When a product brief says "60% drain efficiency at Ku-band," hardware engineers face an immediate question: measured under what conditions, at what output power, at what temperature, and over what bandwidth? The number by itself tells you almost nothing. The methodology behind it tells you everything.
This article works through what drain efficiency actually measures at Ku-band, the measurement conditions that separate a useful spec from a marketing figure, and what a legitimate 60% number means for your phased array thermal budget. We frame Falcomm's own design targets for the FA-2400 series as design goals under validation, not production specifications.
What Drain Efficiency Actually Measures
Drain efficiency (DE) is defined as RF output power divided by DC input power: DE = P_out_RF / P_DC. It does not subtract the RF drive power from the denominator. Power-added efficiency (PAE) does: PAE = (P_out_RF - P_in_RF) / P_DC.
At Ku-band with a typical PA gain of 18-24 dB, the RF drive power is 1.5-4% of the output power. The difference between DE and PAE at those gain levels is 1-3 percentage points. For a two-stage PA chain with lower total gain, the gap widens. Both metrics have legitimate uses; the key is knowing which one is on the datasheet you are reading.
DE is the more conservative figure because it ignores drive power. A device showing 60% DE at 20 dB gain has a PAE of approximately 59.4% at that operating point. The distinction becomes significant if you are building a cascade where the driver PA's efficiency also enters your link budget.
Measurement Conditions That Change the Number
The same device measured under different conditions can produce drain efficiency numbers that differ by 10-15 percentage points. Four variables dominate.
Output Power Operating Point
PA efficiency peaks near saturation and drops at back-off. A device quoted at 60% DE at P_sat may deliver 38-45% DE at P1dB minus 4 dB, which is a more realistic operating point for a carrier running a high-order modulated waveform. Always ask for the efficiency versus output power curve, not just the peak figure.
Case Temperature
GaN device transconductance and electron mobility both decrease with temperature. A PA measured at 25 degrees C case temperature will show 3-6 percentage points higher efficiency than the same device at 65 degrees C case temperature, which is a routine condition for a module in an outdoor unit under direct sun loading. Datasheets that specify DE only at 25 degrees C leave you with an optimistic starting point.
CW Versus Modulated Signal
Continuous-wave (CW) efficiency measurements are cleaner and easier to take but do not reflect the PA behavior under a wideband modulated carrier. A DVB-S2 signal or an OFDM waveform has a peak-to-average power ratio of 6-10 dB, which means the instantaneous envelope swings across a wide range of operating points. Efficiency under a real waveform depends on the AM-AM and AM-PM curves across that swing range, not just efficiency at a single CW power level.
Frequency Point Within the Band
Ku-band spans 12-18 GHz. A PA optimized for 14.5 GHz center may show 55-60% DE at center and 45-50% DE at the band edges. A spec quoted at center frequency without band-edge data is incomplete for a system that must operate across the full allocation.
Translating Drain Efficiency to Phased Array Thermal Budget
The thermal consequence of drain efficiency flows from one identity: power dissipated equals DC input minus RF output. P_diss = P_DC * (1 - DE) = P_out_RF * (1/DE - 1).
For a 64-element Ku-band phased array with each element outputting 2W RF, the thermal budget difference between 40% DE and 60% DE is substantial. At 40% DE: total DC input is 320W, dissipated heat is 192W. At 60% DE: total DC input is 213W, dissipated heat is 85W. The 107W reduction in heat is not a marginal improvement. It determines whether the array can be cooled passively or requires active liquid cooling, and it directly controls how densely the elements can be packed before thermal crosstalk between adjacent elements forces a power derating.
Element spacing in a Ku-band phased array is set by the beamforming requirement: typically lambda/2 or roughly 10-11 mm at 14 GHz. At that pitch, the heat from adjacent elements overlaps in the thermal spreading layer. If each element dissipates 3W at 40% DE, the center of a 64-element array sees a thermal environment roughly 15-25 degrees C hotter than the edge, depending on the thermal via design and the heat spreader conductivity. At 60% DE with 1.3W dissipation per element, the same gradient shrinks to 8-15 degrees C, keeping all elements closer to their calibrated operating conditions and reducing the gain and phase drift across the array.
Back-Off Efficiency: Where the Comparison Really Plays Out
System-level analysis rarely allows a PA to run at P_sat. LEO and MEO uplink terminals, 5G base stations, and radar transmitters all operate at output back-off to meet their modulation quality requirements. At back-off, the efficiency comparison between GaN and GaAs changes character.
GaAs class AB PA efficiency drops steeply with back-off because the AM-AM curve is relatively soft: the device enters its linear region at moderate back-off, where the DC bias continues to draw current that is not producing RF output. GaN class AB devices, operating at higher supply voltage and higher current density, maintain a flatter efficiency curve at moderate back-off. Published industry data for GaN-on-SiC devices at Ku-band consistently show 10-15 percentage point higher back-off efficiency compared to GaAs devices at the same output back-off level.
This matters because a PA running at P1dB minus 4 dB to support a 32APSK waveform is not running at the efficiency peak. If GaAs delivers 28% efficiency at that operating point and GaN delivers 42%, the thermal calculation above shifts by a larger factor than the peak-efficiency comparison suggests.
Reading a Drain Efficiency Spec Critically
Before accepting a drain efficiency claim, ask for five things. First: at what output power, expressed relative to P1dB or P_sat, not in absolute dBm alone. Second: at what case temperature, specifically whether 25 degrees C or a temperature representative of the deployment environment. Third: over what frequency range, specifically whether band-edge performance is included. Fourth: under what signal stimulus, CW or a representative modulated waveform. Fifth: under what supply voltage and gate bias, since efficiency can be tuned by adjusting the bias point in ways that sacrifice linearity.
If those five conditions match your operating environment, the spec is useful. If any of them are optimistic relative to your deployment, the actual efficiency will be lower, and your thermal budget calculation needs to be rerun with the realistic number.
Falcomm's Design Targets for the FA-2400 Series
For the FA-2400 product line, our design target is drain efficiency above 60% at 3 dB output back-off from P1dB, measured at Ku-band center frequency (14.5 GHz), with a case temperature of 40 degrees C. This target is derived from the thermal budget requirements of a 64-element Ku-band phased array designed for LEO ground terminal deployment, where sustained element dissipation must stay below 1.5W to avoid thermal throttling without active cooling.
This is a design goal being validated through our development process. It is based on our GaN-on-SiC architecture choices, the simulation work we have done on the device model, and published achievable ranges for similar device geometries in open literature. We do not publish this as a measured production spec because we have not yet completed the full characterization campaign that would support that claim.
We are not saying GaN solves every phased array problem. For applications requiring very low noise figure on receive, or for frequencies above 40 GHz where InP processes dominate, GaN is not the automatic choice. What GaN changes specifically is the transmit-path thermal budget at Ku and Ka-band, and it does so through a combination of higher supply voltage, better substrate thermal conductivity, and a flatter back-off efficiency curve. Those three factors together are what make the 60% target meaningful rather than aspirational.