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

non linear SSB PA, [WOT]

Expand Messages
  • Art
    Greetings all, I ve recently become interested in using non linear amplifiers to amplify SSB. Obviously, this is a problem, we all know an amplifier for SSB
    Message 1 of 8 , Mar 1, 2007
      Greetings all,

      I've recently become interested in using non linear amplifiers to
      amplify SSB. Obviously, this is a problem, we all know an amplifier
      for SSB has to be linear, or class AB1.

      Even if we bias an AB1 amp properly, some (all) power amps have non
      linear responses. Even if we put the cleanest audio on the planet
      into them, they don't work well. We refer to this problem with many
      different technical terms. But, the problem is non-linearity in the
      PA itself. Whenever the output of a PA is distorted, it uses more
      spectrum than it should and causes interference well outside it's
      expected bandwidth.

      The short version is that the amplifiers output doesn't follow the input.

      But, what about using a sound card to correct this issue? If we feed
      soundcard audio to a PA and measure (or characterize) it's input vs
      output, can we then use that data to modify the audio supplied to the
      PA using the soundcard? In this manner, we would be
      'predistorting' the input audio intentionally, but doing it in such
      a manner as to make the PA respond with proper output.

      I know this might sound radical, and no one would want run a computer
      just for the purpose of doing this. But, many of us already use the
      computer soundcard to generate transmit audio anyway, so the real
      cost of implementing this sort of change means some extra software
      and some additional processor utilization.

      What we might get in return might be a way to use a class B (or C)
      amp as an SSB PA, and it would sound clean and not splatter.
      Depending on how well it works, we might even be able to use a class
      D or E PA for SSB, efficiencies of 90 percent are possible in class D/E.

      I wonder if anyone else has considered this? I sure would like to be
      able to use my 4CX250 amp on ssb without having my neighbors complain
      that my signal was 'wide'.

      Regards,

      Art
    • Dan Andersson
      On Thursday 01 March 2007 21:01:39 Art wrote: Greetings all, I ve recently become interested in using non linear amplifiers to amplify SSB. Obviously,
      Message 2 of 8 , Mar 1, 2007
        On Thursday 01 March 2007 21:01:39 Art wrote:
        > Greetings all,
        >
        > I've recently become interested in using non linear amplifiers to
        > amplify SSB. Obviously, this is a problem, we all know an amplifier
        > for SSB has to be linear, or class AB1.
        >
        > Even if we bias an AB1 amp properly, some (all) power amps have non
        > linear responses. Even if we put the cleanest audio on the planet
        > into them, they don't work well. We refer to this problem with many
        > different technical terms. But, the problem is non-linearity in the
        > PA itself. Whenever the output of a PA is distorted, it uses more
        > spectrum than it should and causes interference well outside it's
        > expected bandwidth.
        >
        > The short version is that the amplifiers output doesn't follow the input.
        >
        > But, what about using a sound card to correct this issue? If we feed
        > soundcard audio to a PA and measure (or characterize) it's input vs
        > output, can we then use that data to modify the audio supplied to the
        > PA using the soundcard? In this manner, we would be
        > 'predistorting' the input audio intentionally, but doing it in such
        > a manner as to make the PA respond with proper output.
        >
        > I know this might sound radical, and no one would want run a computer
        > just for the purpose of doing this. But, many of us already use the
        > computer soundcard to generate transmit audio anyway, so the real
        > cost of implementing this sort of change means some extra software
        > and some additional processor utilization.
        >
        > What we might get in return might be a way to use a class B (or C)
        > amp as an SSB PA, and it would sound clean and not splatter.
        > Depending on how well it works, we might even be able to use a class
        > D or E PA for SSB, efficiencies of 90 percent are possible in class D/E.
        >
        > I wonder if anyone else has considered this? I sure would like to be
        > able to use my 4CX250 amp on ssb without having my neighbors complain
        > that my signal was 'wide'.
        >
        > Regards,
        >
        > Art


        Art,

        Of course it's going to work.

        We do this for microwave applications where we have a wide bandwidth demand on
        the aerials. we compensate based on the known "weak points" and we are also
        actively compensating other unlinearities.

        Regarding the PA unlinearity, this is one of the reasons we use DSP's more and
        more frequently.

        However! You can only compensate for known unlinearities. If your PA changes
        behaviours depending on load and SWR, the you need to add a control mechanism
        and a monitoring receiver to linearise your output.

        It's also a bit easier to compensate for this with SDR as the signal can be
        treated before it reaches the PA. But again, it's the signal at the end of
        the PA that is the main problem.

        But brought to a conclusion, yes, it's done today on commercial and mil
        radios. It takes a lot of umpf to do for us hams but why not...


        //Dan



        --
        Dan Andersson, M0DFI
        dan@...
      • DSNman@comcast.net
        Test please ignore! -- Loren Moline WA7SKT ... From: Dan Andersson ... Art, Of course it s going to work. We do this for microwave
        Message 3 of 8 , Mar 1, 2007
          Test please ignore!

          --
          Loren Moline
          WA7SKT

          -------------- Original message ----------------------
          From: Dan Andersson <dan@...>
          > On Thursday 01 March 2007 21:01:39 Art wrote:
          > > Greetings all,
          > >
          > > I've recently become interested in using non linear amplifiers to
          > > amplify SSB. Obviously, this is a problem, we all know an amplifier
          > > for SSB has to be linear, or class AB1.
          > >
          > > Even if we bias an AB1 amp properly, some (all) power amps have non
          > > linear responses. Even if we put the cleanest audio on the planet
          > > into them, they don't work well. We refer to this problem with many
          > > different technical terms. But, the problem is non-linearity in the
          > > PA itself. Whenever the output of a PA is distorted, it uses more
          > > spectrum than it should and causes interference well outside it's
          > > expected bandwidth.
          > >
          > > The short version is that the amplifiers output doesn't follow the input.
          > >
          > > But, what about using a sound card to correct this issue? If we feed
          > > soundcard audio to a PA and measure (or characterize) it's input vs
          > > output, can we then use that data to modify the audio supplied to the
          > > PA using the soundcard? In this manner, we would be
          > > 'predistorting' the input audio intentionally, but doing it in such
          > > a manner as to make the PA respond with proper output.
          > >
          > > I know this might sound radical, and no one would want run a computer
          > > just for the purpose of doing this. But, many of us already use the
          > > computer soundcard to generate transmit audio anyway, so the real
          > > cost of implementing this sort of change means some extra software
          > > and some additional processor utilization.
          > >
          > > What we might get in return might be a way to use a class B (or C)
          > > amp as an SSB PA, and it would sound clean and not splatter.
          > > Depending on how well it works, we might even be able to use a class
          > > D or E PA for SSB, efficiencies of 90 percent are possible in class D/E.
          > >
          > > I wonder if anyone else has considered this? I sure would like to be
          > > able to use my 4CX250 amp on ssb without having my neighbors complain
          > > that my signal was 'wide'.
          > >
          > > Regards,
          > >
          > > Art
          >
          >
          > Art,
          >
          > Of course it's going to work.
          >
          > We do this for microwave applications where we have a wide bandwidth demand on
          > the aerials. we compensate based on the known "weak points" and we are also
          > actively compensating other unlinearities.
          >
          > Regarding the PA unlinearity, this is one of the reasons we use DSP's more and
          > more frequently.
          >
          > However! You can only compensate for known unlinearities. If your PA changes
          > behaviours depending on load and SWR, the you need to add a control mechanism
          > and a monitoring receiver to linearise your output.
          >
          > It's also a bit easier to compensate for this with SDR as the signal can be
          > treated before it reaches the PA. But again, it's the signal at the end of
          > the PA that is the main problem.
          >
          > But brought to a conclusion, yes, it's done today on commercial and mil
          > radios. It takes a lot of umpf to do for us hams but why not...
          >
          >
          > //Dan
          >
          >
          >
          > --
          > Dan Andersson, M0DFI
          > dan@...
          >
        • Phil Harman
          Hi Art, Your theory is sound but there will be a problem feeding the signal envelope back into the sound card due to the inherent delays in the signal
          Message 4 of 8 , Mar 1, 2007
            Hi Art,

            Your theory is sound but there will be a problem feeding the signal envelope
            back into the sound card due to the inherent delays in the signal
            processing.

            Over on the HPSDR project we are designing an exciter that has the features
            you defined. Rather than using a soundcard/PC we will do the processing in
            an FPGA. Have a look at the block diagram of Penelope at

            < http://hpsdr.org/wiki/index.php?title=PENELOPE >

            As a 1/2w exciter Penelope is intended to drive either a conventional liner
            PA or a non linear PA (Class D/E/F) by providing the relevant phase
            modulated RF and envelope signals.

            73's Phil.... VK6APH




            ----- Original Message -----
            From: "Art" <KY1K@...>
            To: <softrock40@yahoogroups.com>
            Sent: Friday, March 02, 2007 6:01 AM
            Subject: [softrock40] non linear SSB PA, [WOT]


            > Greetings all,
            >
            > I've recently become interested in using non linear amplifiers to
            > amplify SSB. Obviously, this is a problem, we all know an amplifier
            > for SSB has to be linear, or class AB1.
            >
            > Even if we bias an AB1 amp properly, some (all) power amps have non
            > linear responses. Even if we put the cleanest audio on the planet
            > into them, they don't work well. We refer to this problem with many
            > different technical terms. But, the problem is non-linearity in the
            > PA itself. Whenever the output of a PA is distorted, it uses more
            > spectrum than it should and causes interference well outside it's
            > expected bandwidth.
            >
            > The short version is that the amplifiers output doesn't follow the input.
            >
            > But, what about using a sound card to correct this issue? If we feed
            > soundcard audio to a PA and measure (or characterize) it's input vs
            > output, can we then use that data to modify the audio supplied to the
            > PA using the soundcard? In this manner, we would be
            > 'predistorting' the input audio intentionally, but doing it in such
            > a manner as to make the PA respond with proper output.
            >
            > I know this might sound radical, and no one would want run a computer
            > just for the purpose of doing this. But, many of us already use the
            > computer soundcard to generate transmit audio anyway, so the real
            > cost of implementing this sort of change means some extra software
            > and some additional processor utilization.
            >
            > What we might get in return might be a way to use a class B (or C)
            > amp as an SSB PA, and it would sound clean and not splatter.
            > Depending on how well it works, we might even be able to use a class
            > D or E PA for SSB, efficiencies of 90 percent are possible in class D/E.
            >
            > I wonder if anyone else has considered this? I sure would like to be
            > able to use my 4CX250 amp on ssb without having my neighbors complain
            > that my signal was 'wide'.
            >
            > Regards,
            >
            > Art
            >
            >
            >
            >
            >
            > Yahoo! Groups Links
            >
            >
            >
            >
            >
            >
            > --
            > No virus found in this incoming message.
            > Checked by AVG Free Edition.
            > Version: 7.5.446 / Virus Database: 268.18.5/706 - Release Date: 2/28/2007
            > 4:09 PM
            >
            >
          • tom_iphi
            You can even buy an SSB PA which works in class E based on this idea: http://www.sgcworld.com/500watt.html Linear transponders in satellites also use this
            Message 5 of 8 , Mar 1, 2007
              You can even buy an SSB PA which works in class E based on this idea:
              http://www.sgcworld.com/500watt.html

              Linear transponders in satellites also use this method to conserve
              energy.

              73 Tom DG8SAQ

              --- In softrock40@yahoogroups.com, Dan Andersson <dan@...> wrote:
              >
              > On Thursday 01 March 2007 21:01:39 Art wrote:
              > > Greetings all,
              > >
              > > I've recently become interested in using non linear amplifiers to
              > > amplify SSB. Obviously, this is a problem, we all know an
              amplifier
              > > for SSB has to be linear, or class AB1.
              > >
              > > Even if we bias an AB1 amp properly, some (all) power amps have
              non
              > > linear responses. Even if we put the cleanest audio on the planet
              > > into them, they don't work well. We refer to this problem with
              many
              > > different technical terms. But, the problem is non-linearity in
              the
              > > PA itself. Whenever the output of a PA is distorted, it uses more
              > > spectrum than it should and causes interference well outside it's
              > > expected bandwidth.
              > >
              > > The short version is that the amplifiers output doesn't follow
              the input.
              > >
              > > But, what about using a sound card to correct this issue? If we
              feed
              > > soundcard audio to a PA and measure (or characterize) it's input
              vs
              > > output, can we then use that data to modify the audio supplied to
              the
              > > PA using the soundcard? In this manner, we would be
              > > 'predistorting' the input audio intentionally, but doing it in
              such
              > > a manner as to make the PA respond with proper output.
              > >
              > > I know this might sound radical, and no one would want run a
              computer
              > > just for the purpose of doing this. But, many of us already use
              the
              > > computer soundcard to generate transmit audio anyway, so the real
              > > cost of implementing this sort of change means some extra software
              > > and some additional processor utilization.
              > >
              > > What we might get in return might be a way to use a class B (or C)
              > > amp as an SSB PA, and it would sound clean and not splatter.
              > > Depending on how well it works, we might even be able to use a
              class
              > > D or E PA for SSB, efficiencies of 90 percent are possible in
              class D/E.
              > >
              > > I wonder if anyone else has considered this? I sure would like to
              be
              > > able to use my 4CX250 amp on ssb without having my neighbors
              complain
              > > that my signal was 'wide'.
              > >
              > > Regards,
              > >
              > > Art
              >
              >
              > Art,
              >
              > Of course it's going to work.
              >
              > We do this for microwave applications where we have a wide
              bandwidth demand on
              > the aerials. we compensate based on the known "weak points" and we
              are also
              > actively compensating other unlinearities.
              >
              > Regarding the PA unlinearity, this is one of the reasons we use
              DSP's more and
              > more frequently.
              >
              > However! You can only compensate for known unlinearities. If your
              PA changes
              > behaviours depending on load and SWR, the you need to add a control
              mechanism
              > and a monitoring receiver to linearise your output.
              >
              > It's also a bit easier to compensate for this with SDR as the
              signal can be
              > treated before it reaches the PA. But again, it's the signal at the
              end of
              > the PA that is the main problem.
              >
              > But brought to a conclusion, yes, it's done today on commercial and
              mil
              > radios. It takes a lot of umpf to do for us hams but why not...
              >
              >
              > //Dan
              >
              >
              >
              > --
              > Dan Andersson, M0DFI
              > dan@...
              >
            • RAY
              ... From: tom_iphi You can even buy an SSB PA which works in class E based on this idea: http://www.sgcworld.com/500watt.html
              Message 6 of 8 , Mar 1, 2007
                ----- Original Message -----
                From: tom_iphi
                You can even buy an SSB PA which works in class E based on this idea:
                http://www.sgcworld.com/500watt.html
                *************************************************
                The company has a identified what they call a mechanical weakness in
                connectors and
                will not deliver the product till this is resolved. They are also working
                on a self-contained 20M single band only.

                It should be a great product if they get the mechanical thing resolved.

                Please don't shoot the messenger. I want this as bad as anyone else

                73,Ray, N0FY
              • cesco12342000
                ... The solution is EER (Envelope Elimination and Restoration). One problem with soundcard realization of EER is that the Envelope must be a DC coupled signal.
                Message 7 of 8 , Mar 2, 2007
                  > What we might get in return might be a way to use a class B (or C)
                  > amp as an SSB PA, and it would sound clean and not splatter.

                  The solution is EER (Envelope Elimination and Restoration).

                  One problem with soundcard realization of EER is that the Envelope must
                  be a DC coupled signal. Normal soundcards wont do that. You will need a
                  "patched" soundcard, or a dedicated hardware, or...

                  A solution is to modulate the Envelope signal onto an AM carrier, send
                  it out of the soundcard and demodulate it there. The AM frequency would
                  be 24khz, the modulation band going from DC to more than 3khz.

                  For the softrock this means two audio channels IQ phase signal, + one
                  audio channel for the E signal. Since we all got those 5.1 soundcards by
                  now, it IS possible to do !
                • richard
                  ... Yes it possible but its asking for trouble in Ham land, I ve worked on amplifiers that reduce distortion, by prediction the effect a non-linear PA
                  Message 8 of 8 , Mar 2, 2007
                    Art wrote:
                    >
                    > Greetings all,
                    >
                    > I've recently become interested in using non linear amplifiers to
                    > amplify SSB. Obviously, this is a problem, we all know an amplifier
                    > for SSB has to be linear, or class AB1.
                    >
                    > Even if we bias an AB1 amp properly, some (all) power amps have non
                    > linear responses. Even if we put the cleanest audio on the planet
                    > into them, they don't work well. We refer to this problem with many
                    > different technical terms. But, the problem is non-linearity in the
                    > PA itself. Whenever the output of a PA is distorted, it uses more
                    > spectrum than it should and causes interference well outside it's
                    > expected bandwidth.
                    >
                    > The short version is that the amplifiers output doesn't follow the input.
                    >

















                    <snip>

                    Yes it possible but its asking for trouble in Ham land, I've worked on
                    amplifiers that
                    reduce distortion, by prediction the effect a non-linear PA has,
                    intermod, transients etc,
                    and correcting phase and amplitude before it fed to the PA.
                    Its not a technology I feel would be easy to use on Ham applications,
                    Using it for edge and similar
                    modes maybe, but SSB is push the limits hard. you can predict and alter
                    a signal when you know the exact pattern
                    ,speech is much harder


                    Richard
                  Your message has been successfully submitted and would be delivered to recipients shortly.