[SI-LIST] Re: Split GND Plane on PKG
- From: "Geoff Stokes" <GStokes@xxxxxxxxx>
- To: <Alan.Hiltonnickel@xxxxxxx>, <jbtera77@xxxxxxxxx>
- Date: Thu, 8 Jun 2006 10:43:24 +0100
Hello Jongbae
I think Alan is correct in saying you need adequate analysis. However I
think your question is more fundamental: why split ground on the chip
and why split ground on the PCB?
In many cases, there is no need to split the ground on the PCB, but
sometimes there is a need for an analog or converter device on that PCB
to use separate ground connections. Let's look at such a case.
A device with an input and output is sensitive to the input voltage
measured with respect to one of its ground return terminals. The output
drives a current, which flows through a bond wire. The return current
also flows through a bond wire, which has impedance to the signal,
especially the high frequency components of the signal. If the output
and input share the same ground terminal, the voltage drop in the ground
connection impedance is coupled back to the input. This is called
common impedance coupling.=20=20
This leads to the idea that input ground connections should be separated
from output ground. Most designers of high speed circuits will provide
separate chip bond connections, in cases where they expect there will be
an advantage from doing this. (The expectation may come from
calculation, simulation or experience.)=20=20
In some cases, similar arguments apply to the power supplies, which may
require multiple connections to the same supply, which can be by-passed
at the PCB, but by-passing on the chip can only be done in a very
limited way or sometimes it is not practical at all.
On the PCB a similar argument can be applied, but in many cases there is
more advantage to maintaining a solid ground plane. In that case the
split chip grounds we just mentioned are connected separately to the
same PCB ground. There is a need to split the PCB ground plane only in
cases where there is a special sensitivity to circulating ground
currents in the system. For example some oscilloscope input PCB
circuits have used a slot between input channels to achieve very high
analog signal isolation. In that case there may be a need to look at
one signal of several volts and compare with a signal of a few
millivolts, so an isolation of some 80 dB may be required. This is an
extreme case to illustrate the point. There will be various sensor
systems for example which may also require some similar treatment.=20
Geoff Stokes
Systems Engineer
Zetex Semiconductors plc
Zetex Technology Park
Chadderton
Oldham
OL9 9LL
UK
=20
+44-161-622-4857
www.zetex.com
www.zetex.cn
=20
=20
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx]
On Behalf Of Alan.Hiltonnickel@xxxxxxx
Sent: 07 June 2006 19:01
To: jbtera77@xxxxxxxxx
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Split GND Plane on PKG
Jongbae,
Your blanket statement, "most package designers split their ground=20
planes in PKG" is not necessarily true, and is certainly not provable.
I can tell you what a GOOD package designer will do: they will structure
their power and ground carefully in the context of the silicon they are=20
connected to and the PCB environment they are planning to connect that=20
silicon to. I have been know to reconnect isolated ground planes that=20
others felt should be separate from regular "ground", with no ill=20
effect. I have also been known to be very particular about how a=20
separate ground is routed to provide a minimal noise coupling for, say,=20
a PLL.
A designer who splits a ground plane should provide an adequate analysis
(preferably with simulations) to back up such a drastic action, and=20
should also have a strategy to minimize the impact on high speed signals
and EMI.
Alan.
jbtera77@xxxxxxxxx wrote On 06/06/06 22:29,:
>Hi, all
>As i know, most PKG designers split their ground plane in PKG. It is
>usually for seperation of digital with analog. However, i think, even
if we
>split GND on PKG, once the package is connected to board by BGA or
>wirebonding, the effect of split-GND on PKG will not be appear.
Instead, the
>board design is much more dominate than pkg p/g design.
>So, in my opinion, we don't have to split its ground plane on PKG
because it
>doesn't affect to overall p/g design including board (maybe a little)
and
>splitting the ground plane may cause the signal integrity problem.
>
>Could you please let me know the reason designers split the ground
plane in
>package.
>
>rgs,
>Jongbae.
>
>=20=20
>
--=20
Alan Hilton-Nickel
Signal Integrity Engineer
Sun Microsystems Inc.
Netra Systems and Networking
Newark, CA
------------------------------------------------------------------
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:=20=20=20=20=20
//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
=20=20
_________________________________________________________
Zetex Semiconductors - Solutions for an analog world.
http://www.zetex.com
http://www.zetex.cn
E-MAILS are susceptible to interference. You should not assume that
the contents originated from the sender or the Zetex Group or that they=20
have been accurately reproduced from their original form.
Zetex accepts no responsibility for information, errors or omissions in
this e-mail nor for its use or misuse nor for any act committed or
omitted in connection with this communication.
If in doubt, please verify the authenticity with the sender.
_________________________________________________________
=20
------------------------------------------------------------------
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:
