[SI-LIST] Re: AW: fiber weave effect
- From: "Istvan Nagy" <buenos@xxxxxxxxxxx>
- To: <jeff.loyer@xxxxxxxxx>, <Gert.Havermann@xxxxxxxxxxx>, <shlepnev@xxxxxxxxxxxxx>, <si-list@xxxxxxxxxxxxx>
- Date: Fri, 15 Aug 2014 09:04:53 -0700
Hi,
I agree it is statistical. How luckily or unluckily the traces are aligned
with the fiber weave.
With spread glass the chances or probability of unfortunate alignment are
reduced, but not eliminated.
The other thing is for product development we have to ensure a good
production yield, so we have to consider worst-case alignment, instead of a
statistical "mean". For hardware design everything is considered worst-case.
Some high-cost boards with high-cost components (some of our boards have
production cost of $3k-$6k with all soldered components) must have 99-100%
yield, so that have to be considered in the statistical assessment. Some
lower cost ($10-20) boards may afford 90% yield, if they can save big on
material cost.
Have you checked photos of the glass fabrics to see how much percentage of
the area is covered with 2-thread, 1-thread, and no-thread? Also some
materials can be classified as 1D spread, some better ones as 2D spread.
Please check my glass fabric photo collection in this document on the second
tab:
http://www.buenos.extra.hu/download/PCB_MATERIAL_LIBRARY.xls
Istvan Nagy
Principal HW Engineer
Fortinet, Sunnyvale
-----Original Message-----
From: Loyer, Jeff
Sent: Friday, August 15, 2014 8:22 AM
To: Gert.Havermann@xxxxxxxxxxx ; shlepnev@xxxxxxxxxxxxx ;
si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: AW: fiber weave effect
I agree with Gert. I'm also sure there is quite a bit of information about
the effect of spread glass on FWE (Lee Ritchey, Isola, NovaSpeed,
Compunetics?). And it seems intuitive that trace dimensions, relative to
the weave, would influence the skew predictably. To my thinking, having
traces very close together, such that their environment is more similar,
should reduce the skew (since the glass is typically a little larger than
the traces).
I would also stress the need for any FWE study to use many boards with many
instances of test structures on each board to accurately evaluate the
effect. They are very subject to random effects which cannot be controlled:
* Exact trace alignment to glass
* Glass alignment to panel edge
o May be different for warp vs. fill
* Wandering of glass
o Will be different for warp vs. fill
Without many samples, you may draw erroneous conclusions from anecdotal
evidence.
For instance, below is a plot of skew between the "p" and "n" of several
samples on our 10 original test boards (the 10" coupons). Note the
significant difference in skew, depending on the particular sample. If you
happened to measure "Bd10", you would conclude a maximum of 30ps of skew;
"Bd9" would only have 5ps of skew (these boards were part of the same build,
manufactured identically at the same time).
Here's the raw data, in case the figure doesn't come out (2 different
formats, raw text and original format):
0 Degree Rotation with soldermask
Bd1 Bd2 Bd3 Bd4 Bd5 Bd7 Bd8 Bd9 Bd10
Layer Orientation Length Sample Direction Polarity
Sample ID "skew (p2-p1)
(+ = m1 on left)"
1 V 10 01 1 1 LYR01OVL10S01D1 1 1
9 6 -1 -3 -3 5 -4
1 V 10 02 1 1 LYR01OVL10S02D1 5
0 -7 -2 3 3 5 -3 12
1 V 10 03 1 1 LYR01OVL10S03D1 0 2
8 5 -2 0 -4 4 -15
1 V 10 04 1 1 LYR01OVL10S04D1 6
1 -9 -3 5 0 5 -3 20
1 V 10 05 1 1 LYR01OVL10S05D1 -2 0
10 9 -3 0 -3 2 -20
1 V 10 06 1 1 LYR01OVL10S06D1 9
0 -9 -4 5 0 4 2 23
1 V 10 07 1 1 LYR01OVL10S07D1 -1 3
9 7 -2 0 -2 0 -22
1 V 10 08 1 1 LYR01OVL10S08D1 8
1 -5 -4 5 -1 3 3 29
1 V 10 09 1 1 LYR01OVL10S09D1 -4 -2
5 7 -4 3 -3 -1 -26
1 V 10 10 1 1 LYR01OVL10S10D1 10
-3 -2 4 -3 2 6 26
0 Degree Rotation with soldermask Bd1 Bd2 Bd3 Bd4 Bd5
Bd7 Bd8 Bd9 Bd10
Layer Orientation Length Sample Direction Polarity
Sample ID skew (p2-p1)
(+ = m1 on left)
1 V 10 01 1 1 LYR01OVL10S01D1 1 1
9 6 -1 -3 -3 5 -4
1 V 10 02 1 1 LYR01OVL10S02D1 5
0 -7 -2 3 3 5 -3 12
1 V 10 03 1 1 LYR01OVL10S03D1 0 2
8 5 -2 0 -4 4 -15
1 V 10 04 1 1 LYR01OVL10S04D1 6
1 -9 -3 5 0 5 -3 20
1 V 10 05 1 1 LYR01OVL10S05D1 -2 0
10 9 -3 0 -3 2 -20
1 V 10 06 1 1 LYR01OVL10S06D1 9
0 -9 -4 5 0 4 2 23
1 V 10 07 1 1 LYR01OVL10S07D1 -1 3
9 7 -2 0 -2 0 -22
1 V 10 08 1 1 LYR01OVL10S08D1 8
1 -5 -4 5 -1 3 3 29
1 V 10 09 1 1 LYR01OVL10S09D1 -4 -2
5 7 -4 3 -3 -1 -26
1 V 10 10 1 1 LYR01OVL10S10D1 10
-3 -2 4 -3 2 6 26
In my experience, FWE is a statistical beast which doesn't easily lend
itself to precise analysis. We're missing some critical information, such
as the statistical distribution of weave wandering. But, I welcome all
efforts to quantify it better.
Thanks for sharing,
Jeff Loyer
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of Havermann, Gert
Sent: Friday, August 15, 2014 7:35 AM
To: shlepnev@xxxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] AW: fiber weave effect
Hello Yuriy,
I'm not surprised that tight coupling is less attracted to weave effect on
3313 glass.
Mechanically spoken, tighter coupling decreases the dielectric differences
in between traces of a differential pair. Look at Figure 5 of your paper and
imagine what the difference of effective dielectric surrounding the traces
would be when the spacing is close to one tracewidth. The difference would
be very small, especially when flat weave or even flattened weave is used.
BR
Gert
----------------------------------------
Absender ist HARTING Electronics GmbH, Marienwerderstraße 3, D-32339
Espelkamp; Registergericht: Amtsgericht Bad Oeynhausen; Register-Nr.: HRB
8808; Vertretungsberechtigte Geschäftsführer: Dipl.-Kfm. Edgar-Peter Düning,
Dipl.-Ing. Torsten Ratzmann, Dipl.-Wirtschaftsing. Ralf Martin Klein
-----Ursprüngliche Nachricht-----
Von: si-list-bounce@xxxxxxxxxxxxx<mailto:si-list-bounce@xxxxxxxxxxxxx>
[mailto:si-list-bounce@xxxxxxxxxxxxx] Im Auftrag von Yuriy Shlepnev
Gesendet: Freitag, 15. August 2014 16:08
An: si-list@xxxxxxxxxxxxx<mailto:si-list@xxxxxxxxxxxxx>
Betreff: [SI-LIST] fiber weave effect
Hello Everyone,
Some results of our on-going investigation of fiber-weave effect (FWE) were
recently presented at IEEE EMC 2014 symposium (at SIPI section) and the
paper and presentation are now available at
http://www.simberian.com/AppNotes.php - see #2014_04.
One of the interesting outcomes was practically negligible FEW impact on the
tightly coupled traces on spread fiber fabric (voltage coupling coefficient
about 0.2). We have observed it both on strip and micro-strip (one sheet of
dielectric fabric) configurations.
We did not find any published confirmation of this result. Any comments or
thoughts?
Best regards,
Yuriy
Yuriy Shlepnev, Ph.D.
President, Simberian Inc.
3030 S Torrey Pines Dr. Las Vegas, NV 89146, USA Office +1-702-876-2882; Fax
+1-702-482-7903 Cell +1-206-409-2368; Virtual +1-408-627-7706
Skype: shlepnev
www.simberian.com<http://www.simberian.com>
Simbeor - Accurate, Fast, Easy and Affordable Electromagnetic Signal
Integrity Software
2010 and 2011 DesignVision Award Winner
------------------------------------------------------------------
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 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
------------------------------------------------------------------
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 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
------------------------------------------------------------------
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
---
This email is free from viruses and malware because avast! Antivirus protection
is active.
http://www.avast.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 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:
