WO1999066554A1 - System and method for determining the desired decoupling components for power distribution systems using a computer system - Google Patents

System and method for determining the desired decoupling components for power distribution systems using a computer system Download PDF

Info

Publication number
WO1999066554A1
WO1999066554A1 PCT/US1999/013802 US9913802W WO9966554A1 WO 1999066554 A1 WO1999066554 A1 WO 1999066554A1 US 9913802 W US9913802 W US 9913802W WO 9966554 A1 WO9966554 A1 WO 9966554A1
Authority
WO
WIPO (PCT)
Prior art keywords
components
decouplmg
power
dismbution
wherem
Prior art date
Application number
PCT/US1999/013802
Other languages
French (fr)
Inventor
Raymond E. Anderson
Larry D. Smith
Original Assignee
Sun Microsystems, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sun Microsystems, Inc. filed Critical Sun Microsystems, Inc.
Priority to JP2000555293A priority Critical patent/JP2002518763A/en
Priority to AU52045/99A priority patent/AU5204599A/en
Priority to EP99937161A priority patent/EP1088345B1/en
Priority to DE69907187T priority patent/DE69907187T2/en
Priority to KR1020007014265A priority patent/KR20010071488A/en
Priority to AT99937161T priority patent/ATE238607T1/en
Publication of WO1999066554A1 publication Critical patent/WO1999066554A1/en
Priority to HK01104948A priority patent/HK1034601A1/en

Links

Classifications

    • H—ELECTRICITY
    • H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K1/00—Printed circuits
    • H05K1/02—Details
    • H05K1/0213—Electrical arrangements not otherwise provided for
    • H05K1/0216—Reduction of cross-talk, noise or electromagnetic interference
    • H05K1/023—Reduction of cross-talk, noise or electromagnetic interference using auxiliary mounted passive components or auxiliary substances
    • H05K1/0231—Capacitors or dielectric substances
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/30—Circuit design
    • G06F30/36—Circuit design at the analogue level
    • G06F30/367—Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods
    • H—ELECTRICITY
    • H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K2201/00—Indexing scheme relating to printed circuits covered by H05K1/00
    • H05K2201/09—Shape and layout
    • H05K2201/09209—Shape and layout details of conductors
    • H05K2201/0929—Conductive planes
    • H05K2201/09309—Core having two or more power planes; Capacitive laminate of two power planes
    • H—ELECTRICITY
    • H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K3/00—Apparatus or processes for manufacturing printed circuits
    • H05K3/0005—Apparatus or processes for manufacturing printed circuits for designing circuits by computer

Definitions

  • TITLE SYSTEM AND METHOD FOR DETERMINING THE DESIRED DECOUPLING COMPONENTS FOR POWER DISTRIBUTION SYSTEMS USING A COMPUTER SYSTEM
  • This invennon relates to electronic systems, and more particularly to elecmcal interconnecting apparatus having continuous planar conductors.
  • Electronic systems typically employ several different types of elecmcal interconnecting apparatus having planar layers of elect ⁇ cally conductive material (i.e., planar conductors) separated by dielectnc layers
  • a pornon of the conducnve layers may be patterned to form elect ⁇ cally conductive signal lines or "traces"
  • Conducnve traces in different layers are typically connected using contact structures formed in openings in the dielectnc layers (i.e., vias).
  • integrated circuits typically have several layers of conducnve traces which interconnect electronic devices formed upon and within a semiconductor substrate.
  • PCBs Printed circuit boards
  • the conductive traces are used to elect ⁇ cally interconnect terminals of electronic devices mounted upon the PCB.
  • Signals m digital electronic systems typically cany informanon by alternating between two voltage levels (i.e., a low voltage level and a high voltage level).
  • a digital signal cannot transition instantaneously from the low voltage level to the high voltage level, or vice versa.
  • the fmite amount of time dunng which a digital signal transitions from the low voltage level to the high voltage level is called the ⁇ se tune of the signal.
  • the finite amount of tune dunng which a digital signal transinons from the high voltage level to the low voltage level is called the fall tune of the signal.
  • Digital electronic systems are continually being produced which operate at higher signal frequencies (i.e., higher speeds).
  • the rise and fall times of the signals must decrease as signal frequencies mcrease.
  • This decrease in signal transinon tunes i.e., nse and fall tunes
  • a signal launched from a source end of a conducnve trace suffers degradation when a portion of the signal reflected from a load end of the trace arnves at the source end after the transition is complete (i.e., after the nse time or fall tune of the signal).
  • a portion of the signal is reflected back from the load end of the trace when the input impedance of the load does not match the characteristic impedance of the trace
  • the effects of reflections upon signal inteenty i.e., transmission line effects
  • steps should be taken to minimize the degradations of signals conveyed upon the trace due to reflections
  • terminatating the trace For example, the mput impedance of the load mav be altered to match the charactensfic impedance of the trace in order to prevent signal reflection
  • the transition tune (1 e . the ⁇ se or fall time) of the signal decreases so does the length of trace which must be terminated in order to reduce signal degradation
  • a digital signal alternating between the high and low voltage levels includes cont ⁇ butions from a fundamental sinusoidal frequency (1 e , a first harmonic) and integer multiples of the first harmonic
  • the frequency content of a digital signal extends to a frequency equal to the reciprocal of ⁇ times the transition time (l e , nse or fall time) of the signal
  • a digital signal with a 1 nanosecond transition time has a frequency content extending up to about 318 MHz
  • a power supply conductor connects one terminal of an electncal power supply to a power supply terminal of a device, and a ground conductor connects a ground terminal of the power supply to a ground terminal of the device
  • Power supply droop is the term used to desc ⁇ be the decrease in voltage at the power supply terminal of the device due to the flow of transient load cu ⁇ ent through the inductance of the power supply conductor
  • ground bounce is the term used to descnbe the mcrease in voltage at the ground terminal of the device due to the flow of transient load current through the inductance of the ground
  • Power supply droop is commonly reduced by arranging power supply conductors to form a cnsscross network of intersecting power supply conductors (l e , a power supply g ⁇ d) Such a g ⁇ d network has a lower inductance, hence power supply droop is reduced
  • a continuous power supply plane may also be provided which has an even lower inductance than a g ⁇ d network
  • Placing a "bypass" capacitor near the power supply terminal of the device is also used to reduce power supply droop
  • the bypass capacitor supplies a substantial amount of the transient load current, thereby reducmg the amount of transient load cu ⁇ ent flowing through the power supply conductor Ground bounce is reduced by using a low inductance ground conductor g ⁇ d network, or a continuous ground plane having an even lower amount of inductance Power supply and ground g ⁇ ds or planes are commonly placed in close proximity to one another in order to further reduce the inductances of the g ⁇ ds or planes
  • Differential mode EMI is usually reduced by reducmg the areas prosc ⁇ bed by the circuits and the magnitudes of the signal currents. Impedances of power and ground conductors create voltage drops along the conductors, causing the conductors to radiate elect ⁇ c fields m a common mode. Common mode EMI is typically reduced by reducmg the impedances of the power and ground conductors Reducmg the impedances of the power and ground conductors thus reduces EMI as well as power supply droop and ground bounce
  • the elect ⁇ cal impedance between any two parallel conductive planes may vary widely
  • the parallel conductive planes may exhibit multiple elect ⁇ cal resonances, resulting in alternating high and low impedance values
  • Parallel conductive planes tend to radiate a significant amount of differential mode EMI at then bounda ⁇ es (i.e., from their edges)
  • the magnitude of differential mode EMI radiated from the edges of the parallel conductive planes vanes with frequency and is directly proportional to the elect ⁇ cal impedance between the planes
  • Fig. 1 is a perspective view of a pair of 10 m. x 10 in. square conductive planes 110 and 120 separated by a fiberglass-epoxy composite dielectnc layer
  • Each conductive plane is made of copper and is 0.0014 m. thick.
  • the fiberglass-epoxy composite layer separating the planes has a dielectnc constant of 4.0 and is 0004 m. thick. If a 1 ampere constant current is supplied between the centers of the rectangular planes, with a varying frequency of the current, the magnitude of the steady state voltage between the centers of the rectangular planes can be determined 130.
  • the elect ⁇ cal impedance between the parallel conductive planes of Fig. 1 vanes widely at frequencies above about 200 MHz
  • the parallel conductive planes exhibit multiple electncal resonances at frequencies between 100 MHz and 1 GHz and above, resulting m alternating high and low impedance values
  • the parallel conductive planes of Fig. 1 would also radiate substantial amounts of EMI at frequencies where the elect ⁇ cal impedance between the planes anywhere near then- pe ⁇ phe ⁇ es is high.
  • a switching power supply 210 supplies current and voltage to a CMOS chip load 220
  • decouplmg capacitors 215 and the PCB 225 itself In parallel with the power supply 210 and the load 220 are decouplmg capacitors 215 and the PCB 225 itself, with its own capacitance. Smith [1994] teaches that decouplmg capacitors are only necessary up to 200-300 MHz, as the target impedances are rarely exceeded above that frequency.
  • the problems outlined above are in large part solved by a system and method for usmg a computer system to determine the desued decoupling components for stabilizing the elect ⁇ cal impedance m the power dist ⁇ bution system of an elect ⁇ cal interconnecting apparatus including a parr of parallel planar conductors separated by a dielectnc layer
  • the elect ⁇ cal interconnecting apparatus may be, for example, a PCB, a semiconductor device package substrate, or an mtegrated circuit substrate
  • the present method includes creating a model of the power dist ⁇ bution system based upon an MxN g ⁇ d for both the power plane and the ground plane
  • the model is preferably based upon a fixed g ⁇ d that adapts automatically to the actual physical dunensions of the elect ⁇ cal interconnecting apparatus
  • the model preferably also calculates the system response to chosen decouplmg components m both single node and MxN node versions
  • the model receives input from a user and from a database of vanous physical and/or electncal characte ⁇ stics for a plurality of decoupling components
  • Vanous characte ⁇ stics of mterest include physical dunensions, type, and thickness of dielectnc, method and matenals of manufacture, capacitance, mounted inductance, and ESR
  • the desired characte ⁇ stics are preferably saved in a database for co ⁇ echons, additions, deletions, and updates
  • the model of the power dist ⁇ bution system is m a form of a plane mcludmg two dimensional dist ⁇ ubbed transmission lmes
  • the model of the power distribution system may comp ⁇ se a plurality of the following- one or more physical dimensions of the power plane, one or more physical dimensions of the ground plane, physical separation distance of the power plane and the ground plane, composition of a dielectnc separating the power plane and the ground plane, one or more active device charactenstics, one or more power supply characte ⁇ stics, and one or more of the decouplmg components
  • M and N have a uniform value.
  • the active components act as cu ⁇ ent sources or sinks, and may mclude processors, memones, application specific mtegrated cucuits (ASICs), logic ICs, or any device that converts electncal energy mto information
  • a total sum of all values of the cu ⁇ ent sources m the model is scaled to equal one ampere.
  • the model of the power dist ⁇ bution system is operable for determining the decouplmg components for a frequency range above approximately a lowest board resonance frequency In another embodiment, the model of the power dist ⁇ bution system is operable for determining the decouplmg components for a frequency range above a highest resonance frequency from all resonance frequencies of the decouplmg components.
  • the method preferably includes determining a target impedance for the power distnbution system at a desued frequency or over a desired frequency range
  • the target impedance is preferably determined based upon such factors as power supply voltage, total cu ⁇ ent consumption, and allowable voltage ⁇ pple in the power distribution system.
  • determining the target impedance for the power distnbution system comprises the quotient of power supply voltage multiplied by allowable voltage ⁇ pple divided by total cu ⁇ ent.
  • the frequency range may start at dc and ⁇ se to several GHz.
  • the model of the power distribution system is operable for determining the decouplmg components for a frequency range above approximately a lowest board resonance frequency.
  • the model of the power distribution system is operable for determining the decouplmg components for a frequency range above a highest resonance frequency from all resonance frequencies of the decouplmg components.
  • the method preferably selects one or more decouplmg components from a plurality of possible decoupling components.
  • the decoupling components are capacitors with an approximate mounted inductance and an ESR.
  • a range of the values of the capacitors is chosen such that a superposition of impedance profiles provide an impedance at or below the target impedance for the power distnbution system over the frequency range of interest.
  • the impedance profiles of the plurality of possible decoupling components are compared to resonance frequencies for the power distribution system.
  • the decouplmg components have resonance frequencies that substantially co ⁇ espond to the resonance frequencies of the power dist ⁇ bution system m the frequency range of interest.
  • the method preferably determines a number for each of the one or more decoupling components chosen to be included as part of the power distribution model.
  • the number of the vanous decoupling components is chosen based upon the frequency range of interest and the impedance profiles of a plurality of possible decoupling components.
  • the number of a particular one of the decoupling components is chosen to have approximately equal value of a next larger integer of the quotient obtained from dividing the ESR for the particular decoupling components by the target impedance for the power distribution system.
  • the number of particular decoupling components has approximately equal value of impedance to the target impedance for the power distribution system when the number of the particular decoupling components are placed in parallel.
  • determining the number for the each of the decoupling components occurs before effectuating the model of the power distribution system to determine the transfer impedance values as the function of frequency at the one or more specific locations.
  • the method preferably places one or more cu ⁇ ent sources m the model of the power dist ⁇ bution system at one or more spatial locations co ⁇ esponding to one or more locations of active components.
  • the method also preferably places the decoupling components in the model of the power distribution system at nodal points that couple the MxN grid for the power plane and the co ⁇ esponding MxN grid for the ground plane.
  • the method places a power supply in the model of the power distribution system at a fixed location on the power plane.
  • the power supply is preferably compnsed m the model as one or more pole frequencies, one or more zero frequencies, and one or more resistances.
  • the method preferably selects one or more specific locations in the model of the power dist ⁇ bution system to calculate transfer impedance values as a function of frequency.
  • the method preferably effectuates the model of the power dist ⁇ bution system to determine the transfer impedance values as the function of frequency at the one or more specific locations previously chosen.
  • the method then preferably compares the transfer impedance values as the function of frequency at the one or more specific locations to the target impedance for the power distnbution system Preferably, the method determines a bill of goods for the power distnbution system based upon the results of effectuating the model
  • the method for determmmg decouplmg components for a power distnbution system includes determmmg a prefe ⁇ ed or optimum number of decouplmg components for a power distnbution system
  • a prefe ⁇ ed method for determmmg a number of decoupling components for a power dist ⁇ bution system is also disclosed
  • the method for determmmg a number of decouplmg components for a power dist ⁇ bution system comp ⁇ ses selecting a particular one of the decoupling components based upon a mounted inductance of each of the decouplmg components
  • the mounted inductance comp ⁇ ses an indication of a resonance frequency of that particular one of the decouplmg components
  • the method also compares an mdividual decouplmg component impedance of each of the decouplmg components to the target impedance
  • the method selects the number of the particular one of the decouplmg components which provides a total im
  • the method calculates the desued number of the particular decouplmg components in a parallel configuration
  • a plurality of decouplmg components are chosen as necessary to provide a total impedance at or below the target impedance for the given frequency range
  • the method for determmmg decouplmg components for a power distnbution system m cludes determining placement information for prefe ⁇ ed or optimum number of decouplmg components for a power dist ⁇ bution system
  • a prefe ⁇ ed method for determmmg placement of one or more decouplmg components m a power dist ⁇ bution system is also given.
  • each of the one or more decouplmg components mcludes a respective resonance frequency and a respective equivalent se ⁇ es resistance at the respective resonance frequency
  • the power dist ⁇ bution system mcludes a target impedance
  • the electncal interconnecting apparatus has at least a first dimension
  • the method determines one or more board resonance frequencies A first board frequency co ⁇ esponds to the first dimension
  • the method also selects one or more first decouplmg components from a plurality of possible decouplmg components such that the first decouplmg components have theu respective resonance frequency at approximately the f st board resonance frequency.
  • the method places the first decouplmg components on a location of the elect ⁇ cal interconnecting apparatus co ⁇ esponding to the first dimension. Additional dunensions of the elect ⁇ cal interconnecting apparatus may also requue theu own decouplmg components.
  • the fust dimension is preferably an effective length and the second dimension is preferably an effective width.
  • the prefe ⁇ ed location for placmg the decouplmg component for the first dimension comp ⁇ ses a first edge on the effective length, while the prefe ⁇ ed location for placmg the decoupling component for the second dimension compnses a second edge on the effective width
  • the method when the elecmcal interconnecting apparatus has at least one location for at least a first active device, the method further comp ⁇ ses placmg one or more second decouplmg components on the elecmcal interconnecting apparatus at the at least one location for at least the first active device Additional decouplmg components are also placed on the elecmcal interconnecting apparatus as needed for additional active devices.
  • the prefe ⁇ ed location for placmg decouplmg components for active devices is at or near the active devices.
  • the method cludes selecting the decouplmg components from a plurality of possible decouplmg components such that the decouplmg components have the respective resonance frequency at approximately the first operating frequency of the active device. Additional decouplmg components may be selected and placed based upon the harmonics of the operating frequency, as desued.
  • Fig. 1 is a perspective view of a representative elecmcal interconnecting apparatus comp ⁇ smg a p ⁇ or art pau of 10 m. x 10 in. square conductive planes separated by a fiberglass-epoxy composite dielectnc layer;
  • Fig. 2 is an embodiment of a pnor art smgle node model of a power distnbution system
  • Fig. 3 A is a top view of one embodiment of a model of a power dist ⁇ bution system
  • Fig. 3B is an embodiment of a unit cell of the power distnbution system model shown in Fig. 3 A;
  • Fig. 4 is a representative g ⁇ d of the nodal interconnections of the model of the power distnbution system shown m Fig. 3 A;
  • Fig. 5 is a flowchart of an embodiment of a method for determmmg decouplmg components for a power distribution system
  • Fig. 6 is a flowchart of an embodiment of a method for measuring the equivalent se ⁇ es resistance of an electrical device
  • Fig. 7 is a flowchart of an embodiment of a method for placmg decouplmg components m a power distribution system
  • Fig. 8A is a block diagram of an embodiment of a computer system operable to implement the methods of determining the decouplmg components for a power distribution system;
  • Fig. 8B is a flowchart of an embodiment of the method for determmmg decouplmg components for a power distribution system using the computer system of Fig. 8 A;
  • Fig. 9 is a flowchart of another embodiment of the method for determmmg decouplmg components for a power distribution system using the computer system of Fig. 8A. While the invention is susceptible to vanous modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawmgs and will herem be descnbed m detail. It should be understood, however, that the drawmgs and detailed desc ⁇ ption thereto are not mtended to limit the invention to the particular form disclosed, but on the contrary, the mtention is to cover all modifications, equivalents and alternatives falling within the spint and scope of the present mvention as defined by the appended claims.
  • Fig. 3A is a top view of a simplified schematic of one embodiment of a model of a power distribution system.
  • the model compnses a grid 300A of transmission line segments. The segments are grouped into unit cells 350. As shown, there are eight columns labeled "a” through “h”, as well as eight rows labeled, from the bottom to the top, "1" through “8”.
  • the model preferably comp ⁇ ses a SPICE a ⁇ ay of transmission lines in a fixed topology (i.e. in an 8x8 g ⁇ d). The transmission lines are va ⁇ able lengths such that the fixed topology may be used on elecmcal connecting device of any physical dimensions. It is noted that other embodiments of the power distribution system are contemplated, such as an elliptical model based on a "wheel and spoke" geometry.
  • Fig. 3B illustrates a close up view of the unit cell 350 from Fig. 3A.
  • the unit cell 350 is comprised of four transmission lines 355A - 355D.
  • the four transmission lines 355 connect together at nodal point pair 370, also refe ⁇ ed to as node 370.
  • connections to the center conductors represent plane 1, while connections to shield are plane 2.
  • the model is balanced, therefore either plane 1 or plane 2 may be power or ground, as desired.
  • Transmission lines 355A and 355B are preferably described with a width impedance "Zw" and a width time delay " Dw ".
  • Transmission Imes 355C and 355D are preferably described with a length impedance "Z L " and a length time delay "t DL ".
  • ⁇ , and R are resistances.
  • CM metalization thickness
  • ve/c is the speed of light in a vacuum (mches/sec)
  • ⁇ is the copper conductivity (per ohm/mils)
  • H o is the permeability of a vacuum (hen ⁇ es/mil)
  • factor is a calibration factor to compensate for capaci ⁇ ve loading
  • R lA ⁇ LIW)/2)* ⁇ l ⁇ * ⁇ l ⁇ hertz * ⁇ Q ⁇ )))
  • R lB ((L/W)/2) *(l/( ⁇ * cu))
  • the model represents an elecmcal interconnecting apparatus, which may be, for example, a printed cucuit board (PCB), a semiconductor device package substrate, or an integrated cucuit substrate.
  • the present method includes creating a model of the power distnbution system based upon an MxN g ⁇ d for both the power plane and the ground plane.
  • the model is preferably based upon a fixed g ⁇ d that adapts automatically to the actual physical dunensions of the elecmcal interconnecting apparatus.
  • the model preferably also calculates the system response to chosen decouplmg components in both smgle node and MxN node versions.
  • the model receives mput from a user and from a database of vanous physical and/or elecmcal characteristics for a plurality of decouplmg components.
  • Vanous charactenstics of interest include physical dunensions, type and thickness of dielecmc, method and matenals of manufacture, capacitance, mounted inductance, and equivalent se ⁇ es resistance (ESR).
  • ESR se ⁇ es resistance
  • the desued characte ⁇ stics are preferably saved m a database for corrections, additions, deletions, and updates.
  • the model of the power distribution system is m a form of a plane including two dimensional distributed transmission lmes.
  • the model of the power dismbutton system may comp ⁇ se a plurality of the following: one or more physical dimensions of the power plane, one or more physical dimensions of the ground plane, physical separation distance of the power plane and the ground plane, composition of a dielecmc separating the power plane and the ground plane, one or more active device characte ⁇ stics, one or more power supply characteristics, and one or more of the decouplmg components.
  • M and N have a uniform value, 8 as shown.
  • the active components act as cu ⁇ ent sources or sinks, and may include processors, memones, application specific mtegrated cucuits (ASICs), or logic ICs.
  • ASICs application specific mtegrated cucuits
  • a total sum of all values of the cu ⁇ ent sources m the model is scaled to equal one ampere.
  • the model of the power distnbution system is operable for determmmg the decoupling components for a frequency range above approximately a lowest board resonance frequency. Additional mformation may be found with respect to Fig. 5 below.
  • the model of the power distribution system is operable for determmmg the decouplmg components for a frequency range above a highest resonance frequency from all resonance frequencies of the decouplmg components.
  • the model uses a weighting factor in determining a number of a particular decoupling component to include m the model.
  • the weightmg factor is a dimensionless non-zero positive number.
  • the prefe ⁇ ed weighting factor is 0.2.
  • the EMI frequency range is preferably above approximately 200 MHz.
  • the weighting factor is 1.0 m a frequency range where signal integ ⁇ ty is most important.
  • the frequency range where signal integnty is important is preferably approximately 10 MHz up to approximately 200-300 MHz.
  • the weighting factor is preferably 2.0 at all active device operatmg frequencies and harmonics of the active device operating frequencies.
  • the model includes a frequency dependent skm effect loss.
  • Fig. 4 illustrates the 8x8 grid 300B of nodes 370 that are used to model the power and ground planes with the respective designations of al through h8, in a prefe ⁇ ed embodiment.
  • This grid 300B is used to determine the locations of the decouplmg components for the power dist ⁇ bution system.
  • Fig. 5 illustrates a flowchart of an embodiment of a method for determining decoupling components for a power distribution system.
  • the method is shown as a straight through method with no loop-back.
  • the method includes feedback loops at vanous stages to change previous inputs.
  • the method determines a target impedance for the power distribution system 510.
  • the target impedance is preferably determined at a desired frequency or over a desued frequency range.
  • the target impedance is determined based upon such factors as power supply voltage, total cu ⁇ ent consumption, and allowable voltage ripple in the power distribution system.
  • determmmg the target impedance for the power distribution system comprises the quotient of power supply voltage multiplied by allowable voltage ripple divided by total current.
  • the total cu ⁇ ent is normalized to one ampere.
  • the target impedance may be comprised m a group of known system parameters.
  • Other known system parameters may mclude one or more power supply characte ⁇ stics, the allowable voltage ⁇ pple, the total cu ⁇ ent consumption of all devices, one or more physical dimensions of the power distribution system, physical location constraints on where devices may be placed in the power distnbution system, and/or a frequency or frequency range of interest.
  • the method preferably selects a frequency range of interest 515.
  • the frequency range may start at dc and rise up to or above the gigahertz range.
  • the model of the power distribution system is operable for determining the decouplmg components for a frequency range above approximately a lowest board resonance frequency.
  • the model of the power dismbutton system is operable for determining the decouplmg components for a frequency range above a highest resonance frequency from all resonance frequencies of the decoupling components
  • the frequency range of interest may be comp ⁇ sed in the known system parameters
  • the frequency range of terest determines the output of the method by limiting the frequency range over which the system impedance is calculated in the model The method preferably determines the ESR for the plurality of decouplmg components 520.
  • the decouplmg components are preferably capacitors, but other devices with desirable capacinve and inductive values may be used
  • the ESR is preferably included m the database of vanous physical and/or elecmcal characte ⁇ stics for the plurality of decouplmg components Vanous other characte ⁇ s ⁇ cs of mterest may mclude physical dunensions, type and thickness of dielecmc, method and matenals of manufacture, capacitance, and mounted mductance.
  • the desued charactenstics are preferably saved m the database for co ⁇ ections, additions, deletions, and updates. Additional details concemmg determining the ESR for the plurality of decouplmg components 520 is given below with respect to Fig 6
  • the method preferably selects one or more desuable decouplmg components from a plurality of possible decouplmg components 525
  • the decouplmg components are capacitors with an approximate mounted mductance and an ESR
  • a range of the values of the capacitors is chosen such that a superposition of impedance profiles provide an impedance at or below the target impedance for the power dismbution system over the frequency range of mterest.
  • the impedance profiles of the plurality of possible decouplmg components are compared to resonance frequencies for the power dismbution system.
  • the decouplmg components have resonance frequencies, which should substantially co ⁇ espond to the resonance frequencies of the power dismbution system m the frequency range of mterest.
  • Resonance frequencies for the decouplmg components are preferably chosen to approximately co ⁇ espond to board resonance frequencies, operating frequencies and harmonics of active devices, including power supply, on the elecmcal interconnecting apparatus, and interaction resonance frequencies, high frequency response frequencies from interactions of the vanous components m the power dismbution system.
  • the capacitors are selected by the type of manufacture, such as the dielecmc composition, or a physical or elecmcal charactenstic value, such as the mounted mductance
  • the mounted inductance mcludes the geometry and physical couplmg to the elecmcal interconnecting apparatus.
  • the resonance frequency for a capacitor may be calculated from the mounted mductance and the capacitance usmg the following formula:
  • the impedance of the capacitor at the resonance frequency is the ESR It is noted that ceramic capacitors often have a deep cusp at the resonance frequency Tantalum capacitors often have a deep broad bottom with a va ⁇ able slope as a function of frequency
  • the optimum or desued number of each of the particular ones of the decouplmg components are determmed by the method 530
  • the number of each of the particular ones of the decouplmg components are determmed by the method 530 m response to the method selectmg one or more desuable decouplmg components from a plurality of possible decouplmg components 525
  • the method therefore, preferably determines a number (i.e. a counting number. 1. 2. .
  • the method determines how many of each particular decouplmg component to mclude in the model
  • the number of the vanous decouplmg components is chosen based upon the frequency range of interest and the impedance profiles of the plurality of possible decouplmg components
  • the number of a particular one of the decouplmg components is chosen to have approximately equal value of a next larger integer of the quotient obtained from dividing the ESR for the particular decouplmg components by the target impedance for the power dismbution system.
  • the number of a particular decouplmg components has approximately equal value of impedance to the target impedance for the power dismbution system when the number of the particular decouplmg components are placed m parallel.
  • determmmg the number for the each of the decouplmg components 530 occurs before effectuating the model of the power dismbution system to determine the transfer impedance values as the function of frequency at the one or more specific locations 560.
  • the number of a particular one of the one or more decouplmg components has approximately equal value of a next larger mteger of the quotient obtained from dividing an equivalent se ⁇ es resistance for the particular one of the one or more decouplmg components by the target impedance for the power dismbution system.
  • the number of decouplmg components is determmed for all decouplmg components 530 in the plurality of possible decouplmg components (i.e. m the database descnbed above) before selecting the decouplmg components to be used in the model 525.
  • the calculations for selectmg decouplmg components 525 and determining the number of each of the selected decouplmg components 530 are preferably performed by a computer system. Additional details may be gleaned below with respect to Figs. 7-9.
  • the cu ⁇ ent sources are placed at one or more spatial locations co ⁇ espondmg to one or more locations of active components
  • active components mclude processors, memones, application specific mtegrated cucuits (ASICs), or logic mtegrated cucuits (logic ICs).
  • active devices may act as cu ⁇ ent sources or sinks.
  • the total value of the cu ⁇ ent sources is preferably scaled to one ampere
  • the numbers, cu ⁇ ent ratings and strengths, and locations of the cu ⁇ ent sources may be mcluded in the known system parameters. In one embodiment, the placing of the cu ⁇ ent sources is performed by the computer system based on the known system parameters
  • the method places one or more power supplies in the model placed at nodal points 370 representing one or more spatial locations on the elecmcal interconnecting apparatus 545
  • the power supply is comp ⁇ sed in the model as one or more pole frequencies, one or more zero frequencies, and one or more resistances.
  • one pole frequency, one zero frequency, and two resistances are used, along with the output voltage.
  • the parameters are treated as a senes with one resistance m parallel with the zero frequency
  • the parameters and locations of any power supplies are usually pan of the known system parameters
  • placmg the power supply m the model is performed bv the computer svstem Additional details may be found with respect to Fig 8-9
  • the method also preferably places the decouplmg components in the model of the power dismbution system at nodal pomts 370 that couple the MxN g ⁇ d 300 for the power plane and the co ⁇ espondmg MxN gnd for the ground plane 550
  • Particular decouplmg components should optimally be placed as close as possible to those device locations which have resonance frequencies m the frequency range of mterest Resonance frequencies for the power dismbution system should be interpreted m this disclosure to mclude board resonance frequencies, operating frequencies and harmonics of active devices on the elecmcal interconnecting apparatus, and high frequency response frequencies from interactions of the vanous components in the power dismbution svstem High frequency response is often highly spatially dependent
  • Board resonance frequencies are a function of the physical dunensions of the power dismbution system and the dielecmc constant of the dielecmc that separates the power plane from the one or more ground planes
  • the board resonance frequencies of mterest m a prefe ⁇ ed embodiment mclude the half-, full-, three-half-, second-full-, and five-half-wave resonance frequencies for both the length and the width
  • the values for these board resonance frequencies are given by the approp ⁇ ate multiples of vel, L, and W as defined earlier
  • the half wave resonance for the length is ( ⁇ l2)*vel*L
  • the three-half wave resonance for the width is (3/2)*ve/*W
  • the decouplmg components are preferably placed along the edges, at points one-fifth in from each edge, and at pomts two-fifths m from each edge. It is noted that a square elecmcal connecting apparatus the lengthwise and widthwise resonances will be at the same frequencies and have maximums at co ⁇ espondmg locations It is also noted that similar relations are found with respect to an elecmcal connecting apparatus havmg a different geometry, such as elliptical, etc
  • the method preferablv selects one or more specific locations m the model of the power dismbution system to calculate transfer impedance values 555 as a function of frequency
  • the specific locations preferably mclude all 64 nodes on the 8x8 gnd To shorten execution tune of the computer svstem. other numbers of nodes may be chosen. It is noted that as the number of nodes increases, the model will accurately predict the board resonance frequencies up to higher frequencies In one embodiment, the model is run twice, once with a single specific node with all components placed on the smgle specific node and then a second time with the power dismbution system filling the entire 64 nodes of the model
  • the specific locations are usually pan of the known system parameters. It is noted that if fewer numbers of nodes are chosen, the usable bandwidth of the model will be lower.
  • the method preferably effectuates the model of the power dismbution system to determine the transfer impedance values as the function of frequency at the one or more specific locations previously chosen 560.
  • the model is effectuated by running computer programs on the computer system. Additional details may be found m the descnption of Figs 8-9
  • the method then preferablv compares the transfer impedance values as the function of frequency at the one or more specific locations to the target impedance for the power dismbution system 565.
  • one or more graphs are output which illustrates the transfer impedance values as a function of frequency
  • the graphs are computer generated.
  • the method outputs a resultant noise level for the power dismbution system due to the cu ⁇ ent sources and the decouplmg components at the specific locations. In still another embodiment, the method outputs the plurality of resultant impedances at the plurality of specific locations m the power dismbution system dynamically as a function of frequency.
  • FIG. 6 illustrates a flowchart of an embodiment of a method for measuring the ESR of an elecmcal device
  • the method compnses calibrating an impedance tester 610. Calibrating preferably compnses connecting the test heads to the impedance tester pnor to all other work.
  • the impedance tester is a HEWLETT-PACKARD model 4291 A RF Impedance/Mate ⁇ al Analyzer
  • the test heads preferably comp ⁇ se a low impedance test head, an APC7 connector for the test head, and an adapter to couple APC7 to an SMA connector
  • Calibrating preferably involves three test cases using a 50 ⁇ load, a short, and an open circuit
  • the method ve ⁇ fies 620 the calibration performed in 610 before mounting the elecmcal device Ve ⁇ fication preferably comp ⁇ ses comparing the expected smith chart reflection coefficient for each test case with the experimentally determmed reflection coefficient
  • the device is securely coupled to the impedance tester 630
  • securely couplmg the device to the tester compnses soldering the device to an SMA connector by connecting one side of the device to the central post and the other side of the device to the outer connector.
  • securely couplmg the device to the tester compnses mounting the device on the tester in such a fashion that stray capacitances and inductances are mostly eliminated.
  • the SMA connector is then mounted to the impedance tester.
  • the ESR of the device is the magnitude of the impedance at the frequency at which the device has a minimum measured impedance value is 180°.
  • measuring the impedance as a function of frequency is comp ⁇ sed as follows. Set the MAG (
  • Fig. 7 illustrates a flowchart of an embodiment of a method for selecting decouplmg components and placing the decoupling components in a power distribution system.
  • the method first determines resonance frequencies for the electrical interconnection apparatus, the active devices, and the power supply 710.
  • resonance frequency includes the operating frequencies and harmonics of the active devices and the power supply. Integer fractions of these frequencies may also be considered as resonance frequencies.
  • the resonance frequencies of the electrical interconnection apparatus are also described as board resonance frequencies or board frequencies.
  • the method selects approp ⁇ ate decoupling components 715. Approp ⁇ ate decoupling components have approximately co ⁇ esponding resonance frequencies to the system resonance frequencies determined in 710.
  • the method next places the approp ⁇ ate decoupling components in the model at appropriate and co ⁇ espondmg locations for the system resonance frequencies 720. After the model calculations are completed, the approp ⁇ ate decouplmg components will be placed on the elecmcal interconnection apparatus.
  • the method selects first decouplmg components co ⁇ esponding to the board resonance frequencies 715 In another embodiment, the method selects second decouplmg components co ⁇ espondmg to the active device operating frequencies 715 In still another embodiment, the method selects thud decouplmg components co ⁇ espondmg to one or more harmonics of the active device operating frequencies 715 The method may also select additional decouplmg components co ⁇ espondmg to additional board resonance frequencies, active device operating trequencies or harmonics, or interaction resonance frequencies 715
  • the elecmcal interconnection device has approximately a rectangular shape, the fust dimension co ⁇ esponds to an effectn e length and the second dimension co ⁇ esponds to an effective width
  • the prefe ⁇ ed locations for placmg decoupling components co ⁇ espondmg to the board resonance frequencies for the first and second dunensions include the edges along the length and the width
  • a prefe ⁇ ed location along the dimension m cludes the midpoint of the dimension
  • selecting approp ⁇ ate decoupling components with resonance frequencies approximately co ⁇ espondmg to the resonance frequencies of the power dismbution system 715 mcludes selecting the number of each of the decoupling components
  • the number of each of the decouplmg components is chosen in one embodiment based upon the frequency range of mterest and the impedance profiles of the plurality of possible decouplmg components
  • the numbers are chosen by a computer system
  • the computer system may access a database of values on the plurality possible decouplmg components, mcludmg values for physical and/or elecmcal characte ⁇ stics Elecmcal charactenstics mcluded in the database may mclude rated capacitance, equivalent senes resistance, and/or mounted mductance
  • the method for selecting decouplmg components and placmg the decouplmg components in the model further comp ⁇ ses effectuating the model and determining the system unpedance response at one or more selected locations If the system unped
  • the method may mclude comparing an impedance of each particular one of the decouplmg components chosen by the method to the target impedance The method may further select a number of each particular one of a decouplmg components to provide a total impedance at or below the target unpedance as a part of selecting appropnate decouplmg components 715 In yet another embodiment, the method selects decouplmg components above the lowest board resonance frequency In another embodiment, the method also selects decouplmg components above a highest resonance frequency of the decouplmg components Additional details on selectmg particular decouplmg components and the number of each particular one of the decouplmg components may be found elsewhere in this disclosure Although the method is shown in flowchart form, it is noted that po ⁇ ions of Fig 7 mav occur concu ⁇ ently or in different orders
  • Fig 8A illustrates a block diagram of an embodiment of a computer system for selectmg decouplmg components
  • the computer svstem m cludes a local computer 800 and a remote computer 850 coupled by a networking connection 890
  • the local computer 800 and the remote computer 850 are unified as a smgle computer, where the networking connection 890 comp ⁇ ses a bus m the smgle computer
  • Both the local computer 800 and the remote computer 850 are operable to accept mput from a database of physical and/or elecmcal charactenstic data for a plurality of decouplmg components 840
  • the database may be comp ⁇ sed m the local computer 800 or in remote computer 850
  • the database is compnsed m remote computer 850 and accessible to the local computer 800 through the networking connection 890
  • the database 840 is comp ⁇ sed external to both the local computer 800 and the remote computer 850, such as on a database
  • the local computer 800 is operable execute a first program, preferably a web browser 810
  • the web browser 810 is operable to run an teractive applet 820, preferably a JAVA applet, and to accept and display graphical output 830
  • Alternative embodiments may comp ⁇ se a JavaSc ⁇ pt program or HTML code
  • the JAVA applet 820 outputs component and placement data us g the http POST method to the remote computer
  • the CGI scnpt 855 receives the component and placement data and either mcludes or calls a PERL program to build a SPICE deck 860
  • CORBA, remote method invocation (RMI), or other methods may be used
  • the SPICE deck output of the PERL program 860 is sent to a simultaneous-equation-solver program, preferably a SPICE simulator such as HSPICE (available from Avant 1 Corporation, Fremont, California), which executes the SPICE deck 865
  • the HSPICE output is preferably converted by OCTAVE and
  • the graph from 870 is preferably sent from the remote computer 850 to the local computer 800 to be displayed as graphic output 830 m the web browser 810
  • the actions of the CGI scnpt 855 may also be performed by a second program
  • the second program comp ⁇ ses the simultaneous-equation-solver program.
  • the simultaneous equation solver program compnses a cucuit-solver program.
  • Other embodiments of the second programs are also contemplated, mcludmg custom hardware or software
  • Fig. 8B illustrates a flowchart of an embodiment of a method for determmmg decouplmg components for a power dismbution system, preferably usmg the computer system of Fig. 8A
  • Actions 801 (above the lme) preferably take place on the local computer 800
  • Actions 851 (below the lme) preferably take place on the remote computer 850
  • the actions 801 and 851 all take place m a smgle computer system
  • the actions 801 and 851 take place outside the computer system.
  • Systems parameters are defined in 806.
  • the system parameters include power supply voltage, allowable power supply npple, total cu ⁇ ent consumption, power supply poll frequency, power supply zero frequency, first and second power supply resistances, physical dimensions of the elecmcal interconnection device, dielecmc thickness and constant, metalization thickness of the elecmcal interconnection device, and the frequency range of mterest
  • the system parameters defined in 806 are used to calculate values for the target impedance and one or more board resonance frequencies 807
  • Configuration parameters are defined m 821
  • the integration parameters preferably mclude weightmg factors and mounted inductances for the plurality of decoupling components
  • mounted mductance refers to a loop mductance based on the geometry of the decouplmg components, pad geometry, distance to the power planes, and the location on the power planes Values are exuacted from the database of vanous physical and/or elecmcal charactenstic s for a plurality of decouplmg components 841
  • the database preferably mcludes the capacitance and ESR for the plurality of possible decouplmg components
  • the calculated values 807, the configuration definitions 821, and the database values 841 are mput to calculate the decouplmg component resonance frequencies, and the optimum number of each chosen decouplmg component 822 In one embodiment, the optimum number of each chosen decoup
  • Spatial placements for decouplmg components, cu ⁇ ent sources, power supply, and selected locations or probe pomts are chosen m 823, preferably by a user Further details on placmg the decouplmg components m the model of the power dismbution system are given elsewhere m this disclosure
  • Spatial placement data 823 and system parameter definitions 806 are combmed mto spatial placement data, mductance data, elect ⁇ c mterconnection device data, and power supply data 824 to be sent to the remote computer 850
  • the data that were sent to the remote computer 824 are used to build a SPICE deck 861
  • the SPICE deck is used as mput for a SPICE analysis 866, preferably usmg HSPICE Output from the SPICE analysis 866 is processed to create graphical output 871
  • the graph the output returned to the local computer 872, preferably to the web browser 810
  • the graphic display is preferably displayed on the local computer 826, preferably as an HTML page m the web browser 810
  • the HTML page compnses an SGML page, or other program as desued
  • the method is shown m flowchart form, it is noted that portions of Fig 8B may occur concu ⁇ ently or m different orders
  • Fig 9- Another Embodiment of the Computerized Method Fig 9 illustrates a flowchart of an embodiment of a computerized method for determmmg the decouplmg components for a power dismbution system
  • the method calculates the target impedance for the power dismbution system 900
  • the target impedance is preferably calculated as a power supply voltage tunes the allowable power supply ⁇ pple divided by the total cu ⁇ ent In a prefe ⁇ ed embodiment, the total cu ⁇ ent is no ⁇ nalized to one ampere
  • the calculated target unpedance is used to calculate an optimum number of each available decouplmg component 905 The optimum number is preferably defined as the ESR of the decouplmg component divided by the target impedance multiplied by the weighting factor
  • the method also calculates the resonance frequency of each available decouplmg component 910
  • the resonance frequency is preferably calculated as the mverse of two pi multiplied by the square root of the product of the mounted mductance
  • the prefe ⁇ ed cntena for accepting the results of the multi-node analysis are that the system transfer impedance is below the target unpedance over the frequency range of mterest Should results not be acceptable in 950, method modifies the choice of the decouplmg components, the number of each the decouplmg opponents, and/or placement of the decouplmg components 960 and reanalyzes the model usmg multi-node analysis 940 Although the method is shown in flowchart form, it is noted that portions of Fig. 9 may occur concu ⁇ ently or m different orders. Numerous va ⁇ ations and modifications will become apparent to those skilled m the art once the above disclosure is fully appreciated It is intended that the following claims be interpreted to embrace all such vanations and modifications.

Abstract

A system and method for using a computer system to determine the desired decoupling components for stabilizing the electrical impedance in the power distribution system of an electrical interconnecting apparatus, including a method for measuring the ESR for an electrical device, a method for determining a number of desired decoupling components for a power distribution system, and a method for placing the desired decoupling components in the power distribution system. The method creates a model of the power distribution system based upon an MxN grid for both the power plane and the ground plane. The model receives input from a user and from a database of various characteristics for a plurality of decoupling components. The method determines a target impedance over a desired frequency range. The method selects decoupling components. The method determines a number for each of the decoupling components chosen. The method places current sources in the model at spatial locations corresponding to physical locations of active components. The method optionally also places a power supply in the model. The method selects specific locations in the model to calculate transfer impedance values as a function of frequency. The method effectuates the model to determine the transfer impedance values as the function of frequency at the specific locations previously chosen. The method then compares the transfer impedance values as the function of frequency at the specific locations to the target impedance for the power distribution system.

Description

TITLE: SYSTEM AND METHOD FOR DETERMINING THE DESIRED DECOUPLING COMPONENTS FOR POWER DISTRIBUTION SYSTEMS USING A COMPUTER SYSTEM
BACKGROUND OF THE INVENTION
1 Field of the Invennon
This invennon relates to electronic systems, and more particularly to elecmcal interconnecting apparatus having continuous planar conductors.
2. Descπption of the Related Art
Electronic systems typically employ several different types of elecmcal interconnecting apparatus having planar layers of electπcally conductive material (i.e., planar conductors) separated by dielectnc layers A pornon of the conducnve layers may be patterned to form electπcally conductive signal lines or "traces" Conducnve traces in different layers (i.e.. on different levels) are typically connected using contact structures formed in openings in the dielectnc layers (i.e., vias). For example, integrated circuits typically have several layers of conducnve traces which interconnect electronic devices formed upon and within a semiconductor substrate. Each layer is separated from adjacent layers by dielectnc layers Within a semiconductor device package, several layers of conductive traces separated by dielectnc layers may be used to electncally connect bonding pads of an integrated circuit to terminals (e.g., pins or leads) of the device package. Printed circuit boards (PCBs) also typically have several layers of conducnve traces separated by dielectnc layers. The conductive traces are used to electπcally interconnect terminals of electronic devices mounted upon the PCB.
Signals m digital electronic systems typically cany informanon by alternating between two voltage levels (i.e., a low voltage level and a high voltage level). A digital signal cannot transition instantaneously from the low voltage level to the high voltage level, or vice versa. The fmite amount of time dunng which a digital signal transitions from the low voltage level to the high voltage level is called the πse tune of the signal. Similarly, the finite amount of tune dunng which a digital signal transinons from the high voltage level to the low voltage level is called the fall tune of the signal.
Digital electronic systems are continually being produced which operate at higher signal frequencies (i.e., higher speeds). In order for the digital signals within such systems to remain stable for appreciable penods of time between transitions, the rise and fall times of the signals must decrease as signal frequencies mcrease. This decrease in signal transinon tunes (i.e., nse and fall tunes) creates several problems within digital electronic systems, including signal degradation due to reflecnons, power supply "droop", ground "bounce", and increased electromagnenc emissions It is desirable that digital signals be transmitted and received within acceptable tolerances. A signal launched from a source end of a conducnve trace suffers degradation when a portion of the signal reflected from a load end of the trace arnves at the source end after the transition is complete (i.e., after the nse time or fall tune of the signal). A portion of the signal is reflected back from the load end of the trace when the input impedance of the load does not match the characteristic impedance of the trace When the length of a conductive trace is greater than the πse tune divided by three, the effects of reflections upon signal inteenty (i.e., transmission line effects) should be considered If necessary, steps should be taken to minimize the degradations of signals conveyed upon the trace due to reflections The act of altering impedances at the source or load ends of the trace in order to reduce signal reflections is referred to as "terminating the trace For example, the mput impedance of the load mav be altered to match the charactensfic impedance of the trace in order to prevent signal reflection As the transition tune (1 e . the πse or fall time) of the signal decreases so does the length of trace which must be terminated in order to reduce signal degradation
A digital signal alternating between the high and low voltage levels includes contπbutions from a fundamental sinusoidal frequency (1 e , a first harmonic) and integer multiples of the first harmonic As the πse and fall times of a digital signal decrease, the magnitudes of a greater number of the integer multiples of the first harmonic become significant As a general rule, the frequency content of a digital signal extends to a frequency equal to the reciprocal of π times the transition time (l e , nse or fall time) of the signal For example, a digital signal with a 1 nanosecond transition time has a frequency content extending up to about 318 MHz
All conductors have a certain amount of inductance The voltage across the inductance of a conductor is directly proportional to the rate of change of current through the conductor At the high frequencies present in conductors canying digital signals having short transition times, a significant voltage drop occurs across a conductor having even a small inductance A power supply conductor connects one terminal of an electncal power supply to a power supply terminal of a device, and a ground conductor connects a ground terminal of the power supply to a ground terminal of the device When the device generates a digital signal havmg short transition times, high frequency transient load currents flow in the power supply and ground conductors Power supply droop is the term used to descπbe the decrease in voltage at the power supply terminal of the device due to the flow of transient load cuπent through the inductance of the power supply conductor Similarly, ground bounce is the term used to descnbe the mcrease in voltage at the ground terminal of the device due to the flow of transient load current through the inductance of the ground conductor When the device generates several digital signals havmg short transition times simultaneously, the power supply droop and ground bounce effects are additive Sufficient power supply droop and ground bounce can cause the device to fail to function correctlv
Power supply droop is commonly reduced by arranging power supply conductors to form a cnsscross network of intersecting power supply conductors (l e , a power supply gπd) Such a gπd network has a lower inductance, hence power supply droop is reduced A continuous power supply plane may also be provided which has an even lower inductance than a gπd network Placing a "bypass" capacitor near the power supply terminal of the device is also used to reduce power supply droop The bypass capacitor supplies a substantial amount of the transient load current, thereby reducmg the amount of transient load cuπent flowing through the power supply conductor Ground bounce is reduced by using a low inductance ground conductor gπd network, or a continuous ground plane having an even lower amount of inductance Power supply and ground gπds or planes are commonly placed in close proximity to one another in order to further reduce the inductances of the gπds or planes Electromagnetic interference (EMI) is the term used to descπbe unwanted interference energies either conducted as currents or radiated as electromagnetic fields High frequency components present within circuits producing digital signals havmg short transition tunes may be coupled into nearbv electronic systems (e g , radio and television cucuits), disrupting proper operation of these systems The united States Federal Communication Commission has established upper limits for the amounts of EMI products for sale m the United States may generate
Signal circuits form current loops which radiate magnetic fields in a differential mode. Differential mode EMI is usually reduced by reducmg the areas proscπbed by the circuits and the magnitudes of the signal currents. Impedances of power and ground conductors create voltage drops along the conductors, causing the conductors to radiate electπc fields m a common mode. Common mode EMI is typically reduced by reducmg the impedances of the power and ground conductors Reducmg the impedances of the power and ground conductors thus reduces EMI as well as power supply droop and ground bounce
Within the wide frequency range present within electronic systems with digital signals havmg short transition times, the electπcal impedance between any two parallel conductive planes (e.g , adjacent power and ground planes) may vary widely The parallel conductive planes may exhibit multiple electπcal resonances, resulting in alternating high and low impedance values Parallel conductive planes tend to radiate a significant amount of differential mode EMI at then boundaπes (i.e., from their edges) The magnitude of differential mode EMI radiated from the edges of the parallel conductive planes vanes with frequency and is directly proportional to the electπcal impedance between the planes
Fig. 1 is a perspective view of a pair of 10 m. x 10 in. square conductive planes 110 and 120 separated by a fiberglass-epoxy composite dielectnc layer Each conductive plane is made of copper and is 0.0014 m. thick. The fiberglass-epoxy composite layer separating the planes has a dielectnc constant of 4.0 and is 0004 m. thick. If a 1 ampere constant current is supplied between the centers of the rectangular planes, with a varying frequency of the current, the magnitude of the steady state voltage between the centers of the rectangular planes can be determined 130.
The electπcal impedance between the parallel conductive planes of Fig. 1 vanes widely at frequencies above about 200 MHz The parallel conductive planes exhibit multiple electncal resonances at frequencies between 100 MHz and 1 GHz and above, resulting m alternating high and low impedance values The parallel conductive planes of Fig. 1 would also radiate substantial amounts of EMI at frequencies where the electπcal impedance between the planes anywhere near then- peπpheπes is high.
The above problems are currently solved in different ways at different frequency ranges At low frequency, the power supply uses a negative feedback loop to reduce fluctuations. At higher frequencies, large value bypass (i.e. decouplmg) capacitors are placed near devices. At the highest frequencies, up to about 200-300 MHz, very small bypass capacitors are placed very close to devices in an attempt to reduce their parasitic inductance, and thus high frequency impedance, to a minimum value. By November 2, 1994, the practical upper limit remained around 200-300 MHz as shown by Smith [Decouplmg Capacitor Calculations for CMOS Cucuits; pp. 101-105 in Proceedmgs of 3rd Topical Meetmg on Electπcal Performance of Electronic Packagmg of the Institute of Electπcal and Electronics Engineers, Inc.] The power distnbution system was modeled as shown m Fig. 2. A switching power supply 210 supplies current and voltage to a CMOS chip load 220 In parallel with the power supply 210 and the load 220 are decouplmg capacitors 215 and the PCB 225 itself, with its own capacitance. Smith [1994] teaches that decouplmg capacitors are only necessary up to 200-300 MHz, as the target impedances are rarely exceeded above that frequency. This upper limit changes over time as the clock frequencies mcrease and the allowable voltage npple decreases Determining the proper values for decouplmg capacitors and the optimum number of each has been a "tπal and error" process, which relies on the expeπence of the designer There are no known straightforward rules for choosing decouplmg capacitors tor all frequency ranges
It would thus be desirable to have a method for designing the power distπbution system and determining the desued decouplmg components for stabilizing the electπcal impedance in the power distnbution system The method is preferably automatable using a computer system to perform calculations
SUMMARY OF THE INVENTION
The problems outlined above are in large part solved by a system and method for usmg a computer system to determine the desued decoupling components for stabilizing the electπcal impedance m the power distπbution system of an electπcal interconnecting apparatus including a parr of parallel planar conductors separated by a dielectnc layer The electπcal interconnecting apparatus may be, for example, a PCB, a semiconductor device package substrate, or an mtegrated circuit substrate The present method includes creating a model of the power distπbution system based upon an MxN gπd for both the power plane and the ground plane The model is preferably based upon a fixed gπd that adapts automatically to the actual physical dunensions of the electπcal interconnecting apparatus The model preferably also calculates the system response to chosen decouplmg components m both single node and MxN node versions
The model receives input from a user and from a database of vanous physical and/or electncal characteπstics for a plurality of decoupling components Vanous characteπstics of mterest include physical dunensions, type, and thickness of dielectnc, method and matenals of manufacture, capacitance, mounted inductance, and ESR The desired characteπstics are preferably saved in a database for coπechons, additions, deletions, and updates
In one embodiment, the model of the power distπbution system is m a form of a plane mcludmg two dimensional distπbuted transmission lmes The model of the power distribution system may compπse a plurality of the following- one or more physical dimensions of the power plane, one or more physical dimensions of the ground plane, physical separation distance of the power plane and the ground plane, composition of a dielectnc separating the power plane and the ground plane, one or more active device charactenstics, one or more power supply characteπstics, and one or more of the decouplmg components In one embodiment, M and N have a uniform value. In vanous embodiments, the active components act as cuπent sources or sinks, and may mclude processors, memones, application specific mtegrated cucuits (ASICs), logic ICs, or any device that converts electncal energy mto information Preferably, a total sum of all values of the cuπent sources m the model is scaled to equal one ampere.
In one embodiment, the model of the power distπbution system is operable for determining the decouplmg components for a frequency range above approximately a lowest board resonance frequency In another embodiment, the model of the power distπbution system is operable for determining the decouplmg components for a frequency range above a highest resonance frequency from all resonance frequencies of the decouplmg components.
The method preferably includes determining a target impedance for the power distnbution system at a desued frequency or over a desired frequency range The target impedance is preferably determined based upon such factors as power supply voltage, total cuπent consumption, and allowable voltage πpple in the power distribution system. Preferably, determining the target impedance for the power distnbution system comprises the quotient of power supply voltage multiplied by allowable voltage πpple divided by total cuπent.
The frequency range may start at dc and πse to several GHz. In one embodiment, the model of the power distribution system is operable for determining the decouplmg components for a frequency range above approximately a lowest board resonance frequency. In another embodiment, the model of the power distribution system is operable for determining the decouplmg components for a frequency range above a highest resonance frequency from all resonance frequencies of the decouplmg components.
The method preferably selects one or more decouplmg components from a plurality of possible decoupling components. Preferably, the decoupling components are capacitors with an approximate mounted inductance and an ESR. In one embodiment, a range of the values of the capacitors is chosen such that a superposition of impedance profiles provide an impedance at or below the target impedance for the power distnbution system over the frequency range of interest. In one embodiment, the impedance profiles of the plurality of possible decoupling components are compared to resonance frequencies for the power distribution system. The decouplmg components have resonance frequencies that substantially coπespond to the resonance frequencies of the power distπbution system m the frequency range of interest.
The method preferably determines a number for each of the one or more decoupling components chosen to be included as part of the power distribution model. In one embodiment, the number of the vanous decoupling components is chosen based upon the frequency range of interest and the impedance profiles of a plurality of possible decoupling components. In another embodiment, the number of a particular one of the decoupling components is chosen to have approximately equal value of a next larger integer of the quotient obtained from dividing the ESR for the particular decoupling components by the target impedance for the power distribution system. In still another embodiment, the number of particular decoupling components has approximately equal value of impedance to the target impedance for the power distribution system when the number of the particular decoupling components are placed in parallel. In one embodiment, determining the number for the each of the decoupling components occurs before effectuating the model of the power distribution system to determine the transfer impedance values as the function of frequency at the one or more specific locations.
The method preferably places one or more cuπent sources m the model of the power distπbution system at one or more spatial locations coπesponding to one or more locations of active components. The method also preferably places the decoupling components in the model of the power distribution system at nodal points that couple the MxN grid for the power plane and the coπesponding MxN grid for the ground plane. In one embodiment the method places a power supply in the model of the power distribution system at a fixed location on the power plane. The power supply is preferably compnsed m the model as one or more pole frequencies, one or more zero frequencies, and one or more resistances. The method preferably selects one or more specific locations in the model of the power distπbution system to calculate transfer impedance values as a function of frequency. The method preferably effectuates the model of the power distπbution system to determine the transfer impedance values as the function of frequency at the one or more specific locations previously chosen. The method then preferably compares the transfer impedance values as the function of frequency at the one or more specific locations to the target impedance for the power distnbution system Preferably, the method determines a bill of goods for the power distnbution system based upon the results of effectuating the model
In vanous embodiments, the method for determmmg decouplmg components for a power distnbution system includes determmmg a prefeπed or optimum number of decouplmg components for a power distnbution system A prefeπed method for determmmg a number of decoupling components for a power distπbution system is also disclosed For a given frequency or frequency range, the method for determmmg a number of decouplmg components for a power distπbution system compπses selecting a particular one of the decoupling components based upon a mounted inductance of each of the decouplmg components The mounted inductance compπses an indication of a resonance frequency of that particular one of the decouplmg components The method also compares an mdividual decouplmg component impedance of each of the decouplmg components to the target impedance The method then selects the number of the particular one of the decouplmg components which provides a total impedance at or below the target impedance at the given frequency or frequency range
In one embodiment, if the impedance of the particular decouplmg component is greater than the target impedance, then the method calculates the desued number of the particular decouplmg components in a parallel configuration In embodiments that determine a number of each of a plurality of decouplmg components for a power distnbution system for a given frequency range, a plurality of decouplmg components are chosen as necessary to provide a total impedance at or below the target impedance for the given frequency range
In vanous embodiments, the method for determmmg decouplmg components for a power distnbution system mcludes determining placement information for prefeπed or optimum number of decouplmg components for a power distπbution system A prefeπed method for determmmg placement of one or more decouplmg components m a power distπbution system is also given. In one embodiment, each of the one or more decouplmg components mcludes a respective resonance frequency and a respective equivalent seπes resistance at the respective resonance frequency The power distπbution system mcludes a target impedance, and the electncal interconnecting apparatus has at least a first dimension The method determines one or more board resonance frequencies A first board frequency coπesponds to the first dimension The method also selects one or more first decouplmg components from a plurality of possible decouplmg components such that the first decouplmg components have theu respective resonance frequency at approximately the f st board resonance frequency. The method then places the first decouplmg components on a location of the electπcal interconnecting apparatus coπesponding to the first dimension. Additional dunensions of the electπcal interconnecting apparatus may also requue theu own decouplmg components.
In the embodiment where the elecmcal interconnecting apparatus has approximately rectangular dunensions. the fust dimension is preferably an effective length and the second dimension is preferably an effective width. The prefeπed location for placmg the decouplmg component for the first dimension compπses a first edge on the effective length, while the prefeπed location for placmg the decoupling component for the second dimension compnses a second edge on the effective width
In one embodiment, when the elecmcal interconnecting apparatus has at least one location for at least a first active device, the method further compπses placmg one or more second decouplmg components on the elecmcal interconnecting apparatus at the at least one location for at least the first active device Additional decouplmg components are also placed on the elecmcal interconnecting apparatus as needed for additional active devices. The prefeπed location for placmg decouplmg components for active devices is at or near the active devices.
In one embodiment, the method mcludes selecting the decouplmg components from a plurality of possible decouplmg components such that the decouplmg components have the respective resonance frequency at approximately the first operating frequency of the active device. Additional decouplmg components may be selected and placed based upon the harmonics of the operating frequency, as desued.
BRIEF DESCRIPTION OF THE DRAWINGS
Other objects and advantages of the invention will become apparent upon readmg the following detailed descπption and upon reference to the accompanying drawmgs m which:
Fig. 1 is a perspective view of a representative elecmcal interconnecting apparatus compπsmg a pπor art pau of 10 m. x 10 in. square conductive planes separated by a fiberglass-epoxy composite dielectnc layer;
Fig. 2 is an embodiment of a pnor art smgle node model of a power distnbution system;
Fig. 3 A is a top view of one embodiment of a model of a power distπbution system; Fig. 3B is an embodiment of a unit cell of the power distnbution system model shown in Fig. 3 A;
Fig. 4 is a representative gπd of the nodal interconnections of the model of the power distnbution system shown m Fig. 3 A;
Fig. 5 is a flowchart of an embodiment of a method for determmmg decouplmg components for a power distribution system; Fig. 6 is a flowchart of an embodiment of a method for measuring the equivalent seπes resistance of an electrical device;
Fig. 7 is a flowchart of an embodiment of a method for placmg decouplmg components m a power distribution system;
Fig. 8A is a block diagram of an embodiment of a computer system operable to implement the methods of determining the decouplmg components for a power distribution system;
Fig. 8B is a flowchart of an embodiment of the method for determmmg decouplmg components for a power distribution system using the computer system of Fig. 8 A; and
Fig. 9 is a flowchart of another embodiment of the method for determmmg decouplmg components for a power distribution system using the computer system of Fig. 8A. While the invention is susceptible to vanous modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawmgs and will herem be descnbed m detail. It should be understood, however, that the drawmgs and detailed descπption thereto are not mtended to limit the invention to the particular form disclosed, but on the contrary, the mtention is to cover all modifications, equivalents and alternatives falling within the spint and scope of the present mvention as defined by the appended claims.
DETAILED DESCRIPTION OF THE INVENTION
Incorporation by Reference
The following publications are hereby incorporated by reference in their entuery "Decouplmg Capacitor Calculations for CMOS Cucuits" by Larry D. Smith, IEEE Proceedmgs of the 3rd Topical Meeting on Elecmcal Performance of Electronic Packagmg, November 2, 1994; and
"Packagmg and Power Distnbution Design Considerations for a Sun Microsystems Desktop Workstation" by Larry D. Smith, IEEE Proceedmgs of the 6th Topical Meeting on Elecmcal Performance of Electronic Packagmg, October 27, 1997.
Fig. 3 - Power Distnbution System Model
Fig. 3A is a top view of a simplified schematic of one embodiment of a model of a power distribution system. As shown, the model compnses a grid 300A of transmission line segments. The segments are grouped into unit cells 350. As shown, there are eight columns labeled "a" through "h", as well as eight rows labeled, from the bottom to the top, "1" through "8". The model preferably compπses a SPICE aπay of transmission lines in a fixed topology (i.e. in an 8x8 gπd). The transmission lines are vaπable lengths such that the fixed topology may be used on elecmcal connecting device of any physical dimensions. It is noted that other embodiments of the power distribution system are contemplated, such as an elliptical model based on a "wheel and spoke" geometry.
Fig. 3B illustrates a close up view of the unit cell 350 from Fig. 3A. As shown, the unit cell 350 is comprised of four transmission lines 355A - 355D. The four transmission lines 355 connect together at nodal point pair 370, also refeπed to as node 370. As shown, connections to the center conductors represent plane 1, while connections to shield are plane 2. Note that the model is balanced, therefore either plane 1 or plane 2 may be power or ground, as desired. Transmission lines 355A and 355B are preferably described with a width impedance "Zw" and a width time delay " Dw". Transmission Imes 355C and 355D are preferably described with a length impedance "ZL" and a length time delay "tDL". Λ, and R, are resistances. The constants, parameters and variables of interest, as well as the equations that define and relate these quantities, along with the prefeπed units are given below:
"L" is the length of the plane (inches) " W is the width of the plane (inches)
"thk" is the thickness of the dielectnc (mils)
"CM" is the metalization thickness (mils)
"ve/c" is the speed of light in a vacuum (mches/sec)
"hertz" is the frequency variable '%" is the vacuum dielectric constant (permittivity) (picofarads/inch) " is the dielectric constant
"σ" is the copper conductivity (per ohm/mils)
"Ho" is the permeability of a vacuum (henπes/mil)
"ve/" is the velocity of a signal on the elecmcal interconnecting apparatus
Figure imgf000010_0001
"«" is the size of the grid, i.e. 8 as shown
"asp" is the aspect ratio of the grid, asp = L I IV
"factor" is a calibration factor to compensate for capaciπve loading
factor =
Figure imgf000010_0002
"/FL" is the tune of flight for the length dimension, rFL = L I vel
"ffV is the tune of flight for tne width dimension, fFW = W I vel
"tDL" is the transmission line delay time for the length dimension
1 DL = ln !(2n) factor "/Dw" is the transmission line delay time for the width dimension
1 D» = l FW i '(2n) factor "cap" is the parallel plate capacitance of the plane
cap = (ε0εrLW)/(\0'9thk) "ZL" is the impedance m the length duection
ZL = (n l cap)(tFL + asp * tFW ) factor "Zw" is the impedance in the width duection, Zv = ZL I asp
"R " is the smaller of:
RlA = {{LIW)/2)* {\l{σ* {\l^hertz * πμQσ))) RlB = ((L/W)/2) *(l/(σ* cu))
" ?," is the smaller of:
Rlλ = ((WIL)/2) * (\l{σ* {\l.jhertz *πμ0σ)))
R2B = ((W/L)/2)* (l/(σ*cu)) The model represents an elecmcal interconnecting apparatus, which may be, for example, a printed cucuit board (PCB), a semiconductor device package substrate, or an integrated cucuit substrate. The present method includes creating a model of the power distnbution system based upon an MxN gπd for both the power plane and the ground plane. The model is preferably based upon a fixed gπd that adapts automatically to the actual physical dunensions of the elecmcal interconnecting apparatus. The model preferably also calculates the system response to chosen decouplmg components in both smgle node and MxN node versions.
The model receives mput from a user and from a database of vanous physical and/or elecmcal characteristics for a plurality of decouplmg components. Vanous charactenstics of interest include physical dunensions, type and thickness of dielecmc, method and matenals of manufacture, capacitance, mounted inductance, and equivalent seπes resistance (ESR). The desued characteπstics are preferably saved m a database for corrections, additions, deletions, and updates.
In one embodiment, the model of the power distribution system is m a form of a plane including two dimensional distributed transmission lmes. The model of the power dismbutton system may compπse a plurality of the following: one or more physical dimensions of the power plane, one or more physical dimensions of the ground plane, physical separation distance of the power plane and the ground plane, composition of a dielecmc separating the power plane and the ground plane, one or more active device characteπstics, one or more power supply characteristics, and one or more of the decouplmg components. In a prefeπed embodiment, M and N have a uniform value, 8 as shown. In various embodiments, the active components act as cuπent sources or sinks, and may include processors, memones, application specific mtegrated cucuits (ASICs), or logic ICs. Preferably, a total sum of all values of the cuπent sources m the model is scaled to equal one ampere.
In one embodiment, the model of the power distnbution system is operable for determmmg the decoupling components for a frequency range above approximately a lowest board resonance frequency. Additional mformation may be found with respect to Fig. 5 below. In another embodiment, the model of the power distribution system is operable for determmmg the decouplmg components for a frequency range above a highest resonance frequency from all resonance frequencies of the decouplmg components.
In one embodiment, the model uses a weighting factor in determining a number of a particular decoupling component to include m the model. The weightmg factor is a dimensionless non-zero positive number. In the frequency range where EMI is most important, the prefeπed weighting factor is 0.2. The EMI frequency range is preferably above approximately 200 MHz. Preferably, the weighting factor is 1.0 m a frequency range where signal integπty is most important. The frequency range where signal integnty is important is preferably approximately 10 MHz up to approximately 200-300 MHz. The weighting factor is preferably 2.0 at all active device operatmg frequencies and harmonics of the active device operating frequencies. In a prefeπed embodiment, the model includes a frequency dependent skm effect loss.
Fig. 4 illustrates the 8x8 grid 300B of nodes 370 that are used to model the power and ground planes with the respective designations of al through h8, in a prefeπed embodiment. This grid 300B is used to determine the locations of the decouplmg components for the power distπbution system.
Fig. 5 - Method for Determining Decoupling Components
Fig. 5 illustrates a flowchart of an embodiment of a method for determining decoupling components for a power distribution system. The method is shown as a straight through method with no loop-back. In other embodiments, the method includes feedback loops at vanous stages to change previous inputs.
The method determines a target impedance for the power distribution system 510. The target impedance is preferably determined at a desired frequency or over a desued frequency range. The target impedance is determined based upon such factors as power supply voltage, total cuπent consumption, and allowable voltage ripple in the power distribution system. Preferably, determmmg the target impedance for the power distribution system comprises the quotient of power supply voltage multiplied by allowable voltage ripple divided by total current. In a prefeπed embodiment, the total cuπent is normalized to one ampere. The target impedance may be comprised m a group of known system parameters. Other known system parameters may mclude one or more power supply characteπstics, the allowable voltage πpple, the total cuπent consumption of all devices, one or more physical dimensions of the power distribution system, physical location constraints on where devices may be placed in the power distnbution system, and/or a frequency or frequency range of interest.
The method preferably selects a frequency range of interest 515. The frequency range may start at dc and rise up to or above the gigahertz range. In one embodiment, the model of the power distribution system is operable for determining the decouplmg components for a frequency range above approximately a lowest board resonance frequency. In another embodiment, the model of the power dismbutton system is operable for determining the decouplmg components for a frequency range above a highest resonance frequency from all resonance frequencies of the decoupling components As mentioned above, the frequency range of interest may be compπsed in the known system parameters In one embodiment, the frequency range of terest determines the output of the method by limiting the frequency range over which the system impedance is calculated in the model The method preferably determines the ESR for the plurality of decouplmg components 520. The decouplmg components are preferably capacitors, but other devices with desirable capacinve and inductive values may be used The ESR is preferably included m the database of vanous physical and/or elecmcal characteπstics for the plurality of decouplmg components Vanous other characteπsπcs of mterest may mclude physical dunensions, type and thickness of dielecmc, method and matenals of manufacture, capacitance, and mounted mductance. The desued charactenstics are preferably saved m the database for coπections, additions, deletions, and updates. Additional details concemmg determining the ESR for the plurality of decouplmg components 520 is given below with respect to Fig 6
The method preferably selects one or more desuable decouplmg components from a plurality of possible decouplmg components 525 Preferably, the decouplmg components are capacitors with an approximate mounted mductance and an ESR In one embodiment, a range of the values of the capacitors is chosen such that a superposition of impedance profiles provide an impedance at or below the target impedance for the power dismbution system over the frequency range of mterest. In another embodiment, the impedance profiles of the plurality of possible decouplmg components are compared to resonance frequencies for the power dismbution system. The decouplmg components have resonance frequencies, which should substantially coπespond to the resonance frequencies of the power dismbution system m the frequency range of mterest. Resonance frequencies for the decouplmg components are preferably chosen to approximately coπespond to board resonance frequencies, operating frequencies and harmonics of active devices, including power supply, on the elecmcal interconnecting apparatus, and interaction resonance frequencies, high frequency response frequencies from interactions of the vanous components m the power dismbution system. In vanous embodmients, the capacitors are selected by the type of manufacture, such as the dielecmc composition, or a physical or elecmcal charactenstic value, such as the mounted mductance The mounted inductance mcludes the geometry and physical couplmg to the elecmcal interconnecting apparatus. The resonance frequency for a capacitor may be calculated from the mounted mductance and the capacitance usmg the following formula:
Figure imgf000013_0001
The impedance of the capacitor at the resonance frequency is the ESR It is noted that ceramic capacitors often have a deep cusp at the resonance frequency Tantalum capacitors often have a deep broad bottom with a vaπable slope as a function of frequency
Once the desued decouplmg components have been selected, the optimum or desued number of each of the particular ones of the decouplmg components are determmed by the method 530 In one embodiment, the number of each of the particular ones of the decouplmg components are determmed by the method 530 m response to the method selectmg one or more desuable decouplmg components from a plurality of possible decouplmg components 525 The method, therefore, preferably determines a number (i.e. a counting number. 1. 2. . ) for each of the one or more decouplmg components chosen to be mcluded as part of the power dismbution model 530 In other words, the method determines how many of each particular decouplmg component to mclude in the model In one embodiment, the number of the vanous decouplmg components is chosen based upon the frequency range of interest and the impedance profiles of the plurality of possible decouplmg components In another embodiment, the number of a particular one of the decouplmg components is chosen to have approximately equal value of a next larger integer of the quotient obtained from dividing the ESR for the particular decouplmg components by the target impedance for the power dismbution system.
In still another embodiment, the number of a particular decouplmg components has approximately equal value of impedance to the target impedance for the power dismbution system when the number of the particular decouplmg components are placed m parallel. In one embodiment, determmmg the number for the each of the decouplmg components 530 occurs before effectuating the model of the power dismbution system to determine the transfer impedance values as the function of frequency at the one or more specific locations 560. In another embodiment, the number of a particular one of the one or more decouplmg components has approximately equal value of a next larger mteger of the quotient obtained from dividing an equivalent seπes resistance for the particular one of the one or more decouplmg components by the target impedance for the power dismbution system. In still another embodiment, the number of decouplmg components is determmed for all decouplmg components 530 in the plurality of possible decouplmg components (i.e. m the database descnbed above) before selecting the decouplmg components to be used in the model 525. The calculations for selectmg decouplmg components 525 and determining the number of each of the selected decouplmg components 530 are preferably performed by a computer system. Additional details may be gleaned below with respect to Figs. 7-9.
The method creates (i.e. realizes or implements) the model of the power dismbution m 535. The prefeπed model is descnbed above with respect to Fig. 3A and 3B. In a prefeπed embodiment, the model is computerized Additional details may be found elsewhere m this disclosure. The method next populates the model of the power dismbution system That is. the method adds to the model those devices that are coupled to the elecmcal interconnecting apparatus. The method places cuπent sources (or sinks) m the model at nodal points 370 on the MxN gπd 300B m 540. The cuπent sources are placed at one or more spatial locations coπespondmg to one or more locations of active components Examples of active components mclude processors, memones, application specific mtegrated cucuits (ASICs), or logic mtegrated cucuits (logic ICs). It is noted that active devices may act as cuπent sources or sinks. The total value of the cuπent sources is preferably scaled to one ampere The numbers, cuπent ratings and strengths, and locations of the cuπent sources may be mcluded in the known system parameters. In one embodiment, the placing of the cuπent sources is performed by the computer system based on the known system parameters
Optionally, the method places one or more power supplies in the model placed at nodal points 370 representing one or more spatial locations on the elecmcal interconnecting apparatus 545 The power supply is compπsed in the model as one or more pole frequencies, one or more zero frequencies, and one or more resistances. Preferably, one pole frequency, one zero frequency, and two resistances are used, along with the output voltage. Typically, the parameters are treated as a senes with one resistance m parallel with the zero frequency The parameters and locations of any power supplies are usually pan of the known system parameters In one embodiment, placmg the power supply m the model is performed bv the computer svstem Additional details may be found with respect to Fig 8-9
The method also preferably places the decouplmg components in the model of the power dismbution system at nodal pomts 370 that couple the MxN gπd 300 for the power plane and the coπespondmg MxN gnd for the ground plane 550 Particular decouplmg components should optimally be placed as close as possible to those device locations which have resonance frequencies m the frequency range of mterest Resonance frequencies for the power dismbution system should be interpreted m this disclosure to mclude board resonance frequencies, operating frequencies and harmonics of active devices on the elecmcal interconnecting apparatus, and high frequency response frequencies from interactions of the vanous components in the power dismbution svstem High frequency response is often highly spatially dependent
Board resonance frequencies are a function of the physical dunensions of the power dismbution system and the dielecmc constant of the dielecmc that separates the power plane from the one or more ground planes The board resonance frequencies of mterest m a prefeπed embodiment mclude the half-, full-, three-half-, second-full-, and five-half-wave resonance frequencies for both the length and the width The values for these board resonance frequencies are given by the appropπate multiples of vel, L, and W as defined earlier For example, the half wave resonance for the length is (\l2)*vel*L The three-half wave resonance for the width is (3/2)*ve/*W
To suppress the board resonance frequencies, decouplmg components are placed in the power dismbution system at locations that provide a low impedance shunt for high impedance resonance nodes (l e high voltage standing wave pomts) By noting where the board resonance has one or more maximums, the placement follows at or near those coπespondmg locations For a half wave resonance, the decouplmg components should be placed along the edges of the power dismbution system or the elecmcal connecting apparatus Smce the apparatus is not one dimensional, the decouplmg components are placed on the lme resulting from the intersection of the resonance and the plane defining the power dismbution system Therefore, the decouplmg components for the length half- wave resonance are preferably placed along the edges on the width of the power dismbution system For the full wave resonance, the decouplmg components are preferably placed along the edges and along the center axis of the power dismbution system For the three-half-wave resonance, the decouplmg components are preferably placed along the edges and at pomts one-thud m from each edge For the second-full-wave resonance, the decouplmg components are preferably placed along the edges, along the center axis, and at pomts one-fourth m from each edge. For the five-half-wave resonance, the decouplmg components are preferably placed along the edges, at points one-fifth in from each edge, and at pomts two-fifths m from each edge. It is noted that a square elecmcal connecting apparatus the lengthwise and widthwise resonances will be at the same frequencies and have maximums at coπespondmg locations It is also noted that similar relations are found with respect to an elecmcal connecting apparatus havmg a different geometry, such as elliptical, etc
Resonance or operating frequencies for the power supply are usually low enough that the capacitance can be treated as a lumped capacitance Thus decouplmg components for the power supply may be placed anywhere on the elecmcal interconnecting apparatus Spatial limitations on locations must always be observed This means that some decouplmg components will be placed farther away from the noise source than optimum The model will often indicate that additional ones of those decouplmg components will need to be placed on the elecmcal interconnecting apparatus at the farther away location In one embodiment, placement of decouplmg components 550 is input to the computerized model Additional details may be found m the descπptions of Figs. 8-9
The method preferablv selects one or more specific locations m the model of the power dismbution system to calculate transfer impedance values 555 as a function of frequency The specific locations preferably mclude all 64 nodes on the 8x8 gnd To shorten execution tune of the computer svstem. other numbers of nodes may be chosen. It is noted that as the number of nodes increases, the model will accurately predict the board resonance frequencies up to higher frequencies In one embodiment, the model is run twice, once with a single specific node with all components placed on the smgle specific node and then a second time with the power dismbution system filling the entire 64 nodes of the model The specific locations are usually pan of the known system parameters. It is noted that if fewer numbers of nodes are chosen, the usable bandwidth of the model will be lower.
The method preferably effectuates the model of the power dismbution system to determine the transfer impedance values as the function of frequency at the one or more specific locations previously chosen 560. In one embodiment, the model is effectuated by running computer programs on the computer system. Additional details may be found m the descnption of Figs 8-9 The method then preferablv compares the transfer impedance values as the function of frequency at the one or more specific locations to the target impedance for the power dismbution system 565. In one embodiment, one or more graphs are output which illustrates the transfer impedance values as a function of frequency Preferably, the graphs are computer generated. In another embodiment, the method outputs a resultant noise level for the power dismbution system due to the cuπent sources and the decouplmg components at the specific locations. In still another embodiment, the method outputs the plurality of resultant impedances at the plurality of specific locations m the power dismbution system dynamically as a function of frequency.
Preferably, the method determines at least a portion of a "bill of goods" for the power dismbution system based upon the results of effectuating the model 570. The bill of goods lists all relevant mformation from the selecting and placmg of the vanous decouplmg components. The bill of goods is preferably sufficient to allow mass production of the elecmcal interconnecting apparatus modeled to occur with proper decouplmg of the final product. Although the method is shown in flowchart form, it is noted that portions of Fig 5 may occur concuπently or m different orders
Fig. 6 - Method for Measuring ESR Fig. 6 illustrates a flowchart of an embodiment of a method for measuring the ESR of an elecmcal device
The method compnses calibrating an impedance tester 610. Calibrating preferably compnses connecting the test heads to the impedance tester pnor to all other work. In a prefeπed embodiment, the impedance tester is a HEWLETT-PACKARD model 4291 A RF Impedance/Mateπal Analyzer The test heads preferably compπse a low impedance test head, an APC7 connector for the test head, and an adapter to couple APC7 to an SMA connector Calibrating preferably involves three test cases using a 50 Ω load, a short, and an open circuit
The method veπfies 620 the calibration performed in 610 before mounting the elecmcal device Veπfication preferably compπses comparing the expected smith chart reflection coefficient for each test case with the experimentally determmed reflection coefficient After the impedance test passes the calibration, the device is securely coupled to the impedance tester 630 Preferably, securely couplmg the device to the tester compnses soldering the device to an SMA connector by connecting one side of the device to the central post and the other side of the device to the outer connector. In another embodiment, securely couplmg the device to the tester compnses mounting the device on the tester in such a fashion that stray capacitances and inductances are mostly eliminated. The SMA connector is then mounted to the impedance tester. The method measures the impedance of the device as a function of frequency over the desued frequency range 640. In a prefeπed embodiment, both the magnitude and the phase angle of the impedance are measured. Preferably, the measurement is repeated multiple times and the results averaged. The method then verifies the results of the measurements 650. Preferably, verification compnses comparing 180° to the phase angle shift at the frequency at which the device has a minimum measured impedance value. If the phase shift at the frequency at which the device has a minimum measured impedance value is not 180° at an acceptable uncertamty, then the results are discarded and the method performed anew. If the phase shift at the frequency at which the device has a minimum measured impedance value is 180° at an accepted uncertainty, then the ESR of the device is the magnitude of the impedance at the frequency at which the device has a minimum measured impedance value is 180°. Although the method is shown in flowchart form, it is noted that portions of Fig. 6 may occur concuπently or in different orders.
In a prefeπed embodiment, measuring the impedance as a function of frequency is compπsed as follows. Set the MAG (|Z|) and (θz) from "Meas" under the dual parameter key on the HP 4291A. Choose frequency range from 1 MHz to 1.8 GHz under the sweep button menu. Choose sweep type as logarithmic. Choose Marker search under search button and set it to minimum. Set marker search to on. Click on "Bw/Avg" menu under measurement block. Choose Sweep average and set average factor to 20. Hit sweep average start button to start taking measurements as a function of frequency. Note the minimum value after the averaging counter reaches 20. Repeat steps for each device.
Fig. 7 - Method for Selecting and Placing Decoupling Components Fig. 7 illustrates a flowchart of an embodiment of a method for selecting decouplmg components and placing the decoupling components in a power distribution system. The method first determines resonance frequencies for the electrical interconnection apparatus, the active devices, and the power supply 710. Note that "resonance frequency" includes the operating frequencies and harmonics of the active devices and the power supply. Integer fractions of these frequencies may also be considered as resonance frequencies. The resonance frequencies of the electrical interconnection apparatus are also described as board resonance frequencies or board frequencies. The method then selects appropπate decoupling components 715. Appropπate decoupling components have approximately coπesponding resonance frequencies to the system resonance frequencies determined in 710. The method next places the appropπate decoupling components in the model at appropriate and coπespondmg locations for the system resonance frequencies 720. After the model calculations are completed, the appropπate decouplmg components will be placed on the elecmcal interconnection apparatus.
In vanous embodiments, the electrical interconnection apparatus may have one or more board resonance frequencies, with each of the board resonance frequencies coπesponding to one or more particular dunensions of the elecmcal interconnection apparatus. Placement of a decoupling component 720 coπespondmg to a particular board resonance frequency is preferably at a location coπespondmg to the particular dimension m question. In one embodiment the method selects first decouplmg components coπesponding to the board resonance frequencies 715 In another embodiment, the method selects second decouplmg components coπespondmg to the active device operating frequencies 715 In still another embodiment, the method selects thud decouplmg components coπespondmg to one or more harmonics of the active device operating frequencies 715 The method may also select additional decouplmg components coπespondmg to additional board resonance frequencies, active device operating trequencies or harmonics, or interaction resonance frequencies 715
In an embodiment w here the elecmcal interconnection device has approximately a rectangular shape, the fust dimension coπesponds to an effectn e length and the second dimension coπesponds to an effective width The prefeπed locations for placmg decoupling components coπespondmg to the board resonance frequencies for the first and second dunensions include the edges along the length and the width A prefeπed location along the dimension mcludes the midpoint of the dimension
In one embodiment selecting appropπate decoupling components with resonance frequencies approximately coπespondmg to the resonance frequencies of the power dismbution system 715 mcludes selecting the number of each of the decoupling components The number of each of the decouplmg components is chosen in one embodiment based upon the frequency range of mterest and the impedance profiles of the plurality of possible decouplmg components In another embodiment, the numbers are chosen by a computer system The computer system may access a database of values on the plurality possible decouplmg components, mcludmg values for physical and/or elecmcal characteπstics Elecmcal charactenstics mcluded in the database may mclude rated capacitance, equivalent senes resistance, and/or mounted mductance In another embodiment, the method for selecting decouplmg components and placmg the decouplmg components in the model further compπses effectuating the model and determining the system unpedance response at one or more selected locations If the system unpedance response at the one or more selected locations is above a target impedance, the method selects additional decouplmg components m the proper frequency range The method places the additional decouplmg components m available locations The available locations may be consuamed due to existmg devices on the elecmcal interconnection apparatus, mcludmg other decouplmg components
In still another embodiment, the method may mclude comparing an impedance of each particular one of the decouplmg components chosen by the method to the target impedance The method may further select a number of each particular one of a decouplmg components to provide a total impedance at or below the target unpedance as a part of selecting appropnate decouplmg components 715 In yet another embodiment, the method selects decouplmg components above the lowest board resonance frequency In another embodiment, the method also selects decouplmg components above a highest resonance frequency of the decouplmg components Additional details on selectmg particular decouplmg components and the number of each particular one of the decouplmg components may be found elsewhere in this disclosure Although the method is shown in flowchart form, it is noted that poπions of Fig 7 mav occur concuπently or in different orders
Fig 8 - Computer System and Method for Selecting Decouplmg Components
Fig 8A illustrates a block diagram of an embodiment of a computer system for selectmg decouplmg components As shown, the computer svstem mcludes a local computer 800 and a remote computer 850 coupled by a networking connection 890 In one embodiment, the local computer 800 and the remote computer 850 are unified as a smgle computer, where the networking connection 890 compπses a bus m the smgle computer Both the local computer 800 and the remote computer 850 are operable to accept mput from a database of physical and/or elecmcal charactenstic data for a plurality of decouplmg components 840 In vanous embodiments, the database may be compπsed m the local computer 800 or in remote computer 850 In a prefeπed embodiment, the database is compnsed m remote computer 850 and accessible to the local computer 800 through the networking connection 890 In another embodiment, the database 840 is compπsed external to both the local computer 800 and the remote computer 850, such as on a database computer
As shown, the local computer 800 is operable execute a first program, preferably a web browser 810 The web browser 810 is operable to run an teractive applet 820, preferably a JAVA applet, and to accept and display graphical output 830 Alternative embodiments may compπse a JavaScπpt program or HTML code The JAVA applet 820 outputs component and placement data us g the http POST method to the remote computer The CGI scnpt 855 receives the component and placement data and either mcludes or calls a PERL program to build a SPICE deck 860 In other embodiments, CORBA, remote method invocation (RMI), or other methods may be used The SPICE deck output of the PERL program 860 is sent to a simultaneous-equation-solver program, preferably a SPICE simulator such as HSPICE (available from Avant1 Corporation, Fremont, California), which executes the SPICE deck 865 The HSPICE output is preferably converted by OCTAVE and GNUPlot mto a graph 870. The graph from 870 is preferably sent from the remote computer 850 to the local computer 800 to be displayed as graphic output 830 m the web browser 810 The actions of the CGI scnpt 855 may also be performed by a second program In one embodiment, the second program compπses the simultaneous-equation-solver program. In another embodiment, the simultaneous equation solver program compnses a cucuit-solver program. Other embodiments of the second programs are also contemplated, mcludmg custom hardware or software
Fig. 8B illustrates a flowchart of an embodiment of a method for determmmg decouplmg components for a power dismbution system, preferably usmg the computer system of Fig. 8A Actions 801 (above the lme) preferably take place on the local computer 800 Actions 851 (below the lme) preferably take place on the remote computer 850 In one embodiment, the actions 801 and 851 all take place m a smgle computer system In another embodiment, the actions 801 and 851 take place outside the computer system. Systems parameters are defined in 806. Preferably, the system parameters mclude power supply voltage, allowable power supply npple, total cuπent consumption, power supply poll frequency, power supply zero frequency, first and second power supply resistances, physical dimensions of the elecmcal interconnection device, dielecmc thickness and constant, metalization thickness of the elecmcal interconnection device, and the frequency range of mterest
The system parameters defined in 806 are used to calculate values for the target impedance and one or more board resonance frequencies 807 Configuration parameters are defined m 821 The integration parameters preferably mclude weightmg factors and mounted inductances for the plurality of decoupling components For purposes of this disclosure, mounted mductance refers to a loop mductance based on the geometry of the decouplmg components, pad geometry, distance to the power planes, and the location on the power planes Values are exuacted from the database of vanous physical and/or elecmcal charactenstic s for a plurality of decouplmg components 841 As shown, the database preferably mcludes the capacitance and ESR for the plurality of possible decouplmg components The calculated values 807, the configuration definitions 821, and the database values 841 are mput to calculate the decouplmg component resonance frequencies, and the optimum number of each chosen decouplmg component 822 In one embodiment, the optimum number of each chosen decouplmg component chosen for given frequency is the ESR of the decouplmg component divided by the target impedance multiplied by the weighting factor The decouplmg component frequencies are preferably calculated usmg the equation given above
Spatial placements for decouplmg components, cuπent sources, power supply, and selected locations or probe pomts are chosen m 823, preferably by a user Further details on placmg the decouplmg components m the model of the power dismbution system are given elsewhere m this disclosure Spatial placement data 823 and system parameter definitions 806 are combmed mto spatial placement data, mductance data, electπc mterconnection device data, and power supply data 824 to be sent to the remote computer 850
The data that were sent to the remote computer 824 are used to build a SPICE deck 861 The SPICE deck is used as mput for a SPICE analysis 866, preferably usmg HSPICE Output from the SPICE analysis 866 is processed to create graphical output 871 The graph the output returned to the local computer 872, preferably to the web browser 810 The graphic display is preferably displayed on the local computer 826, preferably as an HTML page m the web browser 810 In one embodiment, the HTML page compnses an SGML page, or other program as desued Although the method is shown m flowchart form, it is noted that portions of Fig 8B may occur concuπently or m different orders
Fig 9- Another Embodiment of the Computerized Method Fig 9 illustrates a flowchart of an embodiment of a computerized method for determmmg the decouplmg components for a power dismbution system As shown, the method calculates the target impedance for the power dismbution system 900 The target impedance is preferably calculated as a power supply voltage tunes the allowable power supply πpple divided by the total cuπent In a prefeπed embodiment, the total cuπent is noπnalized to one ampere The calculated target unpedance is used to calculate an optimum number of each available decouplmg component 905 The optimum number is preferably defined as the ESR of the decouplmg component divided by the target impedance multiplied by the weighting factor The method also calculates the resonance frequency of each available decouplmg component 910 The resonance frequency is preferably calculated as the mverse of two pi multiplied by the square root of the product of the mounted mductance and the capacitance of the decoupling component The method also calculates board resonance frequencies 915, preferably based upon the dimensions of the elecmcal mterconnection device and stackup on the elecmcal mterconnection device
The method performs smgle node analysis to compare the composite impedance profile of the elecmcal mterconnection device, including decouplmg components, to target unpedance In smgle node analysis, spatial locations are not taken into account, as m the model illustrated m Fig 2 The method next compares the results of the smgle node analysis to the target unpedance 925 to determine if the composite impedance profile of the elecmcal interconnection device is acceptable Acceptable is preferably defined as the target composite unpedance profile bemg at or below the target impedance If the composite impedance profile of the elecmcal mterconnection device is not acceptable, the method vanes one or more of the mput parameters 930 and agam performs smgle node analysis 920 If the composite impedance profile of the elecmcal mterconnection device after smgle node analysis is acceptable 925, then the method proceeds to spatially place the decouplmg components, the cuπent sources, the power supply, and the specific probe locations in the model 935 The locations chosen for devices placed m the model are preferably influenced by the board resonance frequencies 910 and the capacitor resonance frequencies 915. Additional details on placmg decoupling opponents for the power dismbution system are given elsewhere m this disclosure
The method next performs multi-node analysis 940 In a prefeπed embodiment, multi-node analysis coπesponds to performing HSPICE analysis 866. The results of the multi-node analysis are observed 945 If the results are acceptable m 950, the power dismbution design is considered complete 965. The prefeπed cntena for accepting the results of the multi-node analysis are that the system transfer impedance is below the target unpedance over the frequency range of mterest Should results not be acceptable in 950, method modifies the choice of the decouplmg components, the number of each the decouplmg opponents, and/or placement of the decouplmg components 960 and reanalyzes the model usmg multi-node analysis 940 Although the method is shown in flowchart form, it is noted that portions of Fig. 9 may occur concuπently or m different orders. Numerous vaπations and modifications will become apparent to those skilled m the art once the above disclosure is fully appreciated It is intended that the following claims be interpreted to embrace all such vanations and modifications.

Claims

WHAT IS CLAIMED:
1 A method for determmmg a specific number and placement of decouplmg components for a power dismbution system, the method compπsmg inputting known system parameters for the power dismbution system to a computer system the computer system selecting one or more different decouplmg components from a data base of charactenstic values for a plurality of decouplmg components, wherem said selectmg one or more different decouplmg components is based on the known system parameters for the power dismbution system and the charactenstic values for the plurality of decouplmg components, the computer system calculating a specific number for each selected one or more different decouplmg components based on the known system parameters for the power dismbution system and the charactenstic values for the selected one or more different decouplmg components the computer system calculating the placement of the selected one or more different decouplmg components based on the known system parameters for the power dismbution svstem and the charactenstic values for the selected one or more different decouplmg components
2 The method of claim 1 , wherem said inputting known system parameters for the pow er dismbution system compπses inputting one or more of one or more power supply characteπstics, allowable voltage πpple; total cuπent consumption; one or more physical dunensions for the power dismbution system, one or more dielecmc characteπstics, physical location constraints; or a frequency range
3 The method of claim 2, wherem the specific number for each selected one or more different decouplmg components is chosen based upon the frequency range and the data base of charactenstic values for the plurality of decouplmg components.
4 The method of claim 1, wherem the data base of charactenstic values for the plurality of decouplmg components includes the values of one or more of the following for each of the plurality of decoupling components a rated capacitance, an equivalent senes resistance; or a mounted mductance
5 The method of claim 1 , wherem the specific number of a particular one of the selected one or more different decouplmg components has approximately equal value of unpedance to a target unpedance for the power dismbution system when the specific number of the particular one of the selected one or more different decouplmg components are placed m parallel
6. The method of claim 1. further compπsmg: the computer system outputtmg one or more resultant elecmcal charactenstic values at one or more specific locations in the power dismbutton system based on said computer system calculating the specific number for each selected one or more decouplmg components and said computer system calculating the placement of the selected one or more different decouplmg components.
7. The method of claim 6, wherem said computer system outputtmg one or more resultant elecmcal characteristic values at the one or more specific locations in the power dismbution system compπses the computer system outputtmg a resultant noise level for the power dismbution system due to the cuπent sources or sinks and the selected one or more different decouplmg components.
8. The method of claim 6, wherem said computer system outputting one or more resultant elecmcal characteristic values at the one or more specific locations in the power dismbution system compnses the computer system outputting one or more resultant electrical charactenstic values at the one or more specific locations in the power distribution system dynamically as a function of frequency.
9. The method of claim 1, further comprising: the computer system outputting a resultant bill of goods of the specific number and the placement of the selected one or more different decoupling components in the power distribution system.
10. The method of claim 1, further comprising: changing the specific number for any one or more of the selected one or more different decouplmg components; and the computer system recalculating the placement of the selected one or more different decouplmg components based on the known system parameters for the power distribution system, the electrical charactenstic values for the selected one or more different decoupling components, and said changing the specific number for the any one or more of the selected one or more different decoupling components.
11. The method of claim 1 , further comprising: changing the placement for any one or more of the specific number of the selected one or more different decoupling components; and the computer system outputtmg one or more resultant elecmcal charactenstic values at one or more specific locations m the power distribution system based on said computer system calculating the specific number for each of the selected one or more different decoupling components based on the known system parameters, said computer system calculating the placement of the selected one or more different decouplmg components based on the known system parameters and the elecmcal charactenstic values for the selected one or more different decouplmg components, and said changing the placement for anv one of the specific number of the selected one or more different decouplmg components
12 The method of claim 11 , further compπsmg the computer system outputting a new resultant bill of goods of the specific number and the placement of the selected one or more decouplmg components in the power dismbution system
13 A system for determining decouplmg components for a power dismbution system, the system compπsmg a data base of charactenstic values for a plurality of decouplmg components, and a computer system configured to access the data base of charactenstic values for the plurality of decouplmg components, accept known system parameters for the power dismbutton system, select one or more different decouplmg components based on the known system parameters for the power dismbution system and the entnes m the data base, calculate a specific number for each selected one or more different decouplmg components based on the known system parameters and the entries m the data base, and determine placement data for the selected one or more different decouplmg components based upon the known system parameters and the enmes m the data base
14 The system of claim 13, wherem the computer system is further configured to calculate one or more elecmcal charactenstic values for the decouplmg components based on the known system parameters and en es m the data base
15 The system of claim 13, wherem the computer system is further configured to output one or more elecmcal charactenstic values for the power dismbution system at one or more specified locations
16 The system of claim 13, wherem the computer system is further configured to output a resultant bill of goods of the specific number and the placement of the selected one or more different decouplmg components m the power dismbution system
17 The system of claim 13, wherem the computer system is further configured to accept an mput from a user changmg the specific number for each selected one or more different decouplmg components, and recalculate the placement of the selected one or more different decouplmg components based on the known system parameters for the power dismbution system, the elecmcal charactenstic values for the decouplmg components, and said changing the specific number for each selected one or more decouplmg components
18 The system of claim 13. wherem the computer system is further configured to accept an mput from a user changing the placement for any one or more of the selected one or more different decouplmg components, and recalculate the placement of the selected one or more different decouplmg components based on the known system parameters for the power dismbution system, the elecmcal charactenstic values for the selected one or more different decouplmg components, and said changing the placement for any one or more of the selected one or more different decouplmg components
19. The system of claim 18, wherem the computer system is further configured to output a new resultant bill of goods of the specific number and the placement of the selected one or more decouplmg components in the power dismbution system
20. The system of claim 13, wherem the decouplmg components are capacitors, wherem each capacitor has approximately a mounted mductance and an equivalent senes resistance.
21. The system of claim 20, wherem the specific number of a particular one of the selected one or more different decouplmg components has approximately equal value of a next larger mteger of the quotient obtamed from dividmg a measured equivalent senes resistance for the particular one of the selected one or more different decouplmg components by a target impedance for the power dismbution system
22. A system for determining decouplmg components for a power dismbution system, the system compnsmg: a data base of charactenstic values for a plurality of decouplmg components, a network couplmg the data base, a local computer, and a remote computer, the local computer coupled to the network, wherem the local computer is configured to: access the data base of charactenstic values for the plurality of decouplmg components; accept known system parameters for the power dismbution system, calculate elecmcal charactenstic values for vanous ones of the decouplmg components and the power dismbution system; calculate a specific number for each selected one or more different decouplmg components, determine placement data for the selected one or more different decouplmg components; transfer the elecmcal charactenstic values, the specific number for said each selected one or more different decouplmg components, and the placement data for the decouplmg components to the remote computer over the network; and the remote computer coupled to the network, wherem remote computer is configured to: accept the elecmcal charactenstic values, the specific number for said each selected one or more different decouplmg components, and the placement data for the decouplmg components from the local computer over the network; perform an analysis of the power dismbution system; and output one or more elecmcal charactenstic values for the power dismbution system at one or more locations specified m the known system parameters.
23. The system of claim 22, wherem the known system parameters for the power dismbution system compπses one or more of: one or more power supply characteπstics; allowable voltage πpple; total cuπent consumption; one or more physical dimensions for the power dismbution system; one or more dielecmc characteπstics; physical location constraints; weighting factors; or a frequency range.
24. The system of claim 22, wherem the specific number for each selected one or more different decouplmg components is chosen based upon a frequency range and a measured equivalent senes resistance of the plurality of decoupling components.
25. The system of claim 22, wherem the data base of charactenstic values for the plurality of decouplmg components mcludes one or more of the following for each of the plurality of decouplmg components: a rated capacitance; an equivalent senes resistance; or a mounted inductance.
26. The system of claim 22, wherein the specific number for a particular one of the selected one or more different decoupling has approximately equal value of unpedance to a target unpedance for the power dismbution system when the specific number of the particular one of the selected one or more different decouplmg components are placed in parallel.
27. The system of claim 22, wherem the local computer is configured to execute a spreadsheet program.
28. The system of claim 27, wherein the spreadsheet program compπses an interactive applet.
29. The system of claim 28, wherem the interactive applet is transmitted over the network from the remote computer.
30. The system of claim 22, wherem the local computer and the remote computer are unified.
31. The system of claim 22, wherein the remote computer is configured to execute a simultaneous- equation-solver program.
32. The system of claim 31, wherem executing the simultaneous-equation-solver-program mcludes: creating a model of the power dismbution system, wherem the model of the power distribution system is based upon an MxN grid for a power plane and a coπespondmg MxN grid for a ground plane; placing cuπent sources or sinks in the model of the power distribution system at spatial locations coπesponding to locations of active components; and placing the decoupling components in the model of the power distribution system at nodal points which couple the MxN grid for the power plane and the coπespondmg MxN grid for the ground plane.
33. The system of claim 32, wherem the model of the power distribution system is m a form of a transmission plane.
34. The system of claim 33, wherein the model of the power distribution system is m a form of two dimensional distributed transmission lines.
35. The system of claim 32, wherein the model of the power distribution system comprises one or more of the following: one or more power supply characteristics; allowable voltage ripple; total cuπent consumption; one or more physical dimensions for the power distribution system; one or more dielectric characteristics; physical location constraints; or a frequency range.
36. The system of claim 32, wherein M and N have a uniform value.
37. The system of claim 32, wherem the active components mclude one or more of the following: processors, memones, application specific integrated circuits (ASICs), or logic ICs.
38. The system of claim 32, wherein a total sum of all values of the cuπent sources or sinks is scaled to equal one ampere.
39. The system of claim 32, wherem said executing the simultaneous-equation-solver-program further compπses: fixing a power supply m the model of the power dismbution system at a fixed one or more nodal pomts which couple the MxN gnd for the power plane and the coπespondmg MxN gπd for the ground plane.
40 The system of claim 39, wherein the power supply is comprised in the model as one or more pole frequencies, one or more zero frequencies and one or more resistances.
41. The system of claim 32, wherein the model of the power dismbution system is configured to determme the decouplmg components for a frequency range above approximately a lowest board resonance frequency.
42. The system of claim 32, wherein the model of the power dismbution system is configured to determme the decouplmg components for a frequency range above approximately a highest resonance frequency from all resonance frequencies of the one or more decouplmg components.
43. The system of claim 31, wherem the simultaneous-equation-solver program compnses a cucuit-solver program.
44. The system of claim 43, wherem the cucuit-solver program comprises a SPICE simulator.
45. The system of claim 22, wherein the decouplmg components are capacitors, wherem each capacitor has approximately a mounted inductance and an equivalent senes resistance.
46. The system of claim 45, wherem the known system parameters mclude a target impedance, wherem a number of a particular one of the decouplmg components has approximately equal value of a next larger mteger of the quotient obtamed from dividmg the measured equivalent seπes resistance for the particular one of the decoupling components by the target impedance.
47. A system for determmmg decouplmg components for a power dismbution system, the system compnsmg: measuring one or more elecmcal charactenstic values of a representative sample of a plurality of differing decouplmg components, wherem said one or more elecmcal charactenstic values include mounted mductance; buildmg a data base of said differing decouplmg components, wherein said data base includes said one or more elecmcal charactenstic values measured m said measuring, determmmg a target impedance for the power dismbution system; selecting a frequency range for the target impedance for the power dismbution system: and selecting one or more particular ones of said differmg decouplmg components m said data base based upon said one or more elecmcal charactenstic values of each of the one or more particular ones of said differmg decouplmg components
48. The system of claun 47, wherem the mounted inductance compπses an mdication of a resonance frequency of the particular one of said differmg decouplmg components, the system further compπsmg. comparing an impedance of each of the particular one of said differmg decouplmg components to the target impedance; and selecting a number of the particular one of said differmg decouplmg components to provide a total impedance at or below the target unpedance at the given frequency.
PCT/US1999/013802 1998-06-18 1999-06-18 System and method for determining the desired decoupling components for power distribution systems using a computer system WO1999066554A1 (en)

Priority Applications (7)

Application Number Priority Date Filing Date Title
JP2000555293A JP2002518763A (en) 1998-06-18 1999-06-18 System and method for determining desired decoupling components for a power distribution system using a computer system
AU52045/99A AU5204599A (en) 1998-06-18 1999-06-18 System and method for determining the desired decoupling components for power distribution systems using a computer system
EP99937161A EP1088345B1 (en) 1998-06-18 1999-06-18 System and method for determining the desired decoupling components for power distribution systems using a computer system
DE69907187T DE69907187T2 (en) 1998-06-18 1999-06-18 SYSTEM AND METHOD FOR DETERMINING THE DESIRED DECOUPLING ELEMENTS FOR AN ENERGY DISTRIBUTION SYSTEM USING A COMPUTER SYSTEM
KR1020007014265A KR20010071488A (en) 1998-06-18 1999-06-18 System and method for determining the desired decoupling components for power distribution systems using a computer system
AT99937161T ATE238607T1 (en) 1998-06-18 1999-06-18 SYSTEM AND METHOD FOR DETERMINING DESIRED DISCONNECTING ELEMENTS FOR A POWER DISTRIBUTION SYSTEM USING A COMPUTER SYSTEM
HK01104948A HK1034601A1 (en) 1998-06-18 2001-07-16 System and method for determining the desired decoupling components for power distribution systems using a computer system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US09/099,547 1998-06-18
US09/099,547 US6385565B1 (en) 1998-06-18 1998-06-18 System and method for determining the desired decoupling components for power distribution systems using a computer system

Publications (1)

Publication Number Publication Date
WO1999066554A1 true WO1999066554A1 (en) 1999-12-23

Family

ID=22275532

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1999/013802 WO1999066554A1 (en) 1998-06-18 1999-06-18 System and method for determining the desired decoupling components for power distribution systems using a computer system

Country Status (9)

Country Link
US (1) US6385565B1 (en)
EP (1) EP1088345B1 (en)
JP (1) JP2002518763A (en)
KR (1) KR20010071488A (en)
AT (1) ATE238607T1 (en)
AU (1) AU5204599A (en)
DE (1) DE69907187T2 (en)
HK (1) HK1034601A1 (en)
WO (1) WO1999066554A1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2002086689A2 (en) * 2001-04-24 2002-10-31 Sun Microsystems, Inc. System and method for determining the required decoupling capacitors for a power distribution system using an improved capacitor model
US20120136598A1 (en) * 2010-08-04 2012-05-31 Vladimir Dmitriev-Zdorov Optimization of Decoupling Device Choice for Electronic Design
CN107797001A (en) * 2016-08-30 2018-03-13 北京金风科创风电设备有限公司 Detection method, device and the wind power generating set of capacitive faults

Families Citing this family (35)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6937971B1 (en) * 1999-07-30 2005-08-30 Sun Microsystems, Inc. System and method for determining the desired decoupling components for a power distribution system having a voltage regulator module
US7089577B1 (en) * 2000-01-14 2006-08-08 Terayon Communication Systems, Inc. Process for supplying video-on-demand and other requested programs and services from a headend
US6629044B1 (en) * 2000-03-17 2003-09-30 General Electric Company Electrical distribution analysis method and apparatus
JP4177953B2 (en) * 2000-07-04 2008-11-05 矢崎総業株式会社 Electric wiring simulation apparatus and recording medium recording simulation program in electric wiring simulation apparatus
US6523154B2 (en) * 2000-12-14 2003-02-18 International Business Machines Corporation Method for supply voltage drop analysis during placement phase of chip design
US6571184B2 (en) * 2001-02-06 2003-05-27 Sun Microsystems, Inc. System and method for determining the decoupling capacitors for power distribution systems with a frequency-dependent target impedance
US6938224B2 (en) * 2001-02-21 2005-08-30 Lucent Technologies Inc. Method for modeling noise emitted by digital circuits
DE60235688D1 (en) * 2001-03-26 2010-04-29 Sensormatic Electronics Corp INTELLIGENT BASE TUNING WITH POWER SUPPLY SUPPORT FOR ELECTRONIC ARTICLE MONITORING
US7013254B2 (en) * 2001-03-28 2006-03-14 Sun Microsystems, Inc. Low-complexity, high accuracy model of a CPU power distribution system
JP2002304434A (en) * 2001-04-06 2002-10-18 Nec Corp Generating method, device and program for semiconductor integrated circuit power supply model for emi simulation
US6711719B2 (en) * 2001-08-13 2004-03-23 International Business Machines Corporation Method and apparatus for reducing power consumption in VLSI circuit designs
US6760893B2 (en) * 2001-12-12 2004-07-06 Agilent Technologies, Inc. Using transition time checks to determine noise problems on signal lines of an integrated circuit
US7454733B2 (en) * 2002-03-06 2008-11-18 International Business Machines Corporation Interconnect-aware methodology for integrated circuit design
US6870436B2 (en) * 2002-03-11 2005-03-22 Hewlett-Packard Development Company, L.P. Method and apparatus to attenuate power plane noise on a printed circuit board using high ESR capacitors
US6625791B1 (en) * 2002-05-10 2003-09-23 Sun Microsystems, Inc. Sliding grid based technique for optimal on-chip decap insertion
US7039882B2 (en) * 2002-06-17 2006-05-02 Amar Pal Singh Rana Technology dependent transformations for Silicon-On-Insulator in digital design synthesis
US7003750B2 (en) * 2002-08-01 2006-02-21 Sun Microsystems, Inc. Topology based wire shielding generation
US6789241B2 (en) * 2002-10-31 2004-09-07 Sun Microsystems, Inc. Methodology for determining the placement of decoupling capacitors in a power distribution system
US7080337B2 (en) * 2003-06-05 2006-07-18 Texas Instruments Incorporated Non-uniform decoupling capacitor distribution for providing more uniform noise reduction across chip
US7107555B2 (en) * 2003-07-31 2006-09-12 Taiyo Yuden Co., Ltd. Method and apparatus for designing high-frequency circuit, and display method for use in designing high-frequency circuit
JP4065229B2 (en) * 2003-11-26 2008-03-19 松下電器産業株式会社 Power supply noise analysis method for semiconductor integrated circuit
US7084615B1 (en) 2005-02-18 2006-08-01 International Business Machines Corporation Performance measurement of device dedicated to phase locked loop using second order system approximation
JP4389224B2 (en) * 2005-08-29 2009-12-24 エルピーダメモリ株式会社 Semiconductor device design method, design support system and program, and semiconductor package
US7778039B2 (en) * 2006-05-08 2010-08-17 Micron Technology, Inc. Substrates, systems, and devices including structures for suppressing power and ground plane noise, and methods for suppressing power and ground plane noise
US7633773B2 (en) 2006-05-10 2009-12-15 Micron Technology, Inc. On-die anti-resonance structure for integrated circuit
US7669151B1 (en) 2007-03-07 2010-02-23 Altera Corporation Methods for reducing power supply simultaneous switching noise
US7818704B1 (en) 2007-05-16 2010-10-19 Altera Corporation Capacitive decoupling method and module
JP2009251755A (en) * 2008-04-02 2009-10-29 Nec Electronics Corp Method, program, and apparatus for generating power supply noise model
KR100907020B1 (en) * 2008-02-25 2009-07-08 주식회사 하이닉스반도체 Apparatus for supplying power in semiconductor integrated circuit and input impedance control method of the same
JP2009238130A (en) * 2008-03-28 2009-10-15 Fujitsu Ltd Printed circuit board designing apparatus and printed circuit board designing method
US8090999B2 (en) * 2008-06-10 2012-01-03 Micron Technology, Inc. Memory media characterization for development of signal processors
JP4799651B2 (en) * 2009-08-31 2011-10-26 株式会社東芝 Design support apparatus and design support method
US8195437B2 (en) * 2009-10-03 2012-06-05 Raytheon Company Systems power distribution tool
US8493075B2 (en) 2010-09-08 2013-07-23 International Business Machines Corporation Method and apparatus for preventing circuit failure
US9310432B2 (en) * 2011-07-25 2016-04-12 Cosmin Iorga Method and system for measuring the impedance of the power distribution network in programmable logic device applications

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0472317A1 (en) * 1990-08-09 1992-02-26 AT&T Corp. Sub power plane to provide EMC filtering for VLSI devices
US5566083A (en) * 1994-10-18 1996-10-15 The Research Foundation Of State University Of New York Method for analyzing voltage fluctuations in multilayered electronic packaging structures

Family Cites Families (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4594677A (en) * 1983-11-09 1986-06-10 International Business Machines Corporation System for detecting and diagnosing noise caused by simultaneous current switching
JPS60114968A (en) * 1983-11-28 1985-06-21 Hitachi Ltd Automatic layout forming system
JPH0679319B2 (en) * 1985-04-10 1994-10-05 株式会社日立製作所 Layout update method
US4754408A (en) * 1985-11-21 1988-06-28 International Business Machines Corporation Progressive insertion placement of elements on an integrated circuit
US5027089A (en) 1988-06-10 1991-06-25 Adc Telecommunications, Inc. High frequency noise bypassing
US5031111C1 (en) * 1988-08-08 2001-03-27 Trw Inc Automated circuit design method
EP0393220B1 (en) * 1989-04-20 1994-07-13 International Business Machines Corporation Integrated circuit package
US5490083A (en) * 1992-10-05 1996-02-06 Matsushita Electric Industrial Co., Ltd. Method and apparatus for classifying and evaluating logic circuit
US5422782A (en) 1992-11-24 1995-06-06 Circuit Components Inc. Multiple resonant frequency decoupling capacitor
US5563801A (en) * 1993-10-06 1996-10-08 Nsoft Systems, Inc. Process independent design for gate array devices
WO1996036921A1 (en) * 1995-05-19 1996-11-21 3Com Corporation Method and apparatus for linking computer aided design databases with a numerical control machine database
US5754826A (en) * 1995-08-04 1998-05-19 Synopsys, Inc. CAD and simulation system for targeting IC designs to multiple fabrication processes
KR100439562B1 (en) * 1996-03-07 2005-10-19 마츠시타 덴끼 산교 가부시키가이샤 Transistor Placement Method
US5781446A (en) * 1996-05-07 1998-07-14 Flexible Solutions, Inc. System and method for multi-constraint domain electronic system design mapping
US5984510A (en) * 1996-11-01 1999-11-16 Motorola Inc. Automatic synthesis of standard cell layouts
US5898576A (en) * 1996-11-12 1999-04-27 Bay Networks Inc. Printed circuit board including a terminated power plane and method of manufacturing the same
US5883814A (en) * 1997-03-13 1999-03-16 International Business Machines Corporation System-on-chip layout compilation
JPH10321728A (en) * 1997-05-19 1998-12-04 Fujitsu Ltd Hierarchization wiring processing method and medium for hierarchization wiring processing program recording in layout system of semiconductor integrated circuit
US5798567A (en) * 1997-08-21 1998-08-25 Hewlett-Packard Company Ball grid array integrated circuit package which employs a flip chip integrated circuit and decoupling capacitors
US6195613B1 (en) * 1998-06-18 2001-02-27 Sun Microsystems, Inc. Method and system for measuring equivalent series resistance of capacitors and method for decoupling power distribution systems
US6215372B1 (en) * 1999-06-02 2001-04-10 Sun Microsystems, Inc. Method and apparatus for reducing electrical resonances in power and noise propagation in power distribution circuits employing plane conductors

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0472317A1 (en) * 1990-08-09 1992-02-26 AT&T Corp. Sub power plane to provide EMC filtering for VLSI devices
US5566083A (en) * 1994-10-18 1996-10-15 The Research Foundation Of State University Of New York Method for analyzing voltage fluctuations in multilayered electronic packaging structures

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
CHEN H H ET AL: "On-chip decoupling capacitor optimization for high-performance VLSI design", 1995 INTERNATIONAL SYMPOSIUM ON VLSI TECHNOLOGY, SYSTEMS, AND APPLICATIONS. PROCEEDINGS OF TECHNICAL PAPERS (CAT. NO.95TH8104), 1995 INTERNATIONAL SYMPOSIUM ON VLSI TECHNOLOGY, SYSTEMS, AND APPLICATIONS. PROCEEDINGS OF TECHNICAL PAPERS, TAIPEI, TAIWA, 1995, New York, NY, USA, IEEE, USA, pages 99 - 103, XP002122979, ISBN: 0-7803-2773-X *
CHEN Y ET AL: "OPTIMUM PLACEMENT OF DECOUPLING CAPACITORS ON PACKAGE AND PRINTED CIRCUIT BOARDS UNDER THE GUIDANCE OF ELECTROMAGNETIC FIELD SIMULATION", PROCEEDINGS OF THE ELECTRONIC COMPONENTS AND TECHNOLOGY CONFERENCE,US,NEW YORK, IEEE, vol. CONF. 46, pages 756-760, XP000684956, ISBN: 0-7803-3287-3 *
INGELS M ET AL: "DESIGN STRATEGIES AND DECOUPLING TECHNIQUES FOR REDUCING THE EFFECTS OF ELECTRICAL INTERFERENCE IN MIXED-MODE IC'S", IEEE JOURNAL OF SOLID-STATE CIRCUITS,US,IEEE INC. NEW YORK, vol. 32, no. 7, pages 1136-1141, XP000729375, ISSN: 0018-9200 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2002086689A2 (en) * 2001-04-24 2002-10-31 Sun Microsystems, Inc. System and method for determining the required decoupling capacitors for a power distribution system using an improved capacitor model
WO2002086689A3 (en) * 2001-04-24 2003-12-18 Sun Microsystems Inc System and method for determining the required decoupling capacitors for a power distribution system using an improved capacitor model
US20120136598A1 (en) * 2010-08-04 2012-05-31 Vladimir Dmitriev-Zdorov Optimization of Decoupling Device Choice for Electronic Design
CN107797001A (en) * 2016-08-30 2018-03-13 北京金风科创风电设备有限公司 Detection method, device and the wind power generating set of capacitive faults
CN107797001B (en) * 2016-08-30 2019-01-01 北京金风科创风电设备有限公司 Detection method, device and the wind power generating set of capacitive faults

Also Published As

Publication number Publication date
DE69907187T2 (en) 2004-02-05
JP2002518763A (en) 2002-06-25
US6385565B1 (en) 2002-05-07
KR20010071488A (en) 2001-07-28
ATE238607T1 (en) 2003-05-15
DE69907187D1 (en) 2003-05-28
EP1088345A1 (en) 2001-04-04
EP1088345B1 (en) 2003-04-23
HK1034601A1 (en) 2001-10-26
AU5204599A (en) 2000-01-05

Similar Documents

Publication Publication Date Title
US6532439B2 (en) Method for determining the desired decoupling components for power distribution systems
US6385565B1 (en) System and method for determining the desired decoupling components for power distribution systems using a computer system
US6571184B2 (en) System and method for determining the decoupling capacitors for power distribution systems with a frequency-dependent target impedance
US6937971B1 (en) System and method for determining the desired decoupling components for a power distribution system having a voltage regulator module
US6215373B1 (en) Method for edge termination of parallel conductive planes including estimating the characteristic impedance of the structure
US6441313B1 (en) Printed circuit board employing lossy power distribution network to reduce power plane resonances
US6104258A (en) System and method for edge termination of parallel conductive planes in an electrical interconnecting apparatus
US6674338B2 (en) Adding electrical resistance in series with bypass capacitors to achieve a desired value of electrical impedance between conductors of an electrical power distribution structure
US6850878B2 (en) System and method for determining the required decoupling capacitors for a power distribution system using an improved capacitor model
Xu et al. Estimating the power bus impedance of printed circuit boards with embedded capacitance
US6195613B1 (en) Method and system for measuring equivalent series resistance of capacitors and method for decoupling power distribution systems
Wu et al. Delaunay–Voronoi modeling of power-ground planes with source port correction
US6727774B1 (en) Bypass capacitor methods for achieving a desired value of electrical impedance between parallel planar conductors of an electrical power distribution structure, and associated electrical power distribution structures
Ito et al. Modeling of interconnections and isolation within a multilayered ball grid array package
Zhao et al. System level power integrity analysis with physics-based modeling methodology
US6774641B2 (en) Printed circuit board design support apparatus, method, and program
Fan et al. Quantifying decoupling capacitor location
Feng et al. Analysis of noise coupling result from overlapping power areas within power delivery networks
Shi et al. New efficient method of modeling electronics packages with layered power/ground planes
Low et al. Modeling and experimental verification of the interconnected mesh power system (IMPS) MCM topology
Wu et al. Investigation of crosstalk among vias
Durbhakula et al. Worst-case coupled voltage analysis of printed circuit board traces
Di Febo et al. Impact of planar electromagnetic band-gap structures on IR-DROP and signal integrity in high speed printed circuit boards
Kobayashi et al. An investigation of the effect of chassis connections on radiated EMI from PCBs
Yao et al. Admittance Extraction of Loaded Grid Power Distribution Network with Arbitrary Shapes

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AL AM AT AU AZ BA BB BG BR BY CA CH CN CU CZ DE DK EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT UA UG UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): GH GM KE LS MW SD SL SZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
WWE Wipo information: entry into national phase

Ref document number: 1020007014265

Country of ref document: KR

ENP Entry into the national phase

Ref country code: JP

Ref document number: 2000 555293

Kind code of ref document: A

Format of ref document f/p: F

WWE Wipo information: entry into national phase

Ref document number: 1999937161

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 1999937161

Country of ref document: EP

REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

WWP Wipo information: published in national office

Ref document number: 1020007014265

Country of ref document: KR

WWG Wipo information: grant in national office

Ref document number: 1999937161

Country of ref document: EP

WWW Wipo information: withdrawn in national office

Ref document number: 1020007014265

Country of ref document: KR