LoRa Handheld Device Antenna Optimization in Practice: Iterating from Ceramic to FPC Antennas

In compact LoRa/Sub-GHz device design, the antenna is often the most underestimated yet hardest challenge to overcome. A recent real-world case shared by an open-source community — the RF design process for a compact LoRa/GNSS handheld "Nomad Terminal" — presents a complete picture of antenna evolution from ceramic to FPC antennas, along with the engineering trade-offs behind each decision.

The value of this case isn't in providing a "standard answer," but in revealing an often-overlooked truth: In compact devices, antenna performance is a system-level problem, not a matter of selecting a single component.

Why Antenna Design for Small LoRa Devices Is So Difficult

To understand the challenges of antenna design, you first need to appreciate the unique characteristics of LoRa technology itself. LoRa (Long Range) is an LPWAN physical-layer technology based on Chirp Spread Spectrum (CSS) modulation, operating in the Sub-GHz band — ISM unlicensed frequencies below 1 GHz. Major global frequency bands include 868 MHz in Europe, 915 MHz in North America, and 433 MHz in parts of Asia. Compared to WiFi or Bluetooth at 2.4 GHz, Sub-GHz offers superior diffraction and penetration, with theoretical transmission distances reaching tens of kilometers. But the flip side of this coin is: the lower the frequency, the larger the theoretically optimal antenna size. A quarter wavelength at 868 MHz is approximately 8.6 centimeters, which means in a palm-sized handheld device, the antenna is inevitably in a severely "electrically small" state, with efficiency and bandwidth fundamentally constrained by the laws of physics.

For handheld LoRa devices, the factors affecting final RF performance extend far beyond the antenna itself. PCB layout, ground plane size, battery placement, display, enclosure material, and even how the user holds the device can all significantly alter actual antenna performance.

Among these, the role of the ground plane is especially easy to underestimate. For most small antennas (ceramic, PCB, FPC), the ground plane isn't merely an electrical reference — it's an integral part of the antenna's radiation system. The antenna element acts as one arm of a dipole, while the ground plane serves as the other. The size, shape, and continuity of the ground plane directly determine radiation efficiency, resonant frequency, and radiation pattern. In compact handheld devices with limited PCB area, the ground plane is often less than a quarter wavelength, causing the antenna's radiation impedance to deviate from ideal values, narrowing bandwidth, and reducing efficiency. To make matters worse, if elements like the battery, LCD flex cable, or metal shielding encroach on the keep-out zone, they introduce additional parasitic coupling that further degrades antenna performance.

This means that even if you choose an antenna with excellent nominal specifications, once it's crammed into a tight enclosure near a battery and screen, its resonant frequency, impedance matching, and radiation efficiency can all deteriorate dramatically. This is why many engineers see beautiful lab measurements "fall apart" in actual products. The performance figures in antenna manufacturer datasheets are typically measured on ideal standard evaluation boards with carefully optimized ground plane sizes, keep-out zones, and component layouts — conditions your product can almost never replicate.

The project team tackled this core challenge by evaluating three antenna configurations in succession, progressively converging on a viable engineering solution.

The Iterative Evolution of Three LoRa Antenna Approaches

Approach 1: Ceramic Antenna — Compact but Constrained

The initial design used a ceramic antenna, primarily chosen for its small size and ease of integration. For a compact handheld device, this is almost an instinctive first choice.

The core principle of ceramic antennas involves using high-permittivity ceramic material as the antenna substrate. Electromagnetic wavelength shortens in high-permittivity dielectrics (by a factor roughly equal to the inverse of the square root of the dielectric constant), so an antenna at the same resonant frequency can be made much smaller than its free-space counterpart. Common ceramic antennas have dielectric constants between 6 and 20, enabling antenna sizes to be compressed to a fraction of the free-space wavelength. This miniaturization advantage has made ceramic antennas ubiquitous in consumer electronics such as GPS modules, Bluetooth devices, and WiFi routers. However, miniaturization comes at a cost: according to the Chu-Harrington limit (a fundamental theorem in antenna physics), the smaller an antenna is relative to its operating wavelength, the lower the product of its achievable bandwidth and radiation efficiency. In other words, the laws of physics dictate that you cannot simultaneously have an extremely small size, wide bandwidth, and high efficiency — there is an insurmountable trade-off among the three. While ceramic antennas in the Sub-GHz band are small, they typically have narrow bandwidth, lower peak gain, and are extremely sensitive to the surrounding environment (ground plane size, nearby metals, and dielectrics).

Problems surfaced quickly: the ceramic antenna's placement options were severely constrained. Mounted on the PCB, it couldn't be positioned far from interference sources like the battery and screen, leaving the team with virtually no room to optimize the surrounding RF environment. Under these constraints, the antenna's performance ceiling was firmly locked in. As a surface-mount component, the ceramic antenna's feed and ground points connect directly to PCB traces and the ground plane, often just millimeters from battery connectors, LCD FPC cables, and metal shields — near-field coupling at these distances severely disrupts current distribution and radiation characteristics.

Approach 2: FPC Antenna + Coaxial Cable — Better Performance but Complex Assembly

To break free from placement limitations, the team turned to an FPC (Flexible Printed Circuit) antenna connected via coaxial cable.

An FPC antenna is formed by etching antenna conductor patterns onto a flexible polyimide (commonly known as Kapton) substrate. It's essentially a printed monopole or dipole antenna, but the flexible substrate allows it to be bent and adhered to surfaces, providing exceptional spatial adaptability. FPC antennas are widely used in smartphones, wearables, and IoT terminals, typically adhered to the inner wall of device enclosures. Compared to ceramic antennas, FPC antennas usually have a larger radiating area (since they don't rely on high-permittivity materials for miniaturization), meaning they can often achieve wider bandwidth and higher radiation efficiency at the same frequency band.

The greatest advantage of FPC antennas is layout flexibility — they can be adhered to the enclosure's inner wall, away from the PCB, battery, and display, providing a cleaner RF environment. This step significantly improved the antenna's operating conditions. By "liberating" the antenna from the crowded PCB surface to the enclosure wall, the physical distance between the antenna and interference sources can increase from millimeters to one or two centimeters or more, dramatically reducing near-field coupling and making the antenna's resonant frequency and impedance much easier to control and tune.

But this came with trade-offs: the additional coaxial cable increased mechanical complexity and assembly difficulty. While coaxial cable is the gold standard for RF transmission — its outer conductor provides complete shielding that effectively prevents signal leakage and external interference — in compact handheld devices, the issues are significant: cables require manual soldering or connector installation, routing must avoid sharp bends (which alter characteristic impedance and cause reflections), and the cable itself consumes space and introduces assembly tolerances that affect mass-production consistency. For a handheld device prioritizing compactness and manufacturing consistency, this wasn't an ideal long-term solution.

Approach 3: FPC Antenna + Pogo Pin — The Final Balanced Solution

The final approach retained the FPC antenna's layout flexibility while replacing the coaxial cable with Pogo Pins (spring-loaded contact pins) for the RF connection.

A Pogo Pin is a spring-loaded compression electrical contact consisting of three parts: a plunger, spring, and barrel. When two structural components are mated, the spring is compressed and the plunger maintains stable electrical contact with the opposing pad under spring force. Originally widely used in test fixtures and battery contacts, Pogo Pins have gained increasing popularity in consumer electronics RF connections in recent years — for example, many smartphones use Pogo Pin antenna contacts, and TWS earphone charging cases use them for signal connections. In RF applications, characteristic impedance control requires special attention: by carefully designing the plunger diameter, barrel inner diameter, and ground Pogo Pin layout, the characteristic impedance at the connection point can be controlled close to 50Ω, minimizing return loss at the junction. Some Pogo Pin products specifically designed for RF applications (such as coaxial-structure Pogo Pins) can maintain good impedance matching up to several GHz.

The Pogo Pin approach elegantly balanced two conflicting requirements: it maintained the FPC antenna's freedom to be positioned away from interference sources, while the spring-loaded connection made both the RF path and mechanical structure more compact, reliable, and suitable for automated assembly. Compared to coaxial cable, Pogo Pins eliminated manual soldering — the FPC antenna is adhered to the enclosure, Pogo Pins are placed at the corresponding PCB locations, and the electrical connection is automatically established when the enclosure is closed, greatly simplifying assembly and improving consistency.

This configuration ultimately passed both RF measurement and outdoor field-test validation.

Test Results: S11 Return Loss and Field Link Verification

The final FPC + Pogo Pin solution achieved solid engineering results:

• S11 Return Loss: -11.13 dB at the 868 MHz band, -12.82 dB at the 915 MHz band
• Field Link Test: Achieved a 3-kilometer LoRa communication link outdoors

S11 (return loss) is one of the most commonly used metrics in the S-parameter (Scattering Parameters) family, a fundamental tool in microwave network analysis. In a single-port network, S11 is defined as the complex voltage ratio of the reflected wave to the incident wave, typically expressed in decibels (dB). Physically, S11 reflects the degree of impedance mismatch at the antenna's input port: when the antenna's input impedance perfectly matches the feed system's characteristic impedance (usually 50Ω), all energy is received and radiated by the antenna with no reflection, and S11 approaches negative infinity; when completely mismatched (e.g., open or short circuit), all energy is reflected, and S11 is 0 dB. In practical antenna tuning, engineers typically measure S11 using a vector network analyzer (VNA) and observe the antenna impedance trajectory on a Smith chart as it varies with frequency. If the impedance deviates from the chart's center (the 50Ω normalized point), a matching network — usually an L-type or π-type topology composed of inductors and capacitors — is used to "pull" the antenna impedance back to center. Matching network design is both science and art, as component parasitics, PCB trace distributed capacitance and inductance all affect the final matching result, often requiring iterative tuning.

In general engineering practice, S11 below -10 dB is considered acceptable matching, meaning more than 90% of energy is effectively delivered to the antenna port (note: this refers to energy delivered to the antenna, not all radiated — the antenna's own dielectric and conductor losses consume some energy, with the actual radiated proportion determined by radiation efficiency). S11 below -6 dB (approximately 75% energy transfer) is considered acceptable in some highly constrained applications, but -10 dB is the baseline for most RF designs. This solution consistently exceeded -10 dB across both the 868 and 915 MHz mainstream LoRa bands, and combined with the 3-kilometer real-world communication distance, demonstrates the effectiveness of this system-level optimization.

It's worth noting that LoRa's CSS modulation technology provides an exceptionally strong link budget — a typical LoRa link budget can exceed 154 dB (depending on spreading factor and bandwidth settings) — meaning that even with modest antenna gain, LoRa systems can still achieve long-range communication. The 3-kilometer field-tested distance on a handheld device is a pragmatic result; considering the device's compact size and antenna physical limitations, this distance validates the overall RF link's healthy state.

Core Takeaway: LoRa Antenna Optimization Is a Systems Engineering Challenge

The most important lesson from this case is the repeatedly emphasized point: Antenna performance in compact LoRa devices is a system-level problem.

Antenna selection, placement, PCB layout, and mechanical design must be considered holistically rather than optimized in isolation. Simply swapping in a "better antenna" often fails to solve the fundamental problem; the real breakthrough comes from replanning the entire RF environment.

From ceramic antenna to FPC antenna, from coaxial cable to Pogo Pin — this iterative path is essentially a process of continuously finding the balance point among performance, layout flexibility, mechanical complexity, and mass-production feasibility. This is precisely where embedded RF design is most challenging: there is no single-dimension optimum, only the best system-level compromise.

This methodology has been widely validated in industry. Nearly all successful consumer IoT products — from AirTag's ultra-wideband antenna to various brands' TWS earphone Bluetooth antennas — undergo similar system-level iteration during development. Early collaboration between antenna engineers, mechanical engineers, and industrial designers is invariably far more effective than after-the-fact "antenna rescue" efforts. Reserving adequate keep-out zones and a well-planned ground plane layout for the antenna early in the project yields far greater returns than trying to compensate for structural deficiencies through complex matching networks later.

For engineers developing LoRa or other Sub-GHz devices, this case offers a reusable approach: rather than fixating on the antenna component itself, step back and re-examine the overall environment from a system architecture perspective.

Key Takeaways