HMC994A-Die
Product Specifications
HMC994A-Die
Analog Devices — Analog & Mixed-Signal ICs
Product Overview
The HMC994A-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
HMC994A die: 25 percent power added efficiency from DC to 28 GHz
The HMC994A is a gallium arsenide pHEMT distributed wideband power amplifier covering DC to 28 GHz with 15 dB of gain, 29 dBm of saturated output, 38 dBm of output third order intercept and 25 percent power added efficiency, from a single 10 V supply at 250 mA. The vendor document on file describes the surface mount order code, so the electrical table belongs to the packaged amplifier and the die is the same part before the package adds its transitions. The efficiency claim is the headline, and it is worth checking against the DC point before a thermal budget is set.
Published values
| Parameter | Value |
|---|---|
| Frequency range | DC to 28 GHz, specified in three bands |
| Gain | 14 dB typical DC to 10 GHz, 15 dB 10 to 20 GHz, 16 dB 20 to 28 GHz |
| Gain flatness | plus or minus 0.5 dB inside each band |
| Gain variation over temperature | 0.008 to 0.016 dB per degree |
| Input return loss | 12 dB across the band |
| Output return loss | 25 dB, 22 dB and 20 dB across the bands |
| Output power at 1 dB compression | 28 dBm, 28 dBm and 27 dBm typical |
| Saturated output power | 30 dBm, 29.5 dBm and 28 dBm |
| Output third order intercept | 40 dBm, 38 dBm and 36 dBm at 16 dBm per tone |
| Noise figure | 4.0 dB, 3.5 dB and 4.0 dB |
| Supply | 10 V at 250 mA, 2.5 W of DC |
| Gate bias | VGG2 at 3.5 V typical, VGG1 adjusted for 250 mA |
| Power added efficiency | 25 percent, as stated in the vendor description |
| Package reference | 32 lead 5 mm by 5 mm, 25 square millimetres |
Checking 25 percent efficiency against the DC point
10 V at 250 mA is 2.5 W of DC. The saturated output in the low band is 30 dBm, which is 1.0 W, so the drain efficiency at that point is 40 percent before the input power is subtracted. The vendor headline of 25 percent therefore describes a different operating point, either a backed off one or a different bias, and both numbers can be right at the same time. A thermal budget that assumes 25 percent and a 1 W output will overestimate the heat by about 0.4 W, which on a die with a few square millimetres of area is not a rounding error.
The safe approach is to calculate from the DC point and treat efficiency as a curve rather than a number. At 2.5 W of DC the heat is 1.5 W when the output is 1 W and 2.5 W when the output is zero, so the worst case thermal condition is not full drive but idle. That reverses the usual assumption and it is the reason a distributed power amplifier needs its thermal path sized for the bias and not for the signal.
Forty decibels of intercept at 16 dBm per tone
The third order intercept is specified at 16 dBm per tone, which is a high test level and the right one for a part that is meant to work at 28 dBm. Two tones at 16 dBm each are 24 dB below the 40 dBm intercept in the low band, so third order products land about 48 dB below the carriers. In the top band at 36 dBm the same tones are 20 dB below the intercept and the products are 40 dB down.
That is the number that decides whether the part can sit in front of a multiplexer without a driver. A system needing 45 dB of spurious free range has it up to about 20 GHz and loses it above. It also explains why the vendor quotes the intercept at a fixed tone level: an intercept that is measured at 0 dBm per tone would look better and describe nothing, because the part is never used there.
Die versus packaged, on a part that runs at 250 mA
0.375 W of gate and drain bias current per millimetre of die width, on a device that is a few millimetres wide, is a heat flux that the backside has to carry. On die the path is attach and carrier; in the package it is die attach, lead frame and solder. The vendor quote of 25 percent efficiency and the 25 square millimetre package both come from the packaged measurement, so a die assembly that uses a thick epoxy layer can be worse than the package despite removing a transition.
For a programme choosing between the two, the deciding question is usually volume and the assembly capability rather than the last half decibel. Below a few hundred units a year the packaged part is the cheaper route even with a worse port match, and above that the die becomes worth the attach development, provided the carrier and the void specification are written down.
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 component lead time solutions is what we use when the part on the drawing is not on the shelf.
Two questions buyers ask
What does 25 percent efficiency mean in heat terms? Treat it as a headline for one operating point. At 10 V and 250 mA the DC input is 2.5 W, so the heat is 2.5 W minus the output power, and the worst case is at low drive rather than at saturation.
Is the intercept specified at a usable level? Yes, at 16 dBm per tone, which is close to the working level of the part, so the 40 dBm figure describes the real application rather than a small signal extrapolation.