Forums

Transfer function

Started by User Zero November 23, 2009
Hi 2 all,

is there any program to derive transfer function from circuit? Something 
like PSpice, but instead of numerical to do analytical "computation".
On Mon, 23 Nov 2009 12:19:52 +0000, User Zero wrote:

> Hi 2 all, > > is there any program to derive transfer function from circuit? Something > like PSpice, but instead of numerical to do analytical "computation".
What are you trying to do? You can't choke transfer functions out of Spice, but if you look at it from the right angle you can often get loads of useful information. Most circuits are either simple enough that someone has already cataloged their transfer functions (or you can derive them easily enough yourself), or complicated enough that if you ignore all of the parasitics and nonlinearities you're doing yourself no favors -- yet factoring in the parasitics makes the transfer functions exceedingly opaque, while factoring in the nonlinearities renders the whole concept of transfer function analysis invalid. -- www.wescottdesign.com
Tim Wescott wrote:
> On Mon, 23 Nov 2009 12:19:52 +0000, User Zero wrote: > >> Hi 2 all, >> >> is there any program to derive transfer function from circuit? Something >> like PSpice, but instead of numerical to do analytical "computation". > > What are you trying to do? You can't choke transfer functions out of > Spice, but if you look at it from the right angle you can often get loads > of useful information. > > Most circuits are either simple enough that someone has already cataloged > their transfer functions (or you can derive them easily enough yourself), > or complicated enough that if you ignore all of the parasitics and > nonlinearities you're doing yourself no favors -- yet factoring in the > parasitics makes the transfer functions exceedingly opaque, while > factoring in the nonlinearities renders the whole concept of transfer > function analysis invalid. >
A symbolic algebra program is probably your best bet, or else using something like Octave to fit some suitable curve to the output of a SPICE AC analysis. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal ElectroOptical Innovations 55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
On Mon, 23 Nov 2009 13:07:07 -0500, Phil Hobbs wrote:

> Tim Wescott wrote: >> On Mon, 23 Nov 2009 12:19:52 +0000, User Zero wrote: >> >>> Hi 2 all, >>> >>> is there any program to derive transfer function from circuit? >>> Something like PSpice, but instead of numerical to do analytical >>> "computation". >> >> What are you trying to do? You can't choke transfer functions out of >> Spice, but if you look at it from the right angle you can often get >> loads of useful information. >> >> Most circuits are either simple enough that someone has already >> cataloged their transfer functions (or you can derive them easily >> enough yourself), or complicated enough that if you ignore all of the >> parasitics and nonlinearities you're doing yourself no favors -- yet >> factoring in the parasitics makes the transfer functions exceedingly >> opaque, while factoring in the nonlinearities renders the whole concept >> of transfer function analysis invalid. >> >> > A symbolic algebra program is probably your best bet, or else using > something like Octave to fit some suitable curve to the output of a > SPICE AC analysis. >
Depends on whether you want to plug in numbers or keep the component values as symbolic (which gets _really_ messy _really_ fast). Something like Scilab (and I think Octave) will let you fairly easily reduce a circuit to a transfer function with numeric coefficients (or a state-space representation as polynomials are insanely sensitive to numerical accuracy issues). I've used this when I want to include a filter or something in a larger system model. -- www.wescottdesign.com
On Mon, 23 Nov 2009 13:07:07 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>Tim Wescott wrote: >> On Mon, 23 Nov 2009 12:19:52 +0000, User Zero wrote: >> >>> Hi 2 all, >>> >>> is there any program to derive transfer function from circuit? Something >>> like PSpice, but instead of numerical to do analytical "computation". >> >> What are you trying to do? You can't choke transfer functions out of >> Spice, but if you look at it from the right angle you can often get loads >> of useful information. >> >> Most circuits are either simple enough that someone has already cataloged >> their transfer functions (or you can derive them easily enough yourself), >> or complicated enough that if you ignore all of the parasitics and >> nonlinearities you're doing yourself no favors -- yet factoring in the >> parasitics makes the transfer functions exceedingly opaque, while >> factoring in the nonlinearities renders the whole concept of transfer >> function analysis invalid. >> > >A symbolic algebra program is probably your best bet, or else using >something like Octave to fit some suitable curve to the output of a >SPICE AC analysis. > >Cheers > >Phil Hobbs
PSpice has Laplace behavioral sources. Should be fairly easy to parameterize and fit to a Bode plot. I'd guess that LTspice also has that capability?? ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 |
User Zero wrote:
> Hi 2 all, > > is there any program to derive transfer function from circuit? Something > like PSpice, but instead of numerical to do analytical "computation".
SSpice (symbolic SPICE) is what you are looking for. Pere
On Tue, 24 Nov 2009 17:22:11 +0100, oopere <me@somewhere.net> wrote:

>User Zero wrote: >> Hi 2 all, >> >> is there any program to derive transfer function from circuit? Something >> like PSpice, but instead of numerical to do analytical "computation". > >SSpice (symbolic SPICE) is what you are looking for. > >Pere
Where? I can't find a URL that doesn't roll-over to a book store or similar. ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 |
Jim Thompson wrote:
> On Tue, 24 Nov 2009 17:22:11 +0100, oopere <me@somewhere.net> wrote: > >> User Zero wrote: >>> Hi 2 all, >>> >>> is there any program to derive transfer function from circuit? Something >>> like PSpice, but instead of numerical to do analytical "computation". >> SSpice (symbolic SPICE) is what you are looking for. >> >> Pere > > Where? I can't find a URL that doesn't roll-over to a book store or > similar. > > ...Jim Thompson
I played with an evaluation copy of SSpice some 15 years ago... After a quick search I have found this link http://www.elektroda.pl/rtvforum/download.php?id=349065 which brings up something that vaguely matches what I remember (not very user friendly though). But there are certainly few links on this in the internet! Another option, as someone already suggested, is to make use of the symbolic capabilities of Matlab (or Octave) and compute the relevant determinants from the (hand-written) system of equations. When I was a student I even coded a symbolic determinant function from scratch for exactly this purpose. Pere
That's it, thanks!!!

Quote:

> User Zero wrote: >> Hi 2 all, >> >> is there any program to derive transfer function from circuit? >> Something like PSpice, but instead of numerical to do analytical >> "computation". > > SSpice (symbolic SPICE) is what you are looking for. > > Pere
Late at night, by candle light, Jim Thompson
<To-Email-Use-The-Envelope-Icon@My-Web-Site.com> penned this immortal
opus:

>On Mon, 23 Nov 2009 13:07:07 -0500, Phil Hobbs ><pcdhSpamMeSenseless@electrooptical.net> wrote: > >>Tim Wescott wrote: >>> On Mon, 23 Nov 2009 12:19:52 +0000, User Zero wrote: >>> >>>> Hi 2 all, >>>> >>>> is there any program to derive transfer function from circuit? Something >>>> like PSpice, but instead of numerical to do analytical "computation". >>> >>> What are you trying to do? You can't choke transfer functions out of >>> Spice, but if you look at it from the right angle you can often get loads >>> of useful information. >>> >>> Most circuits are either simple enough that someone has already cataloged >>> their transfer functions (or you can derive them easily enough yourself), >>> or complicated enough that if you ignore all of the parasitics and >>> nonlinearities you're doing yourself no favors -- yet factoring in the >>> parasitics makes the transfer functions exceedingly opaque, while >>> factoring in the nonlinearities renders the whole concept of transfer >>> function analysis invalid. >>> >> >>A symbolic algebra program is probably your best bet, or else using >>something like Octave to fit some suitable curve to the output of a >>SPICE AC analysis. >> >>Cheers >> >>Phil Hobbs > >PSpice has Laplace behavioral sources. Should be fairly easy to >parameterize and fit to a Bode plot. > >I'd guess that LTspice also has that capability?? > > ...Jim Thompson
It does, and very nicely. Use it in the gain statement with dependent voltage or current sources. V=laplace(1/(s-1)) and so on. - YD. -- Remove HAT if replying by mail.