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

Qt2m : B DECODE

Expand Messages
  • EM Sharp
    The reason of the Proto B board Decode failure has been partially discovered. The board recognizes (decodes) packets from one radio and not from another. The
    Message 1 of 6 , Jul 18, 2006
    • 0 Attachment
      The reason of the Proto B board Decode failure has been partially
      discovered. The board recognizes (decodes) packets from one radio and
      not from another. The signals from an ICOM R7000 seem to be OK,
      signals from a HTX202 are not decodable. I am not yet ready to state
      whether the problem is signal level, distortion, noise, or some other
      thing yet to be discovered. But for now signals are being decoded
      with reasonable
      FCS counts (3% of RX). I will measure audio level for the Rx signal
      when I get a chance. I don't know how to measure and/or observe other
      parameters except looking at the signal with a scope. For the time
      being I am running the board with R27 (TUNE) set at the mid-point --
      should I be using the CW max setting?

      Elmer
    • scott@opentrac.org
      I had trouble decoding packets from my KPC-3 initially, before I realized the audio level was set too high. The 2211 doesn t handle twist well. Maybe tweaking
      Message 2 of 6 , Jul 18, 2006
      • 0 Attachment
        I had trouble decoding packets from my KPC-3 initially, before I realized
        the audio level was set too high. The 2211 doesn't handle twist well.
        Maybe tweaking the input filter would help - I'll have to experiment with
        that.

        Try setting the TUNE pot to the max setting, see if it improves anything for
        you.

        > -----Original Message-----
        > From: tracker2@yahoogroups.com
        > [mailto:tracker2@yahoogroups.com] On Behalf Of EM Sharp
        > Sent: Tuesday, July 18, 2006 12:18 PM
        > To: tracker2@yahoogroups.com
        > Subject: [tracker2] Qt2m : B DECODE
        >
        > The reason of the Proto B board Decode failure has been partially
        > discovered. The board recognizes (decodes) packets from one radio and
        > not from another. The signals from an ICOM R7000 seem to be OK,
        > signals from a HTX202 are not decodable. I am not yet ready to state
        > whether the problem is signal level, distortion, noise, or some other
        > thing yet to be discovered. But for now signals are being decoded
        > with reasonable
        > FCS counts (3% of RX). I will measure audio level for the Rx signal
        > when I get a chance. I don't know how to measure and/or observe other
        > parameters except looking at the signal with a scope. For the time
        > being I am running the board with R27 (TUNE) set at the mid-point --
        > should I be using the CW max setting?
        >
        > Elmer
        >
        >
        >
        >
        >
        >
        > Yahoo! Groups Links
        >
        >
        >
        >
        >
        >
        >
        >
        >
      • Ron Patten
        Turned tune for 50% receive duty cycle with cal alt. looped back. Missing approximately 25% packets compared to a reference TNC that uses a TCM3105 demod chip
        Message 3 of 6 , Jul 19, 2006
        • 0 Attachment
          Turned tune for 50% receive duty cycle with cal alt. looped back.
          Missing approximately 25% packets compared to a reference TNC that uses
          a TCM3105 demod chip (same audio source).
          Ran logic output of TNCs TCM3105 demod chip in place of 2211 and copied
          more than reference TNC on the tracker2.
          I have a MX614 (similar to a TCM3105) on order to try some experiments
          with it.

          scott@... wrote:

          > >I had trouble decoding packets from my KPC-3 initially, before I realized
          > >the audio level was set too high. The 2211 doesn't handle twist well.
          > >Maybe tweaking the input filter would help - I'll have to experiment with
          > >that.
          > >
          > >Try setting the TUNE pot to the max setting, see if it improves
          > anything for
          > >you.
          >
        • scott@opentrac.org
          Was cal loopback working? I haven t tested it since I tweaked the FSK code, and that whole thing needs to be reworked. Let me know how your experiments come
          Message 4 of 6 , Jul 19, 2006
          • 0 Attachment
            Was cal loopback working? I haven't tested it since I tweaked the FSK code,
            and that whole thing needs to be reworked.

            Let me know how your experiments come out. It's interesting that the T2
            copied more than the reference TNC with the TCM3105 - must mean my clock
            recovery tweaks worked. What's your test audio source?

            What's in front of the TCM3105? The T2 has a very simple filter network on
            the input to the 2211. From what I read about various TNC2 clones, though,
            the active filter used in some didn't really improve performance, and was
            reducing dynamic range. The T2's filter is nearly identical to the passive
            one on the MFJ-1270B.

            The rest of the 2211 circuit is straight out of the datasheet. I guess I
            ought to look into the collected TNC2 lore and see if any variations were
            found to work better.

            The MX614's not a bad chip, but it's several times more expensive than the
            2211 (more when you include the crystal), it's only available from one
            company, and I didn't need the TX side. I looked at the CMX860 when I
            started on the T2 project, but it would have added $8 to $10 to the cost and
            was only available in a surface mount package. Wouldn't have fit in the
            small case, either. It can do DTMF, but not at the same time as Bell 202.

            The 2211's got a long history in TNCs, and it's dirt cheap. I just need to
            work some more on squeezing as much performance as possible out of it.

            Scott

            > -----Original Message-----
            > From: tracker2@yahoogroups.com
            > [mailto:tracker2@yahoogroups.com] On Behalf Of Ron Patten
            > Sent: Wednesday, July 19, 2006 9:24 AM
            > To: tracker2@yahoogroups.com
            > Subject: Re: [tracker2] Qt2m : B DECODE
            >
            > Turned tune for 50% receive duty cycle with cal alt. looped back.
            > Missing approximately 25% packets compared to a reference TNC
            > that uses
            > a TCM3105 demod chip (same audio source).
            > Ran logic output of TNCs TCM3105 demod chip in place of 2211
            > and copied
            > more than reference TNC on the tracker2.
            > I have a MX614 (similar to a TCM3105) on order to try some
            > experiments
            > with it.
            >
            > scott@... wrote:
            >
            > > >I had trouble decoding packets from my KPC-3 initially,
            > before I realized
            > > >the audio level was set too high. The 2211 doesn't handle
            > twist well.
            > > >Maybe tweaking the input filter would help - I'll have to
            > experiment with
            > > >that.
            > > >
            > > >Try setting the TUNE pot to the max setting, see if it improves
            > > anything for
            > > >you.
            > >
            >
            >
            >
            >
            > ------------------------ Yahoo! Groups Sponsor
            > --------------------~-->
            > Something is new at Yahoo! Groups. Check out the enhanced
            > email design.
            > http://us.click.yahoo.com/SISQkA/gOaOAA/yQLSAA/ELTolB/TM
            > --------------------------------------------------------------
            > ------~->
            >
            >
            > Yahoo! Groups Links
            >
            >
            >
            >
            >
            >
            >
            >
            >
          • scott@opentrac.org
            Here s an easy experiment - pull the 2211, bend pin 6 out of the way, and put it back in the socket. Enable SWDCD and see how your decode performance
            Message 5 of 6 , Jul 19, 2006
            • 0 Attachment
              Here's an easy experiment - pull the 2211, bend pin 6 out of the way, and
              put it back in the socket. Enable SWDCD and see how your decode performance
              compares. I got about a 7% improvement (836 packets decoded vs. 780) but
              I've only run the test once so far.

              Pin 6 is a lock detect output, and serves to keep the data output pulled low
              when no lock is detected. It functions as a carrier detect without tying up
              another I/O pin.

              I think the problem is that the lock detect is dropping during reception of
              marginal signals that the CPU would otherwise still be able to properly
              decode. Disconnecting pin 6 keeps the lock detect from interfering with the
              data output.

              Looks like I'll have to devote an I/O pin to the carrier detect. Either
              that, or just always run in SWDCD mode. That doesn't let you avoid non-AFSK
              traffic, though.

              Scott

              > -----Original Message-----
              > From: tracker2@yahoogroups.com
              > [mailto:tracker2@yahoogroups.com] On Behalf Of Ron Patten
              > Sent: Wednesday, July 19, 2006 9:24 AM
              > To: tracker2@yahoogroups.com
              > Subject: Re: [tracker2] Qt2m : B DECODE
              >
              > Turned tune for 50% receive duty cycle with cal alt. looped back.
              > Missing approximately 25% packets compared to a reference TNC
              > that uses
              > a TCM3105 demod chip (same audio source).
              > Ran logic output of TNCs TCM3105 demod chip in place of 2211
              > and copied
              > more than reference TNC on the tracker2.
              > I have a MX614 (similar to a TCM3105) on order to try some
              > experiments
              > with it.
              >
              > scott@... wrote:
              >
              > > >I had trouble decoding packets from my KPC-3 initially,
              > before I realized
              > > >the audio level was set too high. The 2211 doesn't handle
              > twist well.
              > > >Maybe tweaking the input filter would help - I'll have to
              > experiment with
              > > >that.
              > > >
              > > >Try setting the TUNE pot to the max setting, see if it improves
              > > anything for
              > > >you.
              > >
              >
              >
              >
              >
              > ------------------------ Yahoo! Groups Sponsor
              > --------------------~-->
              > Something is new at Yahoo! Groups. Check out the enhanced
              > email design.
              > http://us.click.yahoo.com/SISQkA/gOaOAA/yQLSAA/ELTolB/TM
              > --------------------------------------------------------------
              > ------~->
              >
              >
              > Yahoo! Groups Links
              >
              >
              >
              >
              >
              >
              >
              >
              >
            • rsp92028
              ... Loopback display function not working, used a scope. ... Local APRS received audio split to the TNCs. ... This is a old AEA PK-12, it has a opamp that
              Message 6 of 6 , Jul 19, 2006
              • 0 Attachment
                --- In tracker2@yahoogroups.com, <scott@...> wrote:
                >
                > Was cal loopback working?

                Loopback display function not working, used a scope.

                > I haven't tested it since I tweaked the FSK code,
                > and that whole thing needs to be reworked.

                >
                > Let me know how your experiments come out. It's interesting that the T2
                > copied more than the reference TNC with the TCM3105 - must mean my clock
                > recovery tweaks worked. What's your test audio source?

                Local APRS received audio split to the TNCs.

                > What's in front of the TCM3105?

                This is a old AEA PK-12, it has a opamp that looks like it's set up to
                audio limit, with no filtering.

                > The T2 has a very simple filter network on
                > the input to the 2211. From what I read about various TNC2 clones,
                though,
                > the active filter used in some didn't really improve performance,
                and was
                > reducing dynamic range. The T2's filter is nearly identical to the
                passive
                > one on the MFJ-1270B.
                >
                > The rest of the 2211 circuit is straight out of the datasheet. I
                guess I
                > ought to look into the collected TNC2 lore and see if any variations
                were
                > found to work better.
                >
                > The MX614's not a bad chip, but it's several times more expensive
                than the
                > 2211 (more when you include the crystal), it's only available from one
                > company, and I didn't need the TX side. I looked at the CMX860 when I
                > started on the T2 project, but it would have added $8 to $10 to the
                cost and
                > was only available in a surface mount package. Wouldn't have fit in the
                > small case, either. It can do DTMF, but not at the same time as
                Bell 202.
                >
                > The 2211's got a long history in TNCs, and it's dirt cheap. I just
                need to
                > work some more on squeezing as much performance as possible out of it.
                >
                > Scott
                >
                > > -----Original Message-----
                > > From: tracker2@yahoogroups.com
                > > [mailto:tracker2@yahoogroups.com] On Behalf Of Ron Patten
                > > Sent: Wednesday, July 19, 2006 9:24 AM
                > > To: tracker2@yahoogroups.com
                > > Subject: Re: [tracker2] Qt2m : B DECODE
                > >
                > > Turned tune for 50% receive duty cycle with cal alt. looped back.
                > > Missing approximately 25% packets compared to a reference TNC
                > > that uses
                > > a TCM3105 demod chip (same audio source).
                > > Ran logic output of TNCs TCM3105 demod chip in place of 2211
                > > and copied
                > > more than reference TNC on the tracker2.
                > > I have a MX614 (similar to a TCM3105) on order to try some
                > > experiments
                > > with it.
                > >
                > > scott@... wrote:
                > >
                > > > >I had trouble decoding packets from my KPC-3 initially,
                > > before I realized
                > > > >the audio level was set too high. The 2211 doesn't handle
                > > twist well.
                > > > >Maybe tweaking the input filter would help - I'll have to
                > > experiment with
                > > > >that.
                > > > >
                > > > >Try setting the TUNE pot to the max setting, see if it improves
                > > > anything for
                > > > >you.
              Your message has been successfully submitted and would be delivered to recipients shortly.