[SI-LIST] Re: Seperate Sparameter for gnd & supply plane !!
- From: Larry Smith <LSMITH@xxxxxxxxxx>
- To: 'steve weir' <weirsi@xxxxxxxxxx>, 'Rajan Hansa' <all.si.list@xxxxxxxxx>
- Date: Wed, 1 Jun 2011 10:04:23 -0700
Rajan - There are some very basic principles that should be considered here.
To the little circuits on the die, the only PDN voltage that is important is
the difference between Vcc and Vss locally at the circuit terminals. The
circuits respond to this voltage and they do not care about the voltage on the
PCB ground plane or the voltage at the center of the earth or any other place.
An S parameter model of the PDN is fully capable of giving Vcc-Vss at the
circuit terminals in a properly set up simulation. The simulation may show you
local Vcc-Vss (differential) at several points in the system (i.e. bump joints,
wire bond pads, package-via tops and bottoms, package balls, pcb vias, pcb
power planes, etc.). At each of these points in the system, the only thing
that is important is the difference between Vcc and Vss voltage. Consider this
the "TEM" for the PDN if you like. The voltage along the ground path is not
defined (voltages are not unique and depends greatly upon the path of
integration). The structures are usually large enough that voltage along a
ground path distance should be considered voltage across time which is not well
defined. S parameters cut through all that and give you the impedance of some
port when a different port is excited, always with respect to a local reference
node.
I've seen a lot of analog designers get really confused about this and try to
talk about "ground noise." But the voltage on a node (particularly ground) is
not defined (what would you measure it with respect to?). Voltage is always
the difference in potential between two nodes (i.e. Vcc-Vss) and is clearly
defined when you can identify a parallel plate capacitor to integrate the E
field through a distance (volts/meter x meters). This is valid when there are
no time varying magnetic fields in the picture. All voltage bets are off when
you get into typical packaging structures where there certainly are time
varying magnetic fields and the integral of dB/dt through a loop area is
important. Once again, the local Vcc-Vss voltage is still valid all along the
PDN path (TEM mode) but voltage along the path (and therefore ground noise) is
not defined. An S parameter model simulation that has a port for Vcc wrt it's
local Vss will tell you everything you need to know about the power quality for
the local circuit.
Now, what about the signals that come away from the die circuit and follow some
path down to a PCB trace? As long as the signals are referenced to ground
along the entire path, the S parameter representation of that path is good.
But if the signal finds itself referenced to Vcc or some other node along the
way (return current is on something other than continuous ground), then you
have to be real careful about mode conversion when the signal return-current is
forced to jump from one reference node to another. Some people will call this
ground bounce but I prefer to think of it as poorly referenced signals.
I hope this does not get you into trouble with your analog engineer. These are
basic power integrity concepts that need to be resolved anytime we talk about
power quality and signal referencing. Any attempt to get S parameters for
ground will simply lead to confusion and trouble.
Regards,
Larry Smith
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of steve weir
Sent: Wednesday, June 01, 2011 4:34 AM
To: Rajan Hansa
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Seperate Sparameter for gnd & supply plane !!
Don't suppose. Analyze. Your real problem is that you analog engineer
wants to understand the extent of noise disturbance that is associated
with the interconnects and PDN. He first needs to define his problem by
deciding for each victim what location in the assembly constitutes his
reference. Once that is done, the problem can be solved with analytical
tools. An equipotential if it exists somewhere convenient reduces the
effort required to get a reasonable answer.
Steve
On 6/1/2011 4:23 AM, Rajan Hansa wrote:
>
> Steve,
>
> The board has many chips including ours and they all have different
> current requirement so it won't be correct to consider it
> equipotential and if I suppose consider them quasi-equipotential, what
> extra things I have to take care while generating s-parameter ?? OR do
> you mean to say for quasi-equipotential, separate vdd & gnd
> s-parameters will take care our problem ?
>
> Rajan
>
> On Wed, Jun 1, 2011 at 4:10 PM, steve weir <weirsi@xxxxxxxxxx
> <mailto:weirsi@xxxxxxxxxx>> wrote:
>
> You have a matrix of current sources attached to a matrix of RLCG
> legs, and you want to know the noise voltage that occurs between
> different point pairs in that matrix. If the product of the
> static and dynamic currents against a section of those legs is
> very small, you can treat that region as a quasi- equipotential.
> Otherwise you have to treat the point pairs as unique.
>
> Steve
>
>
> On 6/1/2011 3:29 AM, Rajan Hansa wrote:
>>
>> Steve,
>>
>> You are right !! Our analog engineer basically wants to
>> see ground bounce w.r.t pcb ground which was not possible with
>> single s-parameter but you said that I have to do whole lot more
>> than just produce s-parameters between one port inside the
>> package and some other virtual port.
>>
>> Can you help me to understand that why separate s-parameter is
>> not sufficient to see ground bounce and what is the right way to
>> handle such problems ??
>> Rajan
>>
>>
>> On Wed, Jun 1, 2011 at 3:30 PM, steve weir <weirsi@xxxxxxxxxx
>> <mailto:weirsi@xxxxxxxxxx>> wrote:
>>
>> What it sounds like your analog engineer wants is to be able
>> to figure out how much bounce will occur in each plane
>> relative to some external reference, such as the PCB Vss
>> plane. In order to figure that out, you have to do a whole
>> lot more than just produce s-parameters between one port
>> inside the package and some other virtual port. Your analog
>> engineer needs to be more specific about what he wants to
>> figure out.
>>
>> Steve.
>>
>>
>> On 6/1/2011 1:41 AM, Rajan Hansa wrote:
>>
>> Guys,
>> I have got a request to provide sparameter for one
>> package but this time my
>> analog engineer wants seperate s-parameter for gnd&
>> supply plane. The
>> reason he gave is that with single sparameter for vdd&
>> gnd plane. He can't
>> see gnd noise as s-parameter is generated with gnd as
>> reference so it'll
>> always be seen as '0' but that problem won't come in
>> separate s-parameters
>> for gnd& supply planes.
>>
>> Though the tool I am using has an option to generate
>> sparameter with
>> reference to some virtual point and I can use that option
>> to generate
>> separate sparameters for gnd& supply but problem is that
>> I have not still
>> understood if using separate s-parameter is a right way
>> to do things and if
>> it's the standard practice in industry.
>>
>> Experts comments please.
>>
>> Rajan
>>
>>
>>
>> ------------------------------------------------------------------
>> To unsubscribe from si-list:
>> si-list-request@xxxxxxxxxxxxx
>> <mailto: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
>> <mailto:si-list-request@xxxxxxxxxxxxx> with 'help' in the
>> Subject field
>>
>>
>> List technical documents are available at:
>> http://www.si-list.net <http://www.si-list.net/>
>>
>> List archives are viewable at:
>> //www.freelists.org/archives/si-list
>>
>> Old (prior to June 6, 2001) list archives are viewable at:
>> http://www.qsl.net/wb6tpu
>>
>>
>>
>>
>>
>> --
>> Steve Weir
>> IPBLOX, LLC
>> 150 N. Center St. #211
>> Reno, NV 89501
>> www.ipblox.com <http://www.ipblox.com/>
>>
>> (775) 299-4236 Business
>> (866) 675-4630 Toll-free
>> (707) 780-1951 Fax
>>
>>
>>
>
>
> --
> Steve Weir
> IPBLOX, LLC
> 150 N. Center St. #211
> Reno, NV 89501
> www.ipblox.com <http://www.ipblox.com/>
>
> (775) 299-4236 Business
> (866) 675-4630 Toll-free
> (707) 780-1951 Fax
>
>
>
--
Steve Weir
IPBLOX, LLC
150 N. Center St. #211
Reno, NV 89501
www.ipblox.com
(775) 299-4236 Business
(866) 675-4630 Toll-free
(707) 780-1951 Fax
------------------------------------------------------------------
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
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Confidentiality Notice.
This message may contain information that is confidential or otherwise
protected from disclosure. If you are not the intended recipient, you are
hereby notified that any use, disclosure, dissemination, distribution, or
copying of this message, or any attachments, is strictly prohibited. If you
have received this message in error, please advise the sender by reply e-mail,
and delete the message and any attachments. Thank you.
------------------------------------------------------------------
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
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts: