Patch Antenna Design Guide for GNSS and Wireless Devices

Patch antenna design comes down to seven decisions: frequency band, patch size, substrate, ground plane, feed and matching, polarization, and everything that sits around the antenna in the finished product. A patch that measures well on a lab fixture can still fail in the device, because the ground plane is small, the plastic cover is close, the coax is long, or a cellular radio sits next to it.

Short answer: the patch length sets the resonant frequency, the substrate sets how small and how wide-band the patch can be, the ground plane sets the radiation pattern and shifts the tuning, and the enclosure plus cable decide what the receiver actually gets. Tune and measure the antenna inside the final mechanics — free-space datasheet numbers do not transfer.

So the useful question is not “which patch has the highest gain”, but “which patch keeps enough margin after installation”. Below you get the judgement criteria: what to specify, which numbers you can calculate yourself to sanity-check a quote, and where GPS/GNSS designs usually lose margin.

What Controls Patch Antenna Design?

• Frequency and size — the resonant length is about half a wavelength in the substrate, so the dielectric constant, not the free-space wavelength, sets the physical size.
• Substrate — dielectric constant and loss tangent decide size, efficiency, and bandwidth.
• Ground plane — controls the pattern and front-to-back ratio, and moves the centre frequency.
• Feed and matching — probe, pin, microstrip, or dual feed; sets impedance and usable bandwidth.
• Polarization — RHCP for GNSS reception, linear for most other links.
• Installation — enclosure, cable, connector, and nearby radios.

For GNSS patch antennas, axial ratio, phase centre stability, LNA placement, cable loss, and nearby transmitters matter as much as the patch itself. Change the patch size without checking the ground plane and the enclosure, and the antenna can resonate on the right frequency and still perform badly in the field.

Core Design Variables

Reference Numbers You Can Check Yourself

• GNSS band edges (ITU allocations): 1559–1610 MHz (GPS L1 / Galileo E1 / BeiDou B1), 1215–1300 MHz (L2 / E6 / B3), and 1164–1215 MHz (L5 / E5 / B2). Centre frequencies: GPS L1 1575.42 MHz, L2 1227.6 MHz, L5 1176.45 MHz, GLONASS L1 1602 MHz (ESA Navipedia, EU PBN Portal).
• First-pass patch length: L ≈ c / (2 · f · √εr). At 1575.42 MHz that is about 95 mm in air, so any smaller patch is buying size with dielectric constant.
• Substrate height floor: keep h above roughly 0.025 λ (about 1/40 of a wavelength) or efficiency degrades (Antenna-Theory).
• Bandwidth scaling: bandwidth is roughly proportional to h / √εr (Microwaves101). Thin plus high dielectric constant means narrow band, and there is no way around it.

These are engineering estimates for checking a quote or a datasheet, not substitutes for full-wave simulation and measurement.

Best Substrate for Patch Antenna: Dielectric Constant and Material Trade-Off

There is no single best substrate for patch antenna design, only the trade you are willing to make. A higher dielectric constant shrinks the patch. A lower loss tangent keeps efficiency. More thickness widens bandwidth but raises surface-wave and mechanical-height issues.

Substrate data from a materials comparison published in the Indian Journal of Engineering & Materials Sciences (CSIR-NIScPR). Laminate-based microstrip antennas typically sit in the εr 2–12 range; ceramic GNSS patches go higher.

Worked example — why GNSS patches are ceramic. A half-wave patch at 1575.42 MHz would be about 95 mm in air. A 25 mm ceramic patch therefore needs √εr ≈ 3.8, i.e. an effective dielectric constant in the mid-teens, far above any normal PCB laminate. Because bandwidth scales with h / √εr, the same choice that makes the patch small also makes it narrow-band and sensitive to anything that shifts its tuning. This is an estimate from the standard resonance formula; fringing fields and the real ground plane move the exact value.

For production devices, ask the supplier for dielectric constant, loss tangent, tolerance, operating temperature range, and the exact measurement conditions behind the datasheet.

Step-by-Step Patch Antenna Design Workflow

1. Define the use case and RF band. Start with the device application, not the antenna shape. A GPS tracker, RTK rover, cellular gateway, and tracking device positioning module do not need the same antenna.
2. Choose the antenna class. For GNSS, decide between passive ceramic patch, active patch module, multiband stacked patch, or external GNSS antenna. For the selection trade-off, use the patch antenna selection guide after the design constraints are clear.
3. Reserve the ground plane early, and treat “bigger is better” as a rough guide only. Peer-reviewed RTK measurements found a hard floor and a plateau: with an added round ground plane, no GNSS reception could be established below about 12 cm diameter for the antenna under test, while a 15 cm round plane gave equal or better 2D position scatter than a 30 × 30 cm plane in 9 of 10 comparisons (Punzet & Eibert, Advances in Radio Science, 2023). What matters as much as area is symmetry: the plane should be flat and symmetrical around the antenna, and the antenna should not sit at a board edge, or the pattern skews and polarization purity drops.
4. Match the feed and impedance. A patch is normally matched to a target system impedance, often 50 ohms. Feed location, feed structure, and the matching network all affect return loss and usable bandwidth.
5. Check polarization. GNSS receiver antennas are commonly RHCP because GNSS satellite signals use right-hand circular polarization. For the full RHCP, LHCP, axial ratio, and multipath discussion, use the RHCP vs LHCP patch antenna guide.
6. Validate the installed antenna. Measure return loss, efficiency, gain pattern, axial ratio if applicable, and receiver performance in the final enclosure. A free-space datasheet is useful, but it is not the final device result.

Patch Antenna Design for GPS and GNSS

GPS and GNSS patch antenna design deserves special attention because satellite signals are weak by the time they reach the receiver. The antenna is not just a metal plate on ceramic. It is part of a receive chain that may include a ground plane, LNA, SAW filter, coaxial cable, connector, and receiver RF front end.

For GPS/GNSS devices, check these items before approving the design:

• Frequency coverage: GPS L1 only, or multiband L1/L2/L5 for higher precision systems.
• Polarization: RHCP for normal GNSS reception.
• Axial ratio: especially important for RTK, surveying, tracking device, and precision agriculture.
• Ground plane: test on the actual PCB size, not only a lab reference board.
• Cable and connector loss: small coaxial cables can reduce signal margin.
• Nearby transmitters: cellular radios can desensitize the GNSS front end if filtering and placement are weak.

Why “One Patch for GPS and GLONASS” Often Backfires

A single-feed circularly polarized patch has a narrow RHCP response, and its cross-polarization rejection is centred on the same tuning frequency. Tune a 25 mm patch to the midpoint between GPS L1 (1575.42 MHz) and GLONASS L1 (1602 MHz), around 1590 MHz, and the rejection peak lands where there are no satellites — so rejection is degraded at both ends and much of the antenna’s ability to reject reflected signals is lost (GPS World). Poor axial ratio also makes the phase response non-linear, which matters for RTK and carrier-phase work.

The practical test is therefore not “is S11 below −10 dB at both frequencies”, but “is the axial ratio still usable at both frequencies”. If the device must cover GPS + GLONASS + Galileo + BeiDou L1, or L1 and L5 together, a dual-feed or stacked structure is not a luxury; it is what keeps axial ratio usable across the band.

If the project needs a ready product path rather than a custom design review, compare the available GPS / RHCP patch antenna options before opening a custom RFQ.

Do the Cable and Enclosure Math Before You Pick a Patch

Cable loss decides active vs passive, and it is a two-minute calculation

RG-174 is the default thin coax inside compact GNSS devices. Published cable specifications put its attenuation at roughly 34 dB per 100 ft, about 1.1 dB/m, at 1 GHz, rising to roughly 1.1–1.3 dB/m across the 1.2–1.6 GHz GNSS bands (RG-174 specification sheet). So:

• 0.15 m pigtail: about 0.2 dB — a passive patch is fine
• 1 m internal run: about 1.2 dB — workable, but count the connectors too
• 3 m external run: about 3.5 dB — you are giving away most of a good patch’s gain
• 5 m roof mount: about 6 dB — a passive patch is the wrong choice

A usable rule: add up the loss between patch and receiver first. Below about 1 dB, passive is fine. Above about 2 dB, put the LNA at the antenna. Do not overcorrect either — total gain should comfortably cover cable loss without pushing the receiver front end toward saturation.

A plastic cover is not RF-neutral

Any dielectric close to the patch lowers the resonant frequency, because it raises the effective dielectric constant around the radiator. Texas Instruments measured an antenna mounted 1 mm from a plastic case: resonance moved from 868 MHz to 862.875 MHz, about −0.6%, VSWR rose from 1.58 to 2.53, and roughly 18% of the power was lost (TI SWRA726). Metal is far worse: severe detuning, or a complete block with a continuous metal shell.

Scale that to GNSS. A 0.6% shift at 1575 MHz is about 9 MHz, the same order as the usable bandwidth of a small ceramic patch. That is the quantitative reason bare-board tuning misleads people. In practice: keep 3–5 mm of air between antenna and enclosure wall where mechanics allow, leave the matching-network pads unpopulated on the PCB, and do the final VNA tuning with the board assembled inside the real enclosure.

Common Patch Antenna Design Mistakes

When to Use a Custom Patch Antenna

A custom patch antenna is worth considering when the device has unusual mechanical constraints, a non-standard ground plane, multiband GNSS requirements, a required cable or connector, or a certification schedule that cannot tolerate late antenna risk.

For B2B projects, provide your antenna supplier with:

• Target frequency bands
• PCB size and ground plane drawing
• Enclosure material and antenna location
• Connector and cable length
• Active or passive preference
• Required samples, datasheet format, and validation tests
• Any cellular, LTE, 4G, or 5G radios near the GNSS antenna

For the full topic hub, product paths, and application map, see the Patch Antennas guide.

FAQ

What is the most important factor in patch antenna design?

The most important factor is matching the antenna design to the final installation, including frequency band, ground plane, enclosure, feed, and nearby components.

How does substrate material affect a patch antenna?

Substrate material affects antenna size, bandwidth, efficiency, and loss. A higher dielectric constant can make the patch smaller, but it may also narrow bandwidth or reduce efficiency depending on the design.

Does a patch antenna need a ground plane?

In most practical designs, yes. The ground plane helps shape the radiation pattern and provides the RF reference needed for stable performance.

Can one patch antenna cover GPS L1, L2, and L5?

Yes, but multiband GNSS usually requires a purpose-built patch structure such as a stacked patch, dual-feed design, or other tuned multiband construction.

Should I choose active or passive for a GPS patch antenna?

Add up the loss between patch and receiver. Below about 1 dB, passive is fine. Above about 2 dB — which a 2 m RG-174 run already reaches — put the LNA at the antenna.

How big should the ground plane be for a GNSS patch antenna?

Bigger than the patch, flat, symmetrical, and validated on the real PCB. RTK measurements showed no usable reception below about 12 cm diameter for the antenna tested, while a 15 cm plane matched or beat a 30 × 30 cm plane in 9 of 10 runs. More metal is not automatically better; the right size is the one you measure.

How much can a plastic cover shift a patch antenna’s frequency?

Enough to matter. In a TI measurement, an antenna 1 mm from a plastic case dropped from 868 MHz to 862.875 MHz, about 0.6%, with VSWR going from 1.58 to 2.53 and roughly 18% power loss. The shift is always downward, and 0.6% at 1575 MHz is about 9 MHz.

Conclusion

Good patch antenna design is not only a geometry problem. It is an integration problem. Start with the target band and device layout, reserve enough ground plane, choose the right polarization, and verify the installed antenna before production.

For GPS/GNSS projects, view related GPS patch antennas or request a quote with your PCB layout, frequency band, and connector requirement.

References

• ESA Navipedia: GNSS signal and frequency bands
• ESA Navipedia: Antennas
• EU PBN Portal: GNSS spectrum allocation
• NTIA: 1559–1610 MHz band allocations
• Antenna-Theory: Microstrip (patch) antennas
• Microwaves101: Microstrip patch antennas
• Punzet & Eibert, Advances in Radio Science: additional antenna ground planes and RTK-GNSS accuracy
• Texas Instruments SWRA726: antenna detuning inside plastic enclosures
• GPS World: Patch antennas for the new GNSS
• Indian Journal of Engineering & Materials Sciences (CSIR-NIScPR): dielectric substrate comparison
• RG-174 coaxial cable attenuation specification
• u-blox GNSS Antennas Application Note
• Orban Microwave: The Basics of Patch Antennas