A high-frequency PCB is a board built to carry GHz-range signals, where the laminate, the etched conductor and the test method decide whether the finished board matches its simulation. Three controls govern the result: the dielectric material, the geometry of the copper and the verification of each unit. The operating band and link budget decide how tightly each control must be held, from the laminate chosen to the tolerance on the finished trace. The sections below show how each of the three controls applies when the board is an RF design.
What changes on the board as frequency rises
At low frequency, a trace connects two points. At RF and microwave frequencies, the trace behaves as a transmission line, and every dimension of the board enters the electrical result.
Loss comes from two sources: Dielectric loss is set by the laminate’s dissipation factor and grows with frequency. Conductor loss grows too, because skin effect pushes current toward the copper surface, where roughness lengthens the signal path.
Geometry carries equal weight: Impedance depends on trace width, dielectric height and dielectric constant (Dk) together. Etching produces a trapezoidal cross-section, so the finished width differs from the drawn width, and the impedance moves with it.
One distinction in terms: High-speed PCB manufacturing addresses fast digital edges and their signal integrity. High-frequency work addresses analogue and RF carrier signals. A board can need both disciplines, yet a digital link usually fails through timing and crosstalk, while an RF link fails through loss, mismatch and phase error.
High-frequency PCB materials and where each laminate fits
Laminate choice sets the loss floor for the whole board. Communication boards typically draw on four laminate families.
- PTFE-based laminates: Materials such as RT/duroid 5880 and RO3003 give very low loss. Designers choose PTFE-based laminates for Ku/Ka satellite and 5G millimetre-wave work to minimise loss.
- Hydrocarbon-ceramic laminates: Grades such as RO4003C and RO4350B serve a wide range of RF, antenna and radar boards. They process closer to conventional FR4 routes.
- Glass-epoxy FR4: FR4 remains the reference material for digital and lower-frequency boards, and it often forms the outer or supporting layers in a hybrid stack-up.
- Hybrid constructions: Rogers cores combined with FR4 layers balance RF performance against cost and layer count.
Two practical points apply to every family. A datasheet Dk is measured under defined test conditions, and the supplier often publishes a separate design Dk for impedance calculations. State which value your model used when you share a stack-up. Stock and thickness availability also shape the schedule, so confirm material at the DFM stage.
Our FR4 and Rogers PCB lists the substrates we manufacture on.
Four process controls in high-frequency PCB manufacturing
Etch compensation and imaging
Laser Direct Imaging (LDI) removes film registration error and delivers sub-25 µm accuracy on outer layers. That accuracy keeps the etch compensation applied to each trace repeatable, so the finished width tracks the width your impedance model assumed.
Copper foil and surface roughness
Rougher copper raises losses at high frequency. Foil type is therefore a specification item for loss-sensitive RF lines, and it belongs on the drawing.
Layer structure and via transitions
We manufacture multilayer boards up to 24 layers, including blind and buried via designs. Via transitions deserve attention, since via-transition loss is among the recurring causes of RF performance problems. A multilayer RF PCB that mixes Rogers and FR4 layers raises lamination and warpage questions, and these are settled in DFM review before material is ordered. See our multilayer PCB and blind and buried via.
Surface finish
We offer ENIG and HASL. ENIG gives a flat surface for fine-pitch RF pads, while HASL leaves a less planar one. The nickel layer in ENIG adds some loss as frequency climbs, so the finish is selected against the operating band.
Verifying an RF board before it ships
Every unit receives 100% inspection: AOI, fly probe testing, impedance measurement and CMI thickness checks. For an impedance-controlled PCB, the tolerance on the drawing is the acceptance limit.
These checks connect to the symptoms RF teams report. Typical problems include excessive insertion loss, abnormal return loss, ground discontinuity and drift after thermal cycling. In manufacturing terms:
- Insertion loss and return loss trace back to laminate, foil and finished trace width.
- Impedance mismatch traces back to dielectric thickness and etch control.
- Drift after thermal cycling traces back to material pairing and lamination.
Insertion loss and return loss are read with a network analyser on the completed module. See our controlled impedance PCB for how impedance control is applied.
What a well-defined RF stack-up specifies
A stack-up drawing that answers these points removes ambiguity before production starts:
- Operating band and bandwidth
- Laminate and the Dk value used in the model
- Dielectric thickness and tolerance
- Copper foil type and weight
- Impedance targets, tolerance and reference layers
- Layer count and via types
- Surface finish
- IPC class and test requirements
Frequently asked questions
1. What is high-frequency PCB manufacturing?
High-frequency PCB manufacturing produces boards for GHz-range signals. It combines low-loss laminates, controlled etching and impedance verification so the finished board matches its RF simulation closely.
2. When does FR4 stop being suitable for RF designs?
The limit depends on the link budget. FR4 loss is higher and its Dk varies more, so designs with tight loss or phase budgets move to low-loss laminates earlier than designs with relaxed budgets.
3. Which laminate families are used for millimetre-wave designs?
PTFE-based laminates such as RT/duroid 5880 and RO3003 are the usual choice at Ku, Ka and millimetre-wave bands because of their very low loss. Hydrocarbon-ceramic grades serve many lower bands.
4. How does copper roughness affect insertion loss?
At high frequency, current flows near the copper surface. A rougher foil lengthens that path and raises conductor loss, so low-profile foil is specified for loss-sensitive RF lines and named on the drawing.
5. Can Rogers and FR4 be combined in one multilayer RF PCB stack-up?
Hybrid stack-ups are common in RF design. They bring lamination and warpage questions from differing materials, so the build is agreed during DFM review before material is ordered.
6. What should an RF PCB stack-up drawing specify?
Operating band, laminate and design Dk, dielectric thickness and tolerance, copper foil type, impedance targets and tolerance, via types and surface finish. Complete data prevents rework later.
7. How is impedance verified on an RF or microwave PCB before delivery?
Impedance is measured on every unit as part of 100% inspection, alongside AOI, fly probe testing and CMI thickness checks, and compared with the tolerance stated on the drawing.
8. Why does laminate availability affect lead time on high-frequency prototypes?
Specialised RF laminates come in defined thicknesses and copper weights, and stock varies. Sourcing can therefore set the delivery date, so material and thickness are confirmed at the DFM stage.
9. Which surface finish suits an RF PCB, ENIG or HASL?
ENIG gives a flat surface for fine-pitch RF pads, while HASL is less planar. ENIG’s nickel layer adds some loss as frequency climbs, so the finish is chosen against the operating band.
10. What credentials matter for RF boards in defence or space programmes?
ISO 9001:2015 and LCSO certification under JSS 52302 are the relevant benchmarks in India, together with in-house production that gives traceability from review to final test.
