Can I change Gnuradio 3.8 to 3.7 ?

[Rachida SAROUI]

Hello everyone,

I'm working with Gnuradio 3.8 but I need a module that is compatible only
with 3.7. Can I change the version of Gnuradio from 3.8 to 3.7?

Thank you

Re: Can I change Gnuradio 3.8 to 3.7 ?

[Josh Morman]

By change the version, do you mean uninstall GNU Radio 3.8 and install 3.7,
or do you mean make the 3.7 modules work with 3.8 - the latter is not
possible because the changes in GR from 3.7 to 3.8 were significant and not
API compatible.

Josh

On Tue, Feb 1, 2022 at 4:47 AM Rachida SAROUI 
wrote:

> Hello everyone,
>
> I'm working with Gnuradio 3.8 but I need a module that is compatible only
> with 3.7. Can I change the version of Gnuradio from 3.8 to 3.7?
>
> Thank you
>

Re: Can I change Gnuradio 3.8 to 3.7 ?

[Rachida SAROUI]

I mean uninstall GNU Radio 3.8 and install 3.7,

Envoyé de mon iPhone

> Le 1 févr. 2022 à 12:02, Josh Morman  a écrit :
> 
> ou mean uninstall GNU Radio 3.8 and install 3.7,

Re: ETTUS N210 spurious

[Steve Hubbard]

Sorry for the late reply. I am guessing you are seeing aliases of the 
audio signal. Sampling a 1 kHz signal at 32 kHz produces aliases around 
multiples of the 32 kHz. In the signal processing chain the sampling 
rate will be increased in stages until the DAC sampling rate is reached. 
This process of interpolation involves inserting zero valued samples 
between the existing samples and then applying a digital low pass filter 
to attenuate the aliases. The stop band of the filter may be -70 dB so 
low level spurii remain. -70 dB is low enough for most intents and 
purposes and probably doesn't need fixing.


Steve

On 19/1/22 1:31 am, Fabien PELLET wrote:

Hello,

When I put a signal generator that produces a sinwave at 1Khz and that 
feeds an USRP Sink with a frequency center at 12MHz for example, I got 
spurs at +/-32Khz, +/-64KHz and +/-96KHz at the output of my N210 
(used with LFTX and LFRX). Whatever the frequency center of the USRP 
and the frequency of the baseband signal, I have to spurs around -70dBc.


Any idea on where it could come from ? Hardware EMC problem ? Digital 
treatment issue ? How to solve it ?


Best regards,

Fabien, F4CTZ

Proper BFSK modulation

[GNU Radio, the Free & Open-Source Toolkit for Software Radio]

Hi Community,
Since I have no one to verify me, maybe there is someone who would find a
minute and verify my flowgraph
for
the BFSK modulation, demodulation and synchronization. My goal is to wrap
bits into packets and create end to end simulation with custom data
transmission (e.g. test or file). My main concern in the attached flowgraph
is the *outer_deviation *parameter. What do you think? I am wondering if
this parameter should be chosen with some earlier calculations (e.g. to
make smooth phase changes in VCO). I have no goal of obtaining any bitrate
or baud rate, just playing with different modulation schemes and getting
familiar with the tool and its features.
Also additional questions:
- is the minimum frequency spacing between changing carriers is 1/2T where
T is the symbol duration?
- is there a rule how to choose proper frequency spacing in the FSK for
higher order (4FSK, 16 FSK etc)?
Thanks in advance
Marcin

https://gist.github.com/marcinsztajn/c713ec33fa3e3c5bdb9c467d8487d4fc

-- 

Marcin Puchlik

Re: ETTUS N210 spurious

[Fabien PELLET]

Hello,

Thanks for the reply. As I wrote, I try several sampling rate on the 
baseband I provide to the USRP without any effects on that spurs so it 
is not related to the sampling rate. I just put a signal source that 
feeds the USRP sink for that test.


It should be EMC related, as far as I understand reading all answer on 
that mailing list.


Thanks,

Best regards,

Fabien, F4CTZ.

Le 01/02/2022 à 12:12, Steve Hubbard a écrit :
Sorry for the late reply. I am guessing you are seeing aliases of the 
audio signal. Sampling a 1 kHz signal at 32 kHz produces aliases 
around multiples of the 32 kHz. In the signal processing chain the 
sampling rate will be increased in stages until the DAC sampling rate 
is reached. This process of interpolation involves inserting zero 
valued samples between the existing samples and then applying a 
digital low pass filter to attenuate the aliases. The stop band of the 
filter may be -70 dB so low level spurii remain. -70 dB is low enough 
for most intents and purposes and probably doesn't need fixing.


Steve

On 19/1/22 1:31 am, Fabien PELLET wrote:

Hello,

When I put a signal generator that produces a sinwave at 1Khz and 
that feeds an USRP Sink with a frequency center at 12MHz for example, 
I got spurs at +/-32Khz, +/-64KHz and +/-96KHz at the output of my 
N210 (used with LFTX and LFRX). Whatever the frequency center of the 
USRP and the frequency of the baseband signal, I have to spurs around 
-70dBc.


Any idea on where it could come from ? Hardware EMC problem ? Digital 
treatment issue ? How to solve it ?


Best regards,

Fabien, F4CTZ

Re: Can I change Gnuradio 3.8 to 3.7 ?

[Marcus Müller]

Dear Rachida,

we don't know how you've installed GNU Radio, hence it's impossible to advise how to 
uninstall it. Chances is: you *really* should not try.
I'd personally say: get used to "containers", and use an old Linux distro that ships your 
desired version of GNU Radio, and install it the distro-typical way (e.g. apt-get, dnf, 
pacman,…).


Do note, though, that GNU Radio 3.7 is officially end-of-life, and the community and the 
developers have less and less experience with it. There's nobody who will fix bugs in 3.7.


So, after you've made your module run, you'd have to port it to a newer version of GNU 
Radio, as the dependencies of GNU Radio 3.7 (Python2.5, cheetah, old versions of Boost, 
many 3.7–versions depend on Qt4…) are unavailable for modern operating systems.


Best regards,
Marcus

On 01.02.22 12:04, Rachida SAROUI wrote:

I mean uninstall GNU Radio 3.8 and install 3.7,

Envoyé de mon iPhone


Le 1 févr. 2022 à 12:02, Josh Morman  a écrit :

ou mean uninstall GNU Radio 3.8 and install 3.7,

Re: ETTUS N210 spurious

[Marcus Müller]

Hi Steve, hi Fabien,

agreed, this might simply insufficient *reconstruction filtering*; note that as Marcus 
Leech said, 70 dB is quite a lot, and I'd be surprised if an analog reconstruction filter 
would be much much better.


Best regards,
Marcus

On 01.02.22 12:12, Steve Hubbard wrote:
Sorry for the late reply. I am guessing you are seeing aliases of the audio signal. 
Sampling a 1 kHz signal at 32 kHz produces aliases around multiples of the 32 kHz. In the 
signal processing chain the sampling rate will be increased in stages until the DAC 
sampling rate is reached. This process of interpolation involves inserting zero valued 
samples between the existing samples and then applying a digital low pass filter to 
attenuate the aliases. The stop band of the filter may be -70 dB so low level spurii 
remain. -70 dB is low enough for most intents and purposes and probably doesn't need fixing.


Steve

On 19/1/22 1:31 am, Fabien PELLET wrote:

Hello,

When I put a signal generator that produces a sinwave at 1Khz and that feeds an USRP 
Sink with a frequency center at 12MHz for example, I got spurs at +/-32Khz, +/-64KHz and 
+/-96KHz at the output of my N210 (used with LFTX and LFRX). Whatever the frequency 
center of the USRP and the frequency of the baseband signal, I have to spurs around -70dBc.


Any idea on where it could come from ? Hardware EMC problem ? Digital treatment issue ? 
How to solve it ?


Best regards,

Fabien, F4CTZ

Re: ETTUS N210 spurious

[Marcus D. Leech]

On 2022-02-01 06:35, Fabien PELLET wrote:

Hello,

Thanks for the reply. As I wrote, I try several sampling rate on the 
baseband I provide to the USRP without any effects on that spurs so it 
is not related to the sampling rate. I just put a signal source that 
feeds the USRP sink for that test.


It should be EMC related, as far as I understand reading all answer on 
that mailing list.


Thanks,

Best regards,

Fabien, F4CTZ.

Could you share the exact flow-graph you're using?

Also, a spur at -70dBc can usually be ignored in any practical system -- 
it's 10,000,000 times weaker than the carrier.