[SI-LIST] Re: DC Drop in PDN simulation
- From: Nithin Kumar <mnithink@xxxxxxxxx>
- To: Feng Wu <wufengthu@xxxxxxxxx>
- Date: Sun, 1 Sep 2013 22:34:46 +0530
Hi Feng,
Yes I'm shorting the VRM port of the PCB model during Z11 analysis.
During transient analysis,I'm connecting an Ideal voltage supply to VRM
port of the PCB s-parameter model.
Sure, as you suggested I will create a lumped model of the PCB and do a
time domain analysis.
Regards,
Nithin
On Sun, Sep 1, 2013 at 8:27 PM, Feng Wu <wufengthu@xxxxxxxxx> wrote:
> [image: Boxbe] <https://www.boxbe.com/overview> Feng Wu (
> wufengthu@xxxxxxxxx) is not on your Guest
> List<https://www.boxbe.com/approved-list?tc_serial025761368&tc_rand99260622&utm_source=stf&utm_medium=email&utm_campaign=ANNO_MWTP&utm_content>
>
> sender<https://www.boxbe.com/anno?tc_serial025761368&tc_rand99260622&utm_source=stf&utm_medium=email&utm_campaign=ANNO_MWTP&utm_content>
>
> domain<https://www.boxbe.com/anno?tc_serial025761368&tc_rand99260622&utm_source=stf&utm_medium=email&utm_campaign=ANNO_MWTP&utm_content>
> Hi Kumar
>
> I assume you short VRM or use a low impedance VRM model, when you check
> impedance (1mohm as you mentioned), in another word,you check the low
> frequency point/DC point of Z11 of the one-port S parameter. I agree that
> this is very helpful to dig out the problems.
>
> I would propose you.
> Replace the S parameter model in the deck, for example PCB/PKG, with a
> simplified /test SPICE netlist. then check the DC drop still exist or not.
> by this way to find which part has the problem.
>
> If the problem is from the S parameter, I would propose you don't short
> the VRM and look at the 2 port S parameter, to see whether the Z11/Z12/Z22
> are high impedance or not.
>
> as Leslie mentioned, S parameter has some problems due to some reason.
> I ever saw Z11 of one port S parameter model show low impedance, while
> the problem was hidden that the power and ground are wrongly low
> impedance/shorted in pkg/pcb in case of open VRM. this will also result in
> the DC drop in the transcient simulation.
>
> Thanks
> Feng
> 2013/8/30 Nithin Kumar <mnithink@xxxxxxxxx>
>
>> Hi Matthew,
>> Thanks.
>> The reference impedance used for extracting s-parameter models is
>> 0.01mohms.
>> I did a freq domain analysis (Z11) connecting die,package and pcb models.
>> The response looks alright with low impedance( of around 1 mohm) near DC.
>> so I would expect a very low DC offset but this isn't the case. I'm not
>> able to figure out the possible reason.
>> Regards,
>> Nithin
>>
>>
>> On Fri, Aug 30, 2013 at 8:14 PM, matthew leslie <mbleslie@xxxxxxxxx>
>> wrote:
>>
>> > [image: Boxbe] <https://www.boxbe.com/overview> matthew leslie (
>> > mbleslie@xxxxxxxxx) is not on your Guest List<
>> https://www.boxbe.com/approved-list?tc_serial012227189&tc_randH9907851&utm_source=stf&utm_medium=email&utm_campaign=ANNO_MWTP&utm_content>
>> sender<https://www.boxbe.com/anno?tc_serial012227189&tc_randH9907851&utm_source=stf&utm_medium=email&utm_campaign=ANNO_MWTP&utm_content>
>> domain<https://www.boxbe.com/anno?tc_serial012227189&tc_randH9907851&utm_source=stf&utm_medium=email&utm_campaign=ANNO_MWTP&utm_content>
>> > In general, s-parameter models are not reliable below a certain
>> frequency
>> > range. It depends on the extraction tool, but typically that range is
>> > around 1-100MHz. Sometimes the tool refuses to extract below a certain
>> > frequency (I believe HFSS does this). Sometimes the tool will agree to
>> > generate a model for low frequencies, but that model is erroneous. For
>> > example, the tool might simply extrapolate "accurate" data in the
>> 1-10MHz
>> > range down to kHz and DC. Other tools switch the field solver techniques
>> > when transitioning to lower frequencies and as a result end up creating
>> > models with discontinuities in insertion/return loss.
>> >
>> > There are some steps you can take to better ensure accuracy at low freq
>> > for s-param model. Make sure you are using a low reference impedance,
>> such
>> > as 1-ohm or 0.1ohm. Also, if your solver permits, try forcing a
>> separate DC
>> > solution into s-param model. Unfortunately, this doesn't guarantee
>> accuracy
>> > from DC up to the 1-100MHz range.
>> >
>> > S-parameter models that are accurate from DC to multi-GHz (as far as I
>> > know) is a big challenge. If I were you, I might move to creating an
>> RLGC
>> > or other type of lumped element model for more accurate low frequency
>> > representation.
>> >
>> >
>> > On Fri, Aug 30, 2013 at 8:26 AM, Nithin Kumar <mnithink@xxxxxxxxx>
>> wrote:
>> >
>> >> Hello Experts,
>> >> I'm doing PDN simulation using DIE model ( from Redhawk), PCB model and
>> >> Package model (s-parameter)
>> >> During transient analysis I'm seeing huge drop at the bga balls as
>> well as
>> >> die bump.
>> >> I have verified the resistance values of package and PCB ( which are
>> very
>> >> small) and varied DIE R as well. But i'm not able to reduce this huge
>> DC
>> >> offset.
>> >> can anybody through some light on possible reasons for such an offset?
>> >>
>> >> --
>> >> -------------------
>> >> Regards,
>> >> Nithin Kumar
>> >>
>> >>
>> >> ------------------------------------------------------------------
>> >> 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 forum is accessible at:
>> >> http://tech.groups.yahoo.com/group/si-list
>> >>
>> >> 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
>> >>
>> >>
>> >>
>> >
>> >
>> > --
>> > Matthew B. Leslie
>> > mbleslie@xxxxxxxxx
>> >
>> >
>>
>>
>> --
>> -------------------
>> Regards,
>> Nithin Kumar
>>
>> ------------------------------------------------------------------
>> 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 forum is accessible at:
>> http://tech.groups.yahoo.com/group/si-list
>>
>> 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
>>
>>
>>
>
>
> --
> B.R.
> Feng Wu (å´æ«)
>
> -------------
> Signal Integrity Engineer
> Cisco Shanghai
> -------------
>
>
--
-------------------
Regards,
Nithin Kumar
------------------------------------------------------------------
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 forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
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:
