EDA365电子论坛网

标题: 好久没发帖,SIlist话题之PCB的损耗可否预测? [打印本页]

作者: stupid    时间: 2012-11-26 22:11
标题: 好久没发帖,SIlist话题之PCB的损耗可否预测?
故事是这样开始的:有人问树脂含量和损耗的关系。
' @& Y5 c/ ~& ~* ~' G' q4 ?# \+ m# ^1 R4 b/ W, ^6 A
Date: Tue, 30 Oct 2012 08:52:23 +0800 (CST) 6 }1 ]0 n# o8 ~% N% z# {3 F8 {+ k

! ~  \9 u) s/ Q' R6 R) E/ ^Hello experts,* M8 u( F5 u) G9 [4 \% {& d
>9 Y4 I; O- F) X, p1 D: N- I
> I'm from PCB house.  Recently we have producted some insertion loss test
! b5 R) e1 `( W% o- e1 p0 z0 h> boards(16L, SET2DIL coupon, IS415/IT150DA/I-Speed Mid/low loss material with
  U0 h$ H1 F+ ]; c+ h$ o> RTF copper foil). We found that the multiply core and high resin PP will ; Z$ w4 W; q7 E' M/ |9 U. r8 o0 U
> result a lower loss result. It's a trouble to MI engineer.  I would like to
' {" r8 f) g0 p0 x/ [" `> know how to predict the loss base on stackup. Please help to suggest (papers,
7 ?: y% \8 ~: P> script, free software etc ). Thanks a lot!
0 g3 t- P$ L& J2 J>
. a9 n* h: D! m6 ~: |>
: U0 O2 [4 Q( W- m+ a$ E& c4 h>
3 j; h' N( [* S# ~> Best regards,
" Q# a  @  A9 P6 F  ]8 F( ?>
: v- r9 ?# E2 Y& @$ O# c$ ]% E> Terry Ho0 k" x$ L/ }0 L5 B9 S
8 K5 e6 \: f6 X9 {
然后 Scott McMorrow ,steve weir,Loyer Jeff 这些活跃分子开始依次发表意见.1 A- ?1 m1 K- w8 V
1 b0 d( y* `$ `: o- |
From: Scott McMorrow <scott@xxxxxxxxxxxxx>7 a, f" A0 d; d, y3 `
Date: Mon, 29 Oct 2012 21:02:36 -0400
& z$ Z! @$ V* n. Z- P) j( z. y* f; s  ^2 g4 G
Hmmm...  I'm in the middle of the middle of Hurricane Sandy.  Power is out.  M7 c7 B+ U& N& K2 W$ [* g. l
Storm surge is causing the river across the street to rise to unprecedented
/ d( |& y" D3 N' H/ F, X: Rlevels.
6 m/ `3 K: S* F3 k9 \8 `& Q, g... and this guy wants us to do his job and suggest free software.& R$ O! @9 G( @7 d. f2 l
0 I) w5 u8 `+ l+ C. {8 H, f
* u& \- r7 U- m/ {. p9 x9 l: I/ D! F: A
From: steve weir <weirsi@xxxxxxxxxx>) y3 O. I- x& L8 _- ~/ c3 M
Date: Mon, 29 Oct 2012 21:23:22 -0700
7 z2 T5 {4 _' t% X" Q/ e1 X( b  e/ ]+ W1 y$ K5 a- {
As a PCB fabricator I think you need to develop in-house material
6 s% u( ~5 |! ~6 c: y2 a) z7 {properties expertise.  Your competitors who understand the materials ! I6 p  f( m# X
they use and their process limits are positioned to get higher yield
$ E8 _; t9 ~7 a0 A8 O( `& spercentages at lower cost because of their knowledge.
. g' t7 I7 _" P0 z, w& e) \% n; m& O  k- k
I appreciate that you don't want to spend unnecessary money, but at
! ]. W/ H- ~- `least spend the time to learn about what you are using.  I am troubled
9 E9 N$ d$ Y3 i' Z( fthat your engineer knows so little about the materials you use that he
2 x2 b$ I, J5 s2 M2 L( l& Qis surprised by common results.  Once your company understands materials
1 s+ _/ V$ V4 Tbetter you may well appreciate the value of commercial stack-up planning
9 f  E% ?0 |& v0 P  Rsoftware.) H2 w5 ]& E- u: Y: a$ w! r3 [3 l' X' A

4 |) y4 Y/ [1 |# nSteve.( i- x$ B4 R/ I% N
1 ]8 d2 o0 q. ]. X6 z3 }
From: "Loyer, Jeff" <jeff.loyer@xxxxxxxxx>- {/ F) e# Y3 c: ^
Date: Wed, 31 Oct 2012 21:33:48 +00000 }' U( W) I$ `: s' s* _+ x) z& n3 J
6 z" L+ z7 b( M4 x

$ g: g/ g0 _* P8 n1 P  \I'm surprised at the tone of the responses to this posting (but perhaps I + \# {1 r) S! d7 N6 y4 Z' t4 m
shouldn't be, unfortunately); I don't see anything untoward in it.  I would 2 y6 F" @* s7 s7 `9 r" f
like to provide some context (with some assumptions on my part) for the message # {. z  P$ e. z4 A
lest other innocent postings meet with similar fates.  I'll also (eventually)
3 V0 t$ _$ V6 c# e5 U3 Q" eprovide my answer to the question, as I understand it.
# [2 s0 o9 ~8 i! E0 D! E1 N: ^  r5 m% x0 o- \* a
3 K# x9 v5 {% ^5 O) d
There is a significant portion (majority?) of the industry which is extremely " E' R6 c7 g+ g& X* I. f
cost constrained.  For instance, to them rotating a design 10 degrees is * Q  w3 [. h7 G: K/ S
impractical, much less 22 or 45 degrees.  Thus, they find other cost-effective
/ X0 T1 c$ v4 ?5 s3 M% `% _yet effective means of solving problems (such as zig-zag routing), even though 4 `9 j- w5 s5 Q3 Q, X9 k8 [
those don't appear efficient to others to whom cost is not an issue.4 P3 j, J8 e  f* K) x' ~7 Q  B

2 v0 @5 p2 a9 ^: d% U* ~) b6 s, a& Q+ |" a  u8 Q
+ e* u8 u3 q1 X2 B2 E' X  e1 [- U; k
There are new pressures being applied to this segment - designers are now not / E# M" ?# G4 F% d; q, h* i% l( e+ Q
only requiring impedance control, but are also insisting on insertion loss 7 m3 j# F/ X6 ]1 V# f6 `
control.  This is a HUGE paradigm shift, very similar to what we encountered
8 B9 f* N( `' t( D' twhen traceable impedance control was first introduced.  That was a very
4 R2 `$ {; \( _! P& W* ^challenging evolution, and this will be also.. j* V: G7 j( G* ^* Q5 ~

& g. n; |9 k8 Q/ d' |7 b/ E* j& V8 e$ L

! [1 e- A. ^* M6 T" T, BAs an example, PCB vendors are now being advised to smooth their copper, after ' H$ H0 c( K$ k1 N* w+ a8 M) K& B
years of purposely roughening it for best mechanical integrity.  It should come . r$ E, t7 G; N  v9 d$ B
as no surprise that this is not a trivial change, considering the effort that * L8 O" \, s3 j: I! u
has gone into ensuring mechanically robust designs.( u( J1 k& ?6 ^: R3 f( n1 L

- G: s% T% R7 z8 o2 u1 `" {, m' f- [. F4 X
0 k: O3 Y. H* A$ d% @$ z, h- t
Likewise, many other basic assumptions that we've been able to apply for years
) ?9 [* i4 v5 |  R. q7 `are now being drawn into question, and PCB vendors are looking for help to
& G- f  m( |5 b: t& lintelligently and cost-effectively explore options - "How much effect does " ]2 d" O1 |4 U6 t2 v
rougher copper have on insertion loss?".   I believe Terry is highlighting the 3 v# g  g( |( [" c# _
fact that, while there are many tools available for impedance prediction, % X  ?# L& x# t
insertion loss modeling is much less accessible.   I don't think it is 7 E! L" L6 V( a' `' _
inappropriate to ask if there are cost-effective, reliable tools available to
# P7 e4 G0 c- M& g4 X5 i3 ?predict insertion loss based on a proposed stackup.: i" F* J! z1 C1 a  D6 `! M
8 X# y' B5 X. @0 [) t

( D3 E, I4 T( C5 C& o* ?* L- W/ U# x6 z
Unfortunately, I believe the answer to the question is that there are no
; M5 D) `5 Z. H9 N% _# _reliable, cheap (~free) modelers available to predict insertion loss.  And, the
2 ?* M: P$ [9 V5 Y# I+ Bones that are available require a great deal more knowledge about the stackup
0 l9 m) `* R7 l5 ethan impedance modeling does, and that information is not easily obtained.  9 Q1 [# X& f7 g& D3 o0 n
There are some of us working with a vendor to test their modeler against a   O' a% l% F' h4 M
variety of stackups and we'll present results at DesignCon.  My personal goal % V# g# k! p. N$ c& J
is not so much to test a specific modeler but to judge how effective a modeler
  j3 e( h, i% u; y4 f2 U* Ocan be given information that can reasonably be gleaned prior to building with
4 _4 B* _" m) g% x  h+ Lvarious materials, copper types, etc.! f4 Z, M5 J5 ^7 @7 W5 \
7 p: q6 M9 i- t

5 b; c& W, V( ?* ?, U
0 d( G6 L: w. J1 n3 ?$ n, [In the absence of a modeling tool, or in addition to one, I believe empirical : Y% w; s# N# o' Q8 T: w" `
data is the best predictor of insertion loss.  To do this, however, you have to : b* H5 j2 H! \% x, R
build a stackup representing the final design, and it's not clear at this point 6 o/ \3 Z; [$ C# K/ T4 k
how broadly you can extrapolate those results to other stackups.  But, I know 1 S& P5 k; f4 b% e
many material vendors and PCB shops are engaged in similar efforts.
# a% N. j1 o! Q( D& u, U
$ g6 l3 U- h( c8 f  T4 u: E  [. E0 @+ H- I8 O, j4 K3 }, d
) }( d5 e/ ?+ U/ R8 R8 v$ N& S  J% r4 y/ S
I think this is very similar to what we went through with impedance control -
1 |# Z9 [/ c; f2 f2 @# O' K, |; _the shops which most quickly were able to predict and control that ' c: R6 F9 G$ ]2 r
characteristic had an advantage.  I think successful PCB vendors will need
8 }: L1 U3 E' l4 areliable modeling software and empirical data on insertion loss for their 4 J3 ~* y. L4 X
particular choices of materials, etc. - they will be able to find the most cost ) P. M7 m! V' q3 E: h
effective solution.
5 y0 I9 P4 L4 A& r2 l4 d4 f4 L6 d# u$ S

% _' z0 z! ]9 [
$ I; R5 f. i$ m9 i/ PBottom line: I doubt a reliable modeling tool is going to be cheap, but is # \* g4 S, D0 i
going to be necessary, and you'll want to compare any tool you do purchase - R5 B0 G2 x2 }: A' J
against empirical data before you trust it.* e% @; F, f5 i; G% \& c

. b% @; R9 ]' L; N" |% |; C4 i. y  c+ y& r0 c$ G
" n% f0 j$ U. i$ a7 n3 R" b4 j
I hope this helps,
# B! O' a' s, C( ]6 e$ n! l* @7 X3 \1 a' T: h6 n
Jeff Loyer
6 O! m7 d# y- {* D
8 v8 y7 m& H/ A3 k) ]) N' ]; _0 H6 M+ R4 @
From: steve weir <weirsi@xxxxxxxxxx>+ k! B# ]! [3 y; p. y" W! n% ^
Date: Wed, 31 Oct 2012 20:14:41 -0700
* P% R% \( d; ]: T; C
1 k2 ?! b* o9 N, a
6 ~3 i# s2 v. V  r5 X' ~6 EJeff, given that the only two responses were Scott and mine, I am
; F; _: }% R4 o/ h6 W) ksurprised that you are disappointed with both.
/ U( v- c! r4 Y
" |+ c9 U( b5 |4 f( {3 w: _. MIn a fabrication market filled with intense competition it is up to
& u' o5 x/ i- B! r$ ]" ]& s# Sindividual players to keep up with the technology requirements of the
+ q/ i& W- d0 f0 u( g; j- L  Cmarket or get left behind.  The task is not simple. Depending on how far 3 l/ D8 W- ^: ^- w- G
up the frequency range one needs to go, dialing in cost effective
; \# _  y! h0 f' J* f' y! q* Kprocess requires substantial skills, time, effort and serious money.  It $ V, X1 I% O& c4 C4 b0 V! G  \
represents competitive advantage to OEMs and their partner pcb fab 4 Y2 s$ v$ o, r8 s% j
houses alike.  Neither who have invested are likely to hand over that 2 T( h& M" Q! r: L; v4 H- A
kind of advantage especially when it is so costly to obtain.
, g( H/ h- n8 `/ B$ r' \: L; e
, `: i8 t5 B  b/ X9 mI don't mind that Terry is looking for a solution on the cheap or free.  
: T7 C- B" R2 ]3 X/ b2 d6 wIf one could obtain such a sweet deal, one would be foolish not to take
  W. c3 e- c  }2 E% M$ |2 Kit.   I am troubled that in this day in age, his organization hopes to
, ?. C  b" ~% ], n4 paddress a sophisticated issue before his technical staff has a grip on
/ _/ r# O# M1 J1 g1 wthe basics.   I fail to understand what you find inappropriate about
' P- |3 ]# q/ B& H4 w9 K2 W4 lthat concern.  I would rather yell at someone headed for a cliff to stop
8 E8 Y  t) ~( g0 M; fthan smile and wave.
  R" k. h' n) b$ Y( c+ n! S/ _0 z5 X4 V! ]# P" C" m9 D; l
Best Regards,( L( a# c, R% N2 [
0 ]  I/ b* ~: E4 O+ {1 f
0 A! T2 C1 R7 u: y7 C
Steve.- W& `6 J. g/ [' C8 z

: ~, ^. Z) g0 z/ f+ xFrom: "Loyer, Jeff" <jeff.loyer@xxxxxxxxx>
& n5 |  z/ d. {% sDate: Fri, 2 Nov 2012 15:37:46 +0000' X( a2 h) F, a- ^

! w6 q0 ?: a8 i& V; d4 u5 D$ ?, A " Z9 V8 l7 V/ t. d2 D
I realized we hadn't answered the basic question - "why does a high resin & S) ~& f# I' V) d: K; a* w7 k
prepreg give lower loss?"  The prediction of loss vs. resin content isn't
: m+ z& q8 O( W) a$ w' mtrivial; as Steve said, a tool which allows you to model loss for the various
. y9 y+ V8 V8 Q6 {0 Uscenarios should be on your Christmas wish list.  Here are the factors that I
# ?2 Z7 V% W9 r2 cknow of (thanks to Richard Kunze for clarifying things for me, and I welcome - U. s* G/ N8 K6 V+ |
others' data/opinions):- _# O5 X# i% r- F. D9 [
* Resin has a lower Er than glass
/ @" ~* u! D' p; K
' ~$ \, P) q- O     * loss is approximately proportional to Df * sqrt(Er), so lowering Er ; F' S7 q. y) ^' f: |: G( R: ?
lowers loss' @0 M" u; K! h* n* j4 Y2 @
5 D) ~# f4 x4 n' g7 y$ ]% p
     * lower Er allows wider traces for the same impedance - this may decrease 8 b! X9 _2 V+ X. o3 Z
loss also
: l: X3 e; {4 [% c7 h7 [( ]5 b* C2 T' z7 K
* But, resin is more lossy than glass, so Df may increase
2 M& @" P+ @1 O* {0 h& I3 V, }# C
! Y, [* j7 m* a3 a! M     * for standard FR4 constructions, this is especially true.  The data sheet
& K% C% j3 p: ~5 ?: ~& u4 N/ ^  @for IS370HR, for instance, shows Df varying from 0.0177 to 0.0247 (1GHz),
" i% p: y+ h+ K# T. e9 v7 Idepending on the resin content6 Z+ L# ^9 k1 I- k% Q/ M' D. D" c: ?

% a2 \! T2 s2 X: x! B     * for low loss materials, this doesn't hold.  The data sheet for Meg6
. }: k: I& ?* cshows Df constant (0.002 @ 1GHz) for all its flavors of prepreg
% s+ h9 N, l1 b$ _0 m  ]# W
2 z5 a6 U& U5 F* Where the factors dominate will depend on your relative conductor vs.
9 W+ q& C% v$ I+ s% E: Adielectric loss effects: for FR4, dielectric loss dominates at >~1GHz; for * d5 A; X. ~2 K) \& w$ ^3 R. ?
low-loss materials, conductor loss dominates up to much higher frequencies (as , ]2 E) d: d; L- ]2 j
much as 10GHz).
- s- o3 H$ r3 B0 j1 D- b1 A% i3 a

: j# L2 F+ ]& L& Q: r
- Z4 D+ G. A! Q9 P( aIn your particular (low loss) case, the lower Er of the resin-rich case is 5 z* ^9 l0 Z4 Z9 r
trumping the Df change (or lack of) so you get lower loss.  r1 [* h: O  t
# }# I( P# J6 B1 R. [, {! ?8 m& m

! z0 Q+ i# I3 R$ j! t. G" g. N) v+ T# S' h+ a
Only a tool which takes into account the properties of the specific material 6 {, O2 r" w7 x
under consideration can be expected to give an accurate prediction of insertion , N5 p) R$ `# v
loss for various resin contents.0 r- ~: h- M/ F( Z. y, L# \
# j' l) _+ w  A
. Q4 [" n9 I5 H9 Y

4 b# H5 D( ?( B& w4 FThere are also environmental effects (I haven't heard or seen these stressed at
2 v0 L/ M( G5 z- J; ~this point, though that may change soon):% P" z. u+ A2 b4 W. M

# j+ b/ E# Y  [! L* Higher resin content will absorb more moisture, and thus your loss will be
3 Q" _% n8 T* \! t2 F0 f0 B* ~more susceptible to humidity effects( Y4 _: ]$ L8 ]4 t$ K: `* i+ S6 @

( g0 H1 n6 C& r* There's a difference in how the various materials' Df changes w/ temperature 5 d+ d- H; I  p9 k& i$ }* a% Y
- more at DesignCon
, v' X3 t) Y/ l  p; z$ B8 V, w) N/ `9 L7 f, K

  \( G( ]( l$ {# ^! b+ X0 S+ W6 s( A+ U0 h; V& s
I hope this helps,
; `( H( m) O4 q
1 Q% t0 e' H. P8 Q. X. tJeff Loyer
7 M2 W3 c5 j' C$ V3 {9 J7 Y6 f+ k+ D0 C
From: Scott McMorrow <scott@xxxxxxxxxxxxx>
4 {8 }2 b$ h" u) L, p3 Q) n% r2 ADate: Thu, 8 Nov 2012 09:12:46 -0500
% C# G( z. D3 O6 l6 g5 j1 @
% T: M9 x+ W: X/ `$ g0 d4 J
; R! V% _+ V; J; }1 j7 DJeff
# e% L" ~; Q% J" w  _& G% U* wA few quick comments.  Although the tanD of Meg 6 is stated to be flat, it
! v" p5 o3 T3 [; H6 Lis not if you measure it.  The manufacturer reported characterization in
% F) t$ C# M8 ]4 m' W7 N4 Z# N4 u, B) j4 zthe data sheet is not correct.  Causality is violated when tanD is flat.
; w5 f; R8 y0 _1 R, y6 L8 A0 M1 L/ E. B) l5 g1 y1 d( M
Loss is generally due to molecular dipole losses in the material.  It can
' r% |1 m  t, x4 X7 E2 s0 jbe low for high Er, as is the case with ceramic.- J9 Z5 d* \, x4 G% p- [9 _
" w' E2 X4 `5 a5 R0 A" c$ n5 ~- M
Hygroscopic loss is due to molecular polarity. Polar molecules "glom" on to
5 c) }# _5 P) Q: L/ p+ i8 swater molecules, which are also polar. Same property makes the material
6 K$ W: `. r9 S. e) }1 ?extremely "sticky."
- m5 {1 I1 T) t* L
+ `$ N5 Z, A0 m& d5 T, B6 GThe paper that Jason Miller of Oracle and I wrote for DesignCon last year* w- a' Y+ m4 \2 C% ~( y- D
covers some of the impact of temperature and humidity on measured losses.
  M* `5 T# V9 l, v I don't have access to my storage server right now, otherwise I'd give a
) m* Z1 l, z- ]/ P- ?0 p$ Hpaper citation.: k7 T- |$ J2 i
% t7 C4 B, X! I& J
regards,: ^2 e  r; Z  k) z' R5 s
8 a% }0 m0 c0 M2 N  n* m2 l( G
Scott
) h0 M2 I- D# D' |! Z  B+ p" ~, y4 U/ C2 p, K
From: Kirby Goulet <kgoulet@xxxxxxxx>
# L2 g3 G+ P# }8 k$ V/ q, uDate: Fri, 9 Nov 2012 11:08:49 -0800 (PST); k" L" T; Z7 `7 v7 F) S( |- f7 T

7 b+ d' q1 p% y$ J! H 5 z) t. u9 f7 z; g! _$ |# q; R
It's not production quality software but you could try the mdtlc calculator to 8 \4 S8 M3 Y' G
experiment.  I tried Jeff's example and it seems to point to an explanation.
1 i$ ^: a9 M+ F: t The source code is available so you might extend it to do what you want if you
1 e* B8 n1 C4 h% W( A+ I: [0 Ahave more time than money.
& J8 I5 j( g& S2 m  V4 wIt looks like a race between loss due to increasing loss due to resin and ! J0 t$ _  v, G. q) o0 N
decreasing loss due to wider traces.  There is a bigger increase in the resin
* H  W! s6 N- I8 Z' c* Y9 Bcontent for the IS370 case over the IS415 case.  Not only that, but the IS370 $ b5 a( S. t) z' K) y
resin is lossier: 0.0169 versus 0.02984 so the winner is increasing loss.  
* C; f% b9 C0 a0 x! m" I
0 _3 U) A9 x# f: N1 f% m9 qFrom the field solver, 9 e2 K! F3 ^0 H5 H, g; ?4 z
; v# n; u$ e3 O5 f
IS370: the effective dielectric loss went up 14.7%.  The perimeter of the # V/ `* L1 z, h! e- O1 q, C& ^
conductor went up 3.6%.  
, @. B& t* g% {; ~' ~- H! [3 yIS415: the effective dielectric loss went up 6.7%. The perimeter of the 5 r4 Y+ I6 e& o9 P9 F) h
conductor went up 5.7%., d' B) m" A( O4 D8 g4 s! n& @/ r

& ]/ _5 D4 l* V, k; U9 K. }/ V  MIn the second case, overall dielectric loss is a smaller fraction than the
% G* \" C/ b+ L0 Kfirst case.  The missing bit of information you need to add is the conductor
# _, U6 \. k* lloss.
  {3 j- Z1 o1 {$ J: A
% h$ R8 o3 Y# ^$ Q- EINPUT PARAMETERS:; W; Y! \7 d% t2 o; W0 w
: k( \$ o! g' K; G
      Layer             Thick   Specifications
$ P0 k# l& \* D      Copper Plane Top   1.30    Opening w=0.0  offset=0.0) {1 w$ D1 L& t6 t0 j
      Laminate Layer 1   3.90    Resin Content  57.0% 3.4-4.9 : O7 Z% y7 x4 D* e3 t
        Signal Layer 1   1.20  4.3-7.2-4.3  Etchback=0.001 @$ P7 v' e; X( F1 @2 D) }
      Laminate Layer 2   3.90    Resin Content  57.0% 3.4-4.9 ! X3 c0 P+ m; T" H0 @  T+ \/ q
   Copper Plane Bottom   1.30    Opening w=0.0  offset=0.0
9 H! x  q& h$ p( B3 F : Z/ ?+ @7 i. G2 c" k. S
      Layer             Thick   Er    Loss Tangent
8 j0 |3 g! D: i3 x5 z+ f- ~      Copper Plane Top   1.30  3.20   
' o3 M. M- A' c( @, ^      Laminate Layer 1   3.90  4.02    0.02100 1 s! V3 T7 P5 `: ~9 i  P
        Signal Layer 1   1.20  3.38    0.02984" R8 U; Y+ D3 F9 j" f- [
      Laminate Layer 2   3.90  4.02    0.021006 R; p* h1 ~+ A; U4 O! L! L) ^9 ^
   Copper Plane Bottom   1.30  3.20( ]0 c. @; ~. m8 c9 {# A8 ]

$ B7 m' I4 ?9 U  e2 p( r$ e DC resistance by dimensions:
' ~4 h( f$ m7 k& ] Rdc_trace_1= 131.53      Rdc_trace_2 = 131.53  milliohms/in 20C
- F; b  z" k1 _& ]; j/ T
: a. B. ~+ P: S DC resistance by pixel count:5 `, n  N3 B. j& O' Z2 u
Rdc_trace_1= 131.531     Rdc_trace_2 = 131.531 milliohm/in! {. p9 G5 w# @/ R  Z
C_odd      =   4.221 pF/in    C_even =   3.968 pF/in0 L6 O! ^" Q1 k$ [' p9 r2 C  X
Er_odd     =   3.923         Er_even =   3.947
6 X5 t& |, b  X! b5 S  f# H: I Loss_tan_o = 0.02212     Loss_tan_e  = 0.02184  " m2 e) c) Q! {$ ]% m- I4 I% q9 D4 i
Delay_odd  = 167.801      Delay_even = 168.314  ps/in.
9 z( q6 |# F" G Z_diff     =  79.501  ohms   Z_comm  =  21.209  ohms% |1 W4 r1 [# E8 p1 J4 R9 ?) x0 d
* F, U/ Q$ G+ r* N
Simulation pix map 122 pixels high by 800 pixels wide.
) n5 y$ e" o  |8 [  q293824 bytes allocated for bmp./ i2 O* z( n! `' O/ t5 N

3 H. q2 Z/ v4 ?4 u' I7 XINPUT PARAMETERS:
1 H1 w, o% ?* V( [, {0 H; K( N 5 h- q6 a1 a; P" n% G& R
      Layer             Thick   Specifications
& g& S' e* L/ Y5 z      Copper Plane Top   1.30    Opening w=0.0  offset=0.0* ^) d7 J9 }0 l- `9 o
      Laminate Layer 1   4.20    Resin Content  75.0% 3.4-4.9
0 c( ?/ ^2 K8 a6 V- b! c( z        Signal Layer 1   1.20  4.5-7.0-4.5  Etchback=0.00, O0 H$ V9 Z  t  f
      Laminate Layer 2   4.20    Resin Content  75.0% 3.4-4.9
( T9 H/ N0 _( I4 |4 n% g# ?   Copper Plane Bottom   1.30    Opening w=0.0  offset=0.0
; H2 S" Q: u; Q# w% }  {  s
+ K& G* t' V0 o8 V      Layer             Thick   Er    Loss Tangent; ^% `+ }6 j* P/ x1 Q
      Copper Plane Top   1.30  3.20   * O! d8 X( w, L2 E: I% g* e
      Laminate Layer 1   4.20  3.75    0.02470
" [; ~6 i" x- N0 Q        Signal Layer 1   1.20  3.38    0.02984
* v/ Z- J9 N; t0 n4 \- z* @2 U      Laminate Layer 2   4.20  3.75    0.02470# U$ _  H" z% X7 B, k, R$ |
   Copper Plane Bottom   1.30  3.20+ |' I9 N/ O6 M' x& M
" w; J( B$ w$ _$ E
DC resistance by dimensions:
: E) S8 _& ~  b7 m: Y Rdc_trace_1= 125.69      Rdc_trace_2 = 125.69  milliohms/in 20C' V, n* ~3 J& x5 G
( F# Z) h! I7 W3 H
DC resistance by pixel count:
  q: w$ f3 b0 `6 p Rdc_trace_1= 125.685     Rdc_trace_2 = 125.685 milliohm/in5 @6 [1 i6 Z) H: T
C_odd      =   3.929 pF/in    C_even =   3.624 pF/in- m9 X+ C* v* D' O( M' n
Er_odd     =   3.694         Er_even =   3.710- f% n- o2 U9 l/ H( _" {8 U
Loss_tan_o = 0.02537     Loss_tan_e  = 0.02518  
" O5 B' M: L/ J Delay_odd  = 162.844      Delay_even = 163.195  ps/in.) j  @' x5 ~$ H$ W3 O* z+ f/ ]
Z_diff     =  82.900  ohms   Z_comm  =  22.519  ohms% j% I0 W( E0 w/ r
% L1 E8 P, r! Q, c, v" W, u7 t
Log file save name:  U8 }5 Y/ [) P+ i
mdtlc_12100946383.txt8 b& }0 t" H) a7 M  e8 I/ T
8 [' B" t( M# n0 K' E! p
Simulation pix map 118 pixels high by 780 pixels wide.
" L. X+ N; r( i! _' t& \' Z9 \. ]& K277144 bytes allocated for bmp.
8 b( N! ]. R, {3 X$ x8 r1 z : c7 p1 m7 r& v- Z5 `
INPUT PARAMETERS:
& f' h5 \8 K0 R) x9 ] * S! D, p$ B: G- g: \+ u
      Layer             Thick   Specifications 5 x* i5 }# I' u7 M8 U
      Copper Plane Top   1.30    Opening w=0.0  offset=0.0
6 n( D1 n: D6 s      Laminate Layer 1   4.00    Resin Content  45.0% 2.6-5.1 / _5 X0 F# E; ~* R4 ], o
        Signal Layer 1   1.20  4.1-7.4-4.1  Etchback=0.00: \! |, z$ U; q& L
      Laminate Layer 2   4.00    Resin Content  45.0% 2.6-5.1
. N/ P2 k8 V8 n$ c9 s   Copper Plane Bottom   1.30    Opening w=0.0  offset=0.0. U! m. d( }$ g) V# s7 |
+ \$ n  K3 `0 `  e4 ^' D' K
      Layer             Thick   Er    Loss Tangent
6 z1 B+ ?0 |; \' E; E& S      Copper Plane Top   1.30  3.20   " V, Z! q) r: ?6 ^2 K7 x* \4 K
      Laminate Layer 1   4.00  3.98    0.01140
1 ~* u4 [. _9 g# c1 s- f        Signal Layer 1   1.20  2.64    0.01690
& a0 z; H- n1 \: p7 m      Laminate Layer 2   4.00  3.98    0.01140- o8 g9 ^9 H5 C' V5 G/ z) K
   Copper Plane Bottom   1.30  3.20
, j! i/ ]3 {+ R8 h8 S
1 d7 l( Q. S( T7 o0 A DC resistance by dimensions:
7 S5 k! Y, K, ?' i Rdc_trace_1= 137.95      Rdc_trace_2 = 137.95  milliohms/in 20C
  J9 Z9 S( o$ U, k' V 7 ?( a. ~6 O$ y  Z4 ]/ ~5 ^. F0 F
DC resistance by pixel count:
) j* [" Q. m6 }" U9 p. p% v Rdc_trace_1= 137.947     Rdc_trace_2 = 137.947 milliohm/in
& x1 S) l/ T3 [6 B. P1 K C_odd      =   3.910 pF/in    C_even =   3.695 pF/in( y9 P) s  h$ n; B4 h
Er_odd     =   3.769         Er_even =   3.817
$ U9 c/ ]6 S2 {5 C/ o' t+ o$ T Loss_tan_o = 0.01202     Loss_tan_e  = 0.01189  
. o. B& T/ D5 y, ~+ S& u( p Delay_odd  = 164.490      Delay_even = 165.524  ps/in.
7 N7 Y8 C+ `) k4 r: ^) V% v Z_diff     =  84.134  ohms   Z_comm  =  22.396  ohms
3 e+ r% I6 ?9 r9 _: L: L. n4 } ! k5 H# k+ K. |. M# U
Simulation pix map 118 pixels high by 795 pixels wide.
$ y' u7 x  L* z8 e& j282454 bytes allocated for bmp.- ?' W8 b5 k4 f5 l
- N" g3 Z& E7 t! |6 U# O' c
INPUT PARAMETERS:
  _" l# w$ q+ w/ i( ]4 y/ p" P. } " G9 d2 K: ^* A' T8 v8 ]
      Layer             Thick   Specifications % ~/ ~3 x, a6 n7 V' |- y
      Copper Plane Top   1.30    Opening w=0.0  offset=0.0
9 n9 S, x, r) n      Laminate Layer 1   4.00    Resin Content  54.0% 2.6-5.1
' R* d8 h7 W4 P2 N9 m: i8 n- k" A/ z        Signal Layer 1   1.20  4.4-7.1-4.4  Etchback=0.00; J: A. |; S7 N% F0 Z% b& @
      Laminate Layer 2   4.00    Resin Content  54.0% 2.6-5.1
/ c2 E" v9 j4 Q4 L1 A+ v  @   Copper Plane Bottom   1.30    Opening w=0.0  offset=0.0; J/ z7 |6 f  S' A' W" K. U- {# H

) ?7 R  y0 L4 Q6 a      Layer             Thick   Er    Loss Tangent
+ e" [0 g$ P0 u1 s. N' Q      Copper Plane Top   1.30  3.20   ( w5 h% Z1 s2 C) G( F
      Laminate Layer 1   4.00  3.76    0.01230
( Q7 p4 t3 t1 |: {1 ^        Signal Layer 1   1.20  2.64    0.01690
: E0 }7 [; ?7 m7 U, v: A      Laminate Layer 2   4.00  3.76    0.012305 o* b' ?8 R& Q
   Copper Plane Bottom   1.30  3.20
9 l) r4 ~6 V& E; x6 D   t0 q/ z+ D% E; H6 N9 L+ G' N4 c& D
DC resistance by dimensions:* c1 b; D* _1 G% i$ I
Rdc_trace_1= 128.54      Rdc_trace_2 = 128.54  milliohms/in 20C
) K6 Y* ]% w# r5 P, Y  g5 _  ^ % S2 ^8 ]3 x- ]  U# t9 y/ \
DC resistance by pixel count:
# I8 K6 u: J8 t" W- I" j Rdc_trace_1= 128.542     Rdc_trace_2 = 128.542 milliohm/in& X. P4 o6 U0 G7 d
C_odd      =   3.865 pF/in    C_even =   3.623 pF/in
, ~+ @/ m" T7 R2 A0 V% q Er_odd     =   3.588         Er_even =   3.631$ B% N, @* X7 {! z5 i' h: U5 N
Loss_tan_o = 0.01283     Loss_tan_e  = 0.01270  
9 j0 P3 t1 P6 w/ a Delay_odd  = 160.480      Delay_even = 161.455  ps/in.6 t- b: ~4 h# J# E; ]
Z_diff     =  83.041  ohms   Z_comm  =  22.280  ohms- C9 V6 m% W. |% R

6 f6 r: e1 C& ~! [% H; x0 K看完之后,我有一事不明,我总是分不清奇偶,不知道这两者到底如何区分,亲,你能告诉我吗?
作者: Nelson    时间: 2012-12-1 14:21
我以前也分不清奇偶,后来强行记住了even是偶,搞的后来一看到odd,就要先想even是“偶”,odd只好是“奇”了,很是麻烦!
9 T+ k* r8 W2 D5 x1 U) S你这一提醒,我以后可以换个记法了:odd是3个字母,“奇”数个字母;even是4个字母,“偶”数个字母。哈哈
作者: hrb011011    时间: 2012-12-23 21:19
晕, 版主居然转贴到这里来了!# g7 S. ~* ?* P/ D) h' [
汗!!!{:soso_e110:}




欢迎光临 EDA365电子论坛网 (https://www.eda365.com/) Powered by Discuz! X3.2