[SI-LIST] Re: [SI-LIST]: Which tool is the best - LINPARdiscussion
- From: Paul Levin <levinpa@xxxxxxxxxxxxx>
- To: raj.raghuram@xxxxxxxxxxxxxxx
- Date: Wed, 21 Apr 2004 11:00:13 -0700
Dear Raj,
LINPAR v2 most definitely handles dielectric loss, permitting a separate
tan-delta for each layer.
Regards,
Paul
____________________
Raj Raghuram wrote:
> Michael,
>
> I do not have a good answer for you. All tools (commerial or otherwise)
> have limitations and it is useful to know where they breakdown.
> Unfortunately, this information is hard to get and this is what I was
> trying to provide. For this, sometimes a knowledge of the numerical
> techniques behind them helps, as you have said.
>
> I have only used the earlier version of LINPAR, and this was very good
> as many have pointed out. For typical traces on boards, it is adequate
> and the price is great. It is, for all practical purposes, like public
> domain software in this regard. On the other hand, if you have a
> commercial tool you know and trust and own, there may be no need to switch.
>
> Some areas ( I do not know about the new version ) where LINPAR may
> fail are:
>
> 1. If you are doing cables (say shielded twisted pair), it may be better
> to use a Finite Difference or Finite Element based more general 2-D
> software.
>
> 2. I do not think it handles dielectric loss.
>
> 3. It does not handle magnetic materials and probably neither do most
> commercial programs.
>
> 4. If accurate skin effect inductance and resistance are important,
> again, other programs which divide the area of cross-section into many
> fine filaments would be needed.
>
> 5. It does not produce a HSPICE w-element compatible model directly.
>
> 6. I have already pointed out a minor problem I had with non-zero
> forward crosstalk for a homogenous medium
>
> If time were not an issue, my opinion (and this can only be an opinion)
> is that it is best to start with a public domain tool or free demo
> version. After you get familiar with it and know its limitations, you
> can move to a commercial tool, if necessary. You will then be able to
> evaluate the commercial tool better.
>
> Warm Regards,
>
> Michael E. Vrbanac wrote:
>
>
>>Raj,
>>
>>I am not particularly well-versed in the intricacies of some of these
>>numerical
>>methods and I do appreciate your explanation. Such academic insight is
>>extremely helpful in developing the tools and explaining their processes and
>>how they might be better used.
>>
>>There are a few questions I'd like to ask. Since there's always pros and
>>cons with
>>just about everything, and very seldom is extreme accuracy warranted, I would
>>be interested in just how much these details might make a very real difference
>>in the outcome on any real design decision based on the data received from
>>them. How might these differences affect my design decisions and what sort
>>of errors might I need to deal with and would any other factors swamp out
>>those
>>differences when applied in an everyday design decision?
>>
>>For instance, I have used ApSimRLGC and got some phenomenally good results
>
>>from it on some pretty critical designs even multi-gigabit stuff. The TDR
>
>>measurements were almost dead on and the cross-sections confirmed the
>>structural definitions. I am hearing that in actual practice Linpar can do
>>this as
>>well. So where would this stuff break down and under what conditions and what
>>would we expect to see when it does? This sort of information is what I was
>>addressing earlier. Its ok to use a tool within its capabilities but when
>>you don't
>>know where that is (a lot of folks don't and are vulnerable), then there's
>>likely to be
>>trouble. A $100K tool used incorrectly can be easily be beat by one costing
>>$1K that is used within its capabilities.
>>
>>I think most of us would agree that up to some point these things really do
>>matter
>>but after that, the differences are insignificant or the end-results would
>>not be substantially
>>different any way you go. This gets to be important if the tool costs go
>
>>from 1%
>
>>accuracy at a few thousand dollars to 0.5% accuracy at $50K. Unless there
>>is some warrant for that additional accuracy, the extra half percent is not
>>likely
>>to be money well spent. What would you advise us that would keep us out of
>>the
>>"more dollars must be better" trap? Where is the "best bang for the buck"?
>>
>>Best Regards,
>>
>>Michael E. Vrbanac
>>
>>
>>------------------------------------------------------------------
>>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
>>
>>
>>
>>
>
>
--
Paul Levin
Senior Principal Engineer
Xyratex Storage Systems
------------------------------------------------------------------
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: