
What Is an Antenna Design Workflow?
An antenna design workflow is the repeatable sequence of engineering stages that transforms an RF specification into a manufactured, tested antenna on a PCB. It typically spans eight phases—requirements, topology, simulation, layout, prototyping, measurement, design transfer, and production—and each phase has defined inputs, outputs, and gate criteria. Following a structured workflow reduces respins (each costing weeks and thousands of dollars in NRE), catches impedance and pattern problems before metal is cut, and keeps the design producible at volume.
Stage 1: Requirements Capture

Every antenna project starts with a requirements document. Ambiguity here is the single largest source of late-stage redesign. Pin down at minimum:
- Operating frequency and bandwidth — center frequency, fractional bandwidth, and whether the spec is defined at −3 dB, −6 dB, or −10 dB return loss.
- Gain and radiation pattern — peak gain in dBi, 3 dB beamwidth, required coverage hemisphere (e.g., upper hemisphere for a GNSS patch, or omnidirectional in azimuth for a BLE device).
- Polarization — linear, circular (RHCP/LHCP), or dual-linear; axial-ratio spec if circular (typically < 3 dB over the coverage cone).
- Impedance and VSWR — almost always 50 Ω; maximum VSWR of 2:1 (return loss ≥ 10 dB) is the de facto industry threshold.
- Size envelope and keep-out zones — the physical box the antenna must fit inside, ground-plane dimensions, and clearance from metal or plastic housing walls.
- Environment — temperature range, IP rating, vibration class, and any radome or enclosure material (ABS, polycarbonate, etc.) that will detune the element.
If you are working within a broader product program, aligning this stage to the RF product development timeline avoids schedule surprises later.
Stage 2: Topology Selection
With requirements locked, choose the antenna topology. The table below covers the most common PCB-compatible options and their sweet spots.
| Topology | Typical Bandwidth | Gain Range | Best For |
|---|---|---|---|
| Microstrip patch | 1–5 % | 5–8 dBi | GNSS, fixed-beam radar, RFID readers |
| Patch array (corporate feed) | 3–10 % | 10–25 dBi | Automotive radar, point-to-point links |
| Planar inverted-F (PIFA) | 5–15 % | 1–3 dBi | Compact IoT, BLE, Wi-Fi |
| Vivaldi / tapered slot | 50–150 % | 5–12 dBi | UWB imaging, GPR, wideband sensing |
| Comb-line array | 5–15 % | 12–20 dBi | Automotive 77 GHz radar, security radar |
A comb-line antenna for automotive radar is a good example of how topology choice is driven by the need for a low-profile, high-gain, series-fed array at millimeter-wave frequencies. For compact consumer products such as BLE wearables, a PIFA or chip antenna on a Bluetooth LE Audio PCB is the more practical starting point.
Stage 3: EM Simulation
Full-wave electromagnetic simulation is non-negotiable for any antenna above a few hundred MHz. The three mainstream solver types each have trade-offs:
- FEM (finite element method) — best for complex 3-D geometries and cavities; memory-heavy. Used by Ansys HFSS.
- FDTD (finite-difference time-domain) — efficient for wideband sweeps; handles large structures well. Used by CST Microwave Studio.
- MoM (method of moments) — fast for planar metallization on layered substrates; ideal for patch arrays. Used by Altium’s integrated solver and Keysight Momentum.
Simulation accuracy depends on correct material models. For a patch antenna on Rogers RO4350B (Dk 3.48 ± 0.05, Df 0.0037 at 10 GHz), entering the wrong Dk by just 0.1 shifts the resonant frequency of a 5.8 GHz patch by roughly 80 MHz—enough to miss the band entirely. Always use the manufacturer’s frequency-dependent Dk/Df curves from the Rogers RO4350B datasheet rather than a single-point value.
Simulate the antenna with its ground plane, feed network, and nearby components (connectors, shields, battery). Omitting the ground plane edge is one of the most frequent antenna design mistakes and leads to pattern distortion that only appears in the lab.
Stage 4: PCB Layout
Translate the simulated geometry into Gerber-ready layout. Key considerations:
- Trace-width tolerance — at 77 GHz a 50 Ω microstrip on RO3003 (Dk 3.00, 5 mil substrate) is roughly 0.30 mm wide; a ±0.025 mm etch tolerance shifts impedance by about 2 Ω. Specify Class 3 per IPC-6012 or tighter.
- Feed transition — coax-to-microstrip, microstrip-to-stripline, or waveguide launch. Each transition adds loss and mismatch; budget 0.1–0.3 dB per transition and validate in simulation.
- Ground stitching — via fences around the antenna feed and between array elements suppress parallel-plate modes. Spacing should be ≤ λ/20 at the highest operating frequency.
- Stackup — antenna layer dielectric thickness sets the patch bandwidth; thicker substrates give wider bandwidth but increase surface-wave losses. A 20 mil RO4350B core is a common compromise for 2.4–5.8 GHz patches.
For designs that route from a transceiver IC to the antenna, the RF transceiver layout guide covers the chip-to-antenna routing discipline in detail.
Stage 5: Prototype Build and Lab Validation
Order prototype boards with the same laminate, stackup, and surface finish planned for production. Using FR-4 as a “quick check” for a design meant for PTFE will give misleading results—FR-4’s Dk of ~4.4 and Df of ~0.02 at 5 GHz are far from RO4350B’s 3.48 / 0.0037.
Measurements to Perform
- Return loss / impedance — VNA, calibrated to the antenna reference plane. Compare S11 to simulation; a frequency shift > 1 % warrants investigation.
- Radiation pattern and gain — anechoic chamber or compact-range. Record E-plane and H-plane cuts, cross-pol, and peak realized gain.
- Efficiency — Wheeler-cap method or gain-comparison method. Total efficiency below 50 % (−3 dB) in a compact IoT antenna is a red flag for excessive ground-plane current or feed loss.
- Isolation (for arrays or MIMO) — port-to-port isolation should meet system spec, typically > 15 dB for a massive MIMO antenna board.
Document every measurement with photos of the setup, cable routing, and fixture details. This data becomes the golden reference for production correlation.
Stage 6: Design Iteration and Change ROI

Prototype results rarely match simulation perfectly on the first spin. Typical adjustments include patch dimension tweaks (±0.1 mm), feed-point repositioning, or adding a parasitic element for bandwidth. Each change costs time and NRE, so evaluate whether the improvement justifies the cost before committing to a respin.
Use the calculator below to estimate the return on investment of a proposed design change. Enter the estimated cost of the change (engineering hours, new tooling, additional prototype run) and the projected savings or performance uplift it enables over the product lifetime.
For example, if a feed-network redesign costs $4,000 in engineering time and a second prototype run, but eliminates a $0.30-per-unit matching network across a 100,000-unit production run, the ROI is strongly positive. Conversely, chasing 0.2 dB of gain improvement on a consumer IoT product with a 5,000-unit forecast rarely pays back. Let the numbers decide.
Stage 7: Design Transfer to Production
This is where many antenna projects stumble. A design that works on the bench must be fully documented so manufacturing can reproduce it without the designer standing over the line. The transfer package should include:
- Final Gerber and drill files, ODB++ preferred for complex stackups.
- Stackup drawing with material call-outs (grade, Dk/Df, thickness tolerance).
- Impedance-controlled trace table with target impedance, trace width, and tolerance.
- Assembly drawing with antenna keep-out zones, no-wash areas, and connector torque specs.
- Test specification: pass/fail limits for return loss, gain, and efficiency, with the reference measurement data from Stage 5.
- Acceptance criteria per IPC-A-610 class and any RF-specific inspection (e.g., via-fill quality under antenna patches).
The design transfer checklist provides a line-by-line template for this handoff. Missing even one item—say, the solder-mask opening spec over a feed via—can cause a production hold.
Stage 8: Production Readiness Checklist
Before releasing the design to volume, run through a formal checklist that covers every deliverable and gate criterion from Stages 1 through 7. The interactive tool below lets you build a customized checklist for your specific antenna project. Select the applicable categories—stackup, simulation, prototype test, transfer documents—and generate a printable list with nothing missing.
A completed checklist for a 24 GHz security radar antenna might include 40+ line items spanning material incoming inspection (Dk verification on a coupon), first-article impedance test, and chamber pattern correlation within ±1 dB of the golden unit. For a 2.4 GHz BLE antenna the list is shorter but still essential—skip it and you risk field returns from detuning caused by housing-material variation.
Common Pitfalls Across the Workflow
- Simulating in free space, testing in an enclosure. A plastic radome with Dk 2.8 and 2 mm wall thickness shifts a 5.8 GHz patch down by 100–200 MHz. Always include the radome in the final simulation.
- Ignoring manufacturing tolerances. A ±0.05 mm etch tolerance on a 77 GHz array element can shift each element’s resonance by 300 MHz. Run Monte Carlo or corner-case simulations.
- Late material substitution. Swapping from RO4350B to a “cost-equivalent” laminate with Dk 3.66 without resimulation guarantees a frequency shift and likely a respin.
- No RF acceptance test in production. Without a go/no-go return-loss test on every board (or at minimum a statistical sample per lot), defective antennas reach the field. Budget for an RF test fixture early—designing it after production starts adds weeks.
Workflow Summary Table
| Stage | Key Output | Typical Duration |
|---|---|---|
| 1. Requirements | Signed-off spec document | 1–2 weeks |
| 2. Topology | Selected antenna type, initial dimensions | 1 week |
| 3. EM Simulation | Optimized geometry, predicted S-parameters and pattern | 2–4 weeks |
| 4. PCB Layout | Gerber-ready design with impedance control | 1–2 weeks |
| 5. Prototype & Test | Measured data, correlation to simulation | 3–5 weeks (incl. fab lead time) |
| 6. Iteration | Final optimized design | 2–6 weeks (0–2 respins) |
| 7. Design Transfer | Complete manufacturing package | 1–2 weeks |
| 8. Production Release | First article approval, volume ramp | 2–4 weeks |
Total timeline from spec to production release typically falls between 13 and 26 weeks, depending on frequency, complexity, and the number of prototype iterations. Millimeter-wave designs (above 24 GHz) trend toward the longer end because tighter tolerances demand more careful correlation between simulation and measurement.
FAQ
How many prototype spins does a typical antenna design need?
Most designs converge in one to two prototype spins. Simple single-element antennas below 6 GHz often get it right on the first build if simulation includes the ground plane and enclosure. Complex arrays above 24 GHz frequently need two spins to nail element spacing and feed-network phase.
Can I use FR-4 for antenna prototyping?
Only if the production design also uses FR-4. FR-4’s Dk varies from roughly 4.2 to 4.6 depending on resin content and frequency, and its Df of ~0.02 at 5 GHz makes it unsuitable for quantitative gain or efficiency measurements intended to correlate with a low-loss laminate.
What is the most important simulation parameter to get right?
Dielectric constant (Dk) of the substrate at the operating frequency. A Dk error of 2–3 % shifts the resonant frequency proportionally, which can move the antenna out of band. Always use frequency-dependent Dk data from the laminate vendor, not a textbook constant.
When should I design the RF test fixture?
Start the fixture design during Stage 5 (prototype validation), not after production release. The fixture needs its own de-embedding characterization, and rushing it leads to poor measurement repeatability—typically worse than ±1 dB, which masks real board-to-board variation. See Altium’s impedance constraint documentation for related layout considerations.
How do I ensure production boards match the prototype?
Lock the material grade, stackup, copper weight, surface finish, and solder-mask type in the fabrication drawing. Require the fabricator to run impedance coupons on every production panel and report Dk back-calculated from the coupon. Any Dk drift beyond the laminate vendor’s tolerance triggers a hold.
Ready for a quotation?
Send your requirements for a material, stackup and manufacturing review. Price and schedule are confirmed for your project.
