Bottom line first: as LEO satellite internet moves from a launch race into constellation build-out and terminal volume, the structural change in PCB procurement lands mainly on phased-array antenna apertures. This is not one difficult board but a board family that has to be delivered with the same stackup, the same process and repeatable results. For the buyer, the real risk is not unit price — it is whether the prototype and the volume build come from the same factory and the same stackup.
Three board categories, three levels of difficulty
- Antenna aperture and feed network: a multilayer hybrid structure, with RF layers on PTFE or hydrocarbon material (Dk around 2.2–3.5) and control and power layers on FR-4 or a mid-Tg material, built up over several press cycles. The aperture can reach hundreds of millimetres on a side with thin dielectric, so flatness and shrinkage control are the main difficulties.
- RF front end and beamforming: an HDI structure with blind and buried vias and 50–75 μm fine lines in common use, where amplitude and phase consistency between channels is demanding.
- Power and digital processing: high-layer-count boards with heavy copper and local thermal structures, where thermal cycling reliability is often tighter than the electrical performance target.
What buyers should notice is that these three categories are usually not split across three suppliers. More channels raise the aperture area and the feed-network layer count at the same time, and the common approach is to zone the RF layers and the control layers and then connect them locally with blind and buried vias — which means hybrid lamination capability and HDI capability are demanded together. A request that only asks "how many layers, what area" rarely returns comparable quotations.
Three engineering numbers you cannot route around
- Hybrid lamination shrinkage and registration: the CTE difference between PTFE and FR-4 is significant, and total shrinkage after multiple press cycles is usually assessed at 0.3–0.8 mm/m. On a large aperture, interlayer registration must carry compensation and cannot reuse the scaling factor of a standard FR-4 build.
- Amplitude and phase consistency: insertion-loss consistency across a multi-channel feed network is often held within ±0.5 dB and phase consistency on the order of ±5°. Both are driven directly by line width, dielectric thickness and impedance tolerance, so specify impedance tolerance of ±8% (against the usual ±10%) and line-width accuracy of ±8 μm.
- Heat removal and metallisation: T/R modules carry high power density, and the combination of ground via array density, barrel copper thickness and heavy-copper layers (finished up to 13oz) decides whether the temperature rise stays under control.
Purchasing calibration: buy a board family, not a board
- Quote by structure — single-layer RF board, hybrid board and HDI digital board priced separately — so that an averaged price does not hide the real cost of the hybrid lamination steps.
- Ask for evidence of process capability: measured shrinkage data, the maximum number of press cycles, and flatness and impedance reports after hybrid lamination. That is worth more than a statement that the factory "can do it".
- Lock the material grade and stackup at sample stage and state in writing that a material change requires re-validation; otherwise a substitution at volume restarts the whole amplitude and phase consistency exercise.
- Write the key parameters — impedance tolerance, amplitude and phase consistency, plated-hole reliability — into the specification, rather than a line saying the board meets IPC Class 2/3.
Timing and risk
Constellation build-out schedules and terminal form factors are both still moving quickly, and fixed, vehicle, maritime and airborne terminals differ widely in board area and heat-removal requirement. The more prudent approach at this stage is to freeze process capability and the supply chain without locking a single high-volume model: complete process validation with multi-variant small batches first, where the prototype cycle can be compressed to 24-hour expedite, and release volume once the specification converges. Doing it the other way round — committing material and capacity to one revision early — leaves dead stock at the first iteration.
Our capability covers 1–44 layers, a maximum panel size of 610×1100 mm, aspect ratios up to 20:1 (25:1 for backplanes) and finished copper up to 13oz, and hybrid builds with multiple press cycles can be assessed together. Raising the hybrid structure, impedance and amplitude/phase requirements in one review saves more time than iterating boards at sample stage.
Send your Gerber files and stackup requirements; a 24-hour expedite channel is available.