[SI-LIST] Re: Maximum Frequency for S-parameter Extraction of Board Power Plane

  • From: Scott McMorrow <scott@xxxxxxxxxxxxx>
  • To: si-list@xxxxxxxxxxxxx
  • Date: Thu, 19 May 2011 10:34:30 -0400

Rohit

To follow along in Steve's thinking I would solve a power system up to 5 
to 10 times Fknee ( .35/edge rate) of the fastest switching edge rate 
that the packages sees at the die interface.  For core power  it would 
depend on the die cutoff.  Lets say was 100 MHz, then I'd model the 
package to at least 1 GHz.  If it was a DDR IO plane, then I'd use 10 
times the DDR switching frequency, or 5 times Fknee of the signal rise time.

For DDR3 1333, 10 times the switching frequency is 6.67 GHz.  If the 
signal rise time were 150 ps, then Fknee is 2.3 GHz. Five times that is 
11.7 GHz.  For goodness, I'd extend my modeling bandwidth to 20 GHz, and 
have a model that matches the full measurement bandwidth of a 20 GHz 
VNA.  Usually Fknee of the signal rise/fall time wins, making an 
extraction bandwidth of between 10 GHz and 20 GHz for parallel 
single-ended I/O systems in packages.  Tr/Tf should be based on the rise 
and fall time of the fastest driven or received signal at the package.  
DDR memory devices are notorious for their fast edge rates.

regards,

Scott



Scott McMorrow
Teraspeed Consulting Group LLC
121 North River Drive
Narragansett, RI 02882
(401) 284-1827 Business
(401) 284-1840 Fax

http://www.teraspeed.com

Teraspeed® is the registered service mark of
Teraspeed Consulting Group LLC


On 5/19/2011 8:38 AM, steve weir wrote:
> As in many things in SI/PI:  it depends.
>
> If you are trying to come up with a hard and fast number, then the safe
> one will be very (and for most cases dramatically overly) conservative,
> such as 10X Fknee.   That would put you down more or less 40dB from the
> knee frequency, and would PROBABLY be very low in most situations.  If
> you want to figure out a number for a given situation then you need some
> sense for the total power involved and what you need out of the package
> you are trying to design or evaluate.
>
> Steve
>
> On 5/19/2011 5:07 AM, Rohit MISHRA wrote:
>> Steve,
>>
>> Thanks for your prompt reply !!
>>
>> I agree that parallel lots of signals will make very effective radio 
>> transmitter even beyond Fknee but do we need to consider this during 
>> s-parameter extraction of power plane and If your answer is yes, then how to 
>> chose a maximum frequency ??
>>
>> Rohit
>>
>>
>>
>>
>>
>> -----Original Message-----
>> From: steve weir [mailto:weirsi@xxxxxxxxxx]
>> Sent: Thursday, May 19, 2011 5:19 PM
>> To: Rohit MISHRA
>> Cc: si-list@xxxxxxxxxxxxx
>> Subject: Re: [SI-LIST] Re: Maximum Frequency for S-parameter Extraction of 
>> Board Power Plane
>>
>> On 5/19/2011 4:28 AM, Rohit MISHRA wrote:
>>> Steve,
>>>
>>> Why do you think you can ignore the package in the frequency ranges that 
>>> you consider the PCB for power delivery?
>>>
>>> I think I was wrong for minimum frequency, As PCB power plane is basically 
>>> a low pass filter so package will also see all low frequency components 
>>> down to 0 Hz, right ?
>> Yes, the package is in series with the PCB.
>>> Why do you think that Fknee is the highest frequency you need to consider?
>>>
>>> Fknee (0.35/Trise) is the maximum frequency where most of the energy of the 
>>> signal exist so we should care bandwidth till knee frequency.
>>>
>>> This is my understanding, I may be wrong but nevertheless I would love to 
>>> hear your views.
>> If you only care about the transmission path for one signal, Fknee gives
>> an APPROXIMATE bandwidth beyond which arbitrary channel response will
>> not introduce MUCH signal distortion.   Parallel lots of signals and
>> even though each signal has only a small percentage power beyond Fknee,
>> the total power beyond Fknee can be substantial and make a very
>> effective radio transmitter.
>>> Rohit
>>>
>>>
>>>
>>> -----Original Message-----
>>> From: steve weir [mailto:weirsi@xxxxxxxxxx]
>>> Sent: Thursday, May 19, 2011 1:56 PM
>>> To: Rohit MISHRA
>>> Cc: si-list@xxxxxxxxxxxxx; Muranyi, Arpad
>>> Subject: Re: [SI-LIST] Re: Maximum Frequency for S-parameter Extraction of 
>>> Board Power Plane
>>>
>>> Why do you think you can ignore the package in the frequency ranges that
>>> you consider the PCB for power delivery?
>>> Why do you think that Fknee is the highest frequency you need to consider?
>>>
>>> Steve.
>>> On 5/19/2011 12:26 AM, Rohit MISHRA wrote:
>>>> Guys,
>>>>
>>>> Interesting discussion, Indeed !!
>>>>
>>>> Now in case if we extract power plane s-parameter for package, what would 
>>>> be the minimum&     maximum frequency ??
>>>>
>>>> As per my understanding :
>>>>
>>>> Minimum frequency = 3 X cut off frequency of board
>>>> Maximum Frequency = Knee frequency
>>>>
>>>> Experts comments please.
>>>>
>>>> Rohit
>>>>
>>>>
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] 
>>>> On Behalf Of Muranyi, Arpad
>>>> Sent: Wednesday, May 18, 2011 10:12 PM
>>>> To: si-list@xxxxxxxxxxxxx
>>>> Subject: [SI-LIST] Re: Maximum Frequency for S-parameter Extraction of 
>>>> Board Power Plane
>>>>
>>>> I might add that one has to make sure that the S-parameter data
>>>> goes down low enough in frequency, preferably to 0 Hz, otherwise
>>>> the simulators will have to guess on how to solve for the DC
>>>> components of the signal...  Based on the "your guess is as good
>>>> as mine" principle your results may vary...
>>>>
>>>> Arpad
>>>> ==================================================================
>>>>
>>>> -----Original Message-----
>>>> From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx]
>>>> On Behalf Of Beal, Weston
>>>> Sent: Wednesday, May 18, 2011 11:32 AM
>>>> To: Rajan Hansa; si-list@xxxxxxxxxxxxx
>>>> Subject: [SI-LIST] Re: Maximum Frequency for S-parameter Extraction of
>>>> Board Power Plane
>>>>
>>>> Make sure that you consider not only the maximum frequency. You also
>>>> need to make sure that your S-parameter set has enough data points to
>>>> correctly model all the resonances in the measured frequency band. If
>>>> you have too few data points, and the graph of the S-parameters over
>>>> frequency looks choppy or jagged, then the model will probably be
>>>> non-causal.
>>>>
>>>> S-parameters are like power tools. A skilled user can get a lot done
>>>> with them, but if used carelessly, they can hurt you.
>>>>
>>>> Weston
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx]
>>>> On Behalf Of Rajan Hansa
>>>> Sent: Wednesday, May 18, 2011 3:44 AM
>>>> To: si-list@xxxxxxxxxxxxx
>>>> Subject: [SI-LIST] Maximum Frequency for S-parameter Extraction of Board
>>>> Power Plane
>>>>
>>>> Hello All,
>>>>
>>>>
>>>> I have got a PCB database of one Ethernet Board and I have to extract
>>>> s-parameter for all power plane from voltage regulator to  power pin of
>>>> chip package. The Board database includes all decap.
>>>>
>>>>
>>>>
>>>> As we know extraction time is directly related with maximum frequency So
>>>> while setting maximum frequency for s-parameter extraction, I am bit
>>>> confused, I would like to know that do I need to set knee frequency (
>>>> Fknee = 0.35/Trise) as maximum frequency OR for board level simulation
>>>> we shouldn't care knee frequency.
>>>>
>>>>
>>>>
>>>> As per my understanding, package inductance won't allow very high
>>>> frequency component (switching current component which depends on Trise
>>>> !!) to travel to board from chip so voltage regulator in board will
>>>> never see those high switching current so I shouldn't simulate/extract
>>>> s-parameter till knee frequency.
>>>>
>>>>
>>>>
>>>> What's expert view on this ?
>>>>
>>>>
>>>>
>>>> Rgds,
>>>>
>>>> Rajan
>>>>
>>>> ------------------------------------------------------------------
>>>> 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
>>>>
>>>>
>>>> ------------------------------------------------------------------
>>>> 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
>>>>
>>>>
>>>> ------------------------------------------------------------------
>>>> 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
>>>>
>>>>
>>>> ------------------------------------------------------------------
>>>> 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
>>>>
>>>>
>>>>
>
------------------------------------------------------------------
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: