ID:	4212
From:	Anthony Tuttle <ke...@gmail.com>
Date:	Tue, 23 Jun 2020 09:20:40 -0500
Subject:	Re: [pcb-rnd] Off topic (maybe) - Spice simulator
in-reply-to:	4209 from stefan <st...@gmail.com>
replies:	4213 from Robin Laing <Me...@TelusPlanet.net>
--00000000000005ef6505a8c11166
Content-Type: text/plain; charset="UTF-8"
 
LTspice has a neon lamp model that can work without being too prone to
convergence issues. I have also used R={expression} as the resistance value
in resistors to simulate switches and this may be workable for an
improvised discharge model.
 
There are spice models for GDTs, you may have to contact the manufacturer
directly to gain access. They are never perfect but there has been some
work on advanced GDT models.
 
On Tue, Jun 23, 2020 at 1:41 AM stefan <stefan.schippers@gmail.com> wrote:
 
>
> Hello,
>
> I am involved with a project that will involve high voltage pulse
> circuit design and I am looking for a circuit simulator that will give
> us an accurate representation of a circuit.
>
> Needs to work with spark gap technology as that is the present design.
> Maybe gas discharge tube as well.
>
> Presently one member of the team is trying to use Multisim from NI but I
> am wondering if there may be a better package.
>
> NGSPICE developers are doing a good job,
> ngspice netlists may contain capacitors, resistors, inductors with
> arbitrary expressions about their value.
> Moreover the 'B source' allows any type of voltage/current source with
> arbitrary expressions.
>
> Combining these elements a model for a gas discharge/breakdown can be
> created.
> The big issue is not the simulator itself, but the quality of the model:
>
> 1) ability to accurately describe the device.
> 2) simulator must converge to a solution.
>
> Breakdown devices (devices with infinite resistance that start conducting
> with very low resistance when a voltage threshold is reached) are famous
> for giving convergence issues. This usually happen due to poor model
> description. A current limit must be included in the equations (no physical
> device can sustain 1e15 Amp of current :-) ).
>
> Stefan
>
 
--00000000000005ef6505a8c11166
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
 
<div dir=3D"ltr">LTspice has a neon lamp model that can work without being =
too prone to convergence issues. I have also used R=3D{expression} as the r=
esistance value in resistors to simulate switches and this may be workable =
for an improvised discharge model.<div><br></div><div>There are spice model=
s for GDTs, you may have to contact the manufacturer directly to gain acces=
s. They are never perfect=C2=A0but there has been some work on advanced GDT=
 models.</div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=
=3D"gmail_attr">On Tue, Jun 23, 2020 at 1:41 AM stefan &lt;<a href=3D"mailt=
o:stefan.schippers@gmail.com">stefan.schippers@gmail.com</a>&gt; wrote:<br>=
</div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;b=
order-left:1px solid rgb(204,204,204);padding-left:1ex">
 =20
   =20
 =20
  <div>
    <div><br>
    </div>
    <blockquote type=3D"cite">Hello,
      <br>
      <br>
      I am involved with a project that will involve high voltage pulse
      <br>
      circuit design and I am looking for a circuit simulator that will
      give
      <br>
      us an accurate representation of a circuit.
      <br>
      <br>
      Needs to work with spark gap technology as that is the present
      design.
      <br>
      Maybe gas discharge tube as well.
      <br>
      <br>
      Presently one member of the team is trying to use Multisim from NI
      but I
      <br>
      am wondering if there may be a better package.
      <br>
      <br>
    </blockquote>
    NGSPICE developers are doing a good job, <br>
    ngspice netlists may contain capacitors, resistors, inductors with
    arbitrary expressions about their value.<br>
    Moreover the &#39;B source&#39; allows any type of voltage/current sour=
ce
    with arbitrary expressions.<br>
    <br>
    Combining these elements a model for a gas discharge/breakdown can
    be created. <br>
    The big issue is not the simulator itself, but the quality of the
    model:<br>
    <br>
    1) ability to accurately describe the device.<br>
    2) simulator must converge to a solution.<br>
    <br>
    Breakdown devices (devices with infinite resistance that start
    conducting with very low resistance when a voltage threshold is
    reached) are famous for giving convergence issues. This usually
    happen due to poor model description. A current limit must be
    included in the equations (no physical device can sustain 1e15 Amp
    of current :-) ).<br>
    <br>
    Stefan<br>
  </div>
 
 
 
</blockquote></div>
 
--00000000000005ef6505a8c11166--
Mailing list archives : pcb-rnd

Navigation
