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

Re: [softrock40] Thoughts on LOs and a (new?) front end?

Expand Messages
  • KD5NWA
    That is still a synthesizer, if instead of having a counter that counts the whole frequency so it locks on one and only one frequency, it compares the raw
    Message 1 of 11 , Dec 4, 2005
    View Source
    • 0 Attachment
      That is still a synthesizer, if instead of having a counter that
      counts the whole frequency so it locks on one and only one frequency,
      it compares the raw frequency against the reference clock. The upshot
      besides being simple to build is that the VCO locks on the nearest
      multiple of the reference within it's range. So if you use a 10 turn
      pot with a stable voltage as you add this voltage to the VCO it tries
      to move but the feedback holds it in place, eventually as enough
      error builds the PLL looses control at frequency X and lock at X+ref
      or X-ref depending on which way you were moving the pot. When you
      design the PLL you need to make sure it looses lock before the next
      step by limiting the frequency of the detectors output to be less
      than the step size.

      If the reference is 10KHz the output of the VCO will be 7.000MHz,
      7.010MHz, 7.020MHz .....etc.

      It was invented as a poor mans version of a divide by N PLL since in
      those days IC's were expensive frequency limited, and bulky.

      Today you can by a IC from National for $2.50 that has the entire PLL
      including the reference dividers and the divide by N counter all you
      need is the VCO Ic and a Op-Amp and you are set to go with the real
      thing that will cover the entire HF spectrum.

      If you look in the archives of this list there was a discussion of
      this subject including this very method about a month ago. You will
      find that a VCO that covers a large frequency range and has low phase
      noise tends to be a very expensive item. The cheapest I have seen in
      a seen that was good enough, was one pointed out by Phil Covington
      and is the Mini-Circuits JTOS-200 which is about $18 in single quantities.

      Of course you can build one yourself from components but that would
      make the clock generator, way more complicated that the whole
      receiver itself if you want it to have a very large frequency range.
      A adjustable current source a couple of bus switches some high speed
      comparator and a good quality capacitor and you are on your way, and
      since many of these components are already in use in the SR-5 you
      have the part numbers already available FST3126 bus switch, LT1719
      comparator would be excellent for creating a VCO .

      In the 70's I've designed and built several of those homemade VCO's
      (there was no choice) and they worked quite well.

      If you find a inexpensive VCO that has low phase noise, do let us
      know, there are several on the list that are interested.

      I have had a copy of the AD8333 in "of interest" folder but have not
      done anything about it, the part looks quite useful. I saw a article
      on the web of someone who used to create a receiver with it and
      reported it worked quite well, he did mention that IP3 figures were
      not as good as claimed but nonetheless were very good. Next step will
      be to ask for a sample from AD they are awfully generous people.

      Sorry for the long rant, my blood sugar must be off.



      At 05:32 AM 12/4/2005, you wrote:
      >While waiting for my SoftRock v5.0 to arrive, I have looked around
      >for various ways to obtain the required 90 degree phaseshift over
      >larger frequency spans, and also an easier way (than a synthesizer)
      >to possibly tune in 50 kHz steps. This latest in order to create a
      >general coverage receiver using programs like Rocky as the last IF.
      >
      >There was an article in No. 145, June 1974 E.B.U. Review, Technical
      >part, by R.C.V Macario, J.F. Craine & D.S. Walters about a receiver
      >design for SSB, ISB and AM medium frequency broadcasts. In describing
      >their Local Oscillator they said: "The local oscillator frequency is
      >controlled by a sampling feedback system. The LO waveform is sampled
      >at the channel spacing frequency and the samples thus obtained are
      >held on C.. , to provide a voltage control signal for the LO. In
      >this way the output frequency may be constrained to lie at a harmonic
      >of the sampling frequency. When the oscillator is manually tuned it
      >will change frequency in a stepwise manner." The Varicap in the LO
      >must be selected to give a somewhat larger capacity shift than
      >absolutely required for the system to be stable (their example was
      >for a MW receiver with 1 kHz steps), but should hopefully be
      >applicable to 50 kHz spacing (when using a wideband phase shifter) or
      >200 kHz spacing when using the divide by four system of phase
      >shifting. If this is indeed feasible it would be an easy way of
      >generating the LO.
      >
      >New Receiver Chip: Has anyone tested the new(?) AD8333 chip. From
      >the Data sheet it seems to be a nearly complete double front end.
      >Requireing a single LO signal at 4xLO, two RF iputs 0-50 MHz, with
      >161 dB dynamic range at the I&Q outputs, IP3 of +30 dBm, it seems
      >like a natural candidate for investigating further. Perfect
      >quadrature demodulation 0.1 degree and 0,005dB. Looks very
      >interesting indeed.
      >
      >73s,
      >Brynjar Berg
      >LA6UY
      >
      >
      >
      >
      >
      >
      >
      >Yahoo! Groups Links
      >
      >
      >
      >


      Cecil Bayona
      KD5NWA
      www.qrpradio.com

      "I fail to see why doing the same thing over and over and getting the
      same results every time is insanity: I've almost proved it isn't;
      only a few more tests now and I'm sure results will differ this time ... "
    • Bill Dumke
      If a DDS board is too noisy or has too many spurs to work with the SoftRock 40 or version 5 baseband receivers, how about this alternative? With a good sound
      Message 2 of 11 , Dec 4, 2005
      View Source
      • 0 Attachment
        If a DDS board is too noisy or has too many spurs to work with the
        SoftRock 40 or version 5 baseband receivers, how about this alternative?

        With a good sound card the SoftRock 40 will give quite good spectrum
        coverage, just as it is. In the case of 96 kHz sampling, it will give a
        frequency range of 96 kHz, single signal. I am able to see this with an
        Audigy 2 sound card, without any spike visible at 7.054 kHz. While it
        is nice to have a panoramic display covering 96 kHz of spectrum, it
        seems kind of a waste to then just use it on one frequency once it is
        driven with a DDS, anyway.

        The real problem is the exciter.

        If Rocky could be modified to do everything in reverse, and be able to
        drive an exciter over the same frequency range, this might solve that
        problem. However, it might not be able to meet the FCC purity
        requirements. The reason I say this, is that with the SoftRock 40, I
        see only about 18 dB in image rejection in receive with the SoftRock 40
        and Rocky v1.3. I first noticed this on some strong signals last night
        on 40 meters, and then checked it with the XG-1 test oscillator. (Note
        that in this case, Rocky has had its IQ calibration only done with
        normal signals on 40 meters, not with a strong signal generator. I
        assume it would be better if done that way instead.) Anybody else see
        this?

        If Rocky in reverse is not feasible, how about just using a DDS-60 or
        something similar for an exciter, at least just on CW, and leave the
        SoftRock, which ever version, as it is?. 3 or 4 crystals could cover
        the entire 40 meter band if needed.

        Or with some more effort in a phone band use a DDS to drive a couple of
        balanced modulators with the appropriate RF and audio phase shift
        networks, to form a SSB phasing rig. The DDS-60 could then be spotted
        to a desired frequency in the SoftRock the old fashioned way. The
        Tayloe quadrature switch method could also be used in place of the
        balanced modulators.

        Of course a conventional VFO could also be used, as Tayloe himself used
        in a recent design.

        Any comments?

        Bill, WB5TCO



        Brynjar Berg wrote:

        > While waiting for my SoftRock v5.0 to arrive, I have looked around
        > for various ways to obtain the required 90 degree phaseshift over
        > larger frequency spans, and also an easier way (than a synthesizer)
        > to possibly tune in 50 kHz steps. This latest in order to create a
        > general coverage receiver using programs like Rocky as the last IF.
        >
        > There was an article in No. 145, June 1974 E.B.U. Review, Technical
        > part, by R.C.V Macario, J.F. Craine & D.S. Walters about a receiver
        > design for SSB, ISB and AM medium frequency broadcasts. In describing
        > their Local Oscillator they said: "The local oscillator frequency is
        > controlled by a sampling feedback system. The LO waveform is sampled
        > at the channel spacing frequency and the samples thus obtained are
        > held on C.. , to provide a voltage control signal for the LO. In
        > this way the output frequency may be constrained to lie at a harmonic
        > of the sampling frequency. When the oscillator is manually tuned it
        > will change frequency in a stepwise manner." The Varicap in the LO
        > must be selected to give a somewhat larger capacity shift than
        > absolutely required for the system to be stable (their example was
        > for a MW receiver with 1 kHz steps), but should hopefully be
        > applicable to 50 kHz spacing (when using a wideband phase shifter) or
        > 200 kHz spacing when using the divide by four system of phase
        > shifting. If this is indeed feasible it would be an easy way of
        > generating the LO.
        >
        > New Receiver Chip: Has anyone tested the new(?) AD8333 chip. From
        > the Data sheet it seems to be a nearly complete double front end.
        > Requireing a single LO signal at 4xLO, two RF iputs 0-50 MHz, with
        > 161 dB dynamic range at the I&Q outputs, IP3 of +30 dBm, it seems
        > like a natural candidate for investigating further. Perfect
        > quadrature demodulation 0.1 degree and 0,005dB. Looks very
        > interesting indeed.
        >
        > 73s,
        > Brynjar Berg
        > LA6UY
        >
        >
        >
        >
        >
        > SPONSORED LINKS
        > Icom ham radio
        > <http://groups.yahoo.com/gads?t=ms&k=Icom+ham+radio&w1=Icom+ham+radio&w2=Shortwave+receivers&w3=Ham+radio&c=3&s=60&.sig=hFU0G_eqoU6djYHaEX8cmQ>
        > Shortwave receivers
        > <http://groups.yahoo.com/gads?t=ms&k=Shortwave+receivers&w1=Icom+ham+radio&w2=Shortwave+receivers&w3=Ham+radio&c=3&s=60&.sig=SnHD-7dqpZF6oUGxGq4Olw>
        > Ham radio
        > <http://groups.yahoo.com/gads?t=ms&k=Ham+radio&w1=Icom+ham+radio&w2=Shortwave+receivers&w3=Ham+radio&c=3&s=60&.sig=39SoBQoTdzg3LjLq8wz-Iw>
        >
        >
        >
        > ------------------------------------------------------------------------
        > YAHOO! GROUPS LINKS
        >
        > * Visit your group "softrock40
        > <http://groups.yahoo.com/group/softrock40>" on the web.
        >
        > * To unsubscribe from this group, send an email to:
        > softrock40-unsubscribe@yahoogroups.com
        > <mailto:softrock40-unsubscribe@yahoogroups.com?subject=Unsubscribe>
        >
        > * Your use of Yahoo! Groups is subject to the Yahoo! Terms of
        > Service <http://docs.yahoo.com/info/terms/>.
        >
        >
        > ------------------------------------------------------------------------
        >
      • KD5NWA
        18 dB rejection is very poor, in two units that I have the rejection is about 70 dB possibly higher, the signal source that I have is -73 dB and the image is
        Message 3 of 11 , Dec 4, 2005
        View Source
        • 0 Attachment
          18 dB rejection is very poor, in two units that I have the rejection
          is about 70 dB possibly higher, the signal source that I have is -73
          dB and the image is not to be seen at all.

          At 06:19 PM 12/4/2005, you wrote:
          >If a DDS board is too noisy or has too many spurs to work with the
          >SoftRock 40 or version 5 baseband receivers, how about this alternative?
          >
          >With a good sound card the SoftRock 40 will give quite good spectrum
          >coverage, just as it is. In the case of 96 kHz sampling, it will give a
          >frequency range of 96 kHz, single signal. I am able to see this with an
          >Audigy 2 sound card, without any spike visible at 7.054 kHz. While it
          >is nice to have a panoramic display covering 96 kHz of spectrum, it
          >seems kind of a waste to then just use it on one frequency once it is
          >driven with a DDS, anyway.
          >
          >The real problem is the exciter.
          >
          >If Rocky could be modified to do everything in reverse, and be able to
          >drive an exciter over the same frequency range, this might solve that
          >problem. However, it might not be able to meet the FCC purity
          >requirements. The reason I say this, is that with the SoftRock 40, I
          >see only about 18 dB in image rejection in receive with the SoftRock 40
          >and Rocky v1.3. I first noticed this on some strong signals last night
          >on 40 meters, and then checked it with the XG-1 test oscillator. (Note
          >that in this case, Rocky has had its IQ calibration only done with
          >normal signals on 40 meters, not with a strong signal generator. I
          >assume it would be better if done that way instead.) Anybody else see
          >this?
          >
          >If Rocky in reverse is not feasible, how about just using a DDS-60 or
          >something similar for an exciter, at least just on CW, and leave the
          >SoftRock, which ever version, as it is?. 3 or 4 crystals could cover
          >the entire 40 meter band if needed.
          >
          >Or with some more effort in a phone band use a DDS to drive a couple of
          >balanced modulators with the appropriate RF and audio phase shift
          >networks, to form a SSB phasing rig. The DDS-60 could then be spotted
          >to a desired frequency in the SoftRock the old fashioned way. The
          >Tayloe quadrature switch method could also be used in place of the
          >balanced modulators.
          >
          >Of course a conventional VFO could also be used, as Tayloe himself used
          >in a recent design.
          >
          >Any comments?
          >
          >Bill, WB5TCO
          >
          >
          >
          >Brynjar Berg wrote:
          >
          > > While waiting for my SoftRock v5.0 to arrive, I have looked around
          > > for various ways to obtain the required 90 degree phaseshift over
          > > larger frequency spans, and also an easier way (than a synthesizer)
          > > to possibly tune in 50 kHz steps. This latest in order to create a
          > > general coverage receiver using programs like Rocky as the last IF.
          > >
          > >SNIP

          > >
          > > 73s,
          > > Brynjar Berg
          > > LA6UY
          > >
          > >

          SNIP



          Cecil Bayona
          KD5NWA
          www.qrpradio.com

          "I fail to see why doing the same thing over and over and getting the
          same results every time is insanity: I've almost proved it isn't;
          only a few more tests now and I'm sure results will differ this time ... "
        • Alex, VE3NEA
          Hi Bill, ... In fact, with a Delta-44 sound card running at 96 kHz the useful bandwidth of the receiver is only 86 KHz, according to my measurements. The last
          Message 4 of 11 , Dec 4, 2005
          View Source
          • 0 Attachment
            Hi Bill,

            > With a good sound card the SoftRock 40 will give quite good spectrum
            > coverage, just as it is. In the case of 96 kHz sampling, it will give a
            > frequency range of 96 kHz, single signal.

            In fact, with a Delta-44 sound card running at 96 kHz the useful bandwidth
            of the receiver is only 86 KHz, according to my measurements. The last 5 kHz
            at each end contain aliases of out-of-band signals: you still can receive
            stations in those segments, but you cannot tell if the signal is a true one
            or an alias.



            > If Rocky could be modified to do everything in reverse, and be able to
            > drive an exciter over the same frequency range, this might solve that
            > problem.

            The iambic keyer in Rocky was tested only by Phil and by me, but it worked
            perfectly in both cases. Once we have test results for different computers,
            we will know if TX I/Q generation in the software is possible. The code that
            generates the sidetone signal can also generate I/Q in the same loop, the
            difference in the CPU load is negligible.

            SSB I/Q generation is a lot easier than CW.




            > However, it might not be able to meet the FCC purity
            > requirements. The reason I say this, is that with the SoftRock 40, I
            > see only about 18 dB in image rejection in receive with the SoftRock 40
            > and Rocky v1.3. I first noticed this on some strong signals last night
            > on 40 meters, and then checked it with the XG-1 test oscillator. (Note
            > that in this case, Rocky has had its IQ calibration only done with
            > normal signals on 40 meters, not with a strong signal generator. I
            > assume it would be better if done that way instead.) Anybody else see
            > this?

            Remember that the I/Q balance error in SoftRock is very sensitive to the
            impedance of the signal source. 18-dB image rejection is what you might
            expect if you calibrated the balancer with the antenna connected and then
            replaced the antenna with a signal generator. If Rocky is calibrated and
            tested with the same signal source, rejection should be a lot better.


            73 Alex VE3NEA
          • n7ve
            As far as an exciter goes, the phase and amplitude setting for a SDR transmitter will not be the same as the phase and amplitude settings used for optimizing
            Message 5 of 11 , Dec 4, 2005
            View Source
            • 0 Attachment
              As far as an exciter goes, the phase and amplitude setting for a SDR
              transmitter will not be the same as the phase and amplitude settings
              used for optimizing the image rejection for the receiver. This is
              because the phase and amplitude adjustments needed to get good image
              rejection on the receive side is a strong function of the interaction
              of the detector with the tuned circuit of the receiver front end.

              For example, the received signal on the low side
              of the LO is usually larger and falls off more slowly with frequency
              than the signal on the high side of the detector. This is not a
              characteristic of the dector, but rather how the changes in the input
              impedence of the detector interact with the input tuned circuit. This
              affect can be flattened using resistive terminations, but that reduces
              the sensitivity of the receiver, something I would rather avoid.

              The transmitter on the other hand seems very frequency flat. The one
              that I have been playing with seems content with a single phase and
              amplitude setting all but the end of the range, unlike the receive
              side. From what I see, third order distortion seems to be more of a
              concern than anything else. I see the third order audio harmonic down
              only 26 db or so which does not seem nearly good enough.

              Has anyone measured the receiver performance of an SR4 or SR5? The
              blocking (113 db BDR) and IP3 performance (92 db IP3DR) of the
              receiver I have been playing with seems just ok. Of course, a very
              wide receiver is much more difficult to get good high level signal
              performance out of than a narrow receiver such as the NC2030 (145 db
              BDR > 20 KHz away, 107-109 db IP3DR >10 KHz away).

              I am using a firebox for audio conversion, which seems to help very
              good high level signal performance. The firebox seems to work well
              with PowerSDR software, but unfortunately does not seem to do anything
              work with Rocky.

              - Dan, N7VE

              --- In softrock40@yahoogroups.com, KD5NWA <KD5NWA@c...> wrote:
              >
              > 18 dB rejection is very poor, in two units that I have the rejection
              > is about 70 dB possibly higher, the signal source that I have is -73
              > dB and the image is not to be seen at all.
              >
              > At 06:19 PM 12/4/2005, you wrote:
              > >If a DDS board is too noisy or has too many spurs to work with the
              > >SoftRock 40 or version 5 baseband receivers, how about this
              alternative?
              > >
              > >With a good sound card the SoftRock 40 will give quite good spectrum
              > >coverage, just as it is. In the case of 96 kHz sampling, it will
              give a
              > >frequency range of 96 kHz, single signal. I am able to see this
              with an
              > >Audigy 2 sound card, without any spike visible at 7.054 kHz. While it
              > >is nice to have a panoramic display covering 96 kHz of spectrum, it
              > >seems kind of a waste to then just use it on one frequency once it is
              > >driven with a DDS, anyway.
              > >
              > >The real problem is the exciter.
              > >
              > >If Rocky could be modified to do everything in reverse, and be able to
              > >drive an exciter over the same frequency range, this might solve that
              > >problem. However, it might not be able to meet the FCC purity
              > >requirements. The reason I say this, is that with the SoftRock 40, I
              > >see only about 18 dB in image rejection in receive with the SoftRock 40
              > >and Rocky v1.3. I first noticed this on some strong signals last night
              > >on 40 meters, and then checked it with the XG-1 test oscillator. (Note
              > >that in this case, Rocky has had its IQ calibration only done with
              > >normal signals on 40 meters, not with a strong signal generator. I
              > >assume it would be better if done that way instead.) Anybody else see
              > >this?
              > >
              > >If Rocky in reverse is not feasible, how about just using a DDS-60 or
              > >something similar for an exciter, at least just on CW, and leave the
              > >SoftRock, which ever version, as it is?. 3 or 4 crystals could cover
              > >the entire 40 meter band if needed.
              > >
              > >Or with some more effort in a phone band use a DDS to drive a couple of
              > >balanced modulators with the appropriate RF and audio phase shift
              > >networks, to form a SSB phasing rig. The DDS-60 could then be spotted
              > >to a desired frequency in the SoftRock the old fashioned way. The
              > >Tayloe quadrature switch method could also be used in place of the
              > >balanced modulators.
              > >
              > >Of course a conventional VFO could also be used, as Tayloe himself used
              > >in a recent design.
              > >
              > >Any comments?
              > >
              > >Bill, WB5TCO
              > >
              > >
              > >
              > >Brynjar Berg wrote:
              > >
              > > > While waiting for my SoftRock v5.0 to arrive, I have looked around
              > > > for various ways to obtain the required 90 degree phaseshift over
              > > > larger frequency spans, and also an easier way (than a synthesizer)
              > > > to possibly tune in 50 kHz steps. This latest in order to create a
              > > > general coverage receiver using programs like Rocky as the last IF.
              > > >
              > > >SNIP
              >
              > > >
              > > > 73s,
              > > > Brynjar Berg
              > > > LA6UY
              > > >
              > > >
              >
              > SNIP
              >
              >
              >
              > Cecil Bayona
              > KD5NWA
              > www.qrpradio.com
              >
              > "I fail to see why doing the same thing over and over and getting the
              > same results every time is insanity: I've almost proved it isn't;
              > only a few more tests now and I'm sure results will differ this time
              ... "
              >
            • Bill Dumke
              Alex, First noticed the problem with strong signals on 40 meters. It was hard to measure the ratio with the signals bouncing up and down with the CW or
              Message 6 of 11 , Dec 5, 2005
              View Source
              • 0 Attachment
                Alex,

                First noticed the problem with strong signals on 40 meters. It was hard
                to measure the ratio with the signals bouncing up and down with the CW
                or speech, in the case of SSB, so that was why I decided to check it
                with the XG-1.

                It is possible there is something wrong with my SoftRock 40, however.
                Ordered another one assembled from Tony, and when I get it I will check
                it, again.

                Several things that are suspicious with my unit.

                1. I see sidebands on a CW carrier from the XG-1, even though I can't
                hear them on my Yaesu FT-757 GX. They look a lot like what I see on
                even slow CW signals.

                2. The SoftRock 40 works great with the antenna connected, but when I
                replace the antenna with a 50 Ohm termination, I see junk all over the
                place. Up to a level of about 27 dB on the left hand scale. And they
                are not legitimate signals. A lot of it appears to be 60 Hz AC hum
                modulating something or other. Also the center 7.054 MHz peak is
                back. I don't see it at all with the antenna connected. My SoftRock
                has the RF input ground isolated from the rest of the ground. And t
                is not in a metal box. Kees tried to help me diagnose the problem, but
                everything checked out. I thought it might be due to some kind of ALC
                in Rocky, but that was just a guess.

                3. The difference between the gains of the I and Q channels is about 1.5
                which seems high to me.

                4. When I try to run two Audigy 2 soundcards to get Mixw to process
                Rocky's output, the CW seems smeared in the waterfall display of Mixw.
                And it has trouble decoding anything. Also the audio from the second
                soundcard seems muffled compared to the first soundcard that gets it's
                signal from the SoftRock. And the noise floor form the second
                soundcard does seem to bounce around a lot. The levels are all set at
                midrange.

                Any ideas?

                BTW, I am glad to hear the good work you are doing with your exciter
                software. Rcoky is a great program.

                Bill, WB5TCO

                Alex, VE3NEA wrote:

                >Hi Bill,
                >
                >
                >
                >>With a good sound card the SoftRock 40 will give quite good spectrum
                >>coverage, just as it is. In the case of 96 kHz sampling, it will give a
                >>frequency range of 96 kHz, single signal.
                >>
                >>
                >
                >In fact, with a Delta-44 sound card running at 96 kHz the useful bandwidth
                >of the receiver is only 86 KHz, according to my measurements. The last 5 kHz
                >at each end contain aliases of out-of-band signals: you still can receive
                >stations in those segments, but you cannot tell if the signal is a true one
                >or an alias.
                >
                >
                >
                >
                >
                >>If Rocky could be modified to do everything in reverse, and be able to
                >>drive an exciter over the same frequency range, this might solve that
                >>problem.
                >>
                >>
                >
                >The iambic keyer in Rocky was tested only by Phil and by me, but it worked
                >perfectly in both cases. Once we have test results for different computers,
                >we will know if TX I/Q generation in the software is possible. The code that
                >generates the sidetone signal can also generate I/Q in the same loop, the
                >difference in the CPU load is negligible.
                >
                >SSB I/Q generation is a lot easier than CW.
                >
                >
                >
                >
                >
                >
                >>However, it might not be able to meet the FCC purity
                >>requirements. The reason I say this, is that with the SoftRock 40, I
                >>see only about 18 dB in image rejection in receive with the SoftRock 40
                >>and Rocky v1.3. I first noticed this on some strong signals last night
                >>on 40 meters, and then checked it with the XG-1 test oscillator. (Note
                >>that in this case, Rocky has had its IQ calibration only done with
                >>normal signals on 40 meters, not with a strong signal generator. I
                >>assume it would be better if done that way instead.) Anybody else see
                >>this?
                >>
                >>
                >
                >Remember that the I/Q balance error in SoftRock is very sensitive to the
                >impedance of the signal source. 18-dB image rejection is what you might
                >expect if you calibrated the balancer with the antenna connected and then
                >replaced the antenna with a signal generator. If Rocky is calibrated and
                >tested with the same signal source, rejection should be a lot better.
                >
                >
                >73 Alex VE3NEA
                >
                >
                >
                >
                >
                >Yahoo! Groups Links
                >
                >
                >
                >
                >
                >
                >
                >
                >
                >
                >
              • KD5NWA
                1. We all see those, although they should be minor, send a picture to Yahoo so people can see how much. The Polyphase FFT tends to exaggerate those a little
                Message 7 of 11 , Dec 5, 2005
                View Source
                • 0 Attachment
                  1. We all see those, although they should be minor, send a picture to
                  Yahoo so people can see how much. The Polyphase FFT tends to
                  exaggerate those a little more when compared to a regular FFT, it's
                  all a matter of degree. The stronger the signal the more crud around it.

                  2. Sounds normal, again send picture

                  3. Seems high, what version of Rocky? Maybe I'm cracking up but older
                  ones read it in terms of percentage off, in which case 1.5% is fine.
                  The latest Rocky is V1.3

                  4. bouncing signal level is a sign of improper signal buffer size of
                  too much CPU load, how high is your CPU load?

                  Have you tried the Flex software? Beta 7 is the latest that supports the SR-40.

                  Have you calibrated the software?, fire a rig into a dummy load with
                  a couple of watts, and make sure you pick it up at least 40dB above
                  the noise floor, starting at the bottom of the band move slowly and
                  stop every KHz until you reach the top of the band.

                  A separate and isolated power supply will help the SR from getting so
                  much crud in the middle, there is a mod to replace the transistor
                  with a 5 volt regulator so you can feed it with a wall wart or a battery.

                  The Audigy although it's a expensive card is one that does the
                  poorest of the batch of sound cards I own in terms of noise floor.

                  It's late I need to back to bed....sorry.

                  At 02:59 AM 12/5/2005, you wrote:
                  >Alex,
                  >
                  >First noticed the problem with strong signals on 40 meters. It was hard
                  >to measure the ratio with the signals bouncing up and down with the CW
                  >or speech, in the case of SSB, so that was why I decided to check it
                  >with the XG-1.
                  >
                  >It is possible there is something wrong with my SoftRock 40, however.
                  >Ordered another one assembled from Tony, and when I get it I will check
                  >it, again.
                  >
                  >Several things that are suspicious with my unit.
                  >
                  >1. I see sidebands on a CW carrier from the XG-1, even though I can't
                  >hear them on my Yaesu FT-757 GX. They look a lot like what I see on
                  >even slow CW signals.
                  >
                  >2. The SoftRock 40 works great with the antenna connected, but when I
                  >replace the antenna with a 50 Ohm termination, I see junk all over the
                  >place. Up to a level of about 27 dB on the left hand scale. And they
                  >are not legitimate signals. A lot of it appears to be 60 Hz AC hum
                  >modulating something or other. Also the center 7.054 MHz peak is
                  >back. I don't see it at all with the antenna connected. My SoftRock
                  >has the RF input ground isolated from the rest of the ground. And t
                  >is not in a metal box. Kees tried to help me diagnose the problem, but
                  >everything checked out. I thought it might be due to some kind of ALC
                  >in Rocky, but that was just a guess.
                  >
                  >3. The difference between the gains of the I and Q channels is about 1.5
                  >which seems high to me.
                  >
                  >4. When I try to run two Audigy 2 soundcards to get Mixw to process
                  >Rocky's output, the CW seems smeared in the waterfall display of Mixw.
                  >And it has trouble decoding anything. Also the audio from the second
                  >soundcard seems muffled compared to the first soundcard that gets it's
                  >signal from the SoftRock. And the noise floor form the second
                  >soundcard does seem to bounce around a lot. The levels are all set at
                  >midrange.
                  >
                  >Any ideas?
                  >
                  >BTW, I am glad to hear the good work you are doing with your exciter
                  >software. Rcoky is a great program.
                  >
                  >Bill, WB5TCO
                  >
                  >Alex, VE3NEA wrote:
                  >
                  > >Hi Bill,
                  > >
                  > >
                  > >
                  > >>With a good sound card the SoftRock 40 will give quite good spectrum
                  > >>coverage, just as it is. In the case of 96 kHz sampling, it will give a
                  > >>frequency range of 96 kHz, single signal.
                  > >>
                  > >>
                  > >
                  > >In fact, with a Delta-44 sound card running at 96 kHz the useful bandwidth
                  > >of the receiver is only 86 KHz, according to my measurements. The
                  > last 5 kHz
                  > >at each end contain aliases of out-of-band signals: you still can receive
                  > >stations in those segments, but you cannot tell if the signal is a true one
                  > >or an alias.
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >>If Rocky could be modified to do everything in reverse, and be able to
                  > >>drive an exciter over the same frequency range, this might solve that
                  > >>problem.
                  > >>
                  > >>
                  > >
                  > >The iambic keyer in Rocky was tested only by Phil and by me, but it worked
                  > >perfectly in both cases. Once we have test results for different computers,
                  > >we will know if TX I/Q generation in the software is possible. The
                  > code that
                  > >generates the sidetone signal can also generate I/Q in the same loop, the
                  > >difference in the CPU load is negligible.
                  > >
                  > >SSB I/Q generation is a lot easier than CW.
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >>However, it might not be able to meet the FCC purity
                  > >>requirements. The reason I say this, is that with the SoftRock 40, I
                  > >>see only about 18 dB in image rejection in receive with the SoftRock 40
                  > >>and Rocky v1.3. I first noticed this on some strong signals last night
                  > >>on 40 meters, and then checked it with the XG-1 test oscillator. (Note
                  > >>that in this case, Rocky has had its IQ calibration only done with
                  > >>normal signals on 40 meters, not with a strong signal generator. I
                  > >>assume it would be better if done that way instead.) Anybody else see
                  > >>this?
                  > >>
                  > >>
                  > >
                  > >Remember that the I/Q balance error in SoftRock is very sensitive to the
                  > >impedance of the signal source. 18-dB image rejection is what you might
                  > >expect if you calibrated the balancer with the antenna connected and then
                  > >replaced the antenna with a signal generator. If Rocky is calibrated and
                  > >tested with the same signal source, rejection should be a lot better.
                  > >
                  > >
                  > >73 Alex VE3NEA
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >Yahoo! Groups Links
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >
                  > >
                  >
                  >
                  >
                  >
                  >Yahoo! Groups Links
                  >
                  >
                  >
                  >


                  Cecil Bayona
                  KD5NWA
                  www.qrpradio.com

                  "I fail to see why doing the same thing over and over and getting the
                  same results every time is insanity: I've almost proved it isn't;
                  only a few more tests now and I'm sure results will differ this time ... "
                • Bill Dumke
                  Cecil, What sound card do you recommend? I don t have time to work on it tonight, but I will try to post the pictures and try your other suggestions tomorrow
                  Message 8 of 11 , Dec 5, 2005
                  View Source
                  • 0 Attachment
                    Cecil,

                    What sound card do you recommend?

                    I don't have time to work on it tonight, but I will try to post the
                    pictures and try your other suggestions tomorrow night. Thanks for the
                    help.

                    Bill WB5TCO

                    KD5NWA wrote:

                    > 1. We all see those, although they should be minor, send a picture to
                    > Yahoo so people can see how much. The Polyphase FFT tends to
                    > exaggerate those a little more when compared to a regular FFT, it's
                    > all a matter of degree. The stronger the signal the more crud around it.
                    >
                    > 2. Sounds normal, again send picture
                    >
                    > 3. Seems high, what version of Rocky? Maybe I'm cracking up but older
                    > ones read it in terms of percentage off, in which case 1.5% is fine.
                    > The latest Rocky is V1.3
                    >
                    > 4. bouncing signal level is a sign of improper signal buffer size of
                    > too much CPU load, how high is your CPU load?
                    >
                    > Have you tried the Flex software? Beta 7 is the latest that supports
                    > the SR-40.
                    >
                    > Have you calibrated the software?, fire a rig into a dummy load with
                    > a couple of watts, and make sure you pick it up at least 40dB above
                    > the noise floor, starting at the bottom of the band move slowly and
                    > stop every KHz until you reach the top of the band.
                    >
                    > A separate and isolated power supply will help the SR from getting so
                    > much crud in the middle, there is a mod to replace the transistor
                    > with a 5 volt regulator so you can feed it with a wall wart or a battery.
                    >
                    > The Audigy although it's a expensive card is one that does the
                    > poorest of the batch of sound cards I own in terms of noise floor.
                    >
                    > It's late I need to back to bed....sorry.
                    >
                    > At 02:59 AM 12/5/2005, you wrote:
                    > >Alex,
                    > >
                    > >First noticed the problem with strong signals on 40 meters. It was hard
                    > >to measure the ratio with the signals bouncing up and down with the CW
                    > >or speech, in the case of SSB, so that was why I decided to check it
                    > >with the XG-1.
                    > >
                    > >It is possible there is something wrong with my SoftRock 40, however.
                    > >Ordered another one assembled from Tony, and when I get it I will check
                    > >it, again.
                    > >
                    > >Several things that are suspicious with my unit.
                    > >
                    > >1. I see sidebands on a CW carrier from the XG-1, even though I can't
                    > >hear them on my Yaesu FT-757 GX. They look a lot like what I see on
                    > >even slow CW signals.
                    > >
                    > >2. The SoftRock 40 works great with the antenna connected, but when I
                    > >replace the antenna with a 50 Ohm termination, I see junk all over the
                    > >place. Up to a level of about 27 dB on the left hand scale. And they
                    > >are not legitimate signals. A lot of it appears to be 60 Hz AC hum
                    > >modulating something or other. Also the center 7.054 MHz peak is
                    > >back. I don't see it at all with the antenna connected. My SoftRock
                    > >has the RF input ground isolated from the rest of the ground. And t
                    > >is not in a metal box. Kees tried to help me diagnose the problem, but
                    > >everything checked out. I thought it might be due to some kind of ALC
                    > >in Rocky, but that was just a guess.
                    > >
                    > >3. The difference between the gains of the I and Q channels is about 1.5
                    > >which seems high to me.
                    > >
                    > >4. When I try to run two Audigy 2 soundcards to get Mixw to process
                    > >Rocky's output, the CW seems smeared in the waterfall display of Mixw.
                    > >And it has trouble decoding anything. Also the audio from the second
                    > >soundcard seems muffled compared to the first soundcard that gets it's
                    > >signal from the SoftRock. And the noise floor form the second
                    > >soundcard does seem to bounce around a lot. The levels are all set at
                    > >midrange.
                    > >
                    > >Any ideas?
                    > >
                    > >BTW, I am glad to hear the good work you are doing with your exciter
                    > >software. Rcoky is a great program.
                    > >
                    > >Bill, WB5TCO
                    > >
                    > >Alex, VE3NEA wrote:
                    > >
                    > > >Hi Bill,
                    > > >
                    > > >
                    > > >
                    > > >>With a good sound card the SoftRock 40 will give quite good spectrum
                    > > >>coverage, just as it is. In the case of 96 kHz sampling, it will
                    > give a
                    > > >>frequency range of 96 kHz, single signal.
                    > > >>
                    > > >>
                    > > >
                    > > >In fact, with a Delta-44 sound card running at 96 kHz the useful
                    > bandwidth
                    > > >of the receiver is only 86 KHz, according to my measurements. The
                    > > last 5 kHz
                    > > >at each end contain aliases of out-of-band signals: you still can
                    > receive
                    > > >stations in those segments, but you cannot tell if the signal is a
                    > true one
                    > > >or an alias.
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >>If Rocky could be modified to do everything in reverse, and be able to
                    > > >>drive an exciter over the same frequency range, this might solve that
                    > > >>problem.
                    > > >>
                    > > >>
                    > > >
                    > > >The iambic keyer in Rocky was tested only by Phil and by me, but it
                    > worked
                    > > >perfectly in both cases. Once we have test results for different
                    > computers,
                    > > >we will know if TX I/Q generation in the software is possible. The
                    > > code that
                    > > >generates the sidetone signal can also generate I/Q in the same
                    > loop, the
                    > > >difference in the CPU load is negligible.
                    > > >
                    > > >SSB I/Q generation is a lot easier than CW.
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >>However, it might not be able to meet the FCC purity
                    > > >>requirements. The reason I say this, is that with the SoftRock 40, I
                    > > >>see only about 18 dB in image rejection in receive with the
                    > SoftRock 40
                    > > >>and Rocky v1.3. I first noticed this on some strong signals last
                    > night
                    > > >>on 40 meters, and then checked it with the XG-1 test oscillator. (Note
                    > > >>that in this case, Rocky has had its IQ calibration only done with
                    > > >>normal signals on 40 meters, not with a strong signal generator. I
                    > > >>assume it would be better if done that way instead.) Anybody else see
                    > > >>this?
                    > > >>
                    > > >>
                    > > >
                    > > >Remember that the I/Q balance error in SoftRock is very sensitive
                    > to the
                    > > >impedance of the signal source. 18-dB image rejection is what you might
                    > > >expect if you calibrated the balancer with the antenna connected
                    > and then
                    > > >replaced the antenna with a signal generator. If Rocky is
                    > calibrated and
                    > > >tested with the same signal source, rejection should be a lot better.
                    > > >
                    > > >
                    > > >73 Alex VE3NEA
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >Yahoo! Groups Links
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > > >
                    > >
                    > >
                    > >
                    > >
                    > >Yahoo! Groups Links
                    > >
                    > >
                    > >
                    > >
                    >
                    >
                    > Cecil Bayona
                    > KD5NWA
                    > www.qrpradio.com
                    >
                    > "I fail to see why doing the same thing over and over and getting the
                    > same results every time is insanity: I've almost proved it isn't;
                    > only a few more tests now and I'm sure results will differ this time
                    > ... "
                    >
                    >
                    > ------------------------------------------------------------------------
                    > YAHOO! GROUPS LINKS
                    >
                    > * Visit your group "softrock40
                    > <http://groups.yahoo.com/group/softrock40>" on the web.
                    >
                    > * To unsubscribe from this group, send an email to:
                    > softrock40-unsubscribe@yahoogroups.com
                    > <mailto:softrock40-unsubscribe@yahoogroups.com?subject=Unsubscribe>
                    >
                    > * Your use of Yahoo! Groups is subject to the Yahoo! Terms of
                    > Service <http://docs.yahoo.com/info/terms/>.
                    >
                    >
                    > ------------------------------------------------------------------------
                    >
                  • KD5NWA
                    Boy that is a tough question, the best card I have is a Delta-44 but it s expensive. The funny part is that some of the built in cards (AC97) don t do too bad
                    Message 9 of 11 , Dec 5, 2005
                    View Source
                    • 0 Attachment
                      Boy that is a tough question, the best card I have is a Delta-44 but
                      it's expensive.

                      The funny part is that some of the built in cards (AC97) don't do too
                      bad in terms of noise and spurs. But I will tell you that I'm no
                      expert on that area, and hopefully some of the long time SDR1000
                      owners or members of the team will chime in. One card that is
                      mentioned as performing well is the Santa Cruz, it's no longer made
                      but you can pick it up on eBay for $10, but it's a 16 bit card. I
                      don't recommend the USB sound cards they eat up a lot of CPU time.

                      In a couple of days I'm going to be testing a new sound card I picked
                      up, it's cheap, and has DMA capability for low overhead and latency.
                      I'll post the results.


                      At 08:30 PM 12/5/2005, you wrote:
                      >Cecil,
                      >
                      >What sound card do you recommend?
                      >
                      >I don't have time to work on it tonight, but I will try to post the
                      >pictures and try your other suggestions tomorrow night. Thanks for the
                      >help.
                      >
                      >Bill WB5TCO
                      >
                      >KD5NWA wrote:
                      >
                      > > 1. We all see those, although they should be minor, send a picture to
                      > > Yahoo so people can see how much. The Polyphase FFT tends to
                      > > exaggerate those a little more when compared to a regular FFT, it's
                      > > all a matter of degree. The stronger the signal the more crud around it.
                      > >
                      > > 2. Sounds normal, again send picture
                      > >
                      > > 3. Seems high, what version of Rocky? Maybe I'm cracking up but older
                      > > ones read it in terms of percentage off, in which case 1.5% is fine.
                      > > The latest Rocky is V1.3
                      > >
                      > > 4. bouncing signal level is a sign of improper signal buffer size of
                      > > too much CPU load, how high is your CPU load?
                      > >
                      > > Have you tried the Flex software? Beta 7 is the latest that supports
                      > > the SR-40.
                      > >
                      > > Have you calibrated the software?, fire a rig into a dummy load with
                      > > a couple of watts, and make sure you pick it up at least 40dB above
                      > > the noise floor, starting at the bottom of the band move slowly and
                      > > stop every KHz until you reach the top of the band.
                      > >
                      > > A separate and isolated power supply will help the SR from getting so
                      > > much crud in the middle, there is a mod to replace the transistor
                      > > with a 5 volt regulator so you can feed it with a wall wart or a battery.
                      > >
                      > > The Audigy although it's a expensive card is one that does the
                      > > poorest of the batch of sound cards I own in terms of noise floor.
                      > >
                      > > It's late I need to back to bed....sorry.
                      > >
                      >SNIP


                      Cecil Bayona
                      KD5NWA
                      www.qrpradio.com

                      "Windows the worlds most successful software virus"
                    • Robert McGwier
                      ... I can confirm this phenomena done by using my 8640B as the generator. It looked like little spurs spread up and down the band. There is nonlinear
                      Message 10 of 11 , Dec 11, 2005
                      View Source
                      • 0 Attachment
                        Bill Dumke wrote:

                        >Alex,
                        >
                        >First noticed the problem with strong signals on 40 meters. It was hard
                        >to measure the ratio with the signals bouncing up and down with the CW
                        >or speech, in the case of SSB, so that was why I decided to check it
                        >with the XG-1.
                        >
                        >It is possible there is something wrong with my SoftRock 40, however.
                        >Ordered another one assembled from Tony, and when I get it I will check
                        >it, again.
                        >
                        >Several things that are suspicious with my unit.
                        >
                        >1. I see sidebands on a CW carrier from the XG-1, even though I can't
                        >hear them on my Yaesu FT-757 GX. They look a lot like what I see on
                        >even slow CW signals.
                        >
                        >
                        >


                        I can confirm this phenomena done by using my 8640B as the generator.
                        It looked like little spurs spread up and down the band. There is
                        nonlinear processing of some type in the power spectra but is was my
                        understanding from conversations with Alex that this was post FFT and
                        done on the magnitude spectra but I could be mistaken. I was using the
                        generator to compare my new polyphase FFT inspired by Alex's for the
                        PowerSDR software and I was comparing them, I noticed the spur
                        generation. I left the country and (writing this from Germany) forgot
                        to write Alex about it.

                        Bob
                        N4HY


                        --
                        Laziness is the number one inspiration for ingenuity. Guilty as charged!
                      Your message has been successfully submitted and would be delivered to recipients shortly.