RF front-end for LEO user terminals - phased array cost drivers?

LEO user terminals need electronic steering but cost targets are brutal. Where does the money actually go - the antenna array, the beamformer ICs, or the modem? Any cost-down tricks that worked?

LEO UT phased-array cost is usually dominated by the front-end count and packaging, not by the beamforming algorithm paper.

Cost drivers that show up on BOMs:
1. T/R or beamformer ICs per element / subarray — hybrid architectures cut converter count but still pay for RF ICs and calibration.
2. Antenna + multilayer board + thermal: mmWave feed networks and heat from PAs often cost more than the digital beam weights.
3. Calibration and test time in production (coupled fixture vs OTA sample) — a cheap silicon choice that needs long OTA every unit loses on factory cost.
4. Mechanical / radome and pointing (electronic only vs hybrid mech+electronic) — pure electronic steering saves motors but raises element count and DC.

If the target is consumer-ish ASP, people usually cut elements, duty cycle, and EIRP before they cut calibration quality. Share band (Ku/Ka), element count goal, and whether TX and RX share the aperture — that frames the cost conversation.

— rfforum 编辑部 / Editorial Team