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

Re: [O_P_E_R_A_] Re: OPERA versus WSPR

Expand Messages
  • Andy Talbot
    So Opera S/N ratios are based on 625Hz noise bandwidth ??? Is this what path-sim adopts, and if so why on earth such an arbitrary value. -36dB in WSPR s JT
    Message 1 of 22 , Jun 30, 2012
    • 0 Attachment
      So Opera S/N ratios are based on 625Hz noise bandwidth  ???   Is this what path-sim adopts, and if so why on earth such an arbitrary value.
       
      -36dB in WSPR's JT standard 2.5kHz means -30dB in 625Hz  (your example below)
       
      Why don't people just settle on normalised S/N in unity 1Hz bandwidth.    Where did this 'magic' 2.5kHz come from?
       
      Andy


       
      On 30 June 2012 19:34, graham787 <g0nbd@...> wrote:
       

      Ok Roger , well we don't want to be seen 'Singing the Praises' :)

      Op s/n is calibrated against the 'sim-path' propagation simulation software. s/n readings are a little subjective, path-sim was felt to give a more realistic value .

      This puts the s/n readings 6 dB less than wspr / fldigi
      Op = -30dB wspr = -36dB

      Op s/n levels are the average of the transmission along the time line and -not- peak values

      With ook , Op has a obvious disadvantage compared to psk/fsk systems , but is not affected by Doppler spread, simplify the Tx path and with the data processing, will decode with up to 50% data loss (random distribution along the time line)

      Good luck with the tests

      G..

      --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@...> wrote:
      >
      > Can someone please remind me of the equivalences between the various OPERA speeds and WSPR?
      >
      > If I am -30/-29dB on OP4 (my reports from ON6EO and G0KTN earlier) what sort of WSPR report would this correspond to?
      >
      > 73s
      > Roger G3XBM
      >


    • G3XBM
      So, roughly, to be clear, a -30dB OP4 signal corresponds to -24dB WSPR? 73s Roger G3XBM
      Message 2 of 22 , Jul 1, 2012
      • 0 Attachment
        So, roughly, to be clear, a -30dB OP4 signal corresponds to -24dB WSPR?

        73s
        Roger G3XBM

        --- In O_P_E_R_A_@yahoogroups.com, Andy Talbot <andy.g4jnt@...> wrote:
        >
        > So Opera S/N ratios are based on 625Hz noise bandwidth ???
      • m0fmt
        Hi All These HAM didgi systems are all based on your average BLACK BOX specs for your average ops as far as I can see, so could it be that they see 2.5kc/s
        Message 3 of 22 , Jul 1, 2012
        • 0 Attachment
          Hi All

          These "HAM" didgi systems are all based on your average BLACK BOX specs for your average ops as far as I can see, so could it be that they see 2.5kc/s as being the typical SSB audio pass band for these boxes therefore basing their specs around it?
          Ignoring elementary units to confuse comparison.

          However Andy have you published in public domain a PIC design for a stand alone WRSP beacon?
          I would like to have another go at one. Now that WSPR doesn't seem to be going away from LF. I have built Johans one which was fine but I could not set deviation to get the tones right it was very touchy. It looked good on paper but I wasn't up to making it work.
          Is there is anying else about for direct keying? If not I will dig it out again but don't hold out much hope for success.

          73 Pete FMT.



          --- In O_P_E_R_A_@yahoogroups.com, Andy Talbot <andy.g4jnt@...> wrote:
          >
          > So Opera S/N ratios are based on 625Hz noise bandwidth ??? Is this what
          > path-sim adopts, and if so why on earth such an arbitrary value.
          >
          > -36dB in WSPR's JT standard 2.5kHz means -30dB in 625Hz (your example
          > below)
          >
          > Why don't people just settle on normalised S/N in unity 1Hz bandwidth.
          > Where did this 'magic' 2.5kHz come from?
          >
          > Andy
          > www.g4jnt.com
          >
          >
          >
          > On 30 June 2012 19:34, graham787 <g0nbd@...> wrote:
          >
          > > **
          > >
          > >
          > > Ok Roger , well we don't want to be seen 'Singing the Praises' :)
          > >
          > > Op s/n is calibrated against the 'sim-path' propagation simulation
          > > software. s/n readings are a little subjective, path-sim was felt to give a
          > > more realistic value .
          > >
          > > This puts the s/n readings 6 dB less than wspr / fldigi
          > > Op = -30dB wspr = -36dB
          > >
          > > Op s/n levels are the average of the transmission along the time line and
          > > -not- peak values
          > >
          > > With ook , Op has a obvious disadvantage compared to psk/fsk systems , but
          > > is not affected by Doppler spread, simplify the Tx path and with the data
          > > processing, will decode with up to 50% data loss (random distribution along
          > > the time line)
          > >
          > > Good luck with the tests
          > >
          > > G..
          > >
          > > --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@> wrote:
          > > >
          > > > Can someone please remind me of the equivalences between the various
          > > OPERA speeds and WSPR?
          > > >
          > > > If I am -30/-29dB on OP4 (my reports from ON6EO and G0KTN earlier) what
          > > sort of WSPR report would this correspond to?
          > > >
          > > > 73s
          > > > Roger G3XBM
          > > >
          > >
          > >
          > >
          >
        • graham787
          Roger , that s the opposite ! This puts the s/n readings 6 dB less than wspr / fldigi In this case , Less is more , for the same power, wspr will report a
          Message 4 of 22 , Jul 1, 2012
          • 0 Attachment
            Roger , that's the opposite !


            This puts the s/n readings 6 dB less than wspr / fldigi

            In this case , Less is more , for the same power, wspr will report a lower signal level .... same power, same link, as Andy says differing measuring parameters

            IE signals levels Op = -30dB, then wspr = -36dB

            OP2 modes is roughly equal to wspr in performance, not quite like for like, due to average / peak , single/fsk but everything is based on stat's , sometimes you will see the advised minimum decode levels exceeded

            The OP system will decode lower than wspr, by simple time integration .. the longer the pulse, the more the average power, the lower the power level needed to achieve a detectable signal.

            Simple in concept, but as the CPU usage indicates , there are a lot of 'decisions' being made to recover the data, cpu loading is a indication as to the level numeric computation and not a pointer to inefficiency in the coding, as Op is under development , loading may change.

            73 -G..


            --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@...> wrote:
            >
            > So, roughly, to be clear, a -30dB OP4 signal corresponds to -24dB WSPR?
            >
            > 73s
            > Roger G3XBM
            >
            > --- In O_P_E_R_A_@yahoogroups.com, Andy Talbot <andy.g4jnt@> wrote:
            > >
            > > So Opera S/N ratios are based on 625Hz noise bandwidth ???
            >
          • Andy Talbot
            Yes... sort of... This one is specifically for generating WSPR using the AD9852 DDS chip http://www.g4jnt.com/WSPRBCNS.ZIP But, take a look at this page.
            Message 5 of 22 , Jul 1, 2012
            • 0 Attachment
              Yes...
              sort of...
              This one is specifically for generating WSPR using the AD9852 DDS chip
               
              But, take a look at this page. http://www.g4jnt.com/JTModesBcns.htm   On there I describe a PIC based DDS that was designed for subcarrier optical comms,  but can equally well generate audio frequencies for upconversion. 
              That low frequency PIC DDS was described in basic form in RadCom Design Notes a  year or so ago.    Different versions will either accept a parallel word suitable for WSPR or JT4   or a serial lowspeed SPI type interface needed for more complex freqeuncy setting, specifically for JT65.
               
              At the moment on the web site I don't include a WSPR generator, just JTxx, but if you wait a couple of hours before downloading, I'll look out the WSPR PIC code and include it too.
               
              Andy G4JNT

              On 1 July 2012 11:57, m0fmt <m0fmt@...> wrote:
               

              Hi All

              However Andy have you published in public domain a PIC design for a stand alone WRSP beacon?
              I would like to have another go at one. Now that WSPR doesn't seem to be going away from LF. I have built Johans one which was fine but I could not set deviation to get the tones right it was very touchy. It looked good on paper but I wasn't up to making it work.
              Is there is anying else about for direct keying? If not I will dig it out again but don't hold out much hope for success.

              73 Pete FMT.

            • booz2m
              Hi all The Next Generation Beacon project DDS multimode S/W can do Opera and WSPR plus JT4, JT65 and PI4 without changing S/W. The only thing that is needed is
              Message 6 of 22 , Jul 1, 2012
              • 0 Attachment
                Hi all

                The Next Generation Beacon project DDS multimode S/W can do Opera and WSPR plus JT4, JT65 and PI4 without changing S/W. The only thing that is needed is to enter the callsign, locator and frequency in plain text via the USB port.

                http://rudius.net/oz2m/ngnb/dds.htm

                The "only" drawback is the time width of the symbols has not yet been calibrated. I expect this to happen soon after the VUSHF Field Day next week-end. When done the S/W will be GNUed.

                73
                Bo, OZ2M
              • Andy Talbot
                So does that mean OP / Sim S/N readings are based on a 10kHz noise bandwidth. That is a bit more logical for an HF simulator tool. Why couldn t Opera use the
                Message 7 of 22 , Jul 1, 2012
                • 0 Attachment
                  So does that mean OP / Sim S/N readings are based on a 10kHz noise bandwidth.  That is a bit more logical for an HF simulator tool.
                   
                  Why couldn't  Opera use the same standard as everything else in thg Amateur communitty- just as we all get to know and understand the 2.5kHz WSJT defined figure, however arbitrary it may be, someone has to come up with another one just to confuse
                  the issue.
                   
                  Andy
                  'jnt


                   
                  On 1 July 2012 12:07, graham787 <g0nbd@...> wrote:
                   

                  Roger , that's the opposite !



                  This puts the s/n readings 6 dB less than wspr / fldigi

                  In this case , Less is more , for the same power, wspr will report a lower signal level .... same power, same link, as Andy says differing measuring parameters

                  IE signals levels Op = -30dB, then wspr = -36dB

                  OP2 modes is roughly equal to wspr in performance, not quite like for like, due to average / peak , single/fsk but everything is based on stat's , sometimes you will see the advised minimum decode levels exceeded

                  The OP system will decode lower than wspr, by simple time integration .. the longer the pulse, the more the average power, the lower the power level needed to achieve a detectable signal.

                  Simple in concept, but as the CPU usage indicates , there are a lot of 'decisions' being made to recover the data, cpu loading is a indication as to the level numeric computation and not a pointer to inefficiency in the coding, as Op is under development , loading may change.

                  73 -G..


                  --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@...> wrote:
                  >
                  > So, roughly, to be clear, a -30dB OP4 signal corresponds to -24dB WSPR?
                  >
                  > 73s
                  > Roger G3XBM
                  >
                  > --- In O_P_E_R_A_@yahoogroups.com, Andy Talbot <andy.g4jnt@> wrote:
                  > >
                  > > So Opera S/N ratios are based on 625Hz noise bandwidth ???
                  >


                • Andy Talbot
                  You may be able to generate PI4; can WSJT decode it ? Andy G4JNT
                  Message 8 of 22 , Jul 1, 2012
                  • 0 Attachment
                    You may be able to generate PI4; can WSJT decode it ?
                    Andy
                    G4JNT


                     
                    On 1 July 2012 12:31, booz2m <booz2m@...> wrote:
                     

                    Hi all

                    The Next Generation Beacon project DDS multimode S/W can do Opera and WSPR plus JT4, JT65 and PI4 without changing S/W. The only thing that is needed is to enter the callsign, locator and frequency in plain text via the USB port.

                    http://rudius.net/oz2m/ngnb/dds.htm

                    The "only" drawback is the time width of the symbols has not yet been calibrated. I expect this to happen soon after the VUSHF Field Day next week-end. When done the S/W will be GNUed.

                    73
                    Bo, OZ2M


                  • Andy Talbot
                    It was already there, but in the optical comms section. http://www.g4jnt.com/OpticalComms/LF_DDS_Beacon_Source.pdf Or the original four frequency parallel
                    Message 9 of 22 , Jul 1, 2012
                    • 0 Attachment
                      It was already there, but in the optical comms section.
                      http://www.g4jnt.com/OpticalComms/LF_DDS_Beacon_Source.pdf

                      Or the original four frequency parallel programmed version which now
                      has the WSPR version too
                      http://www.g4jnt.com/OpticalComms/OPT_SC_Beacon.zip

                      Andy
                      'jnt


                      >>
                      >>
                      >>
                      >> However Andy have you published in public domain a PIC design for a stand
                      >> alone WRSP beacon?
                      >> I would like to have another go at one. Now that WSPR doesn't seem to be
                      >> going away from LF. I have built Johans one which was fine but I could not
                      >> set deviation to get the tones right it was very touchy. It looked good on
                      >> paper but I wasn't up to making it work.
                      >> Is there is anying else about for direct keying? If not I will dig it out
                      >> again but don't hold out much hope for success.
                      >>
                      >> 73 Pete FMT.
                    • booz2m
                      Hi Andy et al No, PI4 is not in WSJT. So that remains. Yet it does not affect the other modes, Opera, WSJT, JT65 and JT4. 73 Bo
                      Message 10 of 22 , Jul 1, 2012
                      • 0 Attachment
                        Hi Andy et al

                        No, PI4 is not in WSJT. So that remains. Yet it does not affect the other modes, Opera, WSJT, JT65 and JT4.

                        73
                        Bo
                      • graham787
                        Dyslexia rules KO ... The software is PathSim http://www.moetronix.com/ae4jy/pathsim.htm We can blame the boys over the pond for this, there was a set of
                        Message 11 of 22 , Jul 1, 2012
                        • 0 Attachment
                          Dyslexia rules KO ... The software is PathSim

                          http://www.moetronix.com/ae4jy/pathsim.htm

                          We can blame the boys over the pond for this, there was a set of tests run , using path-sim as the test engine , testing various data modes , the end result was JR re calibrated the s/n measurements to reflect the path-sim results , where prior to this , fldigi had been used as the yard stick

                          Note : The ROS data modes are also calibrated to path-sim , so if you are using Audio Tx drive and linear frequency translation to say 500K , then if OP is showing -27dB or better -26 , then a qso could be made using ros-mf-2 qso mode, round -20 ros-mf-7 , same speed as pks31 ,may be used

                          G..



                          --- In O_P_E_R_A_@yahoogroups.com, "m0fmt" <m0fmt@...> wrote:
                          >
                          > Hi All
                          >
                          > These "HAM" didgi systems are all based on your average BLACK BOX specs for your average ops as far as I can see, so could it be that they see 2.5kc/s as being the typical SSB audio pass band for these boxes therefore basing their specs around it?
                          > Ignoring elementary units to confuse comparison.
                          >
                          > However Andy have you published in public domain a PIC design for a stand alone WRSP beacon?
                          > I would like to have another go at one. Now that WSPR doesn't seem to be going away from LF. I have built Johans one which was fine but I could not set deviation to get the tones right it was very touchy. It looked good on paper but I wasn't up to making it work.
                          > Is there is anying else about for direct keying? If not I will dig it out again but don't hold out much hope for success.
                          >
                          > 73 Pete FMT.
                          >
                          >
                          >
                          > --- In O_P_E_R_A_@yahoogroups.com, Andy Talbot <andy.g4jnt@> wrote:
                          > >
                          > > So Opera S/N ratios are based on 625Hz noise bandwidth ??? Is this what
                          > > path-sim adopts, and if so why on earth such an arbitrary value.
                          > >
                          > > -36dB in WSPR's JT standard 2.5kHz means -30dB in 625Hz (your example
                          > > below)
                          > >
                          > > Why don't people just settle on normalised S/N in unity 1Hz bandwidth.
                          > > Where did this 'magic' 2.5kHz come from?
                          > >
                          > > Andy
                          > > www.g4jnt.com
                          > >
                          > >
                          > >
                          > > On 30 June 2012 19:34, graham787 <g0nbd@> wrote:
                          > >
                          > > > **
                          > > >
                          > > >
                          > > > Ok Roger , well we don't want to be seen 'Singing the Praises' :)
                          > > >
                          > > > Op s/n is calibrated against the 'sim-path' propagation simulation
                          > > > software. s/n readings are a little subjective, path-sim was felt to give a
                          > > > more realistic value .
                          > > >
                          > > > This puts the s/n readings 6 dB less than wspr / fldigi
                          > > > Op = -30dB wspr = -36dB
                          > > >
                          > > > Op s/n levels are the average of the transmission along the time line and
                          > > > -not- peak values
                          > > >
                          > > > With ook , Op has a obvious disadvantage compared to psk/fsk systems , but
                          > > > is not affected by Doppler spread, simplify the Tx path and with the data
                          > > > processing, will decode with up to 50% data loss (random distribution along
                          > > > the time line)
                          > > >
                          > > > Good luck with the tests
                          > > >
                          > > > G..
                          > > >
                          > > > --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@> wrote:
                          > > > >
                          > > > > Can someone please remind me of the equivalences between the various
                          > > > OPERA speeds and WSPR?
                          > > > >
                          > > > > If I am -30/-29dB on OP4 (my reports from ON6EO and G0KTN earlier) what
                          > > > sort of WSPR report would this correspond to?
                          > > > >
                          > > > > 73s
                          > > > > Roger G3XBM
                          > > > >
                          > > >
                          > > >
                          > > >
                          > >
                          >
                        • Andy Talbot
                          BUT WHAT WAS the Authoritive, offcial, stated, used noise bandwidth figure? If we know that, we can just renormalise to any other bandwidth. jnt
                          Message 12 of 22 , Jul 1, 2012
                          • 0 Attachment
                            BUT WHAT WAS the Authoritive, offcial, stated, used noise bandwidth figure?   If we know that, we can just renormalise to any other bandwidth.
                             
                            'jnt


                             
                            On 1 July 2012 13:25, graham787 <g0nbd@...> wrote:
                             

                            Dyslexia rules KO ... The software is PathSim

                            http://www.moetronix.com/ae4jy/pathsim.htm

                            We can blame the boys over the pond for this, there was a set of tests run , using path-sim as the test engine , testing various data modes , the end result was JR re calibrated the s/n measurements to reflect the path-sim results , where prior to this , fldigi had been used as the yard stick

                            Note : The ROS data modes are also calibrated to path-sim , so if you are using Audio Tx drive and linear frequency translation to say 500K , then if OP is showing -27dB or better -26 , then a qso could be made using ros-mf-2 qso mode, round -20 ros-mf-7 , same speed as pks31 ,may be used

                            G..



                            --- In O_P_E_R_A_@yahoogroups.com, "m0fmt" <m0fmt@...> wrote:
                            >
                            > Hi All
                            >
                            > These "HAM" didgi systems are all based on your average BLACK BOX specs for your average ops as far as I can see, so could it be that they see 2.5kc/s as being the typical SSB audio pass band for these boxes therefore basing their specs around it?
                            > Ignoring elementary units to confuse comparison.
                            >
                            > However Andy have you published in public domain a PIC design for a stand alone WRSP beacon?
                            > I would like to have another go at one. Now that WSPR doesn't seem to be going away from LF. I have built Johans one which was fine but I could not set deviation to get the tones right it was very touchy. It looked good on paper but I wasn't up to making it work.
                            > Is there is anying else about for direct keying? If not I will dig it out again but don't hold out much hope for success.
                            >
                            > 73 Pete FMT.
                            >
                            >
                            >
                            > --- In O_P_E_R_A_@yahoogroups.com, Andy Talbot <andy.g4jnt@> wrote:
                            > >
                            > > So Opera S/N ratios are based on 625Hz noise bandwidth ??? Is this what
                            > > path-sim adopts, and if so why on earth such an arbitrary value.
                            > >
                            > > -36dB in WSPR's JT standard 2.5kHz means -30dB in 625Hz (your example
                            > > below)
                            > >
                            > > Why don't people just settle on normalised S/N in unity 1Hz bandwidth.
                            > > Where did this 'magic' 2.5kHz come from?
                            > >
                            > > Andy
                            > > www.g4jnt.com
                            > >
                            > >
                            > >
                            > > On 30 June 2012 19:34, graham787 <g0nbd@> wrote:
                            > >
                            > > > **
                            > > >
                            > > >
                            > > > Ok Roger , well we don't want to be seen 'Singing the Praises' :)
                            > > >
                            > > > Op s/n is calibrated against the 'sim-path' propagation simulation
                            > > > software. s/n readings are a little subjective, path-sim was felt to give a
                            > > > more realistic value .
                            > > >
                            > > > This puts the s/n readings 6 dB less than wspr / fldigi
                            > > > Op = -30dB wspr = -36dB
                            > > >
                            > > > Op s/n levels are the average of the transmission along the time line and
                            > > > -not- peak values
                            > > >
                            > > > With ook , Op has a obvious disadvantage compared to psk/fsk systems , but
                            > > > is not affected by Doppler spread, simplify the Tx path and with the data
                            > > > processing, will decode with up to 50% data loss (random distribution along
                            > > > the time line)
                            > > >
                            > > > Good luck with the tests
                            > > >
                            > > > G..
                            > > >
                            > > > --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@> wrote:
                            > > > >
                            > > > > Can someone please remind me of the equivalences between the various
                            > > > OPERA speeds and WSPR?
                            > > > >
                            > > > > If I am -30/-29dB on OP4 (my reports from ON6EO and G0KTN earlier) what
                            > > > sort of WSPR report would this correspond to?
                            > > > >
                            > > > > 73s
                            > > > > Roger G3XBM
                            > > > >
                            > > >
                            > > >
                            > > >
                            > >
                            >


                          • G3XBM
                            Graham et al, Sorry if I am being very dense here. What I want to know is, for the same ERP (in my case around 20mW on 500kHz) which will be better WSPR or
                            Message 13 of 22 , Jul 1, 2012
                            • 0 Attachment
                              Graham et al,

                              Sorry if I am being very dense here.

                              What I want to know is, for the same ERP (in my case around 20mW on 500kHz) which will be better WSPR or various flavours of OPERA?

                              If I have a report of marginal WSPR copy at just -30dB S/N as measured in WSPR's normal bandwidth how much better will OP2,4,8 etc be than this? I cannot tell from what has been said so far.


                              Bottom line: will OPERA get me a report when WSPR will not?


                              73s
                              Roger G3XBM
                            • Andy Talbot
                              1) Make the assumption that both Opera and WSPR have equally good demodulators / decoders. There s almost certainly not a lot of difference. Similar
                              Message 14 of 22 , Jul 1, 2012
                              • 0 Attachment
                                1)   Make the assumption that both Opera and WSPR have equally good demodulators  / decoders.   There's almost certainly not a lot of difference.   Similar error correction; both use intelligent power detection algorithms.  
                                 
                                2)  Amplitude shift keying and FSK have identical performance based on MEAN signal power (bear in mind Opera is 50% duty cycle, so a given Tx unit is 3dB down on the same Tx used for WSPR)
                                 
                                3)  Therefore the ultimate sensitivity is determined purely by S/N ratio.
                                 
                                4)    WSPR uses a 1.5Hz noise bandwidth;, that is fixed.
                                 
                                5)  OP4 with 1s symbols is 1Hz bandwidth so is going to be 10.LOG(1.5 / 1) = 1.8dB better than WSPR.
                                6)  OP8 with 2s symbols is 0.5Hz so will be 3dB up on OP4, so 4.8dB better than WSPR .... and pro-rata.
                                 
                                7)  That is all based on MEAN power.   If you run 100 Watts whatever mode, OP8 will only be 1.8dB up on WSPR, and OP4 will be 1.2dB worse.

                                Ady
                                G4JNT

                                 
                                On 1 July 2012 14:47, G3XBM <rogerlapthorn@...> wrote:
                                 

                                Graham et al,

                                Sorry if I am being very dense here.

                                What I want to know is, for the same ERP (in my case around 20mW on 500kHz) which will be better WSPR or various flavours of OPERA?

                                If I have a report of marginal WSPR copy at just -30dB S/N as measured in WSPR's normal bandwidth how much better will OP2,4,8 etc be than this? I cannot tell from what has been said so far.

                                Bottom line: will OPERA get me a report when WSPR will not?

                                73s
                                Roger G3XBM


                              • G3XBM
                                Thanks Andy. That I can understand. 73s Roger
                                Message 15 of 22 , Jul 1, 2012
                                • 0 Attachment
                                  Thanks Andy. That I can understand.

                                  73s Roger
                                • Andy Talbot
                                  We can carry the comparison a bit further looking at data content. Take OP4 and WSPR as being roughly the same (1.5Hz vs. 1Hz, 1.8dB difference, call it the
                                  Message 16 of 22 , Jul 1, 2012
                                  • 0 Attachment
                                    We can carry the comparison a bit further looking at data content.
                                     
                                    Take OP4 and WSPR as being roughly the same (1.5Hz vs. 1Hz,  1.8dB difference, call it the same)
                                     
                                    OP4 takes four minuits to send a callsign only (or a part of a QSO detail ?)
                                    WSPR take 2 minutes to send a callsign,  4 digit locator and a power level.
                                     
                                    The big difference is in the modulation.
                                    WSPR consists of four tones.  These are in effect split equally between carrying clock information and carrying data.   So you 'could' consider it as two tone FSK with inherent clocking, for the full 2 minute period.  The other two tones carry clock only ad no information
                                     
                                    Opera uses Manchester coding where a bit consists of a 0/1 or 1/0 transistion.   This gives the exact 50% duty cycle and multiple transitions needed for reliable OOK detection, and allows the clock to be recovered for each symbol.    But it now means the total number of bits is half the number of symbols.    So the 239 (call it 240)  on-off symbols making up the message actually only carry 120 bits of information in the 4 minute period.   Enough for a heavily error corrected and fully compressed callsign
                                     
                                    WSPR takes 160 symbols but as described above this does mean 160 bits in the 2 minute period.
                                     
                                    So for roughly similar S/N performance, you have 120 bits in 4 minutes, containing just a callsign
                                    or  160 bits in 2 minutes with callsign, locator and power setting
                                     
                                    Because there is more original source information in the WSPR message this allows a sllightly lower overhead  when expanded for error correction,  hence 160 bits in WSPR carrying three items of information against 120 bits in Opera carrying just the one - the callsign.
                                     
                                    Andy
                                    G4JNT
                                     
                                     


                                     
                                    On 1 July 2012 15:11, Andy Talbot <andy.g4jnt@...> wrote:
                                    1)   Make the assumption that both Opera and WSPR have equally good demodulators  / decoders.   There's almost certainly not a lot of difference.   Similar error correction; both use intelligent power detection algorithms.  
                                     
                                    2)  Amplitude shift keying and FSK have identical performance based on MEAN signal power (bear in mind Opera is 50% duty cycle, so a given Tx unit is 3dB down on the same Tx used for WSPR)
                                     
                                    3)  Therefore the ultimate sensitivity is determined purely by S/N ratio.
                                     
                                    4)    WSPR uses a 1.5Hz noise bandwidth;, that is fixed.
                                     
                                    5)  OP4 with 1s symbols is 1Hz bandwidth so is going to be 10.LOG(1.5 / 1) = 1.8dB better than WSPR.
                                    6)  OP8 with 2s symbols is 0.5Hz so will be 3dB up on OP4, so 4.8dB better than WSPR .... and pro-rata.
                                     
                                    7)  That is all based on MEAN power.   If you run 100 Watts whatever mode, OP8 will only be 1.8dB up on WSPR, and OP4 will be 1.2dB worse.

                                    Ady
                                    G4JNT

                                     
                                    On 1 July 2012 14:47, G3XBM <rogerlapthorn@...> wrote:
                                     

                                    Graham et al,

                                    Sorry if I am being very dense here.

                                    What I want to know is, for the same ERP (in my case around 20mW on 500kHz) which will be better WSPR or various flavours of OPERA?

                                    If I have a report of marginal WSPR copy at just -30dB S/N as measured in WSPR's normal bandwidth how much better will OP2,4,8 etc be than this? I cannot tell from what has been said so far.

                                    Bottom line: will OPERA get me a report when WSPR will not?

                                    73s
                                    Roger G3XBM



                                  • graham787
                                    Interesting ... I think the final situation as reached some time ago, was OP2 gave similar performance to wspr I think wspr fails round the -30 dB
                                    Message 17 of 22 , Jul 1, 2012
                                    • 0 Attachment
                                      Interesting ...

                                      I think the final situation as reached some time ago, was OP2 gave similar performance to wspr

                                      I think wspr fails round the -30 dB level as measured by wspr , which is similar to OP2 , ros-mf-2 data mode has roughly the same decode fail point as wspr

                                      In use OP4 will reduce the decode by some 3 or more dB's past this point -29dB expected -30 -32 is sometimes seen .. so you should achieve decodes over the same circuit with -3db less power wspr/op4.

                                      The min decode level drops by -3dB for each X2 in time , as noted, double the time , needs only half the power , The audio OP4H has reached -49 dB -55dB wspr rating.

                                      It must be remembered that only 50% of the data needs to be captured , the loss can be randomly distributed along the time line .

                                      Of course , with psk/fsk another 6 dB may be achieved, but this is a system that is being sent via the most basic telegraphy system invented , (pre dating Mr Morse by 100's of years) with the most simple of Tx equipment and has given some interesting insights into low band propagation , especially 136 ?

                                      73 -G..






                                      --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@...> wrote:
                                      >
                                      > Graham et al,
                                      >
                                      > Sorry if I am being very dense here.
                                      >
                                      > What I want to know is, for the same ERP (in my case around 20mW on 500kHz) which will be better WSPR or various flavours of OPERA?
                                      >
                                      > If I have a report of marginal WSPR copy at just -30dB S/N as measured in WSPR's normal bandwidth how much better will OP2,4,8 etc be than this? I cannot tell from what has been said so far.
                                      >
                                      >
                                      > Bottom line: will OPERA get me a report when WSPR will not?
                                      >
                                      >
                                      > 73s
                                      > Roger G3XBM
                                      >
                                    • graham787
                                      I think this is like comparing the football teams of Italy and Spain, both are football teams , both play foot ball , both do well sometimes ,
                                      Message 18 of 22 , Jul 1, 2012
                                      • 0 Attachment
                                        I think this is like comparing the football teams of Italy and Spain, both are football teams , both play foot ball , both do well sometimes , sometimes one or the other not so well..

                                        EG..Opera is imune to doppler spread so functions in near vertical conditions , where narrow shift systems could have a problem

                                        G..

                                        --- In O_P_E_R_A_@yahoogroups.com, "graham787" <g0nbd@...> wrote:
                                        >
                                        > Interesting ...
                                        >
                                        > I think the final situation as reached some time ago, was OP2 gave similar performance to wspr
                                        >
                                        > I think wspr fails round the -30 dB level as measured by wspr , which is similar to OP2 , ros-mf-2 data mode has roughly the same decode fail point as wspr
                                        >
                                        > In use OP4 will reduce the decode by some 3 or more dB's past this point -29dB expected -30 -32 is sometimes seen .. so you should achieve decodes over the same circuit with -3db less power wspr/op4.
                                        >
                                        > The min decode level drops by -3dB for each X2 in time , as noted, double the time , needs only half the power , The audio OP4H has reached -49 dB -55dB wspr rating.
                                        >
                                        > It must be remembered that only 50% of the data needs to be captured , the loss can be randomly distributed along the time line .
                                        >
                                        > Of course , with psk/fsk another 6 dB may be achieved, but this is a system that is being sent via the most basic telegraphy system invented , (pre dating Mr Morse by 100's of years) with the most simple of Tx equipment and has given some interesting insights into low band propagation , especially 136 ?
                                        >
                                        > 73 -G..
                                        >
                                        >
                                        >
                                        >
                                        >
                                        >
                                        > --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@> wrote:
                                        > >
                                        > > Graham et al,
                                        > >
                                        > > Sorry if I am being very dense here.
                                        > >
                                        > > What I want to know is, for the same ERP (in my case around 20mW on 500kHz) which will be better WSPR or various flavours of OPERA?
                                        > >
                                        > > If I have a report of marginal WSPR copy at just -30dB S/N as measured in WSPR's normal bandwidth how much better will OP2,4,8 etc be than this? I cannot tell from what has been said so far.
                                        > >
                                        > >
                                        > > Bottom line: will OPERA get me a report when WSPR will not?
                                        > >
                                        > >
                                        > > 73s
                                        > > Roger G3XBM
                                        > >
                                        >
                                      • Andy Talbot
                                        football remind me what that is. The little ball being knocked into a hole, whacked between two people or the one with the large egg :-) That s my
                                        Message 19 of 22 , Jul 2, 2012
                                        • 0 Attachment
                                          "football"  remind me what that is.    The little ball being knocked into a hole, whacked between two people or the one with the large egg :-)   That's my level of interest in the "opium for the masses"
                                           
                                          But more seriously -
                                          One advantage of Opera is the ability to chose from a variety of symbol rates to match band conditions.   It is a pity thay the choices are tied to band of operation, but can always be frigged,
                                           
                                          In contrast, although WSPR is fixed symbol rate, you could instead use JT65 or JT4 modulation whh offer faster data with a choice of frequency spans to match teh propgation medium.   JT65 has three spans, although only the slowest is suited to LF.   JT4 has the widest set selectable, from 4.375Hz tone spacing up to 315Hz spacing.  Its coding it is identical to WSPR, but carries less information - 13 characters or a couple of coded callsigns and a report or a few other QSO typecombinations.
                                           
                                          The narrowest JT4a takes about 18Hz overall bandwidth with its 4 tones at 4.375 symbol rate and spacing, taking 48 seconds to send a message using a timed protocol like WSPR.
                                           
                                          The narrowest JT65a is a lot wider at 175 Hz overall span, again 48 seconds to carry the same information content.
                                           
                                          So these two could take over when WSPR can't cope with the Doppler spread.
                                           
                                          Andy
                                          G4JNT
                                           

                                           
                                          On 1 July 2012 21:41, graham787 <g0nbd@...> wrote:
                                           

                                          I think this is like comparing the football teams of Italy and Spain, both are football teams , both play foot ball , both do well sometimes , sometimes one or the other not so well..

                                          EG..Opera is imune to doppler spread so functions in near vertical conditions , where narrow shift systems could have a problem

                                          G..

                                          --- In O_P_E_R_A_@yahoogroups.com, "graham787" <g0nbd@...> wrote:
                                          >
                                          > Interesting ...
                                          >
                                          > I think the final situation as reached some time ago, was OP2 gave similar performance to wspr
                                          >
                                          > I think wspr fails round the -30 dB level as measured by wspr , which is similar to OP2 , ros-mf-2 data mode has roughly the same decode fail point as wspr
                                          >
                                          > In use OP4 will reduce the decode by some 3 or more dB's past this point -29dB expected -30 -32 is sometimes seen .. so you should achieve decodes over the same circuit with -3db less power wspr/op4.
                                          >
                                          > The min decode level drops by -3dB for each X2 in time , as noted, double the time , needs only half the power , The audio OP4H has reached -49 dB -55dB wspr rating.
                                          >
                                          > It must be remembered that only 50% of the data needs to be captured , the loss can be randomly distributed along the time line .
                                          >
                                          > Of course , with psk/fsk another 6 dB may be achieved, but this is a system that is being sent via the most basic telegraphy system invented , (pre dating Mr Morse by 100's of years) with the most simple of Tx equipment and has given some interesting insights into low band propagation , especially 136 ?
                                          >
                                          > 73 -G..
                                          >
                                          >
                                          >
                                          >
                                          >
                                          >
                                          > --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@> wrote:
                                          > >
                                          > > Graham et al,
                                          > >
                                          > > Sorry if I am being very dense here.
                                          > >
                                          > > What I want to know is, for the same ERP (in my case around 20mW on 500kHz) which will be better WSPR or various flavours of OPERA?
                                          > >
                                          > > If I have a report of marginal WSPR copy at just -30dB S/N as measured in WSPR's normal bandwidth how much better will OP2,4,8 etc be than this? I cannot tell from what has been said so far.
                                          > >
                                          > >
                                          > > Bottom line: will OPERA get me a report when WSPR will not?
                                          > >
                                          > >
                                          > > 73s
                                          > > Roger G3XBM
                                          > >
                                          >


                                        • graham787
                                          I think I ll go with the the one with the large egg this time Maybe, maybe, not as our friends from the aquatic north east would, say, these good words
                                          Message 20 of 22 , Jul 2, 2012
                                          • 0 Attachment
                                            I think I'll go with the ''the one with the large egg'' this time

                                            Maybe, maybe, not as our friends from the aquatic north east would, say, these good words are all centered on spot measurements frozen in time, during the initial testing of the mode, problems with 'lab' measurement's where noted.

                                            Missing is the statistical analysis of a real radio path and the influence this has on the over all system performance, as noted the mode is modified by the band..

                                            One of the attributes that enhances the system as a reporting beacon , is that, in path confusion, gain is realized by the ability to sum 'random' decodes along the time line, self clocking as noted.

                                            This method of testing in the final analysis may be more appropriate

                                            So far we have two answers , leaves only the third ..

                                            http://www.youtube.com/watch?v=cV0tCphFMr8

                                            G..



                                            --- In O_P_E_R_A_@yahoogroups.com, Andy Talbot <andy.g4jnt@...> wrote:
                                            >
                                            > "football" remind me what that is. The little ball being knocked into a
                                            > hole, whacked between two people or the one with the large egg :-) That's
                                            > my level of interest in the "opium for the masses"
                                            >
                                            > But more seriously -
                                            > One advantage of Opera is the ability to chose from a variety of symbol
                                            > rates to match band conditions. It is a pity thay the choices are tied to
                                            > band of operation, but can always be frigged,
                                            >
                                            > In contrast, although WSPR is fixed symbol rate, you could instead use JT65
                                            > or JT4 modulation whh offer faster data with a choice of frequency spans to
                                            > match teh propgation medium. JT65 has three spans, although only the
                                            > slowest is suited to LF. JT4 has the widest set selectable, from 4.375Hz
                                            > tone spacing up to 315Hz spacing. Its coding it is identical to WSPR, but
                                            > carries less information - 13 characters or a couple of coded callsigns and
                                            > a report or a few other QSO typecombinations.
                                            >
                                            > The narrowest JT4a takes about 18Hz overall bandwidth with its 4 tones at
                                            > 4.375 symbol rate and spacing, taking 48 seconds to send a message using
                                            > a timed protocol like WSPR.
                                            >
                                            > The narrowest JT65a is a lot wider at 175 Hz overall span, again 48 seconds
                                            > to carry the same information content.
                                            >
                                            > So these two could take over when WSPR can't cope with the Doppler spread.
                                            >
                                            > Andy
                                            > G4JNT
                                            >
                                            >
                                            >
                                            > On 1 July 2012 21:41, graham787 <g0nbd@...> wrote:
                                            >
                                            > > **
                                            > >
                                            > >
                                            > > I think this is like comparing the football teams of Italy and Spain, both
                                            > > are football teams , both play foot ball , both do well sometimes ,
                                            > > sometimes one or the other not so well..
                                            > >
                                            > > EG..Opera is imune to doppler spread so functions in near vertical
                                            > > conditions , where narrow shift systems could have a problem
                                            > >
                                            > > G..
                                            > >
                                            > > --- In O_P_E_R_A_@yahoogroups.com, "graham787" <g0nbd@> wrote:
                                            > > >
                                            > > > Interesting ...
                                            > > >
                                            > > > I think the final situation as reached some time ago, was OP2 gave
                                            > > similar performance to wspr
                                            > > >
                                            > > > I think wspr fails round the -30 dB level as measured by wspr , which is
                                            > > similar to OP2 , ros-mf-2 data mode has roughly the same decode fail point
                                            > > as wspr
                                            > > >
                                            > > > In use OP4 will reduce the decode by some 3 or more dB's past this point
                                            > > -29dB expected -30 -32 is sometimes seen .. so you should achieve decodes
                                            > > over the same circuit with -3db less power wspr/op4.
                                            > > >
                                            > > > The min decode level drops by -3dB for each X2 in time , as noted,
                                            > > double the time , needs only half the power , The audio OP4H has reached
                                            > > -49 dB -55dB wspr rating.
                                            > > >
                                            > > > It must be remembered that only 50% of the data needs to be captured ,
                                            > > the loss can be randomly distributed along the time line .
                                            > > >
                                            > > > Of course , with psk/fsk another 6 dB may be achieved, but this is a
                                            > > system that is being sent via the most basic telegraphy system invented ,
                                            > > (pre dating Mr Morse by 100's of years) with the most simple of Tx
                                            > > equipment and has given some interesting insights into low band propagation
                                            > > , especially 136 ?
                                            > > >
                                            > > > 73 -G..
                                            > > >
                                            > > >
                                            > > >
                                            > > >
                                            > > >
                                            > > >
                                            > > > --- In O_P_E_R_A_@yahoogroups.com, "G3XBM" <rogerlapthorn@> wrote:
                                            > > > >
                                            > > > > Graham et al,
                                            > > > >
                                            > > > > Sorry if I am being very dense here.
                                            > > > >
                                            > > > > What I want to know is, for the same ERP (in my case around 20mW on
                                            > > 500kHz) which will be better WSPR or various flavours of OPERA?
                                            > > > >
                                            > > > > If I have a report of marginal WSPR copy at just -30dB S/N as measured
                                            > > in WSPR's normal bandwidth how much better will OP2,4,8 etc be than this? I
                                            > > cannot tell from what has been said so far.
                                            > > > >
                                            > > > >
                                            > > > > Bottom line: will OPERA get me a report when WSPR will not?
                                            > > > >
                                            > > > >
                                            > > > > 73s
                                            > > > > Roger G3XBM
                                            > > > >
                                            > > >
                                            > >
                                            > >
                                            > >
                                            >
                                          Your message has been successfully submitted and would be delivered to recipients shortly.