I'm designing a system with a GSM modem and GPS receiver. In order to achieve out of band rejection I'm planning on putting a SAW filter right after the antenna. Often on the internet I see a few design options:

  1. LNA - SAW
  2. SAW - LNA
  3. LNA - SAW - LNA

For a standard small system GPS, I think option 3 is excessive. After reading, option 1 can cause non linear amplification of out of band signals due to the high power out of band signals from the GSM transmitter (up to 2W pulses). So that leaves option 2 (SAW -> LNA).

enter image description here

My GPS receiver is a SIM808 module, the documentation on the GPS front end is non existent. My current plan is to use a SAW filter with 0.9dB of insertion loss and forget the LNA all together, since using a 25x25mm patch antenna already gives me +3dB of gain, the IL from the SAW would be similar to using a smaller antenna which is often done in smaller phones I would guess. Also the device is going to be positioned in view of the sky, so the receive signal use case is almost ideal. Is there any real reason to have an LNA to boost the saw output when the SAW filter is less then an inch from the GPS receiver module? The way I see it is that it can only add noise to the system and make matching more difficult (SAW is already 50 Ohm and so is antenna).

Or am I missing something?

Additional Information:
SAW Filter: Digikey
-IL: 0.9dB
-NF: Not Listed.

LNA: Nothing specific picked out yet, I was looking at < 1dB NFs on Digikey and there are a lot available.

As stated in the question, there is no real specs for the GPS receiver built into the cellular modem I am using. I'm safely assuming the LNA is not great. So from that point of view, I guess a better external LNA with a higher IP3 could give better performance. I'm trying to avoid doping on another lead-less package is really what it comes down to.

  • \$\begingroup\$ What's the noise figure for the SAW filter? \$\endgroup\$
    – Andy aka
    Commented Nov 14, 2015 at 12:03
  • \$\begingroup\$ What SAW filter only has 0.9dB IL? \$\endgroup\$
    – johnnymopo
    Commented Nov 14, 2015 at 15:29
  • \$\begingroup\$ Noise figure of an attenuator is equal to its attenuation value \$\endgroup\$
    – tomnexus
    Commented Aug 26, 2021 at 15:48

2 Answers 2


You have to be guided by real use cases and experience, which probably means experiment in the absence of comprehensive data..

SAW -> LNA should provide the best strong signal performance (out of band signals won't overload the LNA and mix with each other into in-band signals).

LNA -> Saw should provide the best weak signal performance (weak signals are amplified before the loss and noise figure of the SAW. But look at its 3rd order intercept point (3IP) : a measure of its linearity to strong signals - can you overload it?

If the GPS receiver is nearly as sensitive - and linear in the presence of strong signals (same 3IP) - as the LNA, then the simplest circuit with fewest noise sources may also be the best performer. (And if the GPS overloads easily, an LNA ahead of it without selectivity will only worsen the situation)

So without more information, no single answer...


The benefits of putting the LNA before the filter is that the filter's insertion loss will not essentially add to the LNA Noise Figure. However since your filter has an incredibly good insertion loss for a saw (checked the datasheet, and had to pick my jaw up off the floor), I would put it in front of the LNA.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.