Patch Antenna Basics: Structure, Specs and Datasheet Guide

If you are new to patch antennas, the fastest way to understand them is to look at the structure. A patch antenna is a flat piece of metal (the patch) sitting above a layer of insulating material (the substrate), with a larger metal sheet underneath (the ground plane). That sandwich is what sends or receives radio signals.

This guide keeps the language plain, explains every technical term the first time it appears, and sticks to numbers that come from published measurements and application notes. The goal is simple: by the end you should be able to read a patch antenna datasheet, understand what will change once the antenna is inside your product, and decide whether a patch is the right choice at all.

Quick Definition

A patch antenna is a low-profile antenna built from a conductive patch over a dielectric substrate and a ground plane. It radiates mainly in one direction, which makes it a good fit for GPS/GNSS receivers and other embedded wireless products.

A few terms worth clearing up right away:

• Dielectric / substrate: the insulating material between the patch and the ground plane. It stores electric energy and helps set the antenna size.
• Ground plane: the metal reference surface under the antenna. It is part of the antenna, not just a piece of PCB.
• Microstrip antenna: the technical name for this patch-over-substrate structure. The patch antenna is the original type of microstrip antenna, first described by Howell in 1972. So when a datasheet says “also known as microstrip antenna,” it usually means exactly this.

Structure: The Parts and What Each One Does

How It Works

The patch and the ground plane together form a short piece of resonant transmission line, roughly half a wavelength long. Radiation comes mainly from the fringing fields at the two radiating edges — the fields that bulge out past the edge of the patch instead of staying neatly between patch and ground.

Those fringing fields have a very practical side effect: they make the patch look electrically longer than it physically is, so the real resonance lands lower than the calculated one.

Real example: a microstrip antenna designed for 100 MHz measured resonance at about 96 MHz. Because of this, patch dimensions are typically trimmed by 2–4% to land on the intended frequency. (Source: Antenna-Theory.com, “Microstrip Antennas: The Patch Antenna”)

That 4% offset is why almost no patch design is right on the first pass. Plan for one tuning round.

Radiation Pattern and Polarization

Pattern. A standard rectangular patch radiates broadside — straight out from the face of the patch — in a fairly wide, hemispherical shape. Very little energy goes out the back, because the ground plane is in the way. That is ideal when the antenna can face the signal source, and a problem when the product tumbles or rotates.

Polarization. Polarization describes how the wave’s electric field is oriented:

• Linear: the field stays in one plane. A plain patch is linearly polarized.
• Circular: the field rotates as the wave travels. Clip a pair of corners or use two feeds 90° apart and the patch becomes circularly polarized.

GNSS satellites transmit right-hand circular polarization (RHCP), so GNSS antennas are built RHCP too. The quality of that circular polarization is measured by axial ratio: 0 dB is perfect, and the common industry target is under 3 dB. If you receive an RHCP signal with a plain linear antenna, you lose about 3 dB in the ideal case — half the signal power — purely from polarization mismatch.

Frequency and Size

Patch size is set by wavelength and by the substrate:

• Patch length is roughly half a wavelength inside the substrate, not in air.
• A higher dielectric constant shrinks the wavelength inside the material, so the patch gets smaller.

This is exactly why GPS patches are made from ceramic rather than ordinary PCB material. Ceramic substrates used for GNSS have a much higher dielectric constant than FR-4, which is how a 1575.42 MHz antenna fits into a 25 × 25 mm or even 12 × 12 mm footprint. The trade-off is unavoidable: smaller size means narrower bandwidth and lower efficiency.

Typical GNSS L1 patch sizes you will see in production:

Gain and Bandwidth

Gain (in dBi, decibels relative to an ideal omnidirectional radiator) tells you how strongly the antenna focuses energy in its best direction. For circularly polarized antennas the unit is often dBic.

Bandwidth here is usually written as a percentage of the centre frequency — 2% at 1575 MHz means roughly 31 MHz of usable range. “Usable” normally means return loss better than −10 dB.

A basic single patch is a narrowband, moderately directive antenna: expect roughly 6–7 dBi of gain and only about 1–3% bandwidth. For comparison, a classic reference figure for patch directivity is approximately 5–7 dB with bandwidth typically around 3% (Source: Antenna-Theory.com). Small ceramic GNSS patches sit lower again, often near 1–3 dBic, because size and high-dielectric material cost you efficiency.

If you need more gain, you combine elements into an array:

• The theoretical ceiling is about 10 × log₁₀(N). Four elements can therefore add up to roughly 6 dB.
• In practice, a 2×2 patch array is often quoted around 14.8 dBi versus 6–7 dBi for a single element, but published 2×2 designs land anywhere from about 8 dBi to 19 dBi depending on feed-network loss, substrate and any added radome or superstrate.
• Every extra dB of array gain costs beamwidth. A high-gain array points at a narrow slice of sky, which is usually wrong for GNSS and right for a fixed point-to-point link.

If someone quotes you a very high patch gain, ask two questions: measured or simulated, and single element or array?

Feed Methods

The feed is how the signal gets onto the patch. Four methods cover almost everything on the market, and they mainly trade bandwidth against manufacturing complexity (overview of microstrip feeding techniques).

For most GNSS and embedded products, you are choosing a pre-tuned probe-fed ceramic patch rather than designing the feed yourself. It still helps to know which one you have, because it explains why the antenna is narrowband.

Types of Patch Antennas

By construction

One clarification that saves confusion in RFQs: a patch antenna and a panel antenna are not the same thing. A patch is a single radiating element; a panel is usually a finished directional product that may contain several patch elements inside. If a supplier quotes 14–18 dBi for a “patch,” you are almost certainly looking at a panel or an array.

Compared with other antenna families

Passive vs Active Patch

A passive patch is only the radiating element. An active patch adds a low-noise amplifier (LNA) — a small amplifier placed right at the antenna — and often a filter.

The important point is that these two “gains” are completely different things, even though both are printed in dB:

• Antenna gain comes from geometry. A typical embedded GNSS patch element measures around 1.1 dBi peak gain at L1 in published component datasheets.
• LNA gain is electronic amplification, commonly 28–33 dB on the same class of part.

So an active antenna’s headline number can look 30 dB better than a passive one while the antenna itself is identical. Amplification protects the signal against cable and connector loss; it does not collect more satellites.

How much LNA gain do you actually need? u-blox states that an LNA gain of 15 dB is usually sufficient even for cable lengths up to 5 m, that there is no need to exceed 26 dB, and that with short cables and gain above 35 dB an overload condition may occur on some receivers (Source: u-blox GNSS antennas application note). Too much gain is a real failure mode, not a bonus.

Rule of thumb:

• Antenna on the same board as the receiver, short trace → passive is often fine.
• Cable longer than about 10 cm, or a remote/roof mount → active, with gain matched to your total path loss.

For the full decision path, see the patch antenna selection guide.

Advantages and Limitations

Advantages

• Flat and low profile, easy to mount inside an enclosure.
• Predictable, mostly one-directional pattern with little backward radiation.
• Straightforward to make circularly polarized, which is what GNSS needs.
• Cheap and repeatable in volume; ceramic parts can be factory pre-tuned.
• Scales into arrays when more gain is required.

Limitations

• Narrow bandwidth — typically a few percent for a single element.
• Performance depends heavily on ground plane size and shape.
• Sensitive to nearby metal, batteries and enclosure materials.
• Directional pattern is a liability if device orientation is random.
• Shrinking the patch costs efficiency and bandwidth, not just size.

Real-World Installation

This is where most projects lose performance, and it has nothing to do with the part number you picked.

1. The ground plane is part of the antenna

A ceramic patch is tuned on a specific reference ground plane. Change that ground plane and you change both the gain and the axial ratio.

Reference figures: for embedded GNSS designs, a 25 × 25 mm patch on a ground plane in the region of 50 × 50 mm is the common baseline, and published curves show gain and axial ratio improving as the ground plane grows toward that size before flattening out. (Source: u-blox GNSS antennas application note)

Practical takeaways:

• Match the ground plane the datasheet was measured on, or expect a shift.
• Keep the patch centred if you can. Off-centre or corner placement skews the pattern.
• A ground plane much smaller than the reference is the single most common cause of “the antenna worked on the eval board but not in our product.”

2. Metal, enclosures and batteries cause detuning

Detuning simply means the frequency the antenna is happiest at has moved away from where you wanted it. Anything conductive or high-dielectric near the antenna does this:

• Metal nearby couples to the antenna and shifts both resonance and pattern. Texas Instruments’ antenna design note is blunt about it: you cannot put metal objects such as crystals close to the antenna without causing detuning, and the case of a phone or other device will also detune the antenna, so some tuning adjustment ability is needed (Source: TI AN-1811).
• Batteries and unshielded wires act as unintended parasitic elements. In engineering practice, the enclosure, a user’s hand, or parts such as batteries and metallic components close to the antenna can cause detuning, and a vector network analyser is used to re-check impedance and adjust the matching network to restore performance (Source: All About Circuits).
• Enclosure plastic is not neutral either. Its dielectric constant pulls the resonance down, which is why an antenna tuned in open air often measures wrong once the lid goes on.

What to do about it:

1. Define a keep-out zone around the antenna before the mechanical design is frozen — not after.
2. Tune with the real enclosure, real battery and real cables in place.
3. Measure S11 or VSWR in the assembled product, not on the bench.
4. If metal is unavoidable, plan for a matching network you can adjust, or move to an antenna type that tolerates the environment better.

3. Orientation

A patch radiates out of its face. For GNSS, that face must point at the sky, and the ground plane should be between the patch and the vehicle or machine below. If your product can be installed in any orientation, either constrain the mounting or reconsider whether a patch is the right form factor.

How to Read a Patch Antenna Datasheet

Datasheet numbers are always measured under specific conditions. Reading them well is mostly about finding those conditions.

Three questions worth asking every supplier:

1. On what ground plane size were these numbers measured?
2. Are the plots simulated or measured, and in what chamber?
3. Does the quoted gain include the LNA?

When to Choose a Patch Antenna

A patch is a good fit when:

• You need circular polarization for GNSS reception.
• The antenna can face a known direction, such as the sky or a fixed base station.
• You have a reasonable ground plane, ideally close to the datasheet reference.
• The product is flat, and height is more constrained than area.
• You want a repeatable, pre-tuned part rather than a custom design.

Look at something else when:

• Device orientation is random → consider a helix or a well-tuned chip antenna.
• Available area is tiny but the ground plane is small too → a chip or FPC antenna may perform better in situ.
• You need wide bandwidth across several distant bands → look at stacked, multi-feed or entirely different antenna types.
• The antenna must sit right next to metal or a battery with no keep-out → solve the mechanical problem first, or the antenna choice will not matter.

Common Beginner Mistakes

1. Assuming every patch is a GPS antenna. “Patch” describes a structure. Frequency and polarization still have to match the application.
2. Choosing by size alone. Smaller patches cost you bandwidth, efficiency and axial ratio.
3. Ignoring the ground plane. It is part of the antenna, not just somewhere to solder it.
4. Tuning before the enclosure exists. The final housing, battery and cabling will move the resonance.
5. Confusing LNA gain with antenna gain. 32 dB of amplifier is not 32 dB of antenna, and too much gain can overload the receiver.

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

FAQ

Is a patch antenna the same as a microstrip antenna?

In most practical cases, yes. The patch antenna is the original type of microstrip antenna, so the two terms are used interchangeably when the structure is a patch over a substrate and ground plane.

What is a patch antenna used for?

Compact, low-profile RF designs: GPS/GNSS receivers, asset trackers, telematics units, timing modules, RTK and surveying receivers, industrial IoT devices, and fixed directional links.

Is a patch antenna directional?

Yes. Radiation is strongest broadside — straight out from the face of the patch — with very little out the back thanks to the ground plane.

How much bandwidth does a patch antenna have?

A single basic patch is narrowband, typically a few percent of its centre frequency. Wider coverage needs stacked patches, multiple feeds or a different antenna type.

Why is my patch antenna resonating at the wrong frequency?

Two usual reasons. First, fringing fields make the patch look electrically longer than it is, which is why designs are trimmed by roughly 2–4%. Second, detuning: nearby metal, the battery, the enclosure or an undersized ground plane has shifted the resonance. Re-measure in the assembled product before changing the antenna.

Do I need an active or a passive patch antenna?

Passive is often enough when the antenna sits on the same board as the receiver with a short trace. Choose active when the cable run is longer than about 10 cm or the antenna is remotely mounted — and size the LNA gain to your cable loss rather than picking the biggest number available.

Why does a GPS patch antenna need RHCP?

GPS/GNSS satellites transmit right-hand circular polarization, so an RHCP antenna matches the incoming signal and rejects reflections. Details are in the RHCP vs LHCP patch antenna guide.

Conclusion

Patch antenna theory is straightforward: a resonant metal patch over a substrate and ground plane, radiating from its fringing fields. Product success, though, comes down to details — trimming for the 2–4% frequency offset, giving the antenna the ground plane it was measured on, keeping metal and batteries away from it, and reading datasheet numbers in context instead of comparing headline gain figures.

If your product is a GPS/GNSS device, pay particular attention to axial ratio and to installed performance rather than bench performance.

View related GPS patch antennas or read the patch antenna selection guide before choosing a model for production.

References

• u-blox, GNSS Antennas Application Note (UBX-15030289) — ground plane, axial ratio and LNA gain guidance
• Antenna-Theory.com, Microstrip (Patch) Antennas and The Patch Antenna measurements — directivity, bandwidth and the 100 MHz / 96 MHz trimming example
• Wikipedia, Patch antenna — structure, history and patch vs panel distinction
• Texas Instruments, AN-1811 Bluetooth Antenna Design — detuning from metal and enclosures
• All About Circuits, Common Antenna Integration Challenges — enclosure, battery and matching-network validation
• Overview of Feeding Techniques of Microstrip Patch Antennas — feed method comparison
• McMaster University, Lecture 21: Microstrip Antennas — transmission-line and cavity models