Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!ub!acsu.buffalo.edu
From: teodor@acsu.buffalo.edu (Dan C. Teodor)
Newsgroups: sci.electronics
Subject: Re: Novice SPICE query (op-amp info)
Message-ID: <59494@eerie.acsu.Buffalo.EDU>
Date: 13 Feb 91 05:36:51 GMT
References: <881@keele.keele.ac.uk>
Sender: news@acsu.Buffalo.EDU
Lines: 45
Nntp-Posting-Host: lictor.acsu.buffalo.edu


In article <881@keele.keele.ac.uk> you write:
>I've just started using SPICE and would greatly appreciate any
>standard subcircuits people could point me to. Specifically I
>need to simulate a TL072 op-amp, among other things. I'm using
>version 2G5 if that makes any difference.
>



Here's what I do. It's kind of messy but it works. First label either
the non-inverting or inverting input to the op-amp as ground (ie.
as node 0). Then leave the non-inverting and inverting input terminals
disconnected on one end (ie. connect to the rest of the network nodes
but leave them "dangling" on the other end). Then, model the op-amp
as a voltage dependent voltage source with a potential of the op-amp's
gain times the potential at the non-ground node that was left "dangling".

I normally use a value to somewhere between 10,000 and 50,000 for the
gain.

Of course all of the above assumes an ideal op-amp with no bias current
running through it's inputs. If you want to model a "realer" one see
what the bias current of your op-amp is and calculate it's Thevenin
equivalent resistance across the inputs (ie. R(th)=V/I). This is 
usually given in an op-amp's specs and no calculation is necessary.
Once you know this equivalent resistance (usually about 1 Megaohm) 
just connect a resistor of this value across those "dangling" terminals
and model as before. It all works out very nicely and it doesn't get 
any more accurate than that.


Good luck.
I'll try and give you some samples if this confuses the shit out of you.


Dan C. Teodor

teodor@sun.acsu.buffalo.edu
v083pzgu@ubvms.cc.buffalo.edu



P.S. I'm sorry to post this in this fashion but the return address on
that message bounced. But this pretty much general info.