[SI-LIST] Re: Hyperlynx modelling of 'wire over ground'
- From: "Jennings, Kevin F" <Kevin.Jennings@xxxxxxxxxx>
- To: "si-list@xxxxxxxxxxxxx" <si-list@xxxxxxxxxxxxx>
- Date: Wed, 20 Feb 2008 15:00:09 -0600
Curt,
Yes your description of the circuit is accurate and yes there is an actual wire
added on the PCB which is why I picked the 'wire over ground' element to model
this (prior to actually putting the wire on the board).
The relatively accurate prediction of signal quality when dealing with PCB
traces but relatively poor prediction when adding another driver and a single
physical wire to the model was rather surprising.
My question here was mainly to see whether or not anyone else had tried to
model a rework wire (and/or multiple drivers) to a PCB and had HyperLynx
accurately predict (or not) the resulting waveform.
I haven't totally ruled out user error on my part, and still need to go back
and re-validate that my LineSim representation is an adequate representation of
the PCB (apparently if I had LineSim EXT I could export the net from BoardSim
and import it into LineSim...sigh). From there maybe try to trace down what
differences between model and reality that could explain the discrepancy, and
if all that fails, open a case with the good folks at Mentor and see if they
can explain the difference.
I definitely agree about the series termination, that's how I design things but
when tasked with trying to salvage something one has to do what one has to do.
Thanks for your input (and to all the others who've responded here and
privately).
Kevin Jennings
-----Original Message-----
From: Curt McNamara [mailto:CurtM@xxxxxxxxxxx]
Sent: Wednesday, February 20, 2008 10:10 AM
To: Jennings, Kevin F; si-list@xxxxxxxxxxxxx
Subject: RE: [SI-LIST] Hyperlynx modelling of 'wire over ground'
There are several possibilities. From your description it appears you have a
net with driver and multiple receivers in LineSim. You then add another
transmission line with another driver and tie it into the first transmission
line. Is that correct? Do you have both drivers active at the same time?
In the real board, is this second transmission line an actual wire on a PCB?
If these are all correct, here are some possibilities:
The actual wire does not have the impedance you model.
The actual wire has significant parasitics which you haven't modeled.
In general we avoid multiple drivers on a single line. There are several
issues: output circuitry may not parallel nicely; propagation delay to inputs
may be different; pulses originating at different line lengths arrive at
receivers at different times, leading to reflections like you observe.
For the case you described we would commonly use series termination at the
driver. Small resistors can often be placed between package leads and solder
pads to prove this out.
Curt
Curt McNamara, P.E. // principal electrical engineer
Logic Product Development
411 Washington Ave. N. Suite 400
Minneapolis, MN 55401
T // 612.436.5178
F // 612.672.9489
www.logicpd.com
/ / / / / / / / / / / / / / / / / / / / / / / / /
This message (including any attachments) contains confidential information
intended for a specific individual and purpose, and is protected by law. If you
are not the intended recipient, you should delete this message and are hereby
notified that any disclosure, copying, or distribution of this message, or the
taking of any action based on it, is strictly prohibited.
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of Jennings, Kevin F
Sent: Tuesday, February 19, 2008 7:59 AM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Hyperlynx modelling of 'wire over ground'
I'm consulting on a signal integrity issue on a board and trying to use Hyp=
erLynx to model proposed changes to the net. The BoardSim model of the PCB=
net appears to be a fairly accurate model of reality. When I construct a =
LineSim model of the net it also seems to be a fairly accurate model of rea=
lity. However, when I then add a 'wire over ground' to the Linesim model t=
he HyperLynx predicted waveforms seem to be nowhere near what is actually s=
een on the reworked PCBA.
The net in question is ~20 inches long, ~40MHz clock driven by a single dri=
ver into ~50 ohm PCB with 7 loads distributed along the line; the single dr=
iver does not have nearly enough oomph to produce a clean edge on the incid=
ent wave. What I was proposing as a rework was a wire from a second driver=
into an appropriate spot in the net. The Linesim model shows a relatively=
clean edge now but adding the wire on the PCBA does not. Instead there is=
a ~1V 5 ns wide dip in the signal after the first incident wave switch tha=
t in some of the loads is dipping back near the Vih spec limit.
Varying the Linesim 'wire over ground' parameters (wire radius and distance=
from ground) suggests that the proposed rework is relatively insensitive t=
o both which then suggests that the major contributor to the change in the =
signal (from the model's perspective) is the change in the net topology and=
available drive that occurs by adding the second driver and wire, which is=
as I expected.
The fact that reality diverges so far from the model basically renders the =
model useless for testing changes from a model perspective.
Does Linesim maybe not do well when modeling multiple drivers on a net?
Does Linesim maybe not model 'wire over ground' very well?
Anything else that I'm missing that might explain the wide discrepancy betw=
een model and reality when wiring in a second driver given that the model w=
as fairly accurate with the unmodified PCBA?
Thanks for your input.
Kevin Jennings
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
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
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
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:
