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

Re: How should I set the input level to my sound card for best RX(TX) performance?

Expand Messages
  • warrenallgyer
    Awwww.... c mon Tony. You just love to watch Alan, David, Roger, and me fight.... admit it! Of course we are having fun! I have said it before: Softrocks are
    Message 1 of 22 , Jan 24, 2013
    • 0 Attachment
      Awwww.... c'mon Tony. You just love to watch Alan, David, Roger, and me fight.... admit it!

      Of course we are having fun! I have said it before: Softrocks are more fun than Beijing KTV and, in the end, a whole lot cheaper! Look at it as a service to the planet, keeping us off the streets.

      Thank you always for your work Tony.

      Best regards,

      Warren Allgyer - W8TOD

      --- In softrock40@yahoogroups.com, "Parks, Tony" wrote:
      >
      > The complex gain of the LT6231 stage is given by
      > (R56/R55)/(1+j6.28*f*R56*C44) where f is the frequency in Hz and j is the
      > sqrt of -1. This results in a break frequency of LP filter at the LT6231
      > of about 1/(6.28*4.99k*220pF)=145kHz. The simplest way to reduce the gain
      > of the amplifier would be to increase the value of R55. The value of R58
      > for the second amplifier stage in the LT6231 should also be increased in
      > the same way R55 is increased.
      >
      > I have enjoyed the discussion concerning dynamic range of the RXTX RX
      > function. I always hope people are having fun with the kits and find them
      > useful.
      >
      > 73,
      > Tony KB9YIG
      >
      > On Thu, Jan 24, 2013 at 5:13 AM, warrenallgyer wrote:
      >
      > > **
      > >
      > >
      > > .... and one more point in favor of the Peaberry:
      > >
      > > With the new firmware I am seeing a device noise floor in the range of
      > > -143 to -146 dBm after S9 calibration at -73. That is unnecessarily low
      > > given antenna noise levels of -120 dBm or worse. There is at least 20 dB of
      > > dynamic range going unused at the low end.
      > >
      > > Warren Allgyer - W8TOD
      > >
      > >
      > > --- In softrock40@yahoogroups.com, "warrenallgyer" wrote:
      > > >
      > > > For these tests I have been using the Realtek analog card inside my MSI
      > > Netbook. I have not tested the iMic.
      > > >
      > > > As a comparison I just took a look at my Peaberry. On the negative side
      > > I see some spurs (9.980 and 10.010) start to appear with a 10 MHz input at
      > > -50 dBm. On the other hand they are about 60 dB down and do not get worse
      > > as the input level is increased until I get to -17 dBm. At -18 dBm the
      > > trace is relatively clean with those spurs 50-60 dB down. At -17 dBm in all
      > > hell breaks loose and the trace is littered with spurs every 5 KHz at -60
      > > to -70 dBm indicated. This is just about the same place the RXTX/Realtek
      > > combo breaks as well.
      > > >
      > > > Warren Allgyer - W8TOD
      > > >
      > > > --- In softrock40@yahoogroups.com, David Turnbull wrote:
      > > > >
      > > > > I looked at my iMic and it seems that it only offers gain and no
      > > > > attenuation on the line in. I didn't test this; it's what the USB
      > > > > enumeration reports. Seems to match your report though. Maybe I got it
      > > > > wrong and most sound cards need it set at minimum, not maximum, or
      > > > > maybe it's just random. I wish Windows had a better volume mixer that
      > > > > would report dB because this information is readily available behind
      > > > > the scenes.
      > > > >
      > > > > If Rocky runs on your computer it will show the noise floor in -dBFS.
      > > > > This is decibels below full scale. Adjust the line-in volume for the
      > > > > lowest number (it's negative so bigger is better).
      > > > >
      > > > > 73 David AE9RB
      > > > > http://AE9RB.com/
      > > > >
      > > > > On Thu, Jan 24, 2013 at 12:41 AM, warrenallgyer wrote:
      > > > > > David
      > > > > >
      > > > > > You and Roger make the same point and I cannot argue whether the
      > > sound card control works in the analog domain prior to ADC or in the
      > > digital domain post ADC. Perhaps some do one and some do the other. I have
      > > not looked at the circuitry, only the behavior.
      > > > > >
      > > > > > However I am very sure that my sound card overloaded at -35 dBM RF
      > > input with the gain at 80 % and at -15 dBM with the gain minimized. And, in
      > > both cases, a -110 dBm signal gave me the same rise above the noise floor.
      > > > > >
      > > > > > I can't see how that can happen if the gain control works post ADC.
      > > > > >
      > > > > > Put another way..... my receiver exhibits 80 dB of dynamic range
      > > between noise floor and overload with the input gain set at 80%. It
      > > exhibits 100 dB of dynamic range with the gain set at minimum. In both
      > > cases the weak signal performance is the same.
      > > > > >
      > > > > > The WSPR decodes I got on 40 meters last night were all with the
      > > input gain at this minimal setting and it was easily the best night in
      > > recent weeks for me.
      > > > > >
      > > > > > Where am I going wrong?
      > > > > >
      > > > > > Warren Allgyer - W8TOD
      > > > > >
      > > > > > --- In softrock40@yahoogroups.com, David Turnbull wrote:
      > > > > >>
      > > > > >> On Wed, Jan 23, 2013 at 7:52 AM, warrenallgyer wrote:
      > > > > >> > How should I set the input level to my sound card for best RX(TX)
      > > performance?...... and why should I bother?
      > > > > >> >
      > > > > >> > The trick is to set the PC input level to the minimum possible
      > > for weak signal reception. This will then allow the largest possible signal
      > > to be received before overloading the sound card.
      > > > > >>
      > > > > >> Warren, I'm pretty sure this isn't doing what you want.
      > > > > >>
      > > > > >> You want to set your sound card line-in at 0dB for the best possible
      > > > > >> dynamic range on a SoftRock or any sound card SDR. For most sound
      > > > > >> cards, this means full volume. Volume settings on modern cards don't
      > > > > >> adjust analog gain; it's all digital. Unfortunately, some cards also
      > > > > >> provide digital gain and I'm not sure how to identify this on
      > > Windows
      > > > > >> (Mac shows volume in dB). If you have a Peaberry you will find that
      > > I
      > > > > >> locked the radio volume settings so it is always 0dB and the
      > > operating
      > > > > >> system can't change it.
      > > > > >>
      > > > > >> The only way to really add dynamic range is to adjust the LT6231
      > > gain.
      > > > > >> This is R59, R56, C45, C44 on a RXTX. Reduce these resistor values
      > > > > >> until you no longer see antenna noise and/or weak signals. The OpAmp
      > > > > >> is an integrator and the formula is f=1/(2*pi*c*r). Given the
      > > resistor
      > > > > >> r(in ohms) find a value for c(in F) where f(in Hz) is greater than
      > > the
      > > > > >> bandwidth of your sound card. By default, these radios are set for
      > > > > >> 96k. If your sound card has an anti-aliasing filter then you can
      > > > > >> simply remove the capacitors and leave them off.
      > > > > >>
      > > > > >> You might find an additional 20dB by using this procedure. It
      > > depends
      > > > > >> on the quality of your sound card and cables. Unfortunately, you
      > > will
      > > > > >> also find that WSPR no longer decodes in I/Q mode and will have to
      > > run
      > > > > >> HDSDR or PowerSDR and connect to WSPR with VAC.
      > > > > >>
      > > > > >> 73 David AE9RB
      > > > > >> http://AE9RB.com/
      > > > >
      > > >
      > >
      > >
      > >
      >
    • warrenallgyer
      I have done some work overnight to try to correlate a WSPR noise level of 0 to an actual noise level on HDSDR. It seems pretty consistent: Using the sound
      Message 2 of 22 , Jan 24, 2013
      • 0 Attachment
        I have done some work overnight to try to correlate a WSPR noise level of "0" to an actual noise level on HDSDR.

        It seems pretty consistent: Using the sound card input level to adjust for "0" during WSPR receive, then switching to HDSDR, the noise floor seems to be about -90 dBm. So it seems WSPR wants to see that level of noise to be comfortable.

        I have not figured an objective way of comparing WSPR decodes at this level versus the minimal setting I am using now as a norm. My gut tells me the minimal setting misses some decodes that I get with having a -90 noise floor but I cannot be sure.

        Warren Allgyer - W8TOD

        --- In softrock40@yahoogroups.com, David Turnbull wrote:
        >
        > On Thu, Jan 24, 2013 at 1:01 AM, Alan wrote:
        > > I'm puzzled by your WSPR-IQ comment. Is this a fault in WSPR?
        >
        > WSPR doesn't seem to find signals that are more than 20dB away from
        > it's arbitrary 0dB level. You can see them clearly in the waterfall.
        > They just don't decode. If you do the OpAmp gain mods then the noise
        > meter in the lower left always stays red with more than -20dB. By
        > using HDSDR or PowerSDR, you can manually set the gain and turn off
        > AGC. This VAC method seems to always outperform the I/Q mode anyways.
        > I suppose if your line-in volume has gain instead of attenuation then
        > you could fix it there too.
        >
        > 73 David AE9RB
        > http://AE9RB.com/
        >
      • warrenallgyer
        Further to this and, hopefully, my final note on the subject, I have run WSPR on 20 meters this afternoon for a little over an hour, alternating every few
        Message 3 of 22 , Jan 24, 2013
        • 0 Attachment
          Further to this and, hopefully, my final note on the subject, I have run WSPR on 20 meters this afternoon for a little over an hour, alternating every few cycles with the sound card input level set to minimum and maximum. At minimum the WSPR noise indicator showed -30, -30 and at maximum it showed -8, -8.

          If you are interested in the results you can check the WSPR database for my reception reports from 02:00 to 03:12 UTC on 1/25. During this period I spotted the same stations multiple times with each setting. I can find no correlation between the sound card setting and the SNR or the number of spots. Both settings seem to work equally well.

          If you read the log here are the times I changed the level:

          02:00 Changed to maximum
          02:12 Changed to minimum
          02:40 Changed to maximum
          02:56 Changed to minimum
          03:18 Changed to maximum

          "Minimum" in this case is one increment above the bottom. At the bottom step the control seems to mute the audio.

          Warren Allgyer - W8TOD

          --- In softrock40@yahoogroups.com, "warrenallgyer" wrote:
          >
          > I have done some work overnight to try to correlate a WSPR noise level of "0" to an actual noise level on HDSDR.
          >
          > It seems pretty consistent: Using the sound card input level to adjust for "0" during WSPR receive, then switching to HDSDR, the noise floor seems to be about -90 dBm. So it seems WSPR wants to see that level of noise to be comfortable.
          >
          > I have not figured an objective way of comparing WSPR decodes at this level versus the minimal setting I am using now as a norm. My gut tells me the minimal setting misses some decodes that I get with having a -90 noise floor but I cannot be sure.
          >
          > Warren Allgyer - W8TOD
          >
          > --- In softrock40@yahoogroups.com, David Turnbull wrote:
          > >
          > > On Thu, Jan 24, 2013 at 1:01 AM, Alan wrote:
          > > > I'm puzzled by your WSPR-IQ comment. Is this a fault in WSPR?
          > >
          > > WSPR doesn't seem to find signals that are more than 20dB away from
          > > it's arbitrary 0dB level. You can see them clearly in the waterfall.
          > > They just don't decode. If you do the OpAmp gain mods then the noise
          > > meter in the lower left always stays red with more than -20dB. By
          > > using HDSDR or PowerSDR, you can manually set the gain and turn off
          > > AGC. This VAC method seems to always outperform the I/Q mode anyways.
          > > I suppose if your line-in volume has gain instead of attenuation then
          > > you could fix it there too.
          > >
          > > 73 David AE9RB
          > > http://AE9RB.com/
          > >
          >
        • warrenallgyer
          I know.... I promised. You can commiserate with my ex-wife. Sorry. A comparison of the dynamic range of: 1) Griffin iMic USB adapter. 16 bits, 48 KHz only.$30
          Message 4 of 22 , Jan 26, 2013
          • 0 Attachment
            I know.... I promised. You can commiserate with my ex-wife. Sorry.

            A comparison of the dynamic range of:
            1) Griffin iMic USB adapter. 16 bits, 48 KHz only.$30 Amazon.com
            2) Creative X-Fi SB Pro USB. 24 bits, 48/92 KHz
            3) Realtec internal audio card, 16 Bits 48/92 KHz MSI Netbook.

            All tests done at 10 MHz with RXTX 30/20/17.

            Griffin iMic
            Noise floor after S9 calibration: -129 dBm
            RF Input level for 1 dB compression: -11 dBm
            Useable Dynamic Range: 118 dB
            Practical Dynamic Range with antenna noise at -110 dBm: 99dB
            Note: Record input level on the PC control panel set to one increment above minimum.

            Creative X-Fi USB
            Noise floor after S9 calibration: -147 dBm
            RF Input level for 1 dB compression: -12 dBm
            Useable Dynamic Range: 135 dB
            Practical Dynamic Range with antenna noise at -110 dBm: 98dB
            Note: Input level is fixed on the PC control panel

            Realtec Internal Card
            Noise floor after S9 calibration: -139 dBm
            RF Input level for 1 dB compression: -13 dBm
            Useable Dynamic Range: 126 dB
            Practical Dynamic Range with antenna noise at -110 dBm: 96dB
            Note: Record input level on the PC control panel set to one increment above minimum.

            My conclusion: The expensive external 24 bit USB card represents no advantage whatsoever over the internal card in the Netbook, and this is how I operate; internal card for IQ in and out, cheap external card for speakers and mic.

            Caveats: I have seen internal cards that do not support stereo and do not support 92 KHz sampling, The stereo is an obvious non-starter and the 92 KHz is operator preference.

            Now I am really finished with this!

            Warren Allgyer - W8TOD
          • warrenallgyer
            (retrieved for a couple of touch-ups) I know.... I promised no more. You can commiserate with my ex-wife on broken promises. Sorry. This is a comparison of the
            Message 5 of 22 , Jan 26, 2013
            • 0 Attachment
              (retrieved for a couple of touch-ups)

              I know.... I promised no more. You can commiserate with my ex-wife on broken promises. Sorry.

              This is a comparison of the dynamic range of:

              1) Griffin iMic USB adapter. 16 bits, 48 KHz only, $30 Amazon.com
              2) Creative X-Fi SB Pro USB. 24 bits, 48/96 KHz
              3) Realtec internal audio card, 16 Bits 48/96 KHz MSI Netbook.

              All tests done at 10 MHz with RXTX 30/20/17.

              Griffin iMic
              Noise floor after S9 calibration: -129 dBm
              RF Input level for 1 dB compression: -11 dBm
              Useable Dynamic Range: 118 dB
              Practical Dynamic Range with antenna noise at -110 dBm: 99dB
              Note: Record input level on the PC control panel set to one increment above minimum.

              Creative X-Fi USB
              Noise floor after S9 calibration: -147 dBm
              RF Input level for 1 dB compression: -12 dBm
              Useable Dynamic Range: 135 dB
              Practical Dynamic Range with antenna noise at -110 dBm: 98dB
              Note: Input level is fixed on the PC control panel

              Realtec Internal Card
              Noise floor after S9 calibration: -139 dBm
              RF Input level for 1 dB compression: -13 dBm
              Useable Dynamic Range: 126 dB
              Practical Dynamic Range with antenna noise at -110 dBm: 96dB
              Note: Record input level on the PC control panel set to one increment above minimum.

              My conclusion: The expensive external 24 bit USB card represents no advantage whatsoever in normal operations over the internal card in the Netbook.

              There is the potential to gain an additional 20 dB or so of headroom with the Soundblaster by adding RF attenuation or reducing the op amp gains, and resetting the software gains. However -11 dBm is more than 60 dB over S9 and that is normally sufficient even in severe overload conditions.

              Caveats: I have seen internal cards that do not support stereo and do not support 96 KHz sampling, The stereo is an obvious non-starter and the 96 KHz is operator preference.

              As I said before: Spend that expensive sound card money on antennas, whiskey, and roses for the wife. Better returns all the way around.

              Now I am really finished with this!

              Warren Allgyer - W8TOD

              Note defining my terms: "RF Input level for 1 dB compression". I found it hard to objectively compare the spurs and baseline disruption among the cards to determine similar levels of overload. Instead I got the input level in the region of overload and then, using the step attenuator, adjusted the RF input level until the indicated signal level stopped increasing. The level recorded was the last one dB increment which resulted in one dB increase in indicated level on HDSDR. Going one dB more gave minimal or no increase in the indicated level.

              --- In softrock40@yahoogroups.com, "warrenallgyer" wrote:
              >
              > Further to this and, hopefully, my final note on the subject,
            Your message has been successfully submitted and would be delivered to recipients shortly.