[SI-LIST] Re: Routing Parallel vs. Paralell
- From: "Jack Olson" <pcbjack@xxxxxxxxx>
- To: SI-LIST <si-list@xxxxxxxxxxxxx>
- Date: Fri, 15 Sep 2006 09:31:35 -0500
I have gotten a lot of private responses
that all say a similar thing: "it depends"
Rather than write everyone back individually,
Please let me ask just one more question:
If I design a test board that has exactly three long traces on it,
- one for a random input signal
- one as a victim in parallel 6mils away
- one as a victim routed broadside 6mils away
put some kind of signal on the input trace
measure whatever happens on both victims
make a ratio of those two numbers
Then change the input signal to something more aggressive
the RATIO will change?
Or that if I modify how grounds are arranged around them
the RATIO will change?
THAT is interesting to me...
Jack
-=-=-=-
On 9/15/06, Jack Olson <pcbjack@xxxxxxxxx> wrote:
>
> That's okay.
> Maybe I shouldn't have asked such a question to this group.
>
> It seems like there should be a ratio between the field strength
> of two random traces routed broadside and the same two traces
> routed parallel (same distance away).
>
> Say I have a long skinny board with nothing special on it,
> 8MHz
> 6 layers
> the engineer wants all the routing on layers 3-4
> between planes,
> no impedance control, no pairs, no nothin'
> and there's one area where the traces overlap for a short
> distance (not on purpose, just no room).
>
> So the question was, how much worse are the overlaps
> than the parallels?
>
> I would assume it would be worse, but the only answer I got
> was a ratio of .60 to .65, which I never would have imagined.
>
> Anyway, my example is far below the kinds of problems this
> group is usually thinking about, maybe its just too ridiculous
> to consider
>
> thanks all the same, though
> Jack (aka "the new guy")
>
>
>
> -=-=-=-
>
>
>
> On 9/15/06, Xilei Liu <xileil@xxxxxxxxxxx> wrote:
> >
> > Jack,
> > Sorry I didn't realize that you were serious. I don't have field solvers
> > in
> > hand either, though I am curious about your question. APLAC can analyze
> > some
> > simple multilayer structures but can not be used for the whole PCB
> > package
> > simulation. Even from this simple model, I can't draw the conclusion
> > that
> > routing broadside is "worse" than routing side by side. It depends on
> > where
> > the grounds are. But I can't see much difference between them in terms
> > of
> > coupling factor from APLAC simulation. Actually, what do you mean by
> > "worse"? This coupling factor gives idea about how strong the coupling
> > signals can be between the two traces, but not how much interference
> > between
> > them. The rule " if you must run tandem, run the lines offset on the two
> > layers" Steve points out gives minimised inteference from the two traces
> >
> > configuration to others. Anyway, I'd better run away now.
> >
> > Regards,
> > Celine
> >
> >
> >
> >
> > >From: steve weir <weirsi@xxxxxxxxxx>
> > >Reply-To: weirsi@xxxxxxxxxx
> > >To: "Jack Olson" <pcbjack@xxxxxxxxx>,"Xilei Liu" <xileil@xxxxxxxxxxx>
> > >CC: si-list@xxxxxxxxxxxxx
> > >Subject: [SI-LIST] Re: Routing Parallel vs. Paralell
> > >Date: Thu, 14 Sep 2006 18:35:38 -0700
> > >
> > >Jack, no that is not a good generalization. To get an accurate
> > >answer for a specific situation you need to use a field
> > >solver. However, if it is a rule that you are looking for, if you
> > >must run tandem, run the lines offset on the two layers. You will
> > >get less total noise and the noise will be less coherent.
> > >Steve.
> > >At 11:27 AM 9/14/2006, Jack Olson wrote:
> > > >I apologize if I didn't phrase the question very well.
> > > >When I said MATH, I just meant I don't know how to get the answer.
> > > >When I used the 1inch vs .2 inch, I just chose random numbers,
> > > >they don't mean anything.
> > > >
> > > >Sorry for not knowing how to ask.
> > > >
> > > >If I have a trace already routed that is 8 inches long, and I have
> > >another
> > > >net not related to it but going to the same general area, and I have
> > a
> > > >choice whether to route it under the first one or next to the first
> > one,
> > > >how much worse is it to route it underneath?
> > > >(even though everyone says "don't do it, put a plane between")
> > > >
> > > >If I am reading your analysis correctly, running traces broadside is
> > > >only slightly worse than running them together on the same layer.
> > > >The difference between .60 and .65 seems a lot closer than I would
> > > >have predicted, but I'm not arguing...
> > > >
> > > >thanks for taking the time to respond,
> > > >Jack
> > > >
> > > >
> > > >On 9/14/06, Xilei Liu <<mailto: xileil@xxxxxxxxxxx>xileil@xxxxxxxxxxx
> > >
> > >wrote:
> > > >Jack,
> > > >Wondering whether I should give your MATH question my PHYSICAL answer
> > or
> > > >not.
> > > >
> > > >Firstly, I assume that the coupling you mentioned refers to the
> > coupling
> > > >between these two traces of different configurations.
> > > >
> > > >Secondly, assuming that you have 1.6mm FR4(er=4.31), double sided,
> > size
> > > >30*30mm^2; by APLAC simulation, the coupling factor(Ze-Zo)/(Ze+Zo)
> > for
> > >6mil
> > > >traces (1/2oz) routed 6mil broadside is about 0.65; and the coupling
> > >factor
> > > >for 6mil traces routed 6mil(edge to edge) side by side is about 0.60;
> > >1oz
> > > >copper doesn't change the results with 1/2oz copper.
> > > >
> > > >So, it sounds that the statement"traces routed 1 inch in parallel
> > will
> > > >couple the same amount as traces routed .2 inches broadside (layer to
> > > >layer)" is incredible, unless there is air or other material between
> > >"layer
> > > >to layer"?
> > > >
> > > >Well, should leave this MATH question to experts. Please give me $3
> > >whenever
> > > >possible ;)
> > > >
> > > >Regards,
> > > >Celine
> > > >
> > > >
> > > >
> > > > >From: steve weir <<mailto: weirsi@xxxxxxxxxx>weirsi@xxxxxxxxxx>
> > > > >Reply-To: <mailto:weirsi@xxxxxxxxxx> weirsi@xxxxxxxxxx
> > > > >To:
> > > >
> > ><mailto:pcbjack@xxxxxxxxx>pcbjack@xxxxxxxxx,<mailto:si-list@xxxxxxxxxxxxx
> > >si-list@xxxxxxxxxxxxx
> > > > >Subject: [SI-LIST] Re: Routing Parallel vs. Paralell
> > > > >Date: Thu, 14 Sep 2006 07:34:38 -0700
> > > > >
> > > > >Jack that is what we have field solvers for. You used to be able to
> > > > >rent time on Polar 8000 for about $20/hr. Otherwise find someone
> > > > >with Polar or Ansoft and run the model.
> > > > >
> > > > >Steve.
> > > > >At 07:29 AM 9/14/2006, Jack Olson wrote:
> > > > > >I got a MATH question I don't know how to solve...
> > > > > >I want to know the difference between running traces
> > > > > >parallel on the same layer (side by side) compared
> > > > > >to running them on different layers right on top of
> > > > > >each other (broadside).
> > > > > >
> > > > > >I know won't be a simple answer,
> > > > > >so I will provide a simple example:
> > > > > >
> > > > > >Say I am using 6mil traces with 6mil clearances on
> > > > > >layers that are 6mil apart using half ounce copper
> > > > > >(which is nearly .6mils) and both layers are internal.
> > > > > >
> > > > > >What distance can I route side by side and get the
> > > > > >same coupling as if I had routed them broadside?
> > > > > >
> > > > > >So I want a simple thing to remember that says
> > > > > >something like, "traces routed 1 inch in parallel will
> > > > > >couple the same amount as traces routed .2 inches
> > > > > >broadside (layer to layer)
> > > > > >
> > > > > >Does anyone know how to make my rule-of-thumb?
> > > > > >and if I change to 1oz copper, will the rule double?
> > > > > >
> > > > > >onward thru the fog,
> > > > > >Jack
> >
> >
> >
> >
>
------------------------------------------------------------------
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 FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
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:
