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

Random loss of image rejection

Expand Messages
  • cctbcn
    After having set it aside for a while, I finally completed my G59/GPA10 combo and am pleased to note that the show stopper that halted my progress earlier -
    Message 1 of 6 , Oct 10, 2012
      After having set it aside for a while, I finally completed my G59/GPA10 combo and am pleased to note that the "show stopper" that halted my progress earlier - a bug in the calculation of the Si570 frequency programming if the calibration was too far off - has been addressed. If you were to dig out a previous email from me, this was manifested by there being "dead spots" when tuning up/down the bands.

      Now that I've gotten everything else together - and in the nice G59 case - I have another head-scratcher, the solution to which has evaded my search of this group's archives.

      What is happening is that I will be listening to the receiver and suddenly the image rejection will go to zero: I have pairs of signals above and below "zero" Hz.

      If the signals are strong and there's no QRM on the image, one may not necessarily notice immediately that anything has happened, aside from a slight "hollow" sound in the audio that results. Of course, if there *is* QRM (say, on a busy band) then one immediately hears other signals appear where they shouldn't.

      When this happens, there's no obvious "click" in the audio and I don't even have to be doing anything on the computer, and it doesn't matter if there are other programs running or not.

      There are two ways to temporarily "fix" this:

      - Hit the stand by button to stop GSDR, and then again to restart it.

      - Tune the frequency off by quite a bit (1 MHz will do) and then back again. Just tuning up and down a few 10's of kHz won't work - and I'm not quite sure how far one really needs to go.

      The other thing that sometimes causes this to happen is when one tunes outside the range of the current passband and the LO has to re-center. This isn't consistent.

      This clearly isn't the G59 itself that's causing this to happen, but something in GSDR. The version that I'm using is the most current (a month or so old) and the sound card is an Asus Xonar D1 on an XP system with quad core, so the CPU usage is under 10%. I'm also running the most recent version of VAC - having upgraded from a previous version with no improvement. The sample rate is 192 kHz and I've tried all sorts of different combinations of buffer sizes and delays, but to no effect. I have yet to try running it at a lower sample rate for a long time to see if the problem occurs at 96 or 48 kHz.

      The most telling thing is the "hollow" sound that appears: It would seem that the two audio channels (L/R or I/Q) are out of sync far enough that an audible delay appears - and this would certainly wreck the image suppression.

      I've not researched it enough, yet, but I *think* that this may also be happening on transmit as well, so it's understandable that I'm reluctant to put the transmitter on the air just yet! For the moment I'm concentrating on the receive problem.

      Any ideas on how to fix this?

      When I get a chance to get back to it, I'll do more testing at lower sample rates (96 kHz) to see if it happens then.

      73,

      Clint
      KA7OEI
    • Hank
      Clint, This only applys if you are using inexpensive audio cables for line in/out, otherwise, just skip my reply. I had similar problems, but maybe not as
      Message 2 of 6 , Oct 10, 2012
        Clint,

        This only applys if you are using inexpensive audio cables for line in/out, otherwise, just skip my reply.

        I had similar problems, but maybe not as consistent. Now and then I would loose my image rejection in GSDR. I finally broke down and bought really good audio cables, double shielded, etc.. This has kept the image problem from returning. What I gather was the mess of cables I had from behind my PC, my power supply, coax to/from radio, and my ladder line antenna all were playing havoc with my inexpensive audio cables going to/from the pc. Not saying that will solve all of your problems, but it certainly will help if you are using the $10 cables for PC line in/out like I was. I was reluctant to buy the cables initially, but I am very happy I did.

        Hank
        WZ3HMT

        --- In GenesisRadio@yahoogroups.com, "cctbcn" <turner@...> wrote:
        >
        > After having set it aside for a while, I finally completed my G59/GPA10 combo and am pleased to note that the "show stopper" that halted my progress earlier - a bug in the calculation of the Si570 frequency programming if the calibration was too far off - has been addressed. If you were to dig out a previous email from me, this was manifested by there being "dead spots" when tuning up/down the bands.
        >
        > Now that I've gotten everything else together - and in the nice G59 case - I have another head-scratcher, the solution to which has evaded my search of this group's archives.
        >
        > What is happening is that I will be listening to the receiver and suddenly the image rejection will go to zero: I have pairs of signals above and below "zero" Hz.
        >
        > If the signals are strong and there's no QRM on the image, one may not necessarily notice immediately that anything has happened, aside from a slight "hollow" sound in the audio that results. Of course, if there *is* QRM (say, on a busy band) then one immediately hears other signals appear where they shouldn't.
        >
        > When this happens, there's no obvious "click" in the audio and I don't even have to be doing anything on the computer, and it doesn't matter if there are other programs running or not.
        >
        > There are two ways to temporarily "fix" this:
        >
        > - Hit the stand by button to stop GSDR, and then again to restart it.
        >
        > - Tune the frequency off by quite a bit (1 MHz will do) and then back again. Just tuning up and down a few 10's of kHz won't work - and I'm not quite sure how far one really needs to go.
        >
        > The other thing that sometimes causes this to happen is when one tunes outside the range of the current passband and the LO has to re-center. This isn't consistent.
        >
        > This clearly isn't the G59 itself that's causing this to happen, but something in GSDR. The version that I'm using is the most current (a month or so old) and the sound card is an Asus Xonar D1 on an XP system with quad core, so the CPU usage is under 10%. I'm also running the most recent version of VAC - having upgraded from a previous version with no improvement. The sample rate is 192 kHz and I've tried all sorts of different combinations of buffer sizes and delays, but to no effect. I have yet to try running it at a lower sample rate for a long time to see if the problem occurs at 96 or 48 kHz.
        >
        > The most telling thing is the "hollow" sound that appears: It would seem that the two audio channels (L/R or I/Q) are out of sync far enough that an audible delay appears - and this would certainly wreck the image suppression.
        >
        > I've not researched it enough, yet, but I *think* that this may also be happening on transmit as well, so it's understandable that I'm reluctant to put the transmitter on the air just yet! For the moment I'm concentrating on the receive problem.
        >
        > Any ideas on how to fix this?
        >
        > When I get a chance to get back to it, I'll do more testing at lower sample rates (96 kHz) to see if it happens then.
        >
        > 73,
        >
        > Clint
        > KA7OEI
        >
      • cctbcn
        Hi Hank, Good thoughts - but (unfortunately - since it would be easy to fix) that s not the problem here. The give-away is that stopping/restarting or retuning
        Message 3 of 6 , Oct 11, 2012
          Hi Hank,

          Good thoughts - but (unfortunately - since it would be easy to fix) that's not the problem here.

          The give-away is that stopping/restarting or retuning seems to "fix" it temporarily - and the cable wouldn't "know" about that.

          The other thing that I notice is that when it is working "normally", stopping WBIR results in the normal hardware image rejection (somewhere in the 45-55 dB area) but when the images appear, there is no difference with WBIR running or not.

          The problem is in the audio chain and, somehow, it appears that the I and Q channels' piplelines getting a few hundred microseconds or so out of sync with each other - this based on the "hollow" sound that I'm hearing in the demodulated audio.

          I'll experiment with other SDR software to see if it happens with that, too: If so, I will know that it's a driver/HW problem.

          73,

          Clint
          KA7OEI


          --- In GenesisRadio@yahoogroups.com, "Hank" <hank@...> wrote:
          >
          > Clint,
          >
          > This only applys if you are using inexpensive audio cables for line in/out, otherwise, just skip my reply.
          >
          > I had similar problems, but maybe not as consistent. Now and then I would loose my image rejection in GSDR. I finally broke down and bought really good audio cables, double shielded, etc.. This has kept the image problem from returning. What I gather was the mess of cables I had from behind my PC, my power supply, coax to/from radio, and my ladder line antenna all were playing havoc with my inexpensive audio cables going to/from the pc. Not saying that will solve all of your problems, but it certainly will help if you are using the $10 cables for PC line in/out like I was. I was reluctant to buy the cables initially, but I am very happy I did.
          >
          > Hank
          > WZ3HMT
          >
        • Bruce KF1Z
          The give-away is that stopping/restarting or retuning seems to fix it temporarily - and the cable wouldn t know about that. ... The cable wouldn t, but
          Message 4 of 6 , Oct 11, 2012
             

            ""The give-away is that stopping/restarting or retuning seems to "fix" it temporarily - and the cable wouldn't "know" about that.""
            -----------------------
             
             
            The cable wouldn't, but the rest of the system would.
             
            Smile
             
            What you describe sounds just like a cable, or other intermittent problem to me.
            Especially caused by sharp pulse type noise ( like a loose audio connection can make)
             
            Once WBIR looses it's way do to this type of interference...
            It needs to be restarted...
            This is done by either stopping and restarting via the standby button..
            Or several other software settings like TX / RX transistion... Af-preamp on then off... band changes ... and some others..
             
             
            Not saying that any of this is perfectly relevant to your particular situation....
            But in short... Yes... the audio cables COULD create a situation where the "fix" is to stop and restart GSDR!
             
             
            Bruce
             


             
          • Bruce KF1Z
            Oh, And yes, test with another piece of software ( HDSDR ) However keep in mind that other software may utilize resources differently, and you may not get the
            Message 5 of 6 , Oct 11, 2012
              Oh,
               
              And yes, test with another piece of software ( HDSDR )
              However keep in mind that other software may utilize resources differently, and you may not get the firm answer you think.   Smile
               
              But, I think you are on the right track...
               
              What are your soundcard settings in GSDR? and are you using VAC at the time?
               
              Easiest would be to email, or post a screenshot of the GSDR soundcard setting tab ( VAC too if used..)
               
              The buffer, and latency settings may be part of the issue... ?
               
               
              Bruce
               
               
              From: cctbcn
              Sent: Thursday, October 11, 2012 5:28 PM
              Subject: [GenesisRadio] Re: Random loss of image rejection
               
               

              Hi Hank,

              Good thoughts - but (unfortunately - since it would be easy to fix) that's not the problem here.

              The give-away is that stopping/restarting or retuning seems to "fix" it temporarily - and the cable wouldn't "know" about that.

              The other thing that I notice is that when it is working "normally", stopping WBIR results in the normal hardware image rejection (somewhere in the 45-55 dB area) but when the images appear, there is no difference with WBIR running or not.

              The problem is in the audio chain and, somehow, it appears that the I and Q channels' piplelines getting a few hundred microseconds or so out of sync with each other - this based on the "hollow" sound that I'm hearing in the demodulated audio.

              I'll experiment with other SDR software to see if it happens with that, too: If so, I will know that it's a driver/HW problem.

              73,

              Clint
              KA7OEI

              --- In mailto:GenesisRadio%40yahoogroups.com, "Hank" <hank@...> wrote:

              >
              > Clint,
              >
              > This only
              applys if you are using inexpensive audio cables for line in/out, otherwise, just skip my reply.
              >
              > I had similar problems, but maybe not as
              consistent. Now and then I would loose my image rejection in GSDR. I finally broke down and bought really good audio cables, double shielded, etc.. This has kept the image problem from returning. What I gather was the mess of cables I had from behind my PC, my power supply, coax to/from radio, and my ladder line antenna all were playing havoc with my inexpensive audio cables going to/from the pc. Not saying that will solve all of your problems, but it certainly will help if you are using the $10 cables for PC line in/out like I was. I was reluctant to buy the cables initially, but I am very happy I did.
              >
              > Hank
              > WZ3HMT
              >

            • billncolo36
              Be sure you do the WBIR cal with a very high input SNR. Going from memory I think I used an input signal level around -30 dBm. When I used an input around S9,
              Message 6 of 6 , Oct 11, 2012
                Be sure you do the WBIR cal with a very high input SNR. Going from memory I think I used an input signal level around -30 dBm. When I used an input around S9, I got results similar to yours.

                Bill N0CU

                --- In GenesisRadio@yahoogroups.com, "cctbcn" <turner@...> wrote:
                >
                > After having set it aside for a while, I finally completed my G59/GPA10 combo and am pleased to note that the "show stopper" that halted my progress earlier - a bug in the calculation of the Si570 frequency programming if the calibration was too far off - has been addressed. If you were to dig out a previous email from me, this was manifested by there being "dead spots" when tuning up/down the bands.
                >
                > Now that I've gotten everything else together - and in the nice G59 case - I have another head-scratcher, the solution to which has evaded my search of this group's archives.
                >
                > What is happening is that I will be listening to the receiver and suddenly the image rejection will go to zero: I have pairs of signals above and below "zero" Hz.
                >
                > If the signals are strong and there's no QRM on the image, one may not necessarily notice immediately that anything has happened, aside from a slight "hollow" sound in the audio that results. Of course, if there *is* QRM (say, on a busy band) then one immediately hears other signals appear where they shouldn't.
                >
                > When this happens, there's no obvious "click" in the audio and I don't even have to be doing anything on the computer, and it doesn't matter if there are other programs running or not.
                >
                > There are two ways to temporarily "fix" this:
                >
                > - Hit the stand by button to stop GSDR, and then again to restart it.
                >
                > - Tune the frequency off by quite a bit (1 MHz will do) and then back again. Just tuning up and down a few 10's of kHz won't work - and I'm not quite sure how far one really needs to go.
                >
                > The other thing that sometimes causes this to happen is when one tunes outside the range of the current passband and the LO has to re-center. This isn't consistent.
                >
                > This clearly isn't the G59 itself that's causing this to happen, but something in GSDR. The version that I'm using is the most current (a month or so old) and the sound card is an Asus Xonar D1 on an XP system with quad core, so the CPU usage is under 10%. I'm also running the most recent version of VAC - having upgraded from a previous version with no improvement. The sample rate is 192 kHz and I've tried all sorts of different combinations of buffer sizes and delays, but to no effect. I have yet to try running it at a lower sample rate for a long time to see if the problem occurs at 96 or 48 kHz.
                >
                > The most telling thing is the "hollow" sound that appears: It would seem that the two audio channels (L/R or I/Q) are out of sync far enough that an audible delay appears - and this would certainly wreck the image suppression.
                >
                > I've not researched it enough, yet, but I *think* that this may also be happening on transmit as well, so it's understandable that I'm reluctant to put the transmitter on the air just yet! For the moment I'm concentrating on the receive problem.
                >
                > Any ideas on how to fix this?
                >
                > When I get a chance to get back to it, I'll do more testing at lower sample rates (96 kHz) to see if it happens then.
                >
                > 73,
                >
                > Clint
                > KA7OEI
                >
              Your message has been successfully submitted and would be delivered to recipients shortly.