[SI-LIST] Re: differential signals spacing

Goutham,

        I don't think "as close as possible" is a good rule of thumb for
differntial pair spacing, although I don't doubt that some guidelines may
actually contain this type of language. While there are advantages to very
closely coupled pairs, there are also reasons for maintaining some degree of
isolation. The major considerations in defining diff pair geometries are
differnetial impedance and loss. You have a limited range of uncoupled
impedance on a PCB, generally 40 to 65 ohms or so, so you have a similiarily
limited range of differential impedance. You don't always have the luxury of
closely coupled pairs if you need to acheive a fairly high differential
impedance. In particular, when using devices with internal termination where
you don't get to choose. You also have to consider loss in determining
miminum trace widths. In the PC world I see alot of 4-5 mil trace with 6-8
mil spaced moderately coupled CMOS clocks, as well as 4-5 mil trace with 4-5
mil space tighly coupled LVDS pairs. These geometries generally yeild
differential impedances between 80 and 100 ohms. I have also seen 7 mil
traces with 7 mil spaces on some I/O interfaces where loss is an issue. In
many cases the pair spacing is not the closest possible spacing based on DFM
rules, but rather spacing is defined to acheive the desired differential
impedance in combination with trace widths within the manurfacturable trace
geometry window of the board, while providing an acceptable level of loss,
and a reasonable degree of coupling. Its just not possible to give a rule of
thumb in my opinion. Given a choice it is nice to have the overall geometry
as compact as possible for routing purposes, especially for wide internal
interfaces and busses, but in the case of individual I/O interfaces this is
not such an issue. Hope that helps some.

Brian P. Moran
Signal Integrity Engineer
Intel Corporation
brian.p.moran@xxxxxxxxx


-----Original Message-----
From: Goutham.S@xxxxxxxxxx [mailto:Goutham.S@xxxxxxxxxx]
Sent: Tuesday, October 09, 2001 9:18 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] differential signals spacing






Hi All,

In the differential trace routing (ethernet, scsi, USB etc.) , Routing
guidelines mentions that the differential pairs should be routed as close as
possible to
effectively utilise the differential signal properties (Noise Cancellation).
They don't mention the spacing in terms of mils or inches.

We were earlier giving the spacing as per the PCB designer's convenience.
The
smallest spacing used is 5mils.
As far as Manufacturering technology is concerned, spacing can be 2.5 mil
and
above.

My doubt is that from Electrical Engineers point of view, what is the
spacing
requirement for differential signals.

Regards
Goutham Sabavat
Force computers India Pvt Ltd


------------------------------------------------------------------
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:
http://www.freelists.org/webpage/si-list

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

List archives are viewable at:     
                http://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:
http://www.freelists.org/webpage/si-list

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

List archives are viewable at:     
                http://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: