Where does the interference come from? (Many people get this wrong）

Many people underestimate the significance of this "-130 dBm" value. From another perspective, this signal is merely a slight increase above the thermal noise floor (-174 dBm/Hz). In other words, even a non-standard high-speed edge on your board, an improperly controlled return path, or even a few-centimeter section of "airborne ground" can generate energy in this frequency band that is many times stronger than the GNSS signal. Once such energy couples into the antenna chain, the receiver's front-end lacks any discernment capability and simply amplifies it all indiscriminately. Many issues related to positioning drift and slow satellite acquisition are not fundamentally due to poor satellite performance, but rather result from the board "covering itself."

DC/DC Switching Power Supply (The Number One Threat)

Switching Frequency: Hundreds of kHz However, there are a large number of harmonics present.

You can climb all the way up to around 1.5 GHz

Typical Path:

DC/DC converter → Power ripple → Ground resistor → Antenna wire → GNSS front-end

Many people mistakenly believe that "with switching frequencies of only a few hundred kHz, it's impossible to affect 1.5 GHz" —this is a common misconception. The actual issue lies in the extremely high dv/dt at the switching nodes: the steeper the edges, the wider the spectral footprint. Theoretically, an ideal square wave contains an infinite number of harmonics; while constrained in practice, extending this range to GHz levels is not surprising. Furthermore, if the SW node area is excessively large, loop paths are loose, or ground return traces are overly long, these high-frequency components can couple to the entire board or even the wiring harness via parasitic capacitance. What appears to be a "power supply issue" actually becomes a source of RF interference.

• L1:1575.42 MHz(most frequently used
• L2:1227.6 MHz
• L5:1176.45 MHz

1) Where does the interference come from? (Many people get this wrong)

DC/DC Switching Power Supply (The Number One Threat)

Switching Frequency: Hundreds of kHz However, there are a large number of harmonics present.

You can climb all the way up to around 1.5 GHz

Typical Path:

DC/DC converter → Power ripple → Ground resistor → Antenna wire → GNSS front-end

2) High-speed interfaces (MIPI/USB/SerDes)

High-speed interfaces use differential signals, which normally cancel out most electromagnetic radiation.

For below point caused improper connectors or cable layout, poor design, unequal differential routing, discountinous ground planes ect, it couldples into the GNSS antenna, RF cable, or LNA,reducing GPS receiver. 

3) Digital noise,Ground return path issues

The place is not clean.  The reflux path is disordered

Differential mode → Convert to common mode → Send to antenna

Tips: This type of issue is often the most challenging to diagnose due to the absence of a "clear component source." Factors such as segmented ground planes, poorly designed zoning, shared return paths among different modules, or abrupt changes in ground reference at interfaces can all force high-frequency return currents to take detours. When the return path length increases, it effectively creates a larger loop area—the larger the loop area, the stronger the radiation capability. Consequently, even originally differential-mode signals may transition to common-mode due to path asymmetry and ultimately be radiated through antennas or cables

Why does GPS become completely useless when it's interfered with?

The LNA is located at the very front.

The design objective of an LNA (Low Noise Amplifier) is to amplify extremely weak signals while introducing as little additional noise as possible. However, it has a significant drawback: it cannot distinguish between the "signal" and "interference." Any signal within the passband or with sufficient amplitude will be amplified together. More critically, strong interference entering the LNA can trigger nonlinear effects such as compression and intermodulation, further contaminating the effective signal. In other words, once a signal enters the LNA, it becomes nearly impossible to fully restore its quality through subsequent filtering processes.

Result :  Useful signals are drowned out  Position Drift / Star Loss

In practical operation, you'll observe several typical phenomena: prolonged cold-start times, sudden deterioration in positioning accuracy, vehicle position drifting when stationary, and particularly weak signals during certain driving directions. These issues stem not merely from weak signals but from a reduced signal-to-noise ratio. When interference levels approach or exceed GNSS signal strength, receivers may misinterpret corresponding peaks, leading to unstable positioning calculations.

How to "Resist Interference Hardly" (Core Solution) 

Start by addressing the "entry point" (most effective) 

Add SAW/BAW filter (optional) 

Function: 

The value of SAW/BAW filters lies in their "selectivity", specifically their out-of -band suppression capacility. A high-quality GNSS fiter exhibits minimal loss near the passband but provides tens of dB of suppression outside the band. 

This means that most interference from non-GNSS frequency bands is attenuated before reaching the LNA. In the complex electromagnetic environment of vehicle systems, this step often makes all-or-nothing rather than a matter of performance quality. The absence of a filter effectively leaves all noise to the LNA for processing.

Engineering suggestion for priority: 

L1 need add necessary 

L1+L5 is more complecated, need double filter. 

The LNA must be located close to the antenna.

GNSS signal is too weak

A crucial concept in radio frequency links is "noise figure cascading." Simply put, losses in subsequent stages are amplified by preceding ones. For instance, if a section of cable between the antenna and the LNA incurs 1–2 dB of loss, this loss directly contributes to the system's noise figure, effectively eroding a portion of the inherently weak signal. However, positioning the LNA close to the antenna and increasing its gain early on significantly reduces the impact of subsequent losses. This is why many automotive antennas are designed as active antennas.

Correct steps:

Antenna→ LNA(touch)→filter→backside 

Antenna power supply must be clean (a key requirement for your project)

Camera Power Supply / Main Power Supply   Noise interference affects the power supply of GNSS systems.

GNSS antennas, particularly active antennas, typically require power supply via  coaxial or dedicated lines. If the power supply is directly sourced from the  system's DC/DC converter without adequate filtering, switching noise can directly propagate along the antenna power line and enter the LNA through power pins. Such interference and signal paths share a common origin, making subsequent separation challenging. Many issues that appear to be RF-related faults ultimately result from power noise contaminating the front-end circuitry through this pathway.

Powered by a single LDO  / Do not use DC/DC directly

π type filtering:  power→L →C →GNSS 

Many GPS-related issues are actually related to power supply problems.

Common-mode interference must be prevented include :

Common-mode inductor,magnetic bead

Application position:  

Antenna cable  Entry : 

The key here is to block the common-mode path. Once high-frequency noise within the system enters the antenna cable via parasitic coupling, it ceases to be a localized issue and spreads through the cable as a large circuit. The role of the common-mode inductor is to increase high-frequency impedance along this path, thereby preventing noise from entering or exiting effectively. Note that its function is not to filter signals but to restrict the formation of current paths—a fundamentally different approach from conventional LC filtering. 

PCB Layout (Most Easily Overlooked)  key points:

GNSS Regional Independence   

Do not rely on DC/DC conversion &Do not rely on the expressway.

Spatial isolation essentially reduces coupling. High-frequency interference primarily couples throughelectric and magnetic fields; the closer the distance and the larger the surface area, the stronger the coupling. Separating the GNSS area from high-noise sources (such as DC/DC converters, motor drivers, and high-speed interfaces) can significantly reduce coupling paths. In many cases, simply “repositioning” components is more effective than adding a host of components.

Ground plane continuity directly determines the return path. If the ground beneath the GNSS area is cut, the return current is forced to take a detour, creating a larger loop area and thereby increasing the probability of radiation and common-mode conversion. In particular, a continuous ground reference must be ensured beneath the antenna feedline; otherwise, the equivalent impedance will change, which may even affect the matching.

Antenna Routing 

Within 50Ω

As short as possible

Impedance control affects not only reflections but also signal integrity and noise coupling. The longer the trace, the greater the loss, and the more likely it is to act as a “receiving antenna” and introduce interference. In actual design, avoid unnecessary vias, keep them away from high-speed lines, and avoid crossing partition areas—these details all directly affect GNSS performance.

GNSS isn't "weakly signal"; it's simply "too vulnerable to being compromised by your own actions

Many GNSS issues ultimately stem not from a "poor external environment," but rather from unwanted noise generated within the system at certain frequency bands that propagates into the reception chain via improper path coupling. By addressing these challenges through three key design approaches—path control, spectrum matching, and front-end protection—the stability of GNSS systems can be significantly enhanced. The real difficulty lies not in adding components, but in identifying and resolving these issues during the design phase.