Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

TEST COUPON DESIGN METHODOLOGY TO MEASURE THE WORST POSITIVE/NEGATIVE SKEW ON WARP AND FILL DIRECTION ON FIBER GLASS WEAVE FOR DIFFERENTIAL PAIR

IP.com Disclosure Number: IPCOM000249398D
Publication Date: 2017-Feb-23
Document File: 7 page(s) / 1M

Publishing Venue

The IP.com Prior Art Database

Related People

Brian Baek: AUTHOR [+4]

Abstract

Described herein is a test coupon design to measure the worst skew for any positive/negative (P/N) differential pair on a printed circuit board by controlling the differential pair-to-pair distance between two adjacent differential pairs. This may be applied to any board design to quantify the performance of fiberglass.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 36% of the total text.

Copyright 2017 Cisco Systems, Inc. 1

TEST COUPON DESIGN METHODOLOGY TO MEASURE THE WORST POSITIVE/NEGATIVE SKEW ON WARP AND FILL DIRECTION ON FIBER

GLASS WEAVE FOR DIFFERENTIAL PAIR

AUTHORS: Brian Baek

Amendra Koul Mike Sapozhnikov

Yaochao Yang

CISCO SYSTEMS, INC.

ABSTRACT

Described herein is a test coupon design to measure the worst skew for any

positive/negative (P/N) differential pair on a printed circuit board by controlling the

differential pair-to-pair distance between two adjacent differential pairs. This may be

applied to any board design to quantify the performance of fiberglass.

DETAILED DESCRIPTION

Introduction

Typical printed circuit board (PCB) dielectric (e.g., core, prepreg, etc.) substrates

are constructed from various woven fiberglass fabrics strengthened and bound together

with epoxy resin. The weaving technology is essentially the same for fiberglass as it is for

any fabric. The warp yarn runs parallel to the length of the fabric roll, while the fill yarn

runs parallel to the width. The routed traces on a PCB can land directly on top of the warp

yarn, directly on top of the fill yarn, centered between the strands of the warp yarns, or

centered between the strands of the fill yarn.

Figure 1 below shows two traces forming a positive/negative (P/N) differential pair.

The two traces have different effective dielectric constants (Dk) because resin and glass

have two different dielectric constants. This results in different propagation velocities and

causes skew between P and N signals of the differential pair. Skew can have a negative

impact on s-parameters, eye margin, and channel budget, and becomes worse as length

increases. Low skew is critical for serializer/deserializer (SerDes) speeds greater than

20Gbps.

Copyright 2017 Cisco Systems, Inc. 2

Figure 1

The differential pair is routed on an arbitrary location of the fiberglass because PCB

core and prepreg are opaque. Also, typical PCBs are constructed from various woven

fiberglass fabrics. A laminate manufacturer layers these fabrics with epoxy to make core

or prepreg sheets. There are many fabric selections to build similar thickness core or

prepreg sheet. As a result, the worst skew of a differential line due to fiberglass is very

difficult to measure for each fabric selection. However, it is desirable to determine this

effect on various PCB designs to understand worst case skew. Currently there is no way to

determine the worst case skew. Therefore, an efficient test coupon design is desired to

capture the worst fiberglass weave skew on warp and fill direction for differential pairs.

Table 1 below shows an IPC (Association Connecting Electronics Industries)

specification for finished fabric woven from e-glass for a PCB.

Described herein is a test coupon design to measure the worst fiberglass weave

skew in the warp and fill directions for a differential pair. The test coupon design has

dimensions based on PCB design requirements, and may capture the worst case skew by

contr...