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

  • From: <Vittal.Balasubramanian@xxxxxxxxxxxxxx>
  • To: <goister@xxxxxxxxx>
  • Date: Thu, 23 Aug 2007 14:37:08 -0400

Goi,=0D=0A=0D=0AYou can find some information on the SATA products offered =
by FCI at the=0D=0Afollowing pages=2E =0D=0Ahttp://portal=2Efciconnect=2Eco=
m/portal/page/portal/FcicntPublic/HighSpeedIO=0D=0AConnectors (Look at the =
I/O serial ATA connectors section on the page)=0D=0A=0D=0Ahttp://portal=2Ef=
FHard+Disk+Drives+Application%2FInput-Output+Connectors=0D=0A(click through=
 to the SATA Series for a list of some part numbers)=2E=0D=0A=0D=0AIf you h=
ave any specific signal integrity questions, feel free to=0D=0Acontact me a=
nd I'll get you the information you need=2E=0D=0A=0D=0ABest Regards,=0D=0AV=
ittal Balasubramanian=0D=0ASr=2E Signal Integrity Engineer=0D=0AFCI USA, In=
c=2E=0D=0A=0D=0A-----Original Message-----=0D=0AFrom: si-list-bounce@freeli=
sts=2Eorg [mailto:si-list-bounce@freelists=2Eorg]=0D=0AOn Behalf Of olaney@=
juno=2Ecom=0D=0ASent: Thursday, August 23, 2007 12:11 PM=0D=0ATo: goister@g=
mail=2Ecom=0D=0ACc: si-list@freelists=2Eorg=0D=0ASubject: [SI-LIST] Re: SAT=
A board to board connector question=0D=0A=0D=0AYes, the connector manufactu=
rer recommendations=2E  Any connector with the=0D=0Aright impedance and ade=
quate data rate specs will do=2E  The same=0D=0Aconnector family can be use=
d for many purposes, so it is not a matter of=0D=0Athe signaling system use=
d as much as the electrical performance and the=0D=0Amechanical need=2E  Ex=
tra pins can be ignored or used for other purposes=2E =0D=0AI don't know if=
 you are trying to connect your boards connector to=0D=0Aconnector direct, =
or through a short cable=2E  In any instance, keep the=0D=0AI/O chip close =
to the connector, and avoid changes in trace widths other=0D=0Athan what mi=
ght be recommended for dealing with proximity effects=0D=0A(sometimes the s=
ignal traces are tapered where they run under connector=0D=0Adielectric)=2E=
  One common signal integrity killer is where the traces are=0D=0Acarefully=
 designed, yet run through coupling caps or other passives=0D=0Ausing the f=
at SMT pads that production loves so much=2E  Sometimes=0D=0Aproduction bal=
ks at the practice of using wider traces or narrower parts=0D=0Ato minimize=
 the discontinuity in microstrip width=2E  Given the choice=0D=0Abetween a =
design that works and one that can be built without hand=0D=0Aoperations or=
 other accommodations, they have a tough time making up=0D=0Atheir minds=2E=
  As an engineer with proper test equipment and adequate=0D=0Atime (That's =
all of us, right?  Right?), you can often meet both goals=2E=0D=0A=0D=0AAny=
way, you might as well check out formal SATA connectors to understand=0D=0A=
them electrically before widening your search=2E  The GSSGSSG layout is=0D=
=0Anot absolutely necessary for short distances, as crosstalk can be=0D=0Ac=
ontrolled with adequate pair separation=2E  I would not assume that a 7=0D=
=0Apin connector is the goal=2E  Think of it as two impedance controlled=0D=
=0Apairs plus whatever ground paths are designed in to ensure impedance and=
=0D=0Acrosstalk control=2E =0D=0AThe ground need not be discrete pins on pa=
r with the signals, in the=0D=0Asame way that a coax shield is not carried =
by a pin like that of the=0D=0Acenter conductor=2E  However, make sure that=
 the ground is carried with=0D=0Athe signals through the same connector!!! =
 Any ground connections=0D=0Aprovided elsewhere in the system might make an=
 ohmmeter happy, but=0D=0Arelying on them =3D death at high frequencies=2E =
 =0D=0A=0D=0AMolex might be another vendor to include in your list=2E  Don'=
t be shy=0D=0Aabout using vendor field app engineers -- that's what they're=
 paid for,=0D=0Aand it's job security for them=2E=0D=0A=0D=0AOrin=0D=0A=0D=
=0AOn Thu, 23 Aug 2007 23:07:46 +0800 "Sihan Goi" <goister@gmail=2Ecom>=0D=
=0Awrites:=0D=0AThanks=2E It's kinda confusing when I go to samtec, FCI or =
amphenol=0D=0Awebsite and they don't show what kind of applications the hig=
h speed=0D=0Aconnectors are meant for, or their impedance rating or whateve=
r=2E=2E=2E=0D=0A=0D=0AAnyway, I'm also wondering about how the signals shou=
ld be routed to the=0D=0Aconnector=2E SATA signals on regular SATA connecto=
r are as follows GND A+=0D=0AA- GND B+ B- GND=0D=0A=0D=0AShould I be follow=
ing this topology or does it even matter? I'm guessing=0D=0Amost of these h=
igh speed connectors won't have exactly 7 pins=2E They're=0D=0Ausually spec=
'ed for 2/3/4 pairs or even more=2E I'm guessing I only need a=0D=0A2 pair =
connector since I only have 2 differential SATA signal pairs=2E I'm=0D=0Awo=
ndering if I even need to route the GND with the high speed connector=0D=0A=
or can it be from another regular connector somewhere else=2E=2E=2E? =0D=0A=
=0D=0ALastly, when you say "pay attention to the manufacturer recommendatio=
ns"=0D=0Awhich manufacturer do you mean? The connector manufacturer?=0D=0A=
=0D=0AAgain, thanks for your reply!=0D=0A=0D=0A=0D=0AOn 8/23/07, olaney@jun=
o=2Ecom <olaney@juno=2Ecom> wrote:=0D=0ABasically, yes, but give your desig=
n some margin=2E  For 3 Gb/s go for 5=0D=0AGHz  or more of frequency respon=
se=2E  Biggest headache tends to be=0D=0Akeeping the impedance constant whe=
re the traces enter the connector=0D=0Alaunch area=2E =0D=0APay attention t=
o the manufacturer recommendations=2E=0D=0A=0D=0AOrin Laney=0D=0A=0D=0AOn T=
hu, 23 Aug 2007 16:08:09 +0800 "Sihan Goi" <goister@gmail=2Ecom>=0D=0Awrite=
s:=0D=0A> Hi,=0D=0A> I have a design where I have to route SATA signals fro=
m a main board =0D=0A> to a daughter board=2E The daughter board will likel=
y have nothing =0D=0A> except for the 7 SATA signals(4 data and 3 GND) conn=
ected to a regular=0D=0A=0D=0A> SATA HDD connector (unless some passives ar=
e needed?)=2E=0D=0A>=0D=0A> In my previous PATA design, I used a normal 44p=
in IDE type connector =0D=0A> pair(pin header + socket), and this worked we=
ll for PATA=2E I'm guessing=0D=0A=0D=0A> this will not work so well with SA=
TA though=2E What kind of connectors =0D=0A> would work for SATA1/2? I know=
 the differential impedance is 100ohms, =0D=0A> and that the trace length d=
ifference must be within 5 mils=2E I'm =0D=0A> guessing I have to get a hig=
h speed connector that has the same =0D=0A> impedance and is able to suppor=
t 3GHz speeds? Is that all I need to be=0D=0A=0D=0A> looking for?=0D=0A>=0D=
=0A> Thanks=2E=0D=0A>=0D=0A> --=0D=0A> - Goi Sihan=0D=0A> goister@gmail=2Ec=
om=0D=0A>=0D=0A>=0D=0A> ---------------------------------------------------=
---------------=0D=0A> To unsubscribe from si-list:=0D=0A> si-list-request@=
freelists=2Eorg with 'unsubscribe' in the Subject field=0D=0A>=0D=0A> or to=
 administer your membership from a web page, go to:=0D=0A> http://www=2Efre=
elists=2Eorg/webpage/si-list=0D=0A>=0D=0A> For help:=0D=0A> si-list-request=
@freelists=2Eorg with 'help' in the Subject field=0D=0A>=0D=0A>=0D=0A> List=
 technical documents are available at: =0D=0A>                 http://www=
=2Esi-list=2Enet=0D=0A>=0D=0A> List archives are viewable at:=0D=0A>       =
          http://www=2Efreelists=2Eorg/archives/si-list=0D=0A> or at our re=
mote archives:=0D=0A>                 http://groups=2Eyahoo=2Ecom/group/si-=
list/messages=0D=0A> Old (prior to June 6, 2001) list archives are viewable=
 at: =0D=0A>                  http://www=2Eqsl=2Enet/wb6tpu=0D=0A>=0D=0A>=
=0D=0A>=0D=0A>=0D=0A=0D=0A=0D=0A=0D=0A=0D=0A--=0D=0A- Goi Sihan=0D=0Agoiste=
r@gmail=2Ecom =0D=0A=0D=0A-------------------------------------------------=
-----------------=0D=0ATo unsubscribe from si-list:=0D=0Asi-list-request@fr=
eelists=2Eorg with 'unsubscribe' in the Subject field=0D=0A=0D=0Aor to admi=
nister your membership from a web page, go to:=0D=0Ahttp://www=2Efreelists=
=2Eorg/webpage/si-list=0D=0A=0D=0AFor help:=0D=0Asi-list-request@freelists=
=2Eorg with 'help' in the Subject field=0D=0A=0D=0A=0D=0AList technical doc=
uments are available at:=0D=0A                http://www=2Esi-list=2Enet=0D=
=0A=0D=0AList archives are viewable at:     =0D=0A              
=2Eorg/archives/si-list=0D=0Aor at our remote archives:=0D=0A           
s=2Eyahoo=2Ecom/group/si-list/messages=0D=0AOld (prior to June 6, 2001) lis=
t archives are viewable at:=0D=0A               
http://www=2Eqsl=2Enet/wb6tpu=0D=0A  =
=0D=0A=0D=0A=0D=0A-----------------------------------------=0D=0A***** Conf=
identiality Statement/Disclaimer *****=0D=0A=0D=0AThis message and any atta=
chments is intended for the sole use of=0Athe intended recipient=2E  It may=
 contain confidential information=2E =0AAny unauthorized use, dissemination=
 or modification is strictly=0Aprohibited=2E  If you are not the intended r=
ecipient, please notify=0Athe sender immediately then delete it from all yo=
ur systems, and do=0Anot copy, use or print=2E  Internet communications are=
 not secure and=0Ait is the responsibility of the recipient to make sure th=
at it is=0Avirus exempt=2E  The company/sender cannot be responsible for an=
y=0Aunauthorized alterations or modifications made to the contents=2E  If=
=0Ayou require any form of confirmation of the contents, please=0Acontact t=
he company/sender=2E  The company/sender is not liable for=0Aany errors or =
omissions in the content of this message
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:

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List technical documents are available at:

List archives are viewable at:     
or at our remote archives:
Old (prior to June 6, 2001) list archives are viewable at:

Other related posts: