[SI-LIST] Re: Simulation Question in Spectra Quest.
- From: Mike LaBonte <mike@xxxxxxxxxxx>
- To: sivakumar@xxxxxxxxxxxxxxx
- Date: Wed, 15 Aug 2001 09:02:44 -0400
Sivakumar,
ibis2signoise runs ibischk3 for you, then runs other checks of its own.
Using MPC8260IBIS.ibs I saw 106 warnings (after fixing the file to
translate at all). Most of the warnings were for non-monotonic waveforms,
with reversals as much as 430mV! The waveforms were collected with
reactive loads, some of them around 62pF and 10nH, which explains the
non-monotonic behavior. Simulators are supposed to handle this, but I
believe IBIS committee members have agreed in the past that reactive
test loads are not a good idea. They are supported in the IBIS specification
to facilitate creating IBIS models from physical measurements of packaged
parts. But it is difficult to create a simulation model from this data.
For what it's worth, HSPICE gives warnings about the models but does
produce a waveform. The 800V output range is somewhat suspect, however :-)
What can be done? I find that simply removing the l_dut parameters
from the IBIS file makes it simulate in SPECCTRAQuest. However, simulating
into a test load that replicates one of the waveform test fixtures results
in an unacceptable match. I don't recommend doing this. It would be nice if
Motorola were to submit this file to the IBIS Model Review committee. I
suspect they would see that the various simulators respond quite differently
to the models.
Mike LaBonte
"Sivakumar S." wrote:
>
> Hi SIexperts,
> I am doing Signal Integrity Analysis using Spectra Quest signal
> explorer. I have a problem in simulating MPC8260 model. First i converted
> MPC8260 IBIS model into ".dml" format using "ibis2signoise" utility and
> there was an error message " TV curve should start at zero". I changed the
> firtst line of the rising and falling waveforms data "zero" and now there
> was no error in the conversion and then i assigned this model to the device.
> When i tried to do the simulation using this I/O model it gives "SERIOUS
> Convergence Problem in Simulation of "DESIGN IO1 1" stimulus Pulse".
> Can anybody help me to solve this problem?
>
> Thanks and regards,
> Sivakumar.S
>
> -----Original Message-----
> From: Heiko Dudek [mailto:heikod@xxxxxxxxxxx]
> Sent: Thursday, August 09, 2001 3:31 PM
> To: we_r_frendz@xxxxxxxxx; si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Simulation Question in Spectra Quest.
>
> Rahul,
>
> >1) The technology that I simulate makes use of
> >Differential strobes. Do I need to set up two strobes
> >in my simulation or a single strobe sampling at both
> >edges.
>
> If your strobe line is layed out as a differential pair, you
> definitely need to model this with a coupled trace element.
> Same for driver / receiver (take care, often differential
> drivers are used together with single receivers). You need
> to use single or differential models, depending on what
> you have in reality. No direct relation to sampling on
> rising, falling or both edges. This again depends on
> whatever technology you're simulating ... RAMBUS for
> example samples on both edges. This way, a 400MHz
> strobe results in 800MT/s.
>
> >2) Does the tool sample at both the edges or only at
> >the rising edges.
>
> whatever you want it to. Select 'Set'->'Strobe Pins'. In
> the 'Strobe Selection Window', you can set 'Active Edge'
> to 'Both', 'Rising' or 'Falling'.
>
> >3)When we make use of Custom measurements, In the IO
> >cell Edit window there is a column for Freq. Does this
> >frequency indicate the sampling frequency ( The rate
> >at which the data is being sampled ). If so what does
> >the Clock freq that appears in the Analysis
> >preferences window indicate.
>
> I assume you're talking about the I/O cell stimulus edit. You
> do have the choice between different stimulus states ...
> there's 'Rising', 'Falling', 'Pulse' and 'Custom' (and 'Quiet
> hi', 'Quiet lo' and 'Tristate' as well). Whereas all I/O cells
> with rising, falling or pulse waveforms are operating in
> sync and with whatever you set in the analysis preference
> window as clock frequency. If you decide an I/O cell to
> have a custom stimulus, you can set the clock frequency
> to whatever you want and individually for all custom I/Os.
> The frequency set in the analysis prefs has no influence
> on custom I/O stimulus.
>
> Hope this helps.
>
> - Heiko
>
> Heiko Dudek
> Technical Marketing Manager, High Speed Systems Design & IC Packaging
> Cadence Design Systems, 270 Billerica Road, Chelmsford, MA 01824
>
> ph (office Chelmsford, USA): (978) 262-6384
> ph (office Feldkirchen, Germany): (089) 4563-0
> fx (office Chelmsford, USA): (978) 446-6798
> fx (office Feldkirchen, Germany): (089) 4563-1800
> mobile world phone: USA - (617) 230 4233 or Europe - (+49) (172) 938 1541
> email: heikod@xxxxxxxxxxx
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
> List archives are viewable at:
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
> List archives are viewable at:
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
