Printer Friendly
The Free Library
14,635,145 articles and books
Member login
User name  
Password 
 
Join us Forgot password?

Topology characteristics of reliable bus design: symmetry, minimized impedance discontinuities and balanced loading are preferred features.


ONE PHASE OF high-speed system design involves exploring a layout solution space to ascertain workable technologies and interconnect architectures for various signal groups. (1,2) Some commonly employed topologies for reliable bus design include point-to-point, daisy chain Connected in series, one after the other. Transmitted signals go to the first device, then to the second and so on.


A SCSI Daisy Chain
Both internal and external SCSI devices are daisy chained together.
, star topology See star network.  and T topology.

Several configurations are illustrated by FIGURE 1. Two important topology design considerations are maintaining symmetry and minimizing impedance discontinuities at each junction. (2) For topologies with few impedance discontinuities, it is possible to solve multiple reflection effects by applying the lattice (bounce) diagram for linear systems or the Bergeron diagram Bergeron diagram method is a method to value the reflection's effects on an electric signal. This graphic method - based on the real line's characteristic - is valid both for linear and non linear models and helps to calculate the delay of an electromagnetic signal on an electric  for non-linear systems Non-Linear Systems is an electronics manufacturing company based in San Diego, California. Non-Linear Systems was founded in 1952, by Andrew Kay, the inventor of the digital voltmeter. Later the company developed miniature digital voltmeters and frequency counters. . (2) However, for more complex net topologies it is preferable to utilize a simulator program such as SPICE for waveform analyses.

In Figure 1, each layout scheme is drawn as unidirectional The transfer or transmission of data in a channel in one direction only. , although they can be applied in bidirectional The ability to move, transfer or transmit in both directions.  designs as well. Furthermore, no terminators are shown but termination may be required (depending on such factors as signal edge rates, trace lengths, etc.) for successful implementation. A point-to-point connection between two components produces optimum signal quality. (3) However, in a majority of cases the nets are multidrop, i.e., several components are connected within a single net.

The daisy chain contains a multidrop net structure connecting three or more pins. (2,3) The route and stub A small software routine placed into a program that provides a common function. Stubs are used for a variety of purposes. For example, a stub might be installed in a client machine, and a counterpart installed in a server, where both are required to resolve some protocol, remote procedure  lengths should be maintained as short as possible. Daisy chains function well when the net delay is short compared to rise time--the net can then be modeled as a capacitive load and the driver scaled accordingly.

Daisy chain is a frequent topology for multidrop buses, including front side bus and memory buses on personal computers. (2) An important disadvantage is the need for stubs stubs

The shares of equity in a firm that is financed almost completely with debt. Stubs are often created when firms go through a leveraged buyout or pay big cash dividends in order to fend off a takeover.
 to connect the middle agent to the main bus trunk. Even for very short stubs, the middle agents' input capacitance can load the bus and lower the line's effective characteristic impedance This article is about impedance in electronics. For characteristic acoustic impedance, see acoustic impedance.

The characteristic impedance or surge impedance of a uniform transmission line, usually written
. (2)

Star topologies are inherently unstable and it is critical that the electrical delay and loading of each leg are identical. Otherwise signal quality can rapidly degrade. For T topologies (2), unidirectionality unidirectionality (yōōˈ·nē·d·rekˈ·sh  and symmetry are crucial, and sometimes the T legs are designed with twice the impedance of the base, minimizing impedance discontinuities.

Radial loading occurs when multiple traces diverge from a common point on a transmission line. (4) Each radial line can be separately treated as distributed, lumped or unloaded depending on its loading features. (5)

Radial lines affect the transmission lines' propagation characteristic by creating the impedance Zrad (at junction point):

Zrad = Zo/N

N = Number of radial lines.

For instance, in FIGURE 1d, the impedance at point X where four traces diverge to the receivers is Zo/4, with Zo representing the impedance of each branch (i.e., T1, T2, T3 and T4). Or consider a differential trace with a main line and a junction from which three line pairs (with impedance of 100 [ohm ohm (ōm) [for G. S. Ohm], unit of electrical resistance, defined as the resistance in a circuit in which a potential difference of one volt creates a current of one ampere; hence, 1 ohm equals 1 volt/ampere. ]) diverge to differential receivers. At the junction point, the effective impedance is then 100/3 = 33.333 [ohm].

The tree architecture depicted by FIGURE 2 is an interesting topology commonly employed for clock signals. For such tree topology, some requirements for minimizing skew (1) The misalignment of a document or punch card in the feed tray or hopper that prohibits it from being scanned or read properly.

(2) In facsimile, the difference in rectangularity between the received and transmitted page.
 include placing the clock inputs close to each other, driving them with same source and balancing the tree topology with identical--and equal numbers of--receivers (gates). (6)

[FIGURE 2 OMITTED]

It can also prove advantageous to utilize low-impedance transmission lines to reduce the sensitivity of driver chip to capacitance of receiver loads. Furthermore, as illustrated by the following example, dummy loads can be employed in some cases towards achieving optimum timing.

Due to effects of loading on signal delays or velocity, sometimes dummy loads are incorporated in a net for timing balance. FIGURE 3 depicts the differential clock net of an unbuffered DIMM (Dual In-Line Memory Module) A printed circuit board that holds memory chips and plugs into a DIMM socket on the motherboard. See memory module.

DIMM - Dual In-Line Memory Module
. The DIMM layout is governed by JEDEC The division of the Electronic Industries Alliance (EIA) that deals with semiconductor standards (officially, the JEDEC Solid State Technology Association of EIA). JEDEC was formed in 1958 when the Joint Electron Tube Engineering Council (JETEC) split into two Joint Electron Device  standards. (7) The input section attaches to a connector, and the output portion contains differential loads (DDR SDRAMs). This illustrates output branches that are fully loaded with SDRAMs; however, it is possible for one or more of these branches to lack ICs.

[FIGURE 3 OMITTED]

The pin assignments of the 184pin DDR (Double Data Rate) Refers to an SDRAM memory chip that increases performance by doubling the effective data rate of the frontside bus. For more details, see SDRAM.

DDR - Double Data Rate Random Access Memory
1 DIMM connector J1 includes three pairs of differential clock inputs, often named (CK0, /CK0), (CK1,/CK1), (CK2,/CK2). (7)

Each clock tree contains six differential output lines for possible connection to SDRAM (Synchronous DRAM) A type of dynamic RAM (DRAM) memory chip that has been widely used since the late 1990s. SDRAM chips eliminated wait states by dividing the chip into two cell blocks and interleaving data between them.  pins. It is then possible for a DIMM to possess a maximum of 18, or a minimum of four SDRAMs.

In a DIMM with fewer than the maximum 18 SDRAMs, the implication is that some of the clock outputs are without IC loads. It is then recommended to insert capacitors in unloaded branches to minimize delay skews for various clock loading conditions (FIGURE 4). The capacitance Ceff for C1, C2, C3, C4, C5 and C6 should equal one-half Cin where Cin represents DDR SDRAM input capacitance. The reason for Ceff = 0.5 * Cin (as opposed to equalling Cin) is that each differential receiver connects to a positive and a negative line.

[FIGURE 4 OMITTED]

FIGURE 5 illustrates how two capacitive loads (CL+ and CL-) of the same capacitance (Cin), with currents (I+ and I-) of identical magnitude but opposite directions, yield an equivalent series circuit of effective capacitance (Ceff = 0.5 * Cin). For special cases when there is no common-mode current, the common/connection point in FIGURE 5b represents virtual ground (a node at zero volts but not directly grounded).

[FIGURE 5 OMITTED]

ACKNOWLEDGEMENTS

Thanks to Peter Arnold
For the marine biologist, see Peter Arnold (biologist).


Peter Arnold is a landscape architect and community designer. His recent projects include: City of Brentwood, College of Marin, Sir Francis Drake High School and Red Hill Park.
, Dean Gonzales, Jeremy Plunkett and Anurag Dhawan for reviewing this column and providing valuable feedback.

REFERENCES

(1.) Jon Powell, "Ensuring Signal Integrity," Printed Circuit Design, June 2000, pp. 12-16.

(2.) Stephen H. Hall, G.W. Hall, J.A. McCall, "High-Speed Digital System Design: A Handbook of Interconnect Theory and Design Practices," John Wiley and Sons Inc. 2000, pp. 21-28, pp. 212-214.

(3.) Brian Young, "Digital Signal Integrity: Modeling and Simulation with Interconnects and Packages," Prentice Hall, 2000, pp. 104-109.

(4.) "Design Guidelines for Electronic Packaging Utilizing High-speed Techniques" IPC-D-317, April 1990, p. 26.

(5.) Abe Riazi, "Loading Effects on Transmission Lines, Part 2," Printed Circuit Design and Manufacture, June 2005, pp. 30, 35.

(6.) Howard Johnson and M. Graham, "High-Speed Digital Design: A Handbook of Black Magic," Prentice Hall, 1993, pp. 346-350.

(7.) "PC1600 and PC2100 DDR SDRAM Unbuffered DIMM Design Specification," Revision 1.1, June 29, 2001, p. 7, p. 14, p. 22 and p. 40.

ABE (ABBAS) RIAZI is a senior signal integrity engineer with ServerWorks (a Broadcom company) in Santa Clara, CA. He can be reached at ariazi@serverworks.com.
COPYRIGHT 2006 UP Media Group, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

 Reader Opinion

Title:

Comment:



 

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:INTERCONNECT STRATEGIES
Author:Riazi, Abe
Publication:Printed Circuit Design & Manufacture
Date:Feb 1, 2006
Words:1080
Previous Article:Implementing RoHS in the U.K.: discussions over exemptions continue, but RoHS will take effect July 1.(COUNTDOWN TO LEAD-FREE)
Next Article:Procuring prototypes using a vendor qualification process: process? Yes, maybe you do need a process.(PROCUREMENT)(Company Profile)



Related Articles
SCSI Cable Interconnect Characteristics Simplified.(small computer system interface)(Technology Information)
Adjacent layer signal routing: avoiding crosstalk and signal reflections takes more than putting adjacent layer conductors at right angles....
What causes reflections? In an ideal world signals would propagate down a line with no distortion. But this is no ideal world. (No Myths Allowed).
Avoiding differential pair: routing violations: take these steps to recognize and avoid layout mistakes when routing high-speed differential...
PCI Express verification: with the right game plan, microwave design doesn't have to be a scary proposition.(Signal Integrity)
Loading effects on transmission lines, Part 2: a minimum loaded line impedance may be essential to your design.(INTERCONNECT STRATEGIES)
TDR for differential pair characterization: not just another abbreviation, TDR is one more tool for analyzing single-ended and differential...
Yes, size does matter; sometimes, shorter is better.(NO MYTHS ALLOWED)
Via modeling: accurate modeling can keep your vias from becoming problems, especially at high speeds. But watch that return current.(INTERCONNECT)
TDR for differential pair characterization, Part 2: single-ended and differential TDR signatures encompass broad signal integrity...

Terms of use | Copyright © 2009 Farlex, Inc. | Feedback | For webmasters | Submit articles