Your GPS receiver is not broken. The satellite signal is fine. The problem is sitting between your antenna and your chip, and it is almost certainly caused by a signal that has nothing to do with GPS.
Out-of-band interference is the most common cause of unexplained GPS degradation, and it is widely misunderstood because the interfering signal is not on the GPS frequency. It does not need to be.
The Problem: Your LNA Has a Breaking Point
GPS L1 sits at 1575.42 MHz. LTE cellular downlink occupies bands from roughly 700 MHz to 2700 MHz. Wi-Fi operates at 2.4 GHz and 5 GHz. These are not the same frequency as GPS, but they are close enough that a poorly designed or poorly filtered receive chain will allow significant power from those signals to reach the first amplifier in the chain: the low-noise amplifier, or LNA.
Here is where the physics turns against you. An LNA is a linear device only up to a point. Push enough power into it and it saturates, a condition called compression. Once an LNA is driven into compression, it stops amplifying cleanly and starts generating noise across the entire band. The GPS signal, which is already extremely weak (typically around -130 dBm at the antenna), gets buried in that self-generated noise floor.
The LNA has not failed. It is doing exactly what physics predicts when overloaded. But from the receiver's perspective, the result is the same as if the GPS signal had disappeared.
This is why you can have a clear sky, a good antenna, and still get poor position accuracy or a complete loss of fix in an urban environment dense with cellular infrastructure. The LTE signal from a nearby tower does not interfere with GPS by jamming it. It interferes by quietly compressing the amplifier that GPS depends on.
The Solutions
Fixing compression-driven GPS degradation requires attacking the problem before the LNA, not after it. Once the LNA is in compression, there is nothing downstream you can do to recover the signal.
The first line of defence is a dedicated GPS bandpass filter placed ahead of the LNA in the signal chain. A good SAW or BAW filter centred on GPS L1 with tight skirts will reject LTE and Wi-Fi power before it ever reaches the amplifier.
This is the most effective single fix and the standard approach in well-designed GPS front ends. The filter needs to have sufficient out-of-band rejection, typically 40 dB or more, to handle the power levels present in dense RF environments.
The second consideration is antenna placement and shielding. An antenna mounted on a ground plane and physically separated from cellular antennas will see less out-of-band power to begin with. If you are designing a system rather than troubleshooting a finished product, keeping the GPS antenna away from LTE and Wi-Fi radiators is worth more than any amount of filtering.
Third, check the LNA's 1dB compression point against the expected interference environment. An LNA with a compression point of -20 dBm may be adequate in a quiet rural setting and completely inadequate on a rooftop surrounded by cellular base stations. Selecting an LNA with a higher compression point, or adding an attenuator ahead of a high-compression-point LNA, can improve dynamic range at the cost of some noise figure.
Finally, if you are using a cascaded gain chain, review where your filtering sits relative to each gain stage. Amplifying the signal before filtering it amplifies the interference too, and pushes the following stages closer to compression. Filter early. Amplify after.
GPS signals are among the weakest in the entire radio spectrum. They have no margin for a compromised receive chain. The cellular networks surrounding them have none of that fragility. Protecting the LNA from out-of-band power is not optional: it is the entire job of the front end.
GPS Bandpass Filter
