Strange behavior near 114Mhz

Hi there!
Playing around limesdr found strang behavior after fm broadcast frequencies. Near 114MHz I saw something like LTE signal.
I connected the generator, set 114Mhz at gqrx and began to increase the frequency.
Start point is FM modulated 200kHz (for better perception) signal:

Look what I got when generator shows 321MHz:


And here is 343MHz on generator:

One more example:
568MHz


571Mhz

Settings for gqrx:
settings

I dont thinks it is normal behavior, this kind of messed up spectrum make it useless on some frequencies.
Did I broke something?

Stuff like that is usually from your own computer system.
Make sure you have ferrite beads on your USB cables and have a shielded enclosure for your LimeSDR or better yet and if at all possible use modified ferrite bead enabled USB cables with only data lines with the ground shield shell removed at the radio side and a -linear- power supply to run the radio.
Computer noise will be all over the spectrums and you will even see some if it moving around or jumping frequencies, computers and monitors are noisy little buggers.

I tried different cables and different computers. I’ll try to move from my lab to test in other envirement, but now its more likely that something not filtered in lime.
Here is 155.060MHz:


and here 465.200:

Looks very similar.
In 100-108MHz I can see tracks on GSM basestations:

small peaks it a wfm stations, but big peak at 105.7, 102.6, 102 is sounds like GSM.

Important note: I don’t see this “basestations” when use rtl dongle in the same computers. So. I’m sure that something misconfigured in my lime.

Have you placed some band pass filtering on the front end?
If it isn’t your computer it is something else in your environment, try a band pass filter to see if it goes away.
Also just changing USB cables isn’t going to help.
You have to actively isolate the lines and grounds from the computer and shield the radio.
If you haven’t done that you wil chase gremlins forever without know where they are coming from.
Computer noise sounds like GSM, you can be sure it isn’t GSM however at those frequencies and signal bandwidths.
I got a LOT of those same kinds of signals as well as stationary and moving spikes and random noise floor uprisings all over the place until I got filters on the front end.
And don’t use an RTL dongle as a baseline, they are practically deaf compared to these at 150MHz, for example I get 5 miles on my RTL for AIS, swapping out to LimeSDR I get out 28 miles.

The key thing to keep in mind here is that the LimeSDR has a wide open 60 MHz front end, if you aren’t shielding and filtering the hell out of it you will get mixing products, harmonics and other noise all over the place. It is just the nature of the beast. Start with shielding and isolating from the computer and its power supply and band pass filters on the antenna, I think you’ll find that will cure nearly all of your random signal issues.

One problem is that you are using bandwidth equal to the sample rate i.e. 10M, giving no space for filter rolloff. See what happens when you set bandwidth to 5M and keep sample rate at 10M. Not sure your problem is actually caused by this ,but it is worth testing for…

You mean lime sdr is mostly unusable as is? You talking about some analog device to be placed in front of limesdr?
(As far as I understand @gasparka talking about digital filter)

While googling I found this thread https://news.ycombinator.com/item?id=16497073 by the way.

I am talking about analog filters inside the LMS7 chip; setting the bandwidth basically programs the analog filter passband (there is a field for this in gqrx settings). Lime also has digital filters, but they are not used by default and are useless if the analog stage is already swamped by aliasing.

I see you don’t understand what you bought.
In analogue terms the LimeSDR is pretty much a glorified high end state of the art mixer with a couple of oscillators attached to it with a super wide tuning range.
It has no real front end filtering, the A/D convertor is out there sucking up 60Mhz of spectrum at a time and downconverting it all, you are just using software to select portions of that downconversion to listen to.
It is NOT a complete, ready to run radio solution, it does indeed require analogue front end filtering for the part of the spectrum you want to listen to or you will get mixing products and images from the whole 60MHz bandwith it is downconverting even if your software is only decoding 10MHz,
The digital filters won’t help, this is all on the front end like any other radio analogue or SDR.
Also because it is a DDC/DUC radio it needs to be in a metal grounded case for shielding purposes and it needs to be electrically isolated as far as possible from the USB ports and power supplies on the computer (they are always giant wide spectrum noise machines no mattter how much you spent on it).
This is what we call a component of a system, the heart of the system actually, it requires a lot of other components outside of it to make it a complete system.

That link you gave is correct, the filters on the stock board if not modified are designed for the GSM/LTE/IoT crowd and not for general SWL or Ham listening. Those of us that are doing that have had to modify our boards filters (usually the low RX filter) to pass more of the lower bands in. This is explained elswhere on this board.

That’s a question! Looks like analog filtering is not working. I see in manuals that limesdr have RXLPF, but cant figure out is it works. @M0GLO says I have to add this filter myself.

Yes, the internal filters are just that, internal, and have no affect on filtering the front end A/D convertor.

Now I get it. :slight_smile: That explains why GSM stations have better SNR than wfm radio stations.

Keep in mind that those weren’t GSM signals at 114MHz: https://en.wikipedia.org/wiki/GSM_frequency_bands
You won’t be seeing any GSM signals lower than 800MHz any more.

Give my PSA a look over, it will help I think:

Thanks for your advices. Still surpised that so promising device useless without front-end. I hoped to work without front-end and have some “not optimal” results in low frequencies but not such a mess.

The VHF world, in any built-up area, is a MESS. Strong signals everywhere, giving LOTs of opportunties for inter-modulation products that appear very early in the analog signal processing, which means they cannot be removed in later digital processing. This is just pretty-normal analog RF stuff.

People tend to think of cheap SDRs as “ready to go for any application at all”. In reality, they are components in an overall RF system design. Just as the CPU chip in your computer is a component in your overall computer system.

It isn’t reasonable (yet, wait a few years) to expect a radio that is tunable over such a wide range to also have low-loss analog filter-banks “up front” that are perfectly tailored to your application. Which means that you end up having to do a bit of RF engineering yourself – think of adding up-front filters as no different than selecting appropriate antennae for an application.

NOW, there ARE SDRs out there that are more application specific, like some of the HF-only SDRs that have switchable filter-banks built-in. But if you try to map the same idea over the entire range of the LimeSDR (or others, like HackRFs, BladeRFs, USRPs, etc), then you end up with a very-large and complex switchable filter bank, much of it likely useless for the applications you have in mind…

1 Like

All commercial SDRs have the filtering added, if you had bought a more expensive one from a commercial supplier it would have some filtering in the case already installed.
The LimeSDR is what is known as a development board, it is a bare bones DDC/DUC demonstrator board for their LMS7002 radio chip. It has some very rudimentary wide band filters on the front end optimised for their target clients (big LTE tower equipment manufacturers and facility repeater designers).
It’s not a radio, it is just part of a radio and though you can suck it up and suffer the slings and arrows of our filthy dirty noisy spectrum with it as is to truly enjoy it as youi want it to be requires some engineering on your part.
Even if it is only buying some filters for where you want to listen.

Note that Lime already has analog filters (RXLPF) before the ADC stage:
Screenshot from 2018-07-05 10-26-57

Also note that the performance of those filters is not that bad:
Screenshot from 2018-07-05 10-30-41

Could anyone explain how adding extra analog filters helps the situation? Is it because filtering is before mixing?

exactly the filter before the first active semiconductor is the one which helps You to get rid of the overload on the amplifiers and mixers from strong out of band signals. This overload generates so called intermodulations which are new generated frequencies. Once generated the intermodulations can no longer be distinguished from wanted signals. So the task is to get rid of the unwanted signals as early as possible in Your signal chain. Already the antenna can bet he first filtering element.

Hopefully this helps in understanding the concept.

Thomas

Note the curves, they are low pass filters.
AND they are already into the mixers (where all intermod and mixing procducts occur).
They won’t help with intermod and overloading.
BAND PASS filtering what is required, just like in every other commercial radio out there.

1 Like

The filter is applied to quadrature signal, turning it into bandbass filter i.e. the lowpass filter also applies to the negative spectrum (with mirrored response).