Loading ...
Sorry, an error occurred while loading the content.

Re: [softrock40] Rocky SDR software

Expand Messages
  • Alex, VE3NEA
    ... { I should have mentioned that if the data from the sound card is first passed through a high pass filter before doing the FFT you can eliminate the data
    Message 1 of 9 , Aug 1 6:55 PM
    • 0 Attachment
      >The author of Rocky for SoftRock-40 at dxatlas.com suggested I ask
      >this question here. I'm hoping a radio guru can answer this.
      >
      >I downloaded and ran the Rocky software without and hardware frontend -
      >I don't have a SoftRock-40 or any other hardware yet.
      >
      >I noticed a signal at 7054.00 Hz. Is that comimg from my soundcard or
      >computer? There's nothing in the soundcard except a stereo out cable
      >to my speakers. The soundcard is a built-on-the-motherboard Intel
      >SigmaTel Audio, i.e., pretty low-end.
      >
      >I've added a screengrab in the photo album "Rocky"
      >
      >TIA

      {
      I should have mentioned that if the data from the sound card is first
      passed through a high pass filter before doing the FFT you can
      eliminate the data at 0 frequency.
      }

      This could be done more easily by zeroing a few FFT bins at the center.
      However, this does not make much sense because the data masked by the noise
      hump cannot be recovered anyway.

      The unusual thing about Fristen's screenshot is that the noise hump is not a
      typical bell shape, it has two peaks at about +/- 400 Hz from the center
      frequency. I thought that someone might be able to identify the source of
      the noise by its spectrum, so I suggested that Fristen posts his question
      here.

      73 Alex VE3NEA
    • k5nwa
      ... I would think an average of the surrounding bins might be better, why? Humans are not rational, that big spike make them worried, a zero on the bins goes
      Message 2 of 9 , Aug 1 8:36 PM
      • 0 Attachment
        At 08:55 PM 8/1/2007, you wrote:


        >This could be done more easily by zeroing a few FFT bins at the center.
        >However, this does not make much sense because the data masked by the noise
        >hump cannot be recovered anyway.
        >
        >The unusual thing about Fristen's screenshot is that the noise hump is not a
        >typical bell shape, it has two peaks at about +/- 400 Hz from the center
        >frequency. I thought that someone might be able to identify the source of
        >the noise by its spectrum, so I suggested that Fristen posts his question
        >here.
        >
        >73 Alex VE3NEA
        >

        I would think an average of the surrounding bins might be better,
        why? Humans are not rational, that big spike make them worried, a
        zero on the bins goes the other way, making it look weird, the
        average makes it look like the rest of the noise.



        Cecil
        K5NWA
        www.softrockradio.org www.qrpradio.com

        "Blessed are the cracked, for they shall let in the light."
      • Dave
        ... I ve seen such a thing when I had audio output coupling into the I and Q inputs (had to do with the grounding configuration of my D44 card at the time:
        Message 3 of 9 , Aug 2 8:04 AM
        • 0 Attachment
          > The unusual thing about Fristen's screenshot is that the noise hump is
          > not a typical bell shape, it has two peaks at about +/- 400 Hz from
          > the center frequency.

          I've seen such a thing when I had audio output coupling into the I and Q
          inputs (had to do with the grounding configuration of my D44 card at the
          time: http://groups.yahoo.com/group/softrock40/message/8216 ). The +/-
          peaks tended to move closer or farther from the center as I tuned across
          the center, changing the audio tone frequency resulting from detecting
          the center peak.

          Dave, WA3OFF
        • mxpklup1
          ... frontend - ... soundcard or ... cable ... first ... center. ... the noise ... hump is not a ... center ... source of ... question ... I uploaded more
          Message 4 of 9 , Aug 2 5:13 PM
          • 0 Attachment
            --- In softrock40@yahoogroups.com, "Alex, VE3NEA" <alshovk@...>
            wrote:
            >
            > >The author of Rocky for SoftRock-40 at dxatlas.com suggested I ask
            > >this question here. I'm hoping a radio guru can answer this.
            > >
            > >I downloaded and ran the Rocky software without and hardware
            frontend -
            > >I don't have a SoftRock-40 or any other hardware yet.
            > >
            > >I noticed a signal at 7054.00 Hz. Is that comimg from my
            soundcard or
            > >computer? There's nothing in the soundcard except a stereo out
            cable
            > >to my speakers. The soundcard is a built-on-the-motherboard Intel
            > >SigmaTel Audio, i.e., pretty low-end.
            > >
            > >I've added a screengrab in the photo album "Rocky"
            > >
            > >TIA
            >
            > {
            > I should have mentioned that if the data from the sound card is
            first
            > passed through a high pass filter before doing the FFT you can
            > eliminate the data at 0 frequency.
            > }
            >
            > This could be done more easily by zeroing a few FFT bins at the
            center.
            > However, this does not make much sense because the data masked by
            the noise
            > hump cannot be recovered anyway.
            >
            > The unusual thing about Fristen's screenshot is that the noise
            hump is not a
            > typical bell shape, it has two peaks at about +/- 400 Hz from the
            center
            > frequency. I thought that someone might be able to identify the
            source of
            > the noise by its spectrum, so I suggested that Fristen posts his
            question
            > here.
            >
            > 73 Alex VE3NEA
            >

            I uploaded more accurate screengrabs, where I slid the "dial"
            directly over the 7054.0 mark. In the first grab it was not - it was
            at 7053.70. Also, the first grab was looking at the signal in BPSK
            mode. The new grab is CW mode. The shape changes from mode to mode
            slightly, and more so when I slide the spectrum, and even more so
            when I change to zoom slider - from 600Hz down to 20Hz. Changing the
            volume to max alters the shape dramaticaly - see grab 7054-3.

            I guess my real question is why that frequency (7054)? And if I
            understand the responses correctly (Cecil), it's just a small
            voltage on the soundcard causing it?

            Thanks again.
          • k5nwa
            ... 7.054MHz where the spike is in this case is the frequency of the LO crystal oscillator, anything deviating from that frequency is either a positive
            Message 5 of 9 , Aug 2 6:16 PM
            • 0 Attachment
              At 07:13 PM 8/2/2007, you wrote:


              >I guess my real question is why that frequency (7054)? And if I
              >understand the responses correctly (Cecil), it's just a small
              >voltage on the soundcard causing it?
              >
              >Thanks again.

              7.054MHz where the spike is in this case is the frequency of the LO
              crystal oscillator, anything deviating from that frequency is either
              a positive frequency for higher frequencies or a negative frequency
              for lower frequencies. That frequency is typically slightly lower
              than the crystal frequency due to the fact that the oscillator used
              places a lot of capacitance on the crystal driving the oscillator
              frequency down several KHz.

              1.25mv (-40dB) offset in the sound card converters and amplifiers
              translates to -80 dB at the antenna (actually a little higher due to
              losses) given that the audio amplifier is running with 40 dB gain
              that would be over +30 dB above the typical noise floor on 40M or
              around -110dB or in plain English a big spike. You can easily have a
              much bigger offset so it doesn't take much to have a large spike at
              the LO frequency.

              That can be frustrating, two people have the sound card one will have
              a huge spike while another will have almost no spike due to the
              variations in offsets, voltages on the sound card, ground loops in
              the PC...etc.



              Cecil
              K5NWA
              www.softrockradio.org www.qrpradio.com

              "Blessed are the cracked, for they shall let in the light."
            • Dave
              ... Yeah, this does look to me like the audio feedback I described. Changing modes probably changes the appearance because of the change to the filter width
              Message 6 of 9 , Aug 3 7:24 AM
              • 0 Attachment
                > The shape changes from mode to mode slightly, and more so when I
                > slide the spectrum, and even more so when I change to zoom slider -
                > from 600Hz down to 20Hz. Changing the volume to max alters the shape
                > dramaticaly - see grab 7054-3.

                Yeah, this does look to me like the audio feedback I described.
                Changing modes probably changes the appearance because of the change
                to the filter width and position. Changes due to change in volume
                are a rather obvious result.

                In my case, both the detected audio out and the I/Q in ran through
                Tony's D44 interface and the original grounding configuration allowed
                audio output current to flow on the I/Q input shields, coupling
                output audio onto the I/Q signal.

                In your case, I believe you are operating with nothing connected to
                the I/Q input, so it probably makes it easy to pick up the audio.
                Try terminating the I/Q input (or grounding these inputs).

                Dave
              Your message has been successfully submitted and would be delivered to recipients shortly.