discuss-gnuradio
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] File source problem


From: mleech
Subject: Re: [Discuss-gnuradio] File source problem
Date: Thu, 16 Jun 2016 11:06:59 -0400
User-agent: Roundcube Webmail/1.1.5

Among SDRs, there is no universal "model" for things like total gain, acceptable baseband magnitudes, etc. Even if there were, expected variations in analog behavior mean that you usually have to tweak things to make it work across multiple hardware.  That's just an attribute of the analog world that "all digital" folks may not be used to.

 

 

 

 

On 2016-06-16 10:09, Olivier Goyette wrote:

****UPDATE****

Even without any Costa loop or clock recovery, it works. It seems that the trouble was only the TX amplitude.

2016-06-16 9:36 GMT-04:00 Olivier Goyette <address@hidden>:
YEAAAAH ! it works now.

Only 2 major things to do :

1st - I multiplied the TX amplitude by 0.5 with a multiply const block

2nd - I used a Costa loop before my demod and it works great.

My video got a tiny bit of error but I guess that with a special coding (reed solomon ) I could get it back in good shape.

Now I gotta to test it but with CPFSK and see if I can get the same result !

Thank you all :) !

2016-06-16 3:36 GMT-04:00 Johannes Demel <address@hidden>:
Hi Olivier,

I guess Nick already pointed out the important bits. Also, the warnings GRC prints to the shell should help you.

Could you solve your file source issues? Your questions here seem to be related to a different issue.

Tx flowgraph: get rid of the throttle block. (also see the warning)
spectrum: this seems inconsistent. the TX spectrum suggests OFDM at first glance. The MDO spectrum looks completely different. You should investigate this.
Most importantly here, the RX spectrum looks a lot like clipping/saturation or alike. In any case non-linear effects seem to cause trouble here.
TX time: It is important to keep the TX values < 1.0! Everything else will be clipped. Non-linear effects.
RX time: looks like a frequency offset. You need synchronization first. Really tight synchronization, FLL, Costas-Loop etc. You may simulate that by multiplying with a signal source with a really low frequency like 10Hz (or maybe even 1Hz, play a bit around).
Rx constellation: you can observe a frequency offset here.
Rx time 2: my guess: it's the same as the other except for a scaling factor. Maybe a different radio channel, maybe because the other hardware behaves a bit differently.

Cheers
Johannes



On 15.06.2016 22:24, Olivier Goyette wrote:
I'm giving a copy of the email I just sent to Ettus Research. I think
that it's been a hardware problem since the beginning but i'm waiting
for their response.

Check it out, maybe some of view have already seen this kind of bug

*************************E-mail*****************************************************

Hi !

My name is Olivier and I'm actually working with your USRP N210 and
GNURadio Companion. It's been a couple of week since I started using
your equipment and yet, I already had issues with it. Doing a simple
modulation/demodulation seems impossible and everything I came up with
until now, suggest me that it's a hardware problem. So, I took quite a
lot of pictures for you to see what's wrong and I'll share the links to
my dropbox so it will be easier like that.

What I'm trying to do is a PSK mod and demod. I know it works because
one of my colleague is using a Nutaq Zepto SDR and he's been using the
same flowgraph as mine in an other project and it works perfectly. So,
what we did is the following setup :
https://www.dropbox.com/s/ottq341sob5z90r/Actual_setup.jpg?dl=0

The upper radio represents the TX link (WBX card) and the one at the
bottom represent the RX link (SBX card).

This is the flowgraph on the TX radio :
https://www.dropbox.com/s/ir5ehp6caiy525a/TX_side.jpg?dl=0

This is the flowgraph on the RX radio :
https://www.dropbox.com/s/tl95cwbstwzl4r8/RX_side.jpg?dl=0

What we're trying to do is to transmit a video stream from one computer
to another. Unfortunately, on the RX computer, the file supposedly
containing the video is empty (0 bytes when you right click on it and go
to the properties panel).

This is the TX spectrum :
https://www.dropbox.com/s/pzsn8oo74sgrehp/TX_spectrum.jpg?dl=0 and this
is the spectrum when you look at it with a spectrum analyser :
https://www.dropbox.com/s/5bks7aksih8ctzh/TX_MDO_spectrum.jpg?dl=0

This is the RX spectrum :
https://www.dropbox.com/s/dlsovfxovv4n9tm/RX_spectrum.jpg?dl=0

This is the TX time domain graph :
https://www.dropbox.com/s/q25sbjukyzsyymz/TX_time_domain.jpg?dl=0

This is the RX time domain graph :
https://www.dropbox.com/s/qetgf0cr4s7d5hb/RX_time_domain_1.jpg?dl=0

This is the TX constellation :
https://www.dropbox.com/s/gflb0f65oa0m1rl/TX_constellation.jpg?dl=0

This is the RX constellation :
https://www.dropbox.com/s/4j0hnhl97v2uj7g/RX_constellation.jpg?dl=0

What seems strange to me is the time domain graph at RX. It's like if
the imaginary part and the real part of the signal were modulated onto
whatever signal that appears to be there. I tried my RX link on one of
the other radio I have and look at the result :
https://www.dropbox.com/s/f1al75vgo5z2wcf/RX_time_domain_2.jpg?dl=0 .
The time domain is completely different but still it is a SBX card and I
haven't touch any settings in my flowgraph to keep consistency.

So, now I'm asking for your help because my knowledge is pretty basic
with this equipment and I'd like to have your input about that.

For your information, I did reload the firmware and the bitfile into the
FPGA before sending this email so it's not necessary to ask me to do it.

I sincerely hope this will lead somewhere as I'm out of options right now.

Thank you for your time and have a nice day !

2016-06-15 13:58 GMT-04:00 Marcus Müller <address@hidden
<mailto:address@hidden>>:

    Hi Olivier,

    what does this mean:
    The 1st file contains only '1' and the other only '0'.
    How did you generate these files. Without knowing better, I'd say
    this is a case for our File Format FAQ:

    http://gnuradio.org/redmine/projects/gnuradio/wiki/FAQ#What-is-the-file-format-of-a-file_sink-How-can-I-read-files-produced-by-a-file-sink

    Best regards,
    Marcus

    On 06/15/2016 12:25 AM, Olivier Goyette wrote:
    This is non sense seriously. This is my flowgraph :
    <https://www.dropbox.com/s/4cfptq5419kvyhz/file_source.png?dl=0>https://www.dropbox.com/s/4cfptq5419kvyhz/file_source.png?dl=0

    When I send only zeros '0', this is what I get :
    <https://www.dropbox.com/s/aipf6wsj5ekwxnb/zeros.png?dl=0>https://www.dropbox.com/s/aipf6wsj5ekwxnb/zeros.png?dl=0

    When I send only ones '1' this is what I get :
    <https://www.dropbox.com/s/g38hgrtc7s6exyt/ones.png?dl=0>https://www.dropbox.com/s/g38hgrtc7s6exyt/ones.png?dl=0

    When I interleave both ones and zeros '1' & '0' , this is what I
    get : https://www.dropbox.com/s/a1cfs0tebmywz0l/interleave.png?dl=0

    What's wrong seriously ? If I can't even modulate properly, how am
    I suppose to make this work ?

    2016-06-14 9:54 GMT-04:00 Olivier Goyette
    <address@hidden <mailto:address@hidden>>:

        Hi

        I have a problem.  On my first setup, I use 2 file source. The
        1st file contains only '1' and the other only '0'. The 2 file
        sources go through a selector and end up in a CPFSK mod block.
        When I run this, switching from '1' to '0' makes the frequency
        drift from + to - FSK_deviation like it should do normally.
        Now, when I use a unique file source containing '1' and '0'
        (actually I wrote up the alphabet in a file) and I run this,
        this is what I get :
        https://www.dropbox.com/s/4tdswnmp79lj67c/unique_file_source.png?dl=0
        . Any Idea why it's doing this. Is it the rate of transmission
        that does this, like if it was to fast ?

        my setup :
        https://www.dropbox.com/s/snamjalklskfdzh/single_file.png?dl=0

        Thank you




    _______________________________________________
    Discuss-gnuradio mailing list
    address@hidden <mailto:address@hidden>
    https://lists.gnu.org/mailman/listinfo/discuss-gnuradio


    _______________________________________________
    Discuss-gnuradio mailing list
    address@hidden <mailto:address@hidden>
    https://lists.gnu.org/mailman/listinfo/discuss-gnuradio




_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio


_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

reply via email to

[Prev in Thread] Current Thread [Next in Thread]