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

Re: [olivia] new crest factor figures - new MT63 test

Expand Messages
  • Tomi Manninen
    ... The symbols in Olivia are shaped. See http://homepage.sunrise.ch/mysunrise/jalocha/mfsk_spec.html -- Tomi Manninen / OH2BNS / KP20JF74
    Message 1 of 25 , Mar 15 10:24 AM
    • 0 Attachment
      On Tue, 2005-03-15 at 16:18, Mark Miller wrote:

      > Question for the group: Why does Olivia have a 7.22 dB crest factor?

      The symbols in Olivia are shaped. See

      http://homepage.sunrise.ch/mysunrise/jalocha/mfsk_spec.html

      --
      Tomi Manninen / OH2BNS / KP20JF74
    • Paul L Schmidt
      ... True, but not *that* much shaping... Here are my numbers: mfsk_tx random.dat random.mfsk sox -t raw -sw -r8000 random.mfsk -t wav /dev/null stat Maximum
      Message 2 of 25 , Mar 15 3:28 PM
      • 0 Attachment
        Tomi Manninen wrote:
        > On Tue, 2005-03-15 at 16:18, Mark Miller wrote:
        >
        >
        >>Question for the group: Why does Olivia have a 7.22 dB crest factor?
        >
        >
        > The symbols in Olivia are shaped. See
        >
        > http://homepage.sunrise.ch/mysunrise/jalocha/mfsk_spec.html
        >

        True, but not *that* much shaping...

        Here are my numbers:

        mfsk_tx random.dat random.mfsk

        sox -t raw -sw -r8000 random.mfsk -t wav /dev/null stat

        Maximum amplitude: 0.538757
        Minimum amplitude: -0.538757
        Midline amplitude: 0.000000
        Mean amplitude: 0.000000
        RMS amplitude: 0.329407

        .538757 / .329407 = 1.63522

        20 log (1.63522) = 4.27 dB ... which needs
        to be dropped by 3 dB since we're comparing an RMS
        voltage to an instantaneous peak (rather than the RMS
        value of the signal at the peak).

        That puts it down to about a 1.63 dB crest factor...

        e.g. if you're running 50 watts RMS output, your
        PEP will be 67 watts... which seems reasonable for
        what I've experienced.
      • Mark Miller
        ... I used the meter and scope method and came up with 1.56 dB, pretty close isn t it Paul. I have posted the oscilloscope picture at:
        Message 3 of 25 , Mar 15 6:45 PM
        • 0 Attachment
          At 05:28 PM 3/15/2005, you wrote:
          >That puts it down to about a 1.63 dB crest factor...


          I used the meter and scope method and came up with 1.56 dB, pretty close
          isn't it Paul. I have posted the oscilloscope picture at:

          http://home.comcast.net/~mdmiller7/crest_factor/olivia_peak.jpg

          73,

          Mark N5RFX
        • Mark Miller
          ... Jose, I agree. When I use FSK I don t see any crest factor. When I use AFSK with MixW I do see a very small envelope. 73, Mark N5RFX
          Message 4 of 25 , Mar 15 6:47 PM
          • 0 Attachment
            At 11:13 AM 3/15/2005, you wrote:
            >There SHOULD BE no difference with FSK, if you are
            >just shifting frequencies and not doing any amplitude
            >modulation. Frequency modulation just shifts energy
            >from the carrier to the sidebands, but total energy
            >should remain the same.


            Jose,

            I agree. When I use FSK I don't see any crest factor. When I use AFSK
            with MixW I do see a very small envelope.

            73,

            Mark N5RFX
          • Paul L Schmidt
            ... What do you see if you hook the sound card output up to the scope or other measurement device -- an envelope or just changing frequency? I see varying
            Message 5 of 25 , Mar 15 7:10 PM
            • 0 Attachment
              Mark Miller wrote:
              > I agree. When I use FSK I don't see any crest factor. When I use AFSK
              > with MixW I do see a very small envelope.

              What do you see if you hook the sound card output up to the 'scope or
              other measurement device -- an envelope or just changing frequency?

              I see varying output depending on frequency when I run 250 Hz Olivia,
              since it puts everything in the 500-1000 Hz audio range going into
              the mic jack. Unfortunately, the audio response of my rig isn't flat -
              it starts dropping off about 700 Hz or so (although not badly until
              down to about 300 Hz).

              73,

              -ps
            • Mark Miller
              ... I used the meter and scope method and came up with 1.56 dB, pretty close isn t it Paul. I have posted the oscilloscope picture at:
              Message 6 of 25 , Mar 15 7:38 PM
              • 0 Attachment
                At 05:28 PM 3/15/2005, you wrote:
                >That puts it down to about a 1.63 dB crest factor...


                I used the meter and scope method and came up with 1.56 dB, pretty close
                isn't it Paul. I have posted the oscilloscope picture at:

                http://home.comcast.net/~mdmiller7/crest_factor/olivia_peak.jpg

                73,

                Mark N5RFX
              • Mark Miller
                ... I see a very tiny envelope. It is so tiny that when I photograph it, it looks like only changing frequency. 73, Mark N5RFX
                Message 7 of 25 , Mar 15 7:44 PM
                • 0 Attachment
                  At 09:10 PM 3/15/2005, you wrote:
                  >What do you see if you hook the sound card output up to the 'scope or
                  >other measurement device -- an envelope or just changing frequency?


                  I see a very tiny envelope. It is so tiny that when I photograph it, it
                  looks like only changing frequency.

                  73,

                  Mark N5RFX
                • Patrick
                  Tomi I do not agree with your interpretation of the Crest-factor. You will find anywhere that the crest-factor of a pure sinewave is 1.414... -this is the peak
                  Message 8 of 25 , Mar 16 1:13 AM
                  • 0 Attachment
                    Tomi
                    I do not agree with your interpretation of the Crest-factor.
                    You will find anywhere that the crest-factor of a pure sinewave is
                    1.414...
                    -this is the peak voltage value, divided by the RMS voltage value

                    If you keep thinking in terms of voltage values converting to dB
                    gives 20log1.414 = 3dB
                    A pure sinewave has a 1.414 or 3dB crest factor...

                    In practice any modulation will have a crest-factor equal(pure-FSK)
                    or superior to 3dB...

                    Now forget about the crest factor and just consider what is
                    interesting for us: mean to peak power values -both being "average
                    power values"- (average power being the square of the RMS voltage
                    divided by the load (*))
                    - In this case the peak (avg) power is the wattmeter reading when
                    you "tune" or when you send a RTTY mark
                    - the mean (avg) power would be a thermical-wattmeter (bolometer)
                    reading... (a normal wattmeter may indicate anything between mean and
                    peak depending among other things on the instrument technology...)

                    I don't have a bolometer, but my conclusion from waveform analysis
                    (and from temperature elevation of my IC706 !)is that the "mean to
                    peak" for Olivia is in the 4dB area, which is something similar to
                    CW... So about 40W with a 100W rated transceiver.

                    (*)As mentionned in ARRL handbook RMS-power is a mathematical
                    curiosity whcih has no physical significance. In practice you'll see
                    everywhere references to RMS-power, but it should be called "average
                    power" (the product of RMS-voltage by RMS-current). The english
                    terminology is extremely confusing...







                    --- In oliviadata@yahoogroups.com, Paul L Schmidt <k9ps@a...> wrote:
                    > Tomi Manninen wrote:
                    > > On Tue, 2005-03-15 at 16:18, Mark Miller wrote:
                    > >
                    > >
                    > >>Question for the group: Why does Olivia have a 7.22 dB crest
                    factor?
                    > >
                    > >
                    > > The symbols in Olivia are shaped. See
                    > >
                    > > http://homepage.sunrise.ch/mysunrise/jalocha/mfsk_spec.html
                    > >
                    >
                    > True, but not *that* much shaping...
                    >
                    > Here are my numbers:
                    >
                    > mfsk_tx random.dat random.mfsk
                    >
                    > sox -t raw -sw -r8000 random.mfsk -t wav /dev/null stat
                    >
                    > Maximum amplitude: 0.538757
                    > Minimum amplitude: -0.538757
                    > Midline amplitude: 0.000000
                    > Mean amplitude: 0.000000
                    > RMS amplitude: 0.329407
                    >
                    > .538757 / .329407 = 1.63522
                    >
                    > 20 log (1.63522) = 4.27 dB ... which needs
                    > to be dropped by 3 dB since we're comparing an RMS
                    > voltage to an instantaneous peak (rather than the RMS
                    > value of the signal at the peak).
                    >
                    > That puts it down to about a 1.63 dB crest factor...
                    >
                    > e.g. if you're running 50 watts RMS output, your
                    > PEP will be 67 watts... which seems reasonable for
                    > what I've experienced.
                  • Patrick
                    Sorry Tomi, this was a reply to Paul s post, not to your... Patrick ... and ... see ... called average
                    Message 9 of 25 , Mar 16 1:38 AM
                    • 0 Attachment
                      Sorry Tomi, this was a reply to Paul's post, not to your...
                      Patrick

                      --- In oliviadata@yahoogroups.com, "Patrick" <f6irf@f...> wrote:
                      >
                      > Tomi
                      > I do not agree with your interpretation of the Crest-factor.
                      > You will find anywhere that the crest-factor of a pure sinewave is
                      > 1.414...
                      > -this is the peak voltage value, divided by the RMS voltage value
                      >
                      > If you keep thinking in terms of voltage values converting to dB
                      > gives 20log1.414 = 3dB
                      > A pure sinewave has a 1.414 or 3dB crest factor...
                      >
                      > In practice any modulation will have a crest-factor equal(pure-FSK)
                      > or superior to 3dB...
                      >
                      > Now forget about the crest factor and just consider what is
                      > interesting for us: mean to peak power values -both being "average
                      > power values"- (average power being the square of the RMS voltage
                      > divided by the load (*))
                      > - In this case the peak (avg) power is the wattmeter reading when
                      > you "tune" or when you send a RTTY mark
                      > - the mean (avg) power would be a thermical-wattmeter (bolometer)
                      > reading... (a normal wattmeter may indicate anything between mean
                      and
                      > peak depending among other things on the instrument technology...)
                      >
                      > I don't have a bolometer, but my conclusion from waveform analysis
                      > (and from temperature elevation of my IC706 !)is that the "mean to
                      > peak" for Olivia is in the 4dB area, which is something similar to
                      > CW... So about 40W with a 100W rated transceiver.
                      >
                      > (*)As mentionned in ARRL handbook RMS-power is a mathematical
                      > curiosity whcih has no physical significance. In practice you'll
                      see
                      > everywhere references to RMS-power, but it should be
                      called "average
                      > power" (the product of RMS-voltage by RMS-current). The english
                      > terminology is extremely confusing...
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      > --- In oliviadata@yahoogroups.com, Paul L Schmidt <k9ps@a...> wrote:
                      > > Tomi Manninen wrote:
                      > > > On Tue, 2005-03-15 at 16:18, Mark Miller wrote:
                      > > >
                      > > >
                      > > >>Question for the group: Why does Olivia have a 7.22 dB crest
                      > factor?
                      > > >
                      > > >
                      > > > The symbols in Olivia are shaped. See
                      > > >
                      > > > http://homepage.sunrise.ch/mysunrise/jalocha/mfsk_spec.html
                      > > >
                      > >
                      > > True, but not *that* much shaping...
                      > >
                      > > Here are my numbers:
                      > >
                      > > mfsk_tx random.dat random.mfsk
                      > >
                      > > sox -t raw -sw -r8000 random.mfsk -t wav /dev/null stat
                      > >
                      > > Maximum amplitude: 0.538757
                      > > Minimum amplitude: -0.538757
                      > > Midline amplitude: 0.000000
                      > > Mean amplitude: 0.000000
                      > > RMS amplitude: 0.329407
                      > >
                      > > .538757 / .329407 = 1.63522
                      > >
                      > > 20 log (1.63522) = 4.27 dB ... which needs
                      > > to be dropped by 3 dB since we're comparing an RMS
                      > > voltage to an instantaneous peak (rather than the RMS
                      > > value of the signal at the peak).
                      > >
                      > > That puts it down to about a 1.63 dB crest factor...
                      > >
                      > > e.g. if you're running 50 watts RMS output, your
                      > > PEP will be 67 watts... which seems reasonable for
                      > > what I've experienced.
                    • Mark Miller
                      ... Patrick, Take a look at this article. http://rfdesign.com/mag/radio_crest_factor_analysis/ Toward the bottom of the article you will see two definitions
                      Message 10 of 25 , Mar 16 1:45 AM
                      • 0 Attachment
                        At 03:13 AM 3/16/2005, you wrote:
                        >Tomi
                        >I do not agree with your interpretation of the Crest-factor.
                        >You will find anywhere that the crest-factor of a pure sinewave is
                        >1.414...
                        >-this is the peak voltage value, divided by the RMS voltage value


                        Patrick,

                        Take a look at this
                        article. http://rfdesign.com/mag/radio_crest_factor_analysis/

                        Toward the bottom of the article you will see two definitions for crest factor.

                        Peak crest factor

                        Next the peak crest factor for the RF signal, W(t), is addressed:

                        This is the ratio of the instantaneous peak voltage squared to the average
                        output power, but for translated signals (IF or RF) it is customary to
                        consider the crest factor for the envelope signal. Because most RF devices
                        are characterized with sinusoidal signals at either specified or measured
                        power levels, having the crest factor as a function of a sinusoidal
                        signal's power level is more appropriate.

                        Envelope crest factor

                        To determine the envelope crest factor, first determine the power of a
                        sinusoidal signal that has the same peak voltage as the modulated signal's
                        peak voltage. This value is then divided by the average power of the
                        modulated signal. The result is defined as the translated signal's envelope
                        crest factor.

                        Also you will find that peak power does not always mean the peak voltage
                        squared. Sometimes it means RMS voltage squared. Take a look at the
                        Alliance for Telecommunications Industry Solutions Web page. It is nicely
                        laid out.

                        http://www.atis.org/tg2k/t1g2k.html

                        crest factor http://www.atis.org/tg2k/_crest_factor.html
                        peak to average ratio http://www.atis.org/tg2k/_peak-to-average_ratio.html
                        peak power http://www.atis.org/tg2k/_peak_power_output.html

                        As you can see peak power is defined as the peak average power.

                        73,

                        Mark N5RFX
                      • Tomi Manninen
                        ... Yes. I didn t look at the figures that closely. My quick tests with gMFSK and octave seem to agree with your figures. -- Tomi Manninen / OH2BNS / KP20JF74
                        Message 11 of 25 , Mar 16 2:15 AM
                        • 0 Attachment
                          On Tue, 15 Mar 2005, Paul L Schmidt wrote:

                          >> The symbols in Olivia are shaped. See
                          >>
                          >> http://homepage.sunrise.ch/mysunrise/jalocha/mfsk_spec.html
                          >
                          > True, but not *that* much shaping...

                          Yes. I didn't look at the figures that closely. My quick
                          tests with gMFSK and octave seem to agree with your figures.

                          --
                          Tomi Manninen / OH2BNS / KP20JF74
                        • Patrick
                          Yes Mark, we agree that the terminology is not very clear... http://www.dliengineering.com/downloads/crest%20factor.pdf http://www.tinaja.com/glib/muse125.pdf
                          Message 12 of 25 , Mar 16 2:40 AM
                          • 0 Attachment
                            Yes Mark, we agree that the terminology is not very clear...

                            http://www.dliengineering.com/downloads/crest%20factor.pdf
                            http://www.tinaja.com/glib/muse125.pdf

                            Even "peak power" may be used to talk about different things (the
                            debate is as old as SSB - i.e. PEP). Not to mention again the use
                            (or abuse) of RMS-power... and/or the frequent confusion between
                            average and mean... (in french we use "puissance efficace" for
                            average power, and "puissance moyenne" for mean power... which I
                            think is less confusing)

                            Why I suggested to forget about the crest-factor, and focus on what
                            is really interesting for us:

                            the ratio between
                            -peak average power (the "tune" wattmeter reading)
                            -mean average power (the bolometer reading)

                            What I called "mean to peak" ratio, which is a about 10dB (10%) for
                            MT63, 6dB (or 25%) for uncompressed-SSB, about 4dB (40%) for CW, and
                            close to 0dB (100%) for FSK.
                            The percentage values are often referred as "duty cycle" by equipment
                            manufacturers, but I also find this appellation a bit confusing...

                            Patrick






                            --- In oliviadata@yahoogroups.com, Mark Miller <kramrellim@c...>
                            wrote:
                            > At 03:13 AM 3/16/2005, you wrote:
                            > >Tomi
                            > >I do not agree with your interpretation of the Crest-factor.
                            > >You will find anywhere that the crest-factor of a pure sinewave is
                            > >1.414...
                            > >-this is the peak voltage value, divided by the RMS voltage value
                            >
                            >
                            > Patrick,
                            >
                            > Take a look at this
                            > article. http://rfdesign.com/mag/radio_crest_factor_analysis/
                            >
                            > Toward the bottom of the article you will see two definitions for
                            crest factor.
                            >
                            > Peak crest factor
                            >
                            > Next the peak crest factor for the RF signal, W(t), is addressed:
                            >
                            > This is the ratio of the instantaneous peak voltage squared to the
                            average
                            > output power, but for translated signals (IF or RF) it is customary
                            to
                            > consider the crest factor for the envelope signal. Because most RF
                            devices
                            > are characterized with sinusoidal signals at either specified or
                            measured
                            > power levels, having the crest factor as a function of a sinusoidal
                            > signal's power level is more appropriate.
                            >
                            > Envelope crest factor
                            >
                            > To determine the envelope crest factor, first determine the power
                            of a
                            > sinusoidal signal that has the same peak voltage as the modulated
                            signal's
                            > peak voltage. This value is then divided by the average power of
                            the
                            > modulated signal. The result is defined as the translated signal's
                            envelope
                            > crest factor.
                            >
                            > Also you will find that peak power does not always mean the peak
                            voltage
                            > squared. Sometimes it means RMS voltage squared. Take a look at
                            the
                            > Alliance for Telecommunications Industry Solutions Web page. It is
                            nicely
                            > laid out.
                            >
                            > http://www.atis.org/tg2k/t1g2k.html
                            >
                            > crest factor http://www.atis.org/tg2k/_crest_factor.html
                            > peak to average ratio http://www.atis.org/tg2k/_peak-to-
                            average_ratio.html
                            > peak power http://www.atis.org/tg2k/_peak_power_output.html
                            >
                            > As you can see peak power is defined as the peak average power.
                            >
                            > 73,
                            >
                            > Mark N5RFX
                          • Paul L Schmidt
                            As stated before, I m not really concerned with what the ratio is called -- I m concerned that John Q. Ham knows how to adjust his power in order to (a)
                            Message 13 of 25 , Mar 16 3:09 AM
                            • 0 Attachment
                              As stated before, I'm not really concerned with what the ratio
                              is called -- I'm concerned that John Q. Ham knows how to adjust
                              his power in order to (a) transmit cleanly, and (b) not fry his rig.

                              The term "crest factor" is apparently ambiguous - the definition
                              that is useful in this case (the one that uses the average power
                              at the peak of the envelope) is the one that makes sense for the
                              stated purpose. If everyone understands what is meant by
                              "crest factor" in this context -- fine. If there is too much
                              ambiguity, call it PEP to Average ratio. Or the previously-
                              suggested "SpongeBob's ratio" (anyone else out there have preschool
                              kids?)

                              73,

                              -ps

                              Patrick wrote:
                              >
                              > Sorry Tomi, this was a reply to Paul's post, not to your...
                              > Patrick
                              >
                              > --- In oliviadata@yahoogroups.com, "Patrick" <f6irf@f...> wrote:
                              >
                              >>Tomi
                              >>I do not agree with your interpretation of the Crest-factor.
                              >>You will find anywhere that the crest-factor of a pure sinewave is
                              >>1.414...
                              >>-this is the peak voltage value, divided by the RMS voltage value
                              >>
                              >
                            • Mark Miller
                              ... Patrick, This is a very good article. I have always enjoyed reading Don Lancaster s work. I miss him from the Radio Electronics days. 73, Mark N5RFX
                              Message 14 of 25 , Mar 16 3:34 AM
                              • 0 Attachment
                                At 04:40 AM 3/16/2005, you wrote:
                                >http://www.tinaja.com/glib/muse125.pdf


                                Patrick,

                                This is a very good article. I have always enjoyed reading Don Lancaster's
                                work. I miss him from the Radio Electronics days.

                                73,

                                Mark N5RFX
                              • Mark Miller
                                ... Gotta love it. Pre-school kids! I am a SpongeBob fan too! 73, Mark N5RFX
                                Message 15 of 25 , Mar 16 3:38 AM
                                • 0 Attachment
                                  At 05:09 AM 3/16/2005, you wrote:

                                  >SpongeBob's ratio


                                  Gotta love it. Pre-school kids! I am a SpongeBob fan too!

                                  73,

                                  Mark N5RFX
                                • Tomi Manninen
                                  ... I was thinking Squidwards clarinet-factor... No, no preschool kids, my son is only 1½ years old but that doesn t prevent _me_ from watching SpongeBob. ;)
                                  Message 16 of 25 , Mar 16 3:40 AM
                                  • 0 Attachment
                                    On Wed, 16 Mar 2005, Paul L Schmidt wrote:

                                    > Or the previously-
                                    > suggested "SpongeBob's ratio" (anyone else out there have preschool
                                    > kids?)

                                    I was thinking Squidwards clarinet-factor... No, no preschool kids,
                                    my son is only 1½ years old but that doesn't prevent _me_ from watching
                                    SpongeBob. ;)

                                    --
                                    Tomi Manninen / OH2BNS / KP20JF74
                                  • Patrick
                                    Paul Looking again at your figures below, at the oscillogram, and at the tone shaping, I would say that you are probably right... Can t explain yet why the
                                    Message 17 of 25 , Mar 16 4:37 AM
                                    • 0 Attachment
                                      Paul
                                      Looking again at your figures below, at the oscillogram, and at the
                                      tone shaping, I would say that you are probably right...
                                      Can't explain yet why the audio sample I used is giving me such
                                      figures (see in file section) but something is not correct
                                      somewhere... 4.22dB "mean to peak" looks too much...
                                      Looking at it in the frequency domain, I can see that the frequency
                                      response of my audio-sample between 500 and 1500 is not flat but that
                                      the lower frequencies are attenuated by some 8dB, so it is probably
                                      why I am finding this wrong value...

                                      Patrick

                                      PS: the same happened to my first MT63 sample, so I think I did
                                      something wrong when recording...
                                      By the way I would also prefer the tones to be in the 1000-2000 range
                                      rather than 500-1500...






                                      --- In oliviadata@yahoogroups.com, Paul L Schmidt <k9ps@a...> wrote:
                                      > Tomi Manninen wrote:
                                      > > On Tue, 2005-03-15 at 16:18, Mark Miller wrote:
                                      > >
                                      > >
                                      > >>Question for the group: Why does Olivia have a 7.22 dB crest
                                      factor?
                                      > >
                                      > >
                                      > > The symbols in Olivia are shaped. See
                                      > >
                                      > > http://homepage.sunrise.ch/mysunrise/jalocha/mfsk_spec.html
                                      > >
                                      >
                                      > True, but not *that* much shaping...
                                      >
                                      > Here are my numbers:
                                      >
                                      > mfsk_tx random.dat random.mfsk
                                      >
                                      > sox -t raw -sw -r8000 random.mfsk -t wav /dev/null stat
                                      >
                                      > Maximum amplitude: 0.538757
                                      > Minimum amplitude: -0.538757
                                      > Midline amplitude: 0.000000
                                      > Mean amplitude: 0.000000
                                      > RMS amplitude: 0.329407
                                      >
                                      > .538757 / .329407 = 1.63522
                                      >
                                      > 20 log (1.63522) = 4.27 dB ... which needs
                                      > to be dropped by 3 dB since we're comparing an RMS
                                      > voltage to an instantaneous peak (rather than the RMS
                                      > value of the signal at the peak).
                                      >
                                      > That puts it down to about a 1.63 dB crest factor...
                                      >
                                      > e.g. if you're running 50 watts RMS output, your
                                      > PEP will be 67 watts... which seems reasonable for
                                      > what I've experienced.
                                    • Mark Miller
                                      I wonder how much of an affect the sound card has on the figures. I use a SoundBlaster Live! and can tell you that its output is not linear and its third
                                      Message 18 of 25 , Mar 16 5:47 AM
                                      • 0 Attachment
                                        I wonder how much of an affect the sound card has on the figures. I use a
                                        SoundBlaster Live! and can tell you that its output is not linear and its
                                        third order IMD when using a two tone test signal is in the -30dB range. I
                                        also have a USB SoundBlaster that does a worse job decoding digital modes,
                                        than the internal SoundBlaster Live!.

                                        Does anyone know of a low noise, low IMD, excellent decoding sound card?

                                        73,

                                        Mark N5RFX

                                        At 06:37 AM 3/16/2005, you wrote:


                                        >Paul
                                        >Looking again at your figures below, at the oscillogram, and at the
                                        >tone shaping, I would say that you are probably right...
                                        >Can't explain yet why the audio sample I used is giving me such
                                        >figures (see in file section) but something is not correct
                                        >somewhere... 4.22dB "mean to peak" looks too much...
                                        >Looking at it in the frequency domain, I can see that the frequency
                                        >response of my audio-sample between 500 and 1500 is not flat but that
                                        >the lower frequencies are attenuated by some 8dB, so it is probably
                                        >why I am finding this wrong value...
                                      • Steve Friis
                                        You are saying basically what I said back on the 13th. I STILL say that the best way to ensure a clean signal is to scope the output of the transmitter. Since
                                        Message 19 of 25 , Mar 16 7:01 AM
                                        • 0 Attachment
                                          You are saying basically what I said back on the 13th.

                                          I STILL say that the best way to ensure a clean signal is to scope the
                                          output of the transmitter. Since all transmitters have a limited output,
                                          simply use a scope to observe the peak voltage that the transmitter can
                                          produce with CW. Then when using the digi modes set the output at just
                                          below this for the peak signal. Your transmitter is now set where it
                                          should be clean. Use the audio control, not the RF power output control.
                                          This could be the mic gain, or the output of the soundcard. Make sure
                                          any processing is also off.

                                          Steve/WM5Z

                                          Patrick wrote:

                                          >
                                          > Tomi
                                          > I do not agree with your interpretation of the Crest-factor.
                                          > You will find anywhere that the crest-factor of a pure sinewave is
                                          > 1.414...
                                          > -this is the peak voltage value, divided by the RMS voltage value
                                          >
                                          > If you keep thinking in terms of voltage values converting to dB
                                          > gives 20log1.414 = 3dB
                                          > A pure sinewave has a 1.414 or 3dB crest factor...
                                          >
                                          > In practice any modulation will have a crest-factor equal(pure-FSK)
                                          > or superior to 3dB...
                                          >
                                          > Now forget about the crest factor and just consider what is
                                          > interesting for us: mean to peak power values -both being "average
                                          > power values"- (average power being the square of the RMS voltage
                                          > divided by the load (*))
                                          > - In this case the peak (avg) power is the wattmeter reading when
                                          > you "tune" or when you send a RTTY mark
                                          > - the mean (avg) power would be a thermical-wattmeter (bolometer)
                                          > reading... (a normal wattmeter may indicate anything between mean and
                                          > peak depending among other things on the instrument technology...)
                                          >
                                          > I don't have a bolometer, but my conclusion from waveform analysis
                                          > (and from temperature elevation of my IC706 !)is that the "mean to
                                          > peak" for Olivia is in the 4dB area, which is something similar to
                                          > CW... So about 40W with a 100W rated transceiver.
                                          >
                                          > (*)As mentionned in ARRL handbook RMS-power is a mathematical
                                          > curiosity whcih has no physical significance. In practice you'll see
                                          > everywhere references to RMS-power, but it should be called "average
                                          > power" (the product of RMS-voltage by RMS-current). The english
                                          > terminology is extremely confusing...
                                          >
                                          >
                                          >
                                          >
                                          >
                                          >
                                          >
                                          > --- In oliviadata@yahoogroups.com, Paul L Schmidt <k9ps@a...> wrote:
                                          > > Tomi Manninen wrote:
                                          > > > On Tue, 2005-03-15 at 16:18, Mark Miller wrote:
                                          > > >
                                          > > >
                                          > > >>Question for the group: Why does Olivia have a 7.22 dB crest
                                          > factor?
                                          > > >
                                          > > >
                                          > > > The symbols in Olivia are shaped. See
                                          > > >
                                          > > > http://homepage.sunrise.ch/mysunrise/jalocha/mfsk_spec.html
                                          > > >
                                          > >
                                          > > True, but not *that* much shaping...
                                          > >
                                          > > Here are my numbers:
                                          > >
                                          > > mfsk_tx random.dat random.mfsk
                                          > >
                                          > > sox -t raw -sw -r8000 random.mfsk -t wav /dev/null stat
                                          > >
                                          > > Maximum amplitude: 0.538757
                                          > > Minimum amplitude: -0.538757
                                          > > Midline amplitude: 0.000000
                                          > > Mean amplitude: 0.000000
                                          > > RMS amplitude: 0.329407
                                          > >
                                          > > .538757 / .329407 = 1.63522
                                          > >
                                          > > 20 log (1.63522) = 4.27 dB ... which needs
                                          > > to be dropped by 3 dB since we're comparing an RMS
                                          > > voltage to an instantaneous peak (rather than the RMS
                                          > > value of the signal at the peak).
                                          > >
                                          > > That puts it down to about a 1.63 dB crest factor...
                                          > >
                                          > > e.g. if you're running 50 watts RMS output, your
                                          > > PEP will be 67 watts... which seems reasonable for
                                          > > what I've experienced.
                                          >
                                          >
                                          >
                                          >
                                          >
                                          > The ARRL now lists Olivia as a "valid" digital mode for
                                          > US hams. Go to the ARRL site for more information.
                                          > http://www.arrl.org/FandES/field/regulations/techchar/
                                          >
                                          >
                                          >
                                          >
                                          >
                                          > *Yahoo! Groups Sponsor*
                                          > ADVERTISEMENT
                                          > click here
                                          > <http://us.ard.yahoo.com/SIG=1297543l5/M=298184.6018725.7038619.3001176/D=groups/S=1705063108:HM/EXP=1111050841/A=2593423/R=0/SIG=11el9gslf/*http://www.netflix.com/Default?mqso=60190075>
                                          >
                                          >
                                          >
                                          > ------------------------------------------------------------------------
                                          > *Yahoo! Groups Links*
                                          >
                                          > * To visit your group on the web, go to:
                                          > http://groups.yahoo.com/group/oliviadata/
                                          >
                                          > * To unsubscribe from this group, send an email to:
                                          > oliviadata-unsubscribe@yahoogroups.com
                                          > <mailto:oliviadata-unsubscribe@yahoogroups.com?subject=Unsubscribe>
                                          >
                                          > * Your use of Yahoo! Groups is subject to the Yahoo! Terms of
                                          > Service <http://docs.yahoo.com/info/terms/>.
                                          >
                                          >
                                        • Patrick
                                          Mark You may browse the RTTY archive on contesting.com... I remember that the subject has been discussed several times. With my previous PC I was using a
                                          Message 20 of 25 , Mar 16 8:04 AM
                                          • 0 Attachment
                                            Mark
                                            You may browse the RTTY archive on contesting.com... I remember that
                                            the subject has been discussed several times.
                                            With my previous PC I was using a "guillemot home studio pro 64"
                                            which had "low noise inputs and outputs"... It was excellent, but it
                                            was an ISA card !
                                            Now I just use the default card which is on the motherboard.
                                            To be honnest I never did IMD measurements before, but as it was an
                                            excellent idea, I immediately did it...

                                            - generated a closed-spaced 2 tones file using cool edit DTMF
                                            generator (1100 and 1200 Hz)
                                            - played it with the windows player
                                            - recorded it using pathsim as digital recorder

                                            I have uploaded the result in the file section... no IMD product
                                            higher than -85dB...
                                            NB I did not use the line input, but I looped the signal internaly
                                            using "out mix" as recorder input. It may be a bit worst looping
                                            externaly with an audio cable, but honnestly don't think that 30dB is
                                            something normal... for a transceiver yes, but definitely not for an
                                            audio device...

                                            Patrick








                                            --- In oliviadata@yahoogroups.com, Mark Miller <kramrellim@c...>
                                            wrote:
                                            > I wonder how much of an affect the sound card has on the figures.
                                            I use a
                                            > SoundBlaster Live! and can tell you that its output is not linear
                                            and its
                                            > third order IMD when using a two tone test signal is in the -30dB
                                            range. I
                                            > also have a USB SoundBlaster that does a worse job decoding digital
                                            modes,
                                            > than the internal SoundBlaster Live!.
                                            >
                                            > Does anyone know of a low noise, low IMD, excellent decoding sound
                                            card?
                                            >
                                            > 73,
                                            >
                                            > Mark N5RFX
                                            >
                                            > At 06:37 AM 3/16/2005, you wrote:
                                            >
                                            >
                                            > >Paul
                                            > >Looking again at your figures below, at the oscillogram, and at the
                                            > >tone shaping, I would say that you are probably right...
                                            > >Can't explain yet why the audio sample I used is giving me such
                                            > >figures (see in file section) but something is not correct
                                            > >somewhere... 4.22dB "mean to peak" looks too much...
                                            > >Looking at it in the frequency domain, I can see that the frequency
                                            > >response of my audio-sample between 500 and 1500 is not flat but
                                            that
                                            > >the lower frequencies are attenuated by some 8dB, so it is probably
                                            > >why I am finding this wrong value...
                                          • Steve Friis
                                            Hmmmm. Interesting finding. I am using a Creative Labs CT4750 ($29.00 US) which produces aproximately 1.25 volts peak output. It is designed to drive amplified
                                            Message 21 of 25 , Mar 16 8:10 AM
                                            • 0 Attachment
                                              Hmmmm. Interesting finding. I am using a Creative Labs CT4750 ($29.00
                                              US) which produces aproximately 1.25 volts peak output. It is designed
                                              to drive amplified speakers, no power amp on board. These boards are
                                              equal to the SB-64, but no speaker driver.

                                              The board was set up to sample at 11025. The single tone distortion is
                                              less than .005% per my HP334A distortion analyzer. Using 2 tones, and an
                                              CT Systems audio spectrum analyzer (NOT FFT), set up with 10 Hz
                                              bandwidth and 100 second sweep my only artifacts are anti-aliasing
                                              errors that are just at -54dB. The dynamic range of this spec-an is
                                              limited to 60 dB. Can not see any IMD within the audio passband of 10 to
                                              8000 hz. The ailising errors are all above 18Khz.

                                              BTW, I did these tests to proof these boards for a project for the US
                                              Army. My samples are all date coded 8/27/04 and the quantity tested is
                                              5. I also tested 3 onboard soundcards using the ASUS P4VP-MX
                                              motherboards. These are Realtek ALC655 sound decoders. My findings there
                                              were very comprable, though the output level max was slightly lower (1.1
                                              Vpeak) The single tone distortion is .007% and the IMD just shows up at
                                              -63 dB. There was no visible anti-aliasing errors.

                                              In all tests, the termination of 600 Ohms was in place, and no
                                              transformer is used between the HP334A and the computer. The case of the
                                              PC was connected to my system ground, and the HP334A was floated which
                                              minimized overall noise. The audio spectrum analyzer was connected to
                                              the output of the HP334A, and also the TERM was engaged on the spec-an.

                                              The test audio was derived from a testfile (SNDTSTxx.WAV) that I have
                                              that comes from a NAB broadcast CD aquired a few years ago. It has
                                              several files at various frequencies, 400Hz, 1000Hz, 2000Hz etc.
                                              single tone. It also has a two-tone test file of 1030 Hz and 900. Note
                                              that these files are WAV files and not just playing through the CD
                                              output. I get the same whether I play the sound files from the CD or
                                              from the hard drive.

                                              I did no tests recording the audio, so can not comment on that. If I get
                                              some time I will set up a PC with one of these and do some tests on the
                                              recording input. Would be interesting.

                                              Hope this helps you.

                                              Steve/WM5Z

                                              Mark Miller wrote:

                                              > I wonder how much of an affect the sound card has on the figures. I
                                              > use a
                                              > SoundBlaster Live! and can tell you that its output is not linear and its
                                              > third order IMD when using a two tone test signal is in the -30dB
                                              > range. I
                                              > also have a USB SoundBlaster that does a worse job decoding digital
                                              > modes,
                                              > than the internal SoundBlaster Live!.
                                              >
                                              > Does anyone know of a low noise, low IMD, excellent decoding sound card?
                                              >
                                              > 73,
                                              >
                                              > Mark N5RFX
                                              >
                                              > At 06:37 AM 3/16/2005, you wrote:
                                              >
                                              >
                                              > >Paul
                                              > >Looking again at your figures below, at the oscillogram, and at the
                                              > >tone shaping, I would say that you are probably right...
                                              > >Can't explain yet why the audio sample I used is giving me such
                                              > >figures (see in file section) but something is not correct
                                              > >somewhere... 4.22dB "mean to peak" looks too much...
                                              > >Looking at it in the frequency domain, I can see that the frequency
                                              > >response of my audio-sample between 500 and 1500 is not flat but that
                                              > >the lower frequencies are attenuated by some 8dB, so it is probably
                                              > >why I am finding this wrong value...
                                              >
                                              >
                                              >
                                              >
                                              > The ARRL now lists Olivia as a "valid" digital mode for
                                              > US hams. Go to the ARRL site for more information.
                                              > http://www.arrl.org/FandES/field/regulations/techchar/
                                              >
                                              >
                                              >
                                              >
                                              >
                                              > *Yahoo! Groups Sponsor*
                                              > ADVERTISEMENT
                                              > click here
                                              > <http://us.ard.yahoo.com/SIG=129p99g40/M=298184.6018725.7038619.3001176/D=groups/S=1705063108:HM/EXP=1111067274/A=2593423/R=0/SIG=11el9gslf/*http://www.netflix.com/Default?mqso=60190075>
                                              >
                                              >
                                              >
                                              > ------------------------------------------------------------------------
                                              > *Yahoo! Groups Links*
                                              >
                                              > * To visit your group on the web, go to:
                                              > http://groups.yahoo.com/group/oliviadata/
                                              >
                                              > * To unsubscribe from this group, send an email to:
                                              > oliviadata-unsubscribe@yahoogroups.com
                                              > <mailto:oliviadata-unsubscribe@yahoogroups.com?subject=Unsubscribe>
                                              >
                                              > * Your use of Yahoo! Groups is subject to the Yahoo! Terms of
                                              > Service <http://docs.yahoo.com/info/terms/>.
                                              >
                                              >
                                            Your message has been successfully submitted and would be delivered to recipients shortly.