[SI-LIST] Re: Fiber weave effect modeling: Stack of materials ...
- From: JASON MILLER <jason.miller@xxxxxxxxxx>
- To: Lee Ritchey <leeritchey@xxxxxxxxxxxxx>
- Date: Wed, 10 Aug 2011 10:37:24 -0400
I just want to mention that there are other considerations when it comes
to weave effects, such as periodic loading of the trace by the glass
weave, which don't get simply resolved by the first approach of routing
at an angle relative to the weave. There are a couple of papers in
recent years that I am aware of which detailed this:
http://www.electrical-integrity.com/Paper_download_files/DC10_7-WA1_Miller-Blando-Novak.pdf
http://www.founderpcb.com/upfile/File/2011/SI_Glassweave_Isola_DesignCon2011%2820110727%29.pdf
On the other hand, tightening the weave or spreading out the glass
fabric *will* improve the uniformity of the dielectric and mitigate the
effect of the glass-weave periodic loading.
Jason Miller
Oracle
Lee Ritchey wrote:
> I've been on vacation, so come late to this discussion. There are three
> glass weaves that are irregular enough to cause skew problems with
> differential pairs as has been shown in papers presented at DesignCon.
> These are 106 ,1080 and 7628. We've tried various ways of nesting multiple
> plies of these to avoid this problem. Doesn't seem to work all that well.
>
> There are some very simple fixes. One is to route the PCB traces at an angle
> to the weave or cock the PCB on the fabrication panel. Both work, but are
> either a hassle or expensive.
>
> A much simpler approach is to use a glass weave with uniformly distributed
> glass. These are readily available. 1067 replaces 106. 1086 replaces
> 1080. 3313 in two plies replaces 7628.
>
> We've done many tests and demonstrated that the skew problem goes away with
> the use of these weaves. I've got a paper that shows the difference between
> 1080 and 3313. It is dramatic!
>
> Lee
>
>
> --------------------------------------------------
> From: "Hermann Ruckerbauer" <hermann.ruckerbauer@xxxxxxxxxxxxx>
> Sent: Friday, August 05, 2011 1:44 AM
> To: <si-list@xxxxxxxxxxxxx>
> Subject: [SI-LIST] Fiber weave effect modeling: Stack of materials ...
>
>
>> Hello *,
>>
>> the original thread is already old, but I would have one more question
>> on this one:
>> Will stacks of multiple materials statistically help to reduce the effect
>> ?
>>
>> "Standard" 1080 looks not too good, but if 2x or 3x 1080 is stacked this
>> could statistically reduce the effect.
>> I would not expect that material production and manufacturing is so
>> accurate that a stack of 3 material will result that always the same
>> structures are overlayed.
>> I would more expect, that there might be a statistical distribution for
>> High volume manufacturing where a part of the final boards will have
>> this worst case, and on other there will be a statistical distribution
>> of different combinations how the materials are overlayed in the stack.
>>
>> Does anybody have more than a feeling on this assumption ?
>>
>> Thanks and regards
>>
>> Hermann
>>
>> EKH - EyeKnowHow
>> Hermann Ruckerbauer
>> www.EyeKnowHow.de
>> Hermann.Ruckerbauer@xxxxxxxxxxxxx
>> Veilchenstrasse 1
>> 94554 Moos
>> Tel.: +49 (0)9938 / 902 083
>> Mobile: +49 (0)176 / 787 787 77
>> Fax: +49 (0)3212 / 121 9008
>>
>>
>> schrieb Lee Ritchey:
>>
>>> The 1086 weave used in laser drilled PCBs is the replacement for 1080 and
>>> the 1067 weave is the replacement for 106 weave. They both look like the
>>> Nova product and are not subject to patents or single sourcing. 3313 is
>>> similar and yields a 4 mil core using a single ply of glass.
>>>
>>> From: bala
>>> Sent: Tuesday, January 18, 2011 12:42 AM
>>> To: Lenkisch, Andreas
>>> Cc: Lee Ritchey ; Loyer, Jeff ; si-list@xxxxxxxxxxxxx
>>> Subject: Re: [SI-LIST] AW: Re: AW: Fiber weave effect modeling
>>>
>>>
>>> http://bethesignal.net/blog/?pB
>>>
>>>
>>> On Tue, Jan 18, 2011 at 2:00 PM, Lenkisch, Andreas
>>> <Andreas.Lenkisch@xxxxxxxxxxx> wrote:
>>>
>>> that's really strange, I got the feedback from the PCB shop (Europe)
>>> that this material is "quite expensive" (about two times more than
>>> traditional glass weave). The answer is already half a year old. I will
>>> ask again.
>>>
>>> Andreas
>>>
>>>
>>>
>>> -----Ursprüngliche Nachricht-----
>>> Von: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx]
>>> Im Auftrag von Lee Ritchey
>>> Gesendet: Montag, 17. Januar 2011 19:42
>>> An: Loyer, Jeff; si-list@xxxxxxxxxxxxx
>>> Betreff: [SI-LIST] Re: AW: Fiber weave effect modeling
>>>
>>>
>>> One thing I forgot to mention in my last response was that the reason
>>> we
>>> switched to 3313 weave was cost reduction, not signal integrity. Prior
>>> to
>>> this material becoming available we achieved 4 mil cores using two
>>> plies of
>>> glass cloth. Our fabricator suggested we switch to 3313 and achieve a
>>> lower
>>> price. I'm not sure why some fabricators would suggest the PCB would
>>> cost
>>> more.
>>>
>>> In the bargain, we got the flat weave and much better impedance
>>> profiles as
>>> well as far lower differential skew.
>>>
>>> My guess is you won't get the new weaves unless you insist on them.
>>> Fabricators don't like to expand their inventories unless they are
>>> forced
>>> to. Guess that is like all manufacturers!
>>>
>>> --------------------------------------------------
>>> From: "Loyer, Jeff" <jeff.loyer@xxxxxxxxx>
>>> Sent: Thursday, January 13, 2011 1:31 PM
>>> To: "Lee Ritchey" <leeritchey@xxxxxxxxxxxxx>; <si-list@xxxxxxxxxxxxx>
>>> Subject: [SI-LIST] Re: AW: Fiber weave effect modeling
>>>
>>> > I don't think the elimination of standard weaves is straightforward.
>>> I
>>> > wholeheartedly agree that "flat" weaves exist and are a very
>>> attractive
>>> > solution but, at the time we wrote our paper, they cost about 2x that
>>> of
>>> > standard material. I don't know if the difference is still that
>>> high, but
>>> > I doubt it's insignificant.
>>> > For many designs, the cost differential is outweighed by the
>>> benefits.
>>> > For others, it is not. 10 degree routing, ugly as it may appear and
>>> as
>>> > time consuming as it is, can be attractive if it saves significant
>>> money.
>>> >
>>> > On the other hand...
>>> > There may soon come a point where bus speeds increase such that it is
>>> > impossible to avoid routing parallel to the board edge for
>>> problematic
>>> > distances. At that point, flat weaves will be a more palatable
>>> option.
>>> > For instance, just breaking out of a large device plus routing into a
>>> > connector might require 2" of length that can't be angled. For a
>>> 40GT/s
>>> > bus, that's probably unacceptable. Then, the choice gets clearer.
>>> >
>>> > For now, many of us are in the grey area where the option of using
>>> flat
>>> > weaves, and getting rid of our funky angled routing, isn't clear-cut.
>>> >
>>> > Jeff Loyer
>>> >
>>> >
>>> > -----Original Message-----
>>> > From: si-list-bounce@xxxxxxxxxxxxx
>>> [mailto:si-list-bounce@xxxxxxxxxxxxx]
>>> > On Behalf Of Lee Ritchey
>>> > Sent: Thursday, January 13, 2011 10:22 AM
>>> > To: Havermann, Gert; si-list@xxxxxxxxxxxxx
>>> > Subject: [SI-LIST] Re: AW: Fiber weave effect modeling
>>> >
>>> > I recommend you use neither of those weaves with high speed serial
>>> links
>>> > due
>>> > to there negative effect on skew. The best weaves are known as
>>> "flat"
>>> > weaves. The best are 1067 (replaces 106), 1086 (replaces 1080), 2113
>>> and
>>> > 3313. These are so uniform you don't need to worry about weave
>>> effects.
>>> > These weaves were developed to make laser drilling blind vias more
>>> uniform
>>> > and happen to be great for SI purposes!
>>> >
>>> > You also don't need to route your PCBs on a 15 degree angle to the
>>> weaves,
>>> > which is painful to do and wastes materials, so long as you stick
>>> with the
>>> > weaves listed above.
>>> >
>>> > Lee Ritchey
>>> >
>>> > --------------------------------------------------
>>> > From: "Havermann, Gert" <Gert.Havermann@xxxxxxxxxxx>
>>> > Sent: Wednesday, January 12, 2011 6:31 AM
>>> > To: <si-list@xxxxxxxxxxxxx>
>>> > Subject: [SI-LIST] AW: Fiber weave effect modeling
>>> >
>>> >> this is a great paper, many thanks for sharing it with the si-list.
>>> >>
>>> >> please allow me a question. I understand that the 106 and 7628
>>> Prepregs
>>> >> are used to predict the dk values of "pure epoxy" and "close
>>> proximity to
>>> >> the Glass bundle".
>>> >> Do you think that the "pure epoxy" value is always the worst case
>>> that I
>>> >> have to expect for my diff pair? If I (for instance) would only use
>>> 7628
>>> >> Style everywhere, will there even be areas "in pure epoxy", or is
>>> the
>>> >> weave dense enough that the worst dk is somewhere between the
>>> calculated
>>> >> min. and max. value?
>>> >>
>>> >> BR
>>> >> Gert
>>> >>
>>> >>
>>> >>
>>> --------------------------------------------------------------------------
>>> >> Absender ist HARTING Electronics GmbH & Co. KG; Sitz der
>>> Gesellschaft:
>>> >> Espelkamp; Registergericht: Bad Oeynhausen; Register-Nr.: HRA 5596;
>>> >> persönlich haftende Gesellschafterin: HARTING Electronics Management
>>> >> GmbH;
>>> >> Sitz der Komplementär-GmbH: Espelkamp; Registergericht der
>>> >> Komplementär-GmbH: Bad Oeynhausen; Register-Nr. der
>>> Komplementär-GmbH:
>>> >> HRB
>>> >> 8808; Geschäftsführer: Edgar-Peter Duening, Torsten Ratzmann, Dr.
>>> >> Alexander Rost
>>> >> -----Ursprüngliche Nachricht-----
>>> >>
>>> >> Von: si-list-bounce@xxxxxxxxxxxxx
>>> [mailto:si-list-bounce@xxxxxxxxxxxxx]
>>> >> Im
>>> >> Auftrag von Bert Simonovich
>>> >> Gesendet: Freitag, 7. Januar 2011 21:55
>>> >> An: si-list@xxxxxxxxxxxxx
>>> >> Betreff: [SI-LIST] Fiber weave effect modeling
>>> >>
>>> >> Hi all,
>>> >>
>>> >> Recently there were discussions on PCB fiber weave effect. I
>>> recently did
>>> >> a study and published a White Paper titled, "Practical Fiber Weave
>>> Effect
>>> >> Modeling".
>>> >>
>>> >> Abstract:
>>> >> Fiber weave effect is becoming more of an issue as bit rates
>>> continue to
>>> >> sore upwards to 5GB/s and beyond. Due to the non-homogenous nature
>>> of
>>> >> printed circuit board laminates, the fiberglass weave pattern causes
>>> >> signals to propagate at different speeds within differential pair
>>> traces;
>>> >> causing timing skew and mode conversion at the receiver; leading to
>>> >> reduced bit-error-rate (BER) performance; and increased EMI
>>> radiation.
>>> >> The
>>> >> relative dielectric constant (Dk) surrounding a trace ultimately
>>> >> determines its propagation delay. This paper delves into the issue
>>> and
>>> >> presents a novel approach to practically establish worst case
>>> min/max
>>> >> values for Dk and use them to model this effect using ADS circuit
>>> >> modeling
>>> >> software. A PCIe CEM
>>> >> Rev2 case study is used to practically demonstrate the model and to
>>> >> explore the design space.
>>> >>
>>> >> Here is the link: http://lamsimenterprises.com/White_Papers.html
>>> >>
>>> >> Thanks to Jeff Loyer, Istvan Novak and Gustavo Blando for there help
>>> in
>>> >> clarifying some results of their prior published work on the
>>> subject.
>>> >>
>>> >> I hope you find it useful.
>>> >>
>>> >> -Bert
>>> >>
>>> >> Lambert (Bert) Simonovich
>>> >> Consultant and Founder
>>> >> LAMSIM Enterprises Inc.
>>> >> Web Site: http://lamsimenterprises.com
>>> >> Blog: http://blog.lamsimenterprises.com/
>>> >>
>>> >>
>>> >>
>>> >>
>>> >> ------------------------------------------------------------------
>>> >> 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
>>>
>>>
>>>
>>>
>>> Schroff GmbH
>>> Langenalber Str. 96-100
>>> D-75334 Straubenhardt
>>> Amtsgericht Mannheim, HRB 503549
>>> Geschäftsführer: Walter Kritikos, Daniel Stirpe
>>>
>>>
>>> Important Notice
>>>
>>> The information contained in this message is confidential and is
>>> intended for the addressee only.
>>>
>>> The unauthorised use, disclosure, copying or alteration of this message
>>> is prohibited and may be unlawful. If you received this message in error
>>> or there are any problems please notify the originator immediately and
>>> contact postmaster@xxxxxxxxxxxxxxxxxxxxxx This mail and any attachments
>>> have been scanned for viruses prior to leaving Pentair's network. Pentair
>>> will not be liable for direct, special, indirect or consequential damages
>>> arising from alteration of the contents of this message by a third party
>>> or as a result of any virus being passed on. All personal views stated in
>>> this message are my own and may not represent the position of, nor is
>>> this message binding upon, Pentair Inc. or any of its subsidiaries.
>>>
>>> ------------------------------------------------------------------
>>> 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
>
>
>
--
Oracle <http://www.oracle.com>
JASON R. MILLER | PRINCIPAL HARDWARE ENGINEER
PHONE: 781.442.2774 | MOBILE: 617.548.7768
35 NETWORK DRIVE | BURLINGTON, MA 01803
------------------------------------------------------------------
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:
