Re: [SI-LIST] : simulating differential pairs in Specctraquest

About this list Date view Thread view Subject view Author view

From: Patrick Riffault (riffault@cisco.com)
Date: Fri Aug 25 2000 - 08:31:24 PDT


At 09:16 AM 08/25/2000 +0530, Aloke Bhattacharya wrote:

My two cents:

>Hi Ken,
> Thanks a lot for the reply. Can you please explain a little more about the
>two alternatives you have
>suggested for the simulation of differential pair? Can you please explain
>exactly how to

>(i)mock up differential pairs interactively in sigxp.

This all assumes you have 13-6 and the cadence default
libraries are loaded.

1) edit -> add part -> differential part

you should see the cadence default diff parts.
put those down and double click on CDSdefaultIO
replace the part with your own IO
set the parameters according to what your need (thresholds and stuff)
repeat as necessary depending on the number of different parts

If you need a '"packaged device" look at the IbisDevice diffpair that is in
the default
library

2) edit -> add part -> interconnect

I would chose the stripline-2 or microstrip-2 depending on requirements.
Those are
trace models, rather than lossless T-lines.

set parameters as needed. Look at the xsection information for the impedance,
velocity, diff-impedance ...

I would do it like that in order to ensure the coupling that you are design
for.

3) put down the topology

4) Double click on the wanted driver where it says tri-state enter the
wanted stimulus

5) hit the simulate button and all should be simulated with two drivers 180
phase
difference and the output combined.

>(ii)simulate differential pair at the board level.
>It would be nice if you can explain the method in terms of Specctraquest
>menus/commands.

IBIS dependent. If you go and set the diffpair pins in your IBIS model
it will be done for you automatically when you do a board level sim.

I don't do it pre-route. I am not sure what assumptions the pre-route
algorithm does based on the conductors that need to be next to
each other and obviously are not at that point.

I find that it is also better to set the diffpair rules

constraint -> diffpair

before doing the analysis.

Patrick

>Regarding the board level simulation of differential pair, you have mentioned
>only about post route
>analysis- is pre route analysis not possible for diff pairs?
>
>After the board level simulation as you have mentioned in the second
>alternative, is it possible to get
>the wavefoms(I means the 180 degree phase shifted waveforms of the two lines)
>in sigwave?
>
>With regards,
>Aloke
>
>Ken Willis wrote:
>
> > Hi Aloke,
> >
> > Today, you can't extract coupled-line circuits into Sigxp. You
> > can build coupled-line topologies interactively in Sigxp and
> > simulate them, or simulate coupled-line stuff to your heart's
> > content at the board lavel, but today you can only extract
> > single-line topologies directly from Allegro into Sigxp. The
> > reason for this is that, besides being complicated (try to come
> > up with an algorithm to auto-place all those symbols in the Sigxp
> > canvas legibly!), for a given geometry window you may end up with
> > an enormous coupled circuit. It isn't uncommon to drag
> > 20 neighbor nets into a crosstalk simulation. While this is routinely
> > done at the board-level simulation with SPECCTRAQuest, extraction
> > of coupled stuff into Sigxp as topologies won't be supported until
> > some kind of intelligent pruining algorithms are written, for example
> > to just extract the 2 diff pair nets into Sigxp and nothing else. I
> > know this was being kicked around a couple of months back, but don't know
> > the status.
> >
> > The way I do it today is two-fold:
> >
> > - mock up your diff pairs interactively in Sigxp and
> > figure out your wiring rules
> >
> > - simulate the diff pairs at the board level post-route
> > and look at the skews, waveforms, and eye patterns
> >
> > It works pretty well, I've gotten good correlation with lab
> > measurement well up over 600 MHz for diff pairs.
> >
> > Ken Willis
> > Sirocco Systems
> >
> > -----Original Message-----
> > From: Aloke Bhattacharya [mailto:aloke.bhattacharya@wipro.com]
> > Sent: Thursday, August 24, 2000 12:31 PM
> > To: si-list@silab.eng.sun.com
> > Subject: [SI-LIST] : simulating differential pairs in Specctraquest
> >
> > Hello,
> > I am facing a problem with simulation of differencial pairs in
> > Specctraquest. When I try to extract the topology of a net which is
> > defined as part of a differential pair, Sigxp gives error saying
> > "differential pair extraction is currently not supported". Does anybody
> > know how to simulate a differential pair
> > in Specctraquest?
> >
> > Thanks in advance,
> > With regards,
> > Aloke
> >
> >
> >
> >
> > ****
>
>--
>**********************************************************************
>* Aloke Bhattacharya, *
>* Senior Engineer-VLSI/System Design,
>* *
>* Wipro Infotech, *
>* Global R&D, *
>* 88, M.G. Road, 5th Floor, *
>* S.B. Towers, *
>* Bangalore- 560 001 , *
>* INDIA *
>* Tel : 91-80-5588422(Ext. 520) *
>* email: aloke.bhattacharya@wipro.com
>* *
>**********************************************************************
>
>
>
>**** To unsubscribe from si-list or si-list-digest: send e-mail to
>majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
>si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
>si-list archives are accessible at http://www.qsl.net/wb6tpu
>****

**** To unsubscribe from si-list or si-list-digest: send e-mail to
majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Tue May 08 2001 - 14:29:20 PDT