level should be checked at the top of the temperature range, and a source with automatic level control covers the drift in hardware.
The 2.7 to 3.5 GHz applications
The S band radar allocation is the primary use for this window, and 10 W with a 12 dBm drive is a useful level for a short range surveillance radar, a radar altimeter or a target simulator. The continuous wave rating suits a simulator, which transmits steadily rather than in pulses.
The 3.3 to 3.4 GHz band is used for broadband wireless access in several regions, and a 10 W stage with 28 dB gain fits an outdoor unit where the drive comes from the modem and the antenna is a sector panel. The 800 MHz window covers the whole allocation with a margin at both edges.
The 2.7 to 3.5 GHz point to point links are also inside the window. For those the linearity at backoff matters more than the saturated output, and with no compression point published the backoff has to be established by measurement rather than read from a curve.
Where the product is a radar or an electronic warfare module, the traceability around the part matters as much as the electrical data. Our defense grade components page covers the documentation and the chain of custody that such a program normally asks for.
Bias, heat and the first article
Twelve watts of heat on a carrier offering 5 C per watt rises 61 C over the case, and on 3 C per watt it rises 37 C. At a 70 C ambient the first case sits at 131 C and the second at 107 C, so the carrier choice decides whether the part is inside a 100 C limit or outside it.
No junction to case figure is published, so the thermal answer is a measurement. A thirty minute run at saturation with a thermocouple on the flange and one on the sink gives the number that matters, and it also shows whether the bolt torque and the thermal interface are doing their job.
The gate bias point is not published, and with 28 dB of gain the bias is more sensitive than on a low gain part. The first article should record a bias sweep against drain current and against gain, and the production test should check the quiescent current at the chosen point rather than the gain alone.
No moisture level, packing or marking is published either. For a production program the factory should confirm all three before the assembly flow is fixed, because a part that needs baking changes the schedule. The one stop RF supply route covers the confirmation and the holding arrangement for a build that runs over several months.
Choosing inside the high gain family
The family has nine members and the choice is driven by band and output together. On 2.7 to 3.1 GHz the 36 dBm member is HXNM40003 with 25 dB gain, and on 2.7 to 3.5 GHz this part gives 40 dBm with 28 dB. The wider part is also the more powerful one, which is not always the case in a family.
Above this part the family stops. The highest output in the high gain group is 44 dBm at 9.3 to 9.5 GHz, and at lower frequencies 43 dBm on 4.4 to 5 GHz. A requirement above that has to move to the standard family, which reaches 58 dBm and gives up gain to do it.
Below this part the family steps down by 4 dB to HXNM40003 at 36 dBm and then to 33 dBm at 5.3 to 5.9 GHz. The gain stays between 25 and 28 dB across those steps, so a product line can move up and down in power without changing the driver.
For a replacement study the fields to compare are the rail, the mode, the band, the gain and the drive level. Our amplifier cross reference page sets out those fields and the bench work needed when a candidate differs on any of them.
Questions buyers ask about HXNM40004
Is the gain figure small signal or at saturation. The vendor prints a single gain column with a minimum, and the industry convention is that the number is the large signal gain at the rated output. A small signal gain is normally higher, so a link budget built on 28 dB is the conservative reading.
Can it be used below 2.7 GHz. Nothing on the page says so, and the match is synthesised for the stated window. Operation outside the window is untested by the vendor and would have to be characterised before it could be relied on.
What is the compression point. Not published. For a modulated signal the backoff has to be established on a bench, and the usual starting point for a part with this much gain is 6 to 8 dB below saturation.
How is it ordered. There is no addendum and no packing column, so the purchase order names the band, the rail and the quantity. The part is held with the rest of the S band family through the RF component sourcing route, and a build that needs several members can be consolidated into one shipment.