• current-fed push-pull dc/dc converter

    From john larkin @21:1/5 to All on Fri Jul 19 13:35:37 2024
    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=https%
    3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From piglet@21:1/5 to All on Fri Jul 19 21:43:00 2024
    john larkin <jlarkin_highland_tech> wrote:

    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=
    https%3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.



    What do you find nasty about that topology, it works well , or does it
    remind you too much of Baxandall?


    --
    piglet

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From bitrex@21:1/5 to john larkin on Fri Jul 19 17:28:20 2024
    On 7/19/2024 4:35 PM, john larkin wrote:

    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=
    https%3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.



    With respect to magnetics it follows the "first rule of government
    spending": why have one when you can have two, at twice the price."

    The area product of the choke is on the same order as the area product
    of the transformer. It's annoying enough when you have to use one
    magnetics, much less two...and there are other isolated topologies e.g.
    Cuk that offer a constant input current.

    But as they remark in the abstract it's nice that they give a thorough
    design procedure though and doesn't use any exotic parts, for some
    applications particularly perhaps in some underprivileged areas of the
    world, the "best" converter is the one you can build with what you have available, the tolerances for this one don't look too critical, unlike
    the Cuk it doesn't look like it need a PhD to design a loop for.

    --
    This email has been checked for viruses by Avast antivirus software. www.avast.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From john larkin @21:1/5 to erichpwagner@hotmail.com on Fri Jul 19 20:11:04 2024
    On Fri, 19 Jul 2024 21:43:00 -0000 (UTC), piglet
    <erichpwagner@hotmail.com> wrote:

    john larkin <jlarkin_highland_tech> wrote:

    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=
    https%3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.



    What do you find nasty about that topology, it works well , or does it
    remind you too much of Baxandall?

    When a fet turns off, it will spike (to kilovolts in simulation) and
    ring like a churchbell. Now we need snubbers.

    Worse, if the fet turnons are slightly asymmetric, you get DC in the transformer primary and potential saturation. That can cause flux
    walking, The Devil's Staircase.

    Only half of the primary is used at a time, which uses the copper inefficiently.

    It needs another inductor. The waveform at the transformer CT is
    ghastly.

    Certain Parties want to replace my design with this horror. I'm using
    a 4-fet full bridge driving both ends of my primary, from the 48v bulk
    supply. Nothing goes hi-z, the copper is fully used, and I can add a
    series cap to keep DC out of the transformer primary.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From bitrex@21:1/5 to john larkin on Fri Jul 19 23:35:36 2024
    On 7/19/2024 11:11 PM, john larkin wrote:
    \
    What do you find nasty about that topology, it works well , or does it
    remind you too much of Baxandall?

    When a fet turns off, it will spike (to kilovolts in simulation) and
    ring like a churchbell. Now we need snubbers.

    Worse, if the fet turnons are slightly asymmetric, you get DC in the transformer primary and potential saturation. That can cause flux
    walking, The Devil's Staircase.

    Only half of the primary is used at a time, which uses the copper inefficiently.

    It needs another inductor. The waveform at the transformer CT is
    ghastly.

    Certain Parties want to replace my design with this horror. I'm using
    a 4-fet full bridge driving both ends of my primary, from the 48v bulk supply. Nothing goes hi-z, the copper is fully used, and I can add a
    series cap to keep DC out of the transformer primary.


    It says in the text the transistors are supposed to be turned on
    simultaneously for a fraction of the cycle.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Bill Sloman@21:1/5 to piglet on Sat Jul 20 13:25:31 2024
    On 20/07/2024 7:43 am, piglet wrote:
    john larkin <jlarkin_highland_tech> wrote:

    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=
    https%3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.

    What do you find nasty about that topology, it works well , or does it
    remind you too much of Baxandall?

    It does look exactly like Baxandall's 1959 configuration. Admittedly
    MOSFets hadn't been invented back then.

    http://sophia-electronica.com/Baxandall1959JM.pdf

    but they work rather better than bipolar transistors in the circuit,
    though they are a bit trickier to bias.

    Somebody has re-invented the wheel.

    --
    Bill Sloman, Sydney



    --
    This email has been checked for viruses by Norton antivirus software. www.norton.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Bill Sloman@21:1/5 to bitrex on Sat Jul 20 13:40:34 2024
    On 20/07/2024 7:28 am, bitrex wrote:
    On 7/19/2024 4:35 PM, john larkin wrote:

    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=
    https%3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.

    With respect to magnetics it follows the "first rule of government
    spending": why have one when you can have two, at twice the price."

    It does actually buy you better performance, and lets you run the
    converter at higher frequencies where both the magnetic components can
    be more compact - and cheaper - than they would be for a converter
    running at a lower frequency.

    The area product of the choke is on the same order as the area product
    of the transformer. It's annoying enough when you have to use one
    magnetics, much less two...and there are other isolated topologies e.g.
    Cuk that offer a constant input current.

    But as they remark in the abstract it's nice that they give a thorough
    design procedure though and doesn't use any exotic parts, for some applications particularly perhaps in some underprivileged areas of the
    world, the "best" converter is the one you can build with what you have available, the tolerances for this one don't look too critical, unlike
    the Cuk it doesn't look like it need a PhD to design a loop for.

    There is a variant that is claimed to be ripple free, though when I
    simulated it this was only approximately true - there was still some
    switching transient ripple though not a lot of it.

    https://www.researchgate.net/publication/3281291_A_Resonant_DC-DC_Transformer_With_Zero_Current_Ripple

    It has been discussed here before.

    --
    Bill Sloman, Sydney




    --
    This email has been checked for viruses by Norton antivirus software. www.norton.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Bill Sloman@21:1/5 to john larkin on Sat Jul 20 15:50:03 2024
    On 20/07/2024 1:11 pm, john larkin wrote:
    On Fri, 19 Jul 2024 21:43:00 -0000 (UTC), piglet
    <erichpwagner@hotmail.com> wrote:

    john larkin <jlarkin_highland_tech> wrote:

    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=
    https%3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.



    What do you find nasty about that topology, it works well , or does it
    remind you too much of Baxandall?

    When a fet turns off, it will spike (to kilovolts in simulation) and
    ring like a churchbell. Now we need snubbers.

    We know you don't bother putting in a parallel capacitance for
    inductors, let alone an an extra windings to represent the currents in
    induced in the somewhat conductive ferrite cores.

    Those omission can produce spikes in a simulation

    Simulations don't work very well if you don't simulate actual features
    of the real circuit

    Worse, if the fet turn-ons are slightly asymmetric, you get DC in the transformer primary and potential saturation.

    It's hard to get that with real components

    That can cause flux walking, The Devil's Staircase.

    Really?

    Only half of the primary is used at a time, which uses the copper inefficiently.

    The point about the Baxandall circuit is that it copes well with the
    parallel capacitances of the primary and secondary winding, which drive
    current through the inactive winding even when the relevant Fet is off.

    If you don't bother putting these capacitances into your simulation you
    will miss this.

    As Jim Williams pointed out you can get quite high efficiences out of
    the Baxandall inverter (though he never called it that).

    It needs another inductor. The waveform at the transformer CT is
    ghastly.

    I tend to add a few ferrite beads. They have much lower parallel
    capacitances than wound inductors, and tend to look like dissipative
    resistor a really high frequencies

    Certain Parties want to replace my design with this horror. I'm using
    a 4-fet full bridge driving both ends of my primary, from the 48v bulk supply. Nothing goes hi-z, the copper is fully used, and I can add a
    series cap to keep DC out of the transformer primary.

    It's not as horrible as you like to think - you seem to be suffering
    from a dose of not-invented-here.

    --
    Bill Sloman, Sydney



    --
    This email has been checked for viruses by Norton antivirus software. www.norton.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From boB@21:1/5 to bitrex on Sat Jul 20 00:02:23 2024
    On Fri, 19 Jul 2024 23:35:36 -0400, bitrex <user@example.net> wrote:

    On 7/19/2024 11:11 PM, john larkin wrote:
    \
    What do you find nasty about that topology, it works well , or does it
    remind you too much of Baxandall?

    When a fet turns off, it will spike (to kilovolts in simulation) and
    ring like a churchbell. Now we need snubbers.

    Worse, if the fet turnons are slightly asymmetric, you get DC in the
    transformer primary and potential saturation. That can cause flux
    walking, The Devil's Staircase.

    Only half of the primary is used at a time, which uses the copper
    inefficiently.

    It needs another inductor. The waveform at the transformer CT is
    ghastly.

    Certain Parties want to replace my design with this horror. I'm using
    a 4-fet full bridge driving both ends of my primary, from the 48v bulk
    supply. Nothing goes hi-z, the copper is fully used, and I can add a
    series cap to keep DC out of the transformer primary.


    It says in the text the transistors are supposed to be turned on >simultaneously for a fraction of the cycle.


    This is exactly true ! Current fed so that you can do that.

    I think this topology is also called a Clarke Converter.


    boB

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From legg@21:1/5 to All on Sat Jul 20 09:10:27 2024
    On Fri, 19 Jul 2024 13:35:37 -0700, john larkin
    <jlarkin_highland_tech> wrote:


    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=
    https%3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.

    Fairly common and usefull in it's day.

    Feed with buck regulator and ensure overlap during
    commutation.

    Bipolars did it first and can be self-oscillating
    if the gates are transformer-coupled.

    A million Boscherts camn't be wrong.


    RL

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Lasse Langwadt@21:1/5 to john larkin on Sat Jul 20 17:09:46 2024
    On 7/19/24 22:35, john larkin wrote:

    https://www.google.com/imgres?q=current%20fed%20push%20pull%20converter&imgurl=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F266522114%2Ffigure%2Ffig1%2FAS%3A392191802396676%401470517220704%2FCurrent-fed-Push-Pull-DC-DC-converter.png&imgrefurl=
    https%3A%2F%2Fwww.researchgate.net%2Ffigure%2FCurrent-fed-Push-Pull-DC-DC-converter_fig1_266522114&docid=2dgnjh28EeLzCM&tbnid=_McR7CYsN3_9eM&w=405&h=225&hcb=2


    Looks nasty to me.

    And TinyURL doesn't seem to work lately.


    similar to how most RF push-pull amplifiers are done, or with a cap
    across the primary how many self oscillating induction heater or high
    voltage generators are made

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From john larkin @21:1/5 to bitrex on Sat Jul 20 07:57:12 2024
    On Fri, 19 Jul 2024 23:35:36 -0400, bitrex <user@example.net> wrote:

    On 7/19/2024 11:11 PM, john larkin wrote:
    \
    What do you find nasty about that topology, it works well , or does it
    remind you too much of Baxandall?

    When a fet turns off, it will spike (to kilovolts in simulation) and
    ring like a churchbell. Now we need snubbers.

    Worse, if the fet turnons are slightly asymmetric, you get DC in the
    transformer primary and potential saturation. That can cause flux
    walking, The Devil's Staircase.

    Only half of the primary is used at a time, which uses the copper
    inefficiently.

    It needs another inductor. The waveform at the transformer CT is
    ghastly.

    Certain Parties want to replace my design with this horror. I'm using
    a 4-fet full bridge driving both ends of my primary, from the 48v bulk
    supply. Nothing goes hi-z, the copper is fully used, and I can add a
    series cap to keep DC out of the transformer primary.


    It says in the text the transistors are supposed to be turned on >simultaneously for a fraction of the cycle.

    There's still a huge fly-up spike when one turns off.

    But I will sim the overlap case.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)