HMC464-Die
Product Specifications
HMC464-Die
Analog Devices — Analog & Mixed-Signal ICs
Product Overview
The HMC464-Die from Analog Devices is a high-performance analog / mixed-signal IC providing precision signal acquisition, conversion, and conditioning. Analog Devices (ADI) is a world leader in high-performance analog, mixed-signal, and DSP ICs for instrumentation, healthcare, automotive, and communications.
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-resolution data conversion with excellent SNR and THD
- Wide dynamic range and high linearity across operating conditions
- Integrated voltage reference with low temperature drift
- Flexible digital interface — SPI, I2C, or parallel options
- Industrial/automotive temperature range with guaranteed specs
- Low power consumption with power-down modes
Key Specifications
| Resolution | bits |
| Channels | — |
| Sample Rate | kSPS |
| Interface | — |
| Supply Voltage | V |
| Operating Temp | °C |
| Package | — |
| INL | LSB |
| SNR | dB |
| THD | dB |
Applications
Request Quote — In Stock, 100% New & Original, Full Warranty
HMC464 die: 2.32 W of DC for 26.5 dBm at 2 GHz to 20 GHz
The HMC464 is a gallium arsenide pHEMT distributed power amplifier die covering 2 GHz to 20 GHz with 14 dB of gain, 26.5 dBm at one decibel compression, 27.5 dBm saturated and 32 dBm of output third order intercept, from a single 8 V supply at 290 mA. It carries two gates, VGG1 for the drain current and VGG2 for the bias of the output stage, and VGG2 is referenced to the drain, with an absolute maximum range from VDD minus 8 V to VDD. The published table was taken in a 50 ohm fixture, and the thermal figures in the vendor document are given channel to ground paddle, which is the packaged path and not the die path.
Published values
| Parameter | Value |
|---|---|
| Frequency range | 2 GHz to 20 GHz, specified in three bands |
| Gain | 14 dB typical 2 to 6 GHz, 13.5 dB 6 to 16 GHz, 11 dB 16 to 20 GHz |
| Gain flatness | plus or minus 0.5 dB to 16 GHz, plus or minus 1.0 dB to 20 GHz |
| Gain variation over temperature | 0.025 to 0.06 dB per degree depending on band |
| Noise figure | 4.0 dB typical 2 to 16 GHz, 6.0 dB 16 to 20 GHz |
| Input return loss | 15 dB low band, 10 dB mid band, 7 dB top band |
| Output return loss | 15 dB, 9 dB and 11 dB across the three bands |
| Output power at 1 dB compression | 26.5 dBm typical 2 to 6 GHz, 25 dBm 6 to 16 GHz, 21 dBm 16 to 20 GHz |
| Saturated output power | 27.5 dBm, 26 dBm and 24 dBm across the three bands |
| Output third order intercept | 32 dBm, 26 dBm and 22 dBm across the three bands |
| Supply | 8 V at 290 mA, 9 V absolute maximum drain |
| Gate bias | VGG1 sets 290 mA, VGG2 at 3 V typical, range VDD minus 8 V to VDD |
| RF input power | 20 dBm at 8 V drain, absolute maximum |
| Thermal resistance | 19.4 degrees per watt channel to ground paddle |
| Channel temperature | 150 degrees maximum |
| Continuous dissipation | 3.35 W at an 85 degree base, derate 51.5 mW per degree |
Two point three watts of DC and the arithmetic of the output stage
8 V at 290 mA is 2.32 W of DC going in. At the 26.5 dBm compression point the output is 447 mW, so 1.87 W becomes heat, and the saturated point at 27.5 dBm gives 562 mW out and 1.76 W of heat. That is the opposite of a small signal part: the dissipation is set by the bias and moves very little with drive, so the thermal answer has to be right before the first board is built rather than measured afterwards.
The 19.4 degrees per watt figure is quoted channel to ground paddle and belongs to the packaged order code, so on die the number is different and depends on the attach. What the die does inherit is the 150 degree channel limit and the 3.35 W dissipation at an 85 degree base, which is 65 degrees of allowed rise and therefore 19.4 degrees per watt for the whole 2.32 W. Any attach that adds more than a couple of degrees per watt of its own takes the margin below what the vendor assumed, which is why void percentage under a die this size is a process parameter and not a cosmetic check.
VGG2 is referenced to the drain, and that changes the bias sequence
The absolute maximum for VGG2 is written as VDD minus 8 V to VDD, so at 8 V of drain the gate may sit anywhere from 0 V to 8 V and the operating point is 3 V. That is a cascode gate, not a second control like the AGC pin on the smaller amplifier, and it has to come up in the right order. The safe sequence sets VGG1 at its negative end, brings VDD to 8 V, brings VGG2 to 3 V and then walks VGG1 up until the drain meter reads 290 mA.
Getting that order wrong is the usual way a first article dies. If VGG2 is left at 0 V while VDD is applied, the cascode device is off and nothing bad happens, which tempts an assembly to treat the pin as optional. The failure comes later, when the bias is reworked and VGG1 is at its nominal value while VGG2 is driven from a logic rail that comes up last; for the interval between the two the output device sees full drain voltage with no gate control, and the part conducts whatever it can until the second rail arrives.
Where the 2 GHz to 20 GHz table stops being usable
The three bands are not one amplifier with small tolerances. Compression falls from 26.5 dBm to 21 dBm and the intercept from 32 dBm to 22 dBm between the low and top bands, and the noise figure rises from 4 dB to 6 dB at the same time. A link budget written on the headline row of this table is 4.5 dB optimistic in power and 10 dB optimistic in linearity at 20 GHz, so the band of operation has to be fixed before a part is chosen.
The top band is also where the return loss is weakest, 7 dB at the input over 16 GHz to 20 GHz, which means a 2:1 voltage standing wave ratio. In a chain where the preceding stage is a filter with its own reflection, that mismatch turns into ripple in the gain response, and the ripple then reads as a spurious modulation on a wideband signal. A pad or an isolator at that interface is cheap insurance on a part that dissipates 1.9 W.
Sourcing notes for this family: the RF power transistor cross reference holds the cross references, how to choose an RF amplifier is the pre substitution check, and the RF component sourcing page is what we use when the part on the drawing is not on the shelf.
Two questions buyers ask
Can the HMC464 be run from a single positive supply? No. VGG1 sits below the source to set 290 mA and VGG2 sits at 3 V referenced to the drain, so a two rail bias is needed with the gate applied first.
Is 27.5 dBm available across 20 GHz? No. That is the low band figure. It falls to 26 dBm from 6 GHz to 16 GHz and 24 dBm from 16 GHz to 20 GHz.