GQRX - unexpected plot when changing frequency range

Main forum to discuss RTL-SDR related topics.
KevinJones
Posts: 9
Joined: Mon May 04, 2020 4:38 pm

GQRX - unexpected plot when changing frequency range

Post by KevinJones » Mon May 04, 2020 5:36 pm

I have been attempting to analyze signals in GQRX. I will try to describe my observations as best I can, but the included screenshots should help.

The frequency range of the FFT plot is about 1.5 MHz. Thus when changing the center frequency by 1 MHz, there is some overlap of frequencies, where some of the frequencies in one plot also appear in the other. I would expect that in both cases the "shifted over" signals should appear the same in both plots, as long as I am not using any automatic correction such as AGC or squelch.

However, here is what I have observed:

1) In some cases, signals that appear in the graph when the center frequency is set to a given frequency, do not appear when the center frequency is increased or decreased by 1 MHz, even though the scale in both cases includes the frequencies where those signals appeared.

2) In some cases, when changing the center frequency by 1 MHz, there appears to be a compression/expansion algorithm applied which causes the amplitude of the plot in one case to be quite different than the other, appearing like the gain had been boosted.

Note that I am not using AGC, and squelch is reset (-150 dB). I have gone over all the settings the best I understand them, and I cannot find a reason for the described behaviour.

The following examples refer to the screenshots of the FFT plots/waterfalls.

Compare the screenshots taken when center frequencies were at 878 MHz and 879 MHz. At f0 = 878 MHz, there is a prominent signal occurring at 878.8 MHz, with slightly less prominent signals occurring at 878.34, 878.5, and 878.65 MHz. Compare with screenshot where f0 = 879 MHz, there is no distinction at those frequencies.

Also notice between those two graphs that the gain appears to be boosted, if you compare the range of frequencies which appear on both graphs.

The transition between 879 and 880 MHz appears to be more like expected, the detected signals are shifted to the left when increasing f0 by 1 MHz. The gain between the two graphs appear to be congruent.

The transition between 880 and 881 MHz again demonstrates the sudden change in gain, as it drops back down to a level that appears the same as the 878 MHz plot.

Image

Image

Image

Image

I've included shots of my settings:

Image

Image

Image

alanzfq
Posts: 140
Joined: Thu Oct 04, 2018 11:18 am

Re: GQRX - unexpected plot when changing frequency range

Post by alanzfq » Tue May 05, 2020 10:04 am

What signals are you trying to analise?
It seems all you show are spurii, maybe from all the electronic devices around. Maybe from overloading from some powerful transmitter.
If so there is no real way of predicting where they will be with a different LO frequency.
You need real, known signals to be sure of anything.
Alan

hotpaw2
Posts: 143
Joined: Sat Jan 14, 2017 11:07 pm
Contact:

Re: GQRX - unexpected plot when changing frequency range

Post by hotpaw2 » Tue May 05, 2020 10:31 pm

The signals in your waterfall screenshots look more like spurious intermodulation products or oscillator birdies than real RF signals. Try again after you find a "real" signal, e.g. one you can listen to or decode. Shift up or down a quarter of a frame after you find a signal with a station ID or something, and see if it stays in the waterfall.

KevinJones
Posts: 9
Joined: Mon May 04, 2020 4:38 pm

Re: GQRX - unexpected plot when changing frequency range

Post by KevinJones » Wed May 06, 2020 1:31 pm

Thanks for replying.

I am trying to analyze the output from a cellular repeater. The repeater downlink output is connected to an antenna, and the sdr input is connected to a receiving antenna. The antennas are in close proximity, about 2 feet. So maybe yes, it is overloading.

As far as using real known signals, the point of what I'm doing is to be able to detect and analyze what frequencies contain the data. Hence, if I already knew what the signals were, I wouldn't need to do this.

KevinJones
Posts: 9
Joined: Mon May 04, 2020 4:38 pm

Re: GQRX - unexpected plot when changing frequency range

Post by KevinJones » Wed May 06, 2020 3:41 pm

A little more background:

First of all I am very new to sdr.

I am investigating setting up a cell repeater. I live in a rural area and I am in a "hole". I can get cell signals from a few spots higher up, though the better the signal I get, the farther I am from the desired area. I would like to be able to tell the signal strength of the cell signals at various points.

One requirement I have is that I need to be able to get 3G Verizon for a particular device which will only operate on that. The nearest tower does offer 3G Verizon as the device works when within range. However I do not know whether it is CDMA, GSM, or ?, so I don't know what frequency band the tower is using for 3G Verizon, as the tower also operates on 2G and LTE (psuedo 4G). From what I can gather from research on the internet, Verizon may use either 850 or 1900 MHz band for 3G. And I don't even know how comprehensive that information is.

So the first step is I am taking a spectrum analyzer approach to start by identifying whether there are some signals in the either of those bands (or possibly others.)

Later, if I can do that, maybe at some point I can see if I can identify what network they are on, though that would be a much more advanced exercise.

hotpaw2
Posts: 143
Joined: Sat Jan 14, 2017 11:07 pm
Contact:

Re: GQRX - unexpected plot when changing frequency range

Post by hotpaw2 » Wed May 06, 2020 5:56 pm

The first step is being able to detect real signals is to test against real signals and their absence and make sure your equipment is calibrated to tell the difference. You might even need some sort of signal generator and some attenuators, etc. to do this.

alanzfq
Posts: 140
Joined: Thu Oct 04, 2018 11:18 am

Re: GQRX - unexpected plot when changing frequency range

Post by alanzfq » Thu May 07, 2020 7:01 am

Antennas 2 feet apart? You are likely to fry the receiver.
Local transmissions should be received strongly even with a dummy load.
Keep well away from the transmitter.
Alan

KevinJones
Posts: 9
Joined: Mon May 04, 2020 4:38 pm

Re: GQRX - unexpected plot when changing frequency range

Post by KevinJones » Fri May 08, 2020 2:43 pm

hotpaw2 wrote:
Wed May 06, 2020 5:56 pm
The first step is being able to detect real signals is to test against real signals and their absence and make sure your equipment is calibrated to tell the difference. You might even need some sort of signal generator and some attenuators, etc. to do this.
Can you recommend a signal generator ala Ali Express/Banggood that would be good for doing this? I don't have a lot in the budget...

KevinJones
Posts: 9
Joined: Mon May 04, 2020 4:38 pm

Re: GQRX - unexpected plot when changing frequency range

Post by KevinJones » Fri May 08, 2020 7:06 pm

alanzfq wrote:
Thu May 07, 2020 7:01 am
Antennas 2 feet apart? You are likely to fry the receiver.
Local transmissions should be received strongly even with a dummy load.
Keep well away from the transmitter.
Alan
So max input power is 10 dBm, is that correct?

That seems to be what I'm interpreting from here:

http://superkuh.com/gnuradio/R820T_data ... locked.pdf

alanzfq
Posts: 140
Joined: Thu Oct 04, 2018 11:18 am

Re: GQRX - unexpected plot when changing frequency range

Post by alanzfq » Sat May 09, 2020 7:07 am

You do not mention powers, how much power is transmitted? How much power is sent to the RTL?
Is 10mW the maximum before damage is done? It is a very big signal compared with normal antenna levels microvolls, microwatts.
As I say a dummy load is usually leaky enough to receive signals strongly from local transmitters. Or, just no antenna at all.
Alan

Post Reply