[SI-LIST] Re: SATA board to board connector question

  • From: olaney@xxxxxxxx
  • To: weirsi@xxxxxxxxxx
  • Date: Mon, 27 Aug 2007 16:35:57 -0700

I absolutely agree.  One of the things that interchange on this forum
accomplishes is a more complete picture than can be provided by any one
person.  Keep it coming.

Orin

On Mon, 27 Aug 2007 16:31:25 -0700 steve weir <weirsi@xxxxxxxxxx> writes:
> Orin, when it comes to fast signals, Lee is absolutely correct when 
> he 
> suggests looking at not just the underlying resin, but the laminate 
> weave as well. Paying attention to the weave selection and routing 
> in 
> light of the selected weave can buy a lot of performance for the 
> price 
> of some homework. The extra loss in strip line versus microstrip can 
> be 
> as much of a good thing as bad. It just depends.
> 
> Regards,
> 
> 
> Steve.
> 
> olaney@xxxxxxxx wrote:
> > But that's backward from what I said.  All else being equal, 
> microstrip
> > has the edge.  Here's one citation: 
> > Microstrip versus balanced stripline
> > Masse, D.   Pucel, R.A.   Hartwig, C.P.   
> > This paper appears in: Proceedings of the IEEE
> > Publication Date: Aug. 1968
> > Volume: 56,  Issue: 8
> > On page(s): 1359- 1360
> > Abstract
> > Microstrip is found to exhibit lower loss than balanced strip-line 
> of a
> > comparable physical structure. This property is traced to 
> conduction loss
> > theoretically and is confirmed experimentally.
> > Here's another:
> > //www.freelists.org/archives/si-list/12-2005/msg00155.html
> > Orin
> >
> > On Mon, 27 Aug 2007 14:07:31 -0700 "Loyer, Jeff" 
> <jeff.loyer@xxxxxxxxx>
> > writes:
> > Thanks for your kind words regarding the article.  I?d point out 
> that my
> > results were based on the effects not only of Nelco N4000-SI (with 
> a
> > corresponding Er = 3.4, Df = 0.007), but also of more common 
> (cheap) FR4
> > materials with Er = 4.2 and Df as high as 0.017.
> >  
> > The only assertion I was questioning was that microstrip is more 
> lossy
> > than stripline ? do you have any data supporting this?
> > Jeff Loyer 
> >
> >
> >
> >
> >
> > From: olaney@xxxxxxxx [mailto:olaney@xxxxxxxx] 
> > Sent: Monday, August 27, 2007 1:33 PM
> > To: Loyer, Jeff
> > Cc: si-list@xxxxxxxxxxxxx
> > Subject: Re: [SI-LIST] Re: SATA board to board connector question
> >  
> > No question, there are many decent materials on the market.  They 
> fab
> > like FR4, and often are called FR4 for marketing purposes.  FR4 is 
> a
> > concept, not a product spec per se, so the wise designer will do 
> the
> > homework and specify a vendor part # for consistent high 
> performance
> > results.  The fact remains that there is nasty FR4 and there is 
> nice FR4.
> >  If nothing but FR4 is called out, and your board house buys at 
> lowest
> > cost, you are rolling the dice regarding what you'll get back.  
> The idea
> > that you can run umpty Gb/s over something called FR4 is 
> gratifying, but
> > incomplete.  The drop in dielectric constant with rising frequency 
> for
> > the nasty stuff is well known, as is the rise in the loss tangent, 
> for
> > classic, nasty FR4.  Digital signals are forgiving, and when 
> combined
> > with equalization, extensive simulation and other tricks, can be 
> made to
> > do amazing things on "FR4", or polyimide, or oak leaves laminated 
> with
> > racoon poop if need be.  But if it was easy, the demand for SI
> > consultants would be rather lower than it is.  The very need for 
> care and
> > expertise underscores that the effect of the laminate is 
> nontrivial.  
> >  
> > Having rescued clients by scrapping the tricks and simply using a 
> decent
> > microwave capable material for critical circuit portions, I stand 
> behind
> > my observations.  I also totally believe the results in your 
> article.  Of
> > course, the material upon which the article is based is "Nelco 
> N4000-13
> > SI, a readily available (though relatively expensive, compared to 
> FR-4)
> > material."  Good stuff.  I like the article.  Your advice 
> concerning the
> > transmission path implementation is right on: "Proper PCB stackup 
> and
> > trace geometry design are key elements in the fight to lower 
> losses".  So
> > is a short path.  The best advice for a SATA connection is simple: 
> use
> > the right connector and put the phy chip next to it.  That way, 
> even
> > those using the nasty type of FR4 (whose name is Legion, for they 
> are
> > many) have a fighting chance of success, without need to do double 
> back
> > flips to get there.  
> >  
> > Regards,
> >             Orin
> >  
> > On Mon, 27 Aug 2007 09:57:16 -0700 "Loyer, Jeff" 
> <jeff.loyer@xxxxxxxxx>
> > writes:
> >   
> >> Hi Orin,
> >> Do you have some specific data on the difference in performance 
> >> between
> >> microstrip and stripline?  While I agree with your first premise 
> >> (that
> >> stripline introduces the parasitics of the via to get down to the
> >> layer), my studies (see url below) have indicated equal loss 
> between 
> >> the
> >> 2 mediums (stripline vs. microstrip), with the caveat that 
> >> microstrip
> >> often allows wider traces with resultant lower copper loss.  But,
> >> impedance control on the inner layers (with no plating) is 
> usually
> >> better, and crosstalk is reduced.
> >>
> >> I've often heard similar arguments to your "snorkeling" metaphor, 
> 
> >> but
> >> haven't seen data supporting it.
> >>
> >> This leads me to favor lower layers of stripline for long lengths 
> 
> >> of
> >> high-speed signals, when I have a choice (a rare luxury).
> >>
> >> Perhaps you've had some other experience or have data to share?
> >>
> >> url to my study on microstrip vs. stripline loss:
> >> http://pcdandm.com/cms/content/view/2572/95/
> >>
> >> Note to all: there have been many postings on "Microstrip vs. 
> >> Stripline"
> >> on this forum, worthy of perusing if you're interested.
> >>
> >> Disclaimer:
> >> The content of this message is my personal opinion only and 
> although 
> >> I
> >> am an employee of Intel, the statements I make here in no way 
> >> represent
> >> Intel's position on the issue, nor am I authorized to speak on 
> >> behalf of
> >> Intel on this matter.
> >>
> >> Jeff Loyer
> >>
> >> -----Original Message-----
> >> From: si-list-bounce@xxxxxxxxxxxxx 
> >> [mailto:si-list-bounce@xxxxxxxxxxxxx]
> >> On Behalf Of olaney@xxxxxxxx
> >> Sent: Monday, August 27, 2007 8:56 AM
> >> To: nareshd@xxxxxxxxxxxxx
> >> Cc: goister@xxxxxxxxx; si-list@xxxxxxxxxxxxx
> >> Subject: [SI-LIST] Re: SATA board to board connector question
> >>
> >> Usually I prefer microstrip over stripline for signals as fast as 
> 
> >> SATA. 
> >> The reasoning boils down to two considerations.  One is that 
> >> transitions
> >> between buried stripline and other layers brings its own set of
> >> problems,
> >> and the other is that in a world where (censored) insist on using 
> 
> >> FR4 in
> >> the microwave region, the less contact a high speed signal has 
> with 
> >> the
> >> stuff, the better.  Air is a reasonable microwave dielectric.  
> The 
> >> FR in
> >> FR4 means "fire retardent", which is not a microwave concept.
> >> Microstrip
> >> steps in the poop on only one side of the trace, where stripline 
> is 
> >> like
> >> snorkling in it.  This makes a difference in how much signal 
> >> quality
> >> gets
> >> eaten per unit length.
> >>
> >> Don't get me wrong.  Stripline is excellent microwave practice 
> when
> >> proper dielectrics are used and connectorization is competently
> >> implemented.  But, that level of attention seems reserved for a
> >> different
> >> world filled with directional couplers, 90 degree hybrids, and 
> >> other
> >> exotic animals never seen on this forum.
> >>
> >> Orin
> >>
> >>     
> >
> > ------------------------------------------------------------------
> > 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
> >   
> >
> >
> >
> >   
> 
> 
> -- 
> Steve Weir
> Teraspeed Consulting Group LLC 
> 121 North River Drive 
> Narragansett, RI 02882 
> 
> California office
> (408) 884-3985 Business
> (707) 780-1951 Fax
> 
> Main office
> (401) 284-1827 Business 
> (401) 284-1840 Fax 
> 
> Oregon office
> (503) 430-1065 Business
> (503) 430-1285 Fax
> 
> http://www.teraspeed.com
> This e-mail contains proprietary and confidential intellectual 
> property of Teraspeed Consulting Group LLC
>
-------------------------------------------------------------------------
-----------------------------
> Teraspeed(R) is the registered service mark of Teraspeed Consulting 
> Group LLC
> 
> 
> 
------------------------------------------------------------------
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: