Re: Galileo Frequency Shifting and Filtering

[Marcus D. Leech]

On 01/03/2023 13:01, Kyeong Su Shin wrote:

Hello all:

Regarding the signal amplification:

• Amplifiers CAN improve the SNR, when used properly. The idea is to minimize the noise figure of the receive chain by putting a good LNA and filters (prevents LNA overload) in front of the chain (Friis law for noise). Since wideband SDRs often do not have great Rx frontend (claimed noise figures are often not met in real-world environments, due to the lack of preselector filtering), these amplifiers in active GNSS antennas do actually improve the signal quality.

    I'll point out that:
        (A) Most "active" GPS/GNSS antennas have an LNA, and the antenna
    structure is a printed-on-high-Er-ceramic patch antenna.
              Such antennas are quite narrow-band, so, there's already
    filtering ahead of the pre-amp.   Such antennas are also,
              unfortunately, lossy, even within the pass-band.

• That said, the gain level of such amplifiers is not that important (as long as the level is somewhere around 20 - 30 dB). Their noise figure is. Unfortunately, sellers and manufacturers often do not publish the typical noise figure of the circuits.

    If you care about noise figure, then your receiver noise-figure
    doesn't really matter as much as the LNA that physics and
      coaxial cable dictates you have right up at the antenna.  Unless
    your receiver is literally dangling off the antenna outdoors,
      its noise figure (as long as it isn't totally horrific) doesn't
    actually matter that much because you'll have an LNA right up
      at the antenna.    Active GPS/GNSS antennas satisfy this
    requirement already to a great extent.  The problem is that in
      most environments (except perhaps rural and middle-of-nowhere),
    you *need* some filtering ahead of the LNA, which
      necessarily eats into the noise figure of the LNA.  (See, as you
    say, Friis noise chain analysis).

• Still (as mentioned several times already), you are probably not going to see the real signals on your spectrum plot. Matched filtering is necessary (or maybe cyclostationary detectors?) . Using simulated signals is not a bad idea, I think, since that would make debugging a bit easier.
  

    The only situation in which you'd see the signal (as an extended
    increase in the apparent noise floor) is when your
      front-end noise figure is really really low, and perhaps you're
    pointing at only a single GPS satellite.
    
    The "trick" in GPS and similar spread-spectrum systems is that if
    you know the spreading code and offset, you can
      "despread" the signal, vastly increasing the effective SNR on the
    underlying BPSK signal and proceed from there.  But
      the signal architecture was *deliberately* designed so that you
    couldn't "see" it with ordinary equipment if you didn't
      know the spreading code.

Regards,

Kyeong Su Shin

---

보낸 사람: It's a cat <luoshumymail@gmail.com> 대신 discuss-gnuradio-bounces+ksshin=postech.ac.kr@gnu.org <discuss-gnuradio-bounces+ksshin=postech.ac.kr@gnu.org>
보낸 날짜: 2023년 3월 1일 수요일 오후 10:55
받는 사람: David Dima <daviddimavd@gmail.com>
참조: GNURadio Discussion List <discuss-gnuradio@gnu.org>
제목: Re: Galileo Frequency Shifting and Filtering

 

Amplifier only increase power (signal plus noise), they can't give you better SNR.

Anyway, It's much better to use a CW signal generator to help you debug the frequency converter chain (remember to keep output power low, e.g. -60dBm). You can use freq sink to check every stage of your flowgraph, and use Range / slider widget to control block parameters to search and check. If the signal disappears, or the passband doesn't meet your requirement, the problem can be easily located.

David Dima <daviddimavd@gmail.com> 于2023年3月1日周三 21:50写道：

Hi,

Sure, but the power of this kind of signals is below noise level. Current antenna amplifies 30dB but it still not enough to see anything in the spectrum (I have checked it alredy).

Regards,

David

El mié., 1 mar. 2023 7:27, It's a cat <luoshumymail@gmail.com> escribió:

If you have a signal generator, GNSS simulator or a high gain antenna, you can verify your settings by adding a spectrum display (Qt Freq Sink) or something like that to your receiver path.

Because many things can go wrong in your setup, if you misunderstand something. Just try to verify this directly. If there's a signal you should be able to see it on the spectrum display.

David Dima <daviddimavd@gmail.com> 于2023年2月28日周二 02:00写道：

Hi guys,

I am trying to simultaneously record Galileo's E5a and E5b signals as follows:

I have tried to record them separately by tuning the local oscillator in the middle of the signals bandwidths and sampling at 25MHz and it worked. But when I try to shift and filter to get the signals in a different file, it does not work. I have no idea what am i doing wrong.

Galileo's Signal Plan is as follows: