HX339030
Product Specifications
HX339030
Hisiwell — Amplifiers & Comparators
Product Overview
The HX339030 from Hisiwell is a precision amplifier / comparator delivering high gain, wide bandwidth, and low distortion for signal conditioning. 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
- High gain-bandwidth product for wideband signal amplification
- Low input offset voltage and drift for precision DC applications
- EMI-hardened design for robust operation in noisy environments
- Rail-to-rail input/output for maximum dynamic range
- Low quiescent current for power-sensitive designs
- Unity-gain stable with capacitive load drive capability
Key Specifications
| Channels | — |
| GBWP | MHz |
| Slew Rate | V/µs |
| Vos | µV |
| Iq per Channel | µA |
| Input Type | — |
| Supply Voltage | V |
| Operating Temp | °C |
| Package | — |
Applications
HX339030: 36 dB of gain from one die, and the isolation that comes with it
Two parts in this series cover the same 0.9 to 1.3 GHz window. The HX339030 is the high gain one: 36 dB of small signal gain from a single GaAs die, with plus or minus 0.5 dB flatness and a 3.2 dB noise figure. That much gain in one stage changes what the rest of the chain has to do, and it also changes what can go wrong.
Published values
| Parameter | Value |
|---|---|
| Frequency range | 0.9 to 1.3 GHz |
| Small signal gain | 36 dB, flatness plus or minus 0.5 dB |
| Output P-1dB | 25.5 dBm |
| Noise figure | 3.2 dB |
| Input and output VSWR | 1.8:1 both ports |
| Supply | +8 V at 162 mA |
| Form | Bare die |
What 36 dB does to the drive requirement
36 dB of gain means the input needed for 25.5 dBm of output is minus 10.5 dBm, which is a level most local oscillators and DAC outputs can reach without an extra buffer. The trade is that any signal arriving at the input, wanted or not, comes out 36 dB stronger. A minus 30 dBm interferer at the band edge leaves the part at plus 6 dBm, and two such tones can generate products well above the noise floor. This is a part for a clean, filtered input, not for an antenna port.
Current against gain
+8 V at 162 mA is 1.30 W of DC. At 25.5 dBm, or 0.355 W of RF, the stage converts 27 percent of its DC input at compression. Per milliamp the gain is high: about 2.2 dB for every 10 mA, which is why buyers using this part are usually trading board space and stage count rather than current.
Die level handling
This part ships as bare die, so the assembly decisions land on your side of the order: die attach method, wire bond length at the RF ports, and the ground via pattern under the die. Bond wire inductance at 1.3 GHz is small but not zero, so the vendor recommends keeping the RF bonds short and symmetrical. SUPERB can quote the die together with the packaged alternatives in the same series so you can compare assembly cost against a QFN part.
Two questions buyers ask
Can this replace two 18 dB stages? Electrically it can, and it saves one bias network. It does not save the filtering that two stages would have allowed between them, so check the spurious budget before you collapse the chain.
What about reverse isolation? The vendor table does not publish S12. If your chain has a reflective load after this stage, ask for the S parameters before you commit, because 36 dB of forward gain usually comes with a reverse path that is not negligible.
SUPERB supplies the HX339030 and the rest of the HX339 series with datasheet, die drawing and lot traceability per shipment. If you are qualifying a second source for an existing 0.9 to 1.3 GHz stage, start with the amplifier cross reference hub and check the four parameters that decide a swap. For lead time against a build schedule, see component lead time solutions.
The full series index, including the packaged members of this band, is on the amplifiers and comparators page.
HX339030: 36 dB in one stage means the reverse path is a specification too
The published row is unusually strong for a single stage: 36 dB of gain with plus or minus 0.5 dB across 0.9 to 1.3 GHz, 25.5 dBm at compression and a 3.2 dB noise figure, from 8 V at 162 mA as a bare die. What the row does not publish is reverse isolation. At 36 dB of forward gain, the reverse path stops being an academic parameter and becomes the thing that decides whether the stage is stable into a load that reflects.
Stability is a property of the pair, not of the die
An amplifier with 36 dB of forward gain and 1.8 to 1 on both ports has a round trip loop through whatever sits at its output. If the reverse isolation is 18 dB, the loop gain through a load that reflects 10 percent of the power and a source that reflects 10 percent is 36 minus 18 minus 20, which is minus 2 dB at one frequency and could be positive at another. Nothing in the published table lets a buyer compute that number, so the only responsible route is to measure it: terminate the output in a mismatched load, sweep the input return loss, and look for the region where the input return loss goes above 0 dB. Any part that shows a negative resistance region at a band edge will oscillate on some boards and not on others, and the difference is usually a bond wire length rather than a circuit change.
| Item | Published | Consequence |
|---|---|---|
| Gain | 36 dB, plus or minus 0.5 dB | drive for 25.5 dBm is minus 10.5 dBm |
| Noise figure | 3.2 dB | 316 K, usable as a first stage behind a limiter |
| Output at compression | 25.5 dBm | 0.355 W, 27 percent of the 1.30 W DC input |
| Match | 1.8 to 1 both ports | 10.9 dB return loss, 0.352 dB mismatch loss per port |
| Reverse isolation | not published | to be measured with a mismatched load |
Thirty six dB of gain moves the whole level plan into the input connector
With minus 10.5 dBm needed for full output, any signal that reaches the input comes out 36 dB stronger. A minus 20 dBm signal from an adjacent channel, a local oscillator leakage or a test signal left connected arrives at plus 16 dBm, which is 9 dB below compression and therefore still linear and still invisible in a quick check. This is why the part wants filtered input rather than an antenna port, and why an input limiter or a switched attenuator in front of it is cheap insurance. On the receive side the same 36 dB makes the following stage irrelevant: a second stage with a 5 dB noise figure contributes 5 minus 36 dB, which is nothing, so the system figure is set by this die and everything ahead of it.
Bias at 8 V and 162 mA from a bare die
1.30 W of DC with 0.355 W of RF leaves 0.94 W to remove through a die that has no package to spread it. That is the number the attach and the via field have to carry, and at L band the physical size of the die is large enough that the spread is not the limiting factor, provided the ground under the source is continuous. The 8 V rail is worth noting: it is above the 5 V that most digital boards carry, so the supply arrangement is a separate regulator rather than a tap off the logic supply, and the decoupling has to hold at 0.9 GHz, which means a capacitor close to the die and a second one at the feed point rather than a single bulk part at the edge of the board. Parts in this class and the supply position on them are collected on RF component sourcing, and the comparison route when a second source is needed for an L band stage is on how to verify a drop in replacement.
Two questions buyers ask
Does the flatness figure hold at the band edges? Plus or minus 0.5 dB across 400 MHz is a tight figure for one stage and the row does not say where it was measured. Ask for the sweep, and if the edges are 0.3 dB worse than centre the part is still within the row and a level plan that assumed flat across the whole band will be 0.3 dB out at two frequencies, which matters in a calibration loop but not in a link budget.
Can it be cascaded with itself for 72 dB? Mechanically yes, and there is no reason to. Two stages at 36 dB would put 2.6 W of DC into a chain whose first stage already sets the noise figure, and the second stage would contribute nothing to the system figure. If more gain is needed, the cheaper answer is one of these followed by a lower current block, which is the comparison our one stop RF supply page covers when several stages of one family are quoted together.
Request Quote — In Stock, 100% New & Original, Full Warranty