US20010000428A1 - Termination resistance independent system for impedance matching in high speed input-output chip interfacing - Google Patents

Termination resistance independent system for impedance matching in high speed input-output chip interfacing Download PDF

Info

Publication number
US20010000428A1
US20010000428A1 US09/735,679 US73567900A US2001000428A1 US 20010000428 A1 US20010000428 A1 US 20010000428A1 US 73567900 A US73567900 A US 73567900A US 2001000428 A1 US2001000428 A1 US 2001000428A1
Authority
US
United States
Prior art keywords
impedance matching
driver circuit
transmission line
node
output
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
US09/735,679
Other versions
US6278339B2 (en
Inventor
Wagdi Abadeer
John Connor
Patrick Hansen
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Individual
Original Assignee
Individual
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
Priority claimed from US09/255,997 external-priority patent/US6140885A/en
Application filed by Individual filed Critical Individual
Priority to US09/735,679 priority Critical patent/US6278339B2/en
Publication of US20010000428A1 publication Critical patent/US20010000428A1/en
Application granted granted Critical
Publication of US6278339B2 publication Critical patent/US6278339B2/en
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00Baseband systems
    • H04L25/02Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0264Arrangements for coupling to transmission lines
    • H04L25/0278Arrangements for impedance matching

Definitions

  • the present invention relates generally to fabrication of electronic devices and, more particularly, to a system and structure for impedance matching in very high speed input-output (I/O) chip interfacing to improve performance and reliability of I/O chips and systems that contain I/O chips.
  • I/O input-output
  • CMOS Complementary Metal-Oxide-Semiconductor
  • Transmission line effects become significant when the round trip propagation delay from the sending chip to the receiving chip is greater than the rise time of the transmitted signal. This condition is almost always met for modern CMOS-based digital systems. When this is the case, reflections occur on the signal line due to impedance mis-matches between the source, the transmission line, and the load. These reflections are superimposed on the transmitted signal, causing significant overshoot, undershoot, and system-wide noise. These conditions can cause both performance problems and reliability problems.
  • the performance problem is a reduction in the valid-data window of transmitted pulses with respect to the system clock.
  • the reliability problem involves devices with ultra-thin gate dielectric layers, less than 5 nm thickness, for which overshoots and undershoots can have a significant effect in reducing the effective useful life of the thin gate dielectric. This can lead to catastrophic breakdown.
  • the voltage overshoots and undershoots can also cause latch-up in CMOS devices leading to serious reliability problems. For high speed CMOS applications where the data rates are in the range of 1 GHz to 10 GHz and above, these performance and reliability problems can cause severe design limitations and prove costly.
  • a new system and structure for impedance matching in high-speed input-output chip interfacing is provided. It is an object of the present invention to provide a system and structure to achieve impedance matching at a driver circuit output, thereby preventing and not merely compensating for high voltage transients (overshoots) and low voltage transients (undershoots). It is another object of the present invention to provide a system and structure which can be manually adjusted to provide impedance matching at a driver circuit output during system set-up. It is yet another object of the present invention to provide an economical and manufacturable system and structure which can be manually adjusted to provide impedance matching at a driver circuit output for actual operating conditions of a specific system.
  • Another object of the present invention is to provide a control circuit for an impedance matching system which detects overshoots and undershoots on the driver circuit output and automatically provides a control voltage to an adjustment mechanism to adjust the length of an adjustable-length transmission line providing impedance matching based on actual overshoots and undershoots.
  • Still another object of the present invention is to provide an impedance matching system capable of performing its functions under a wide range of values for the various components comprising both the on-chip driver circuit and the off-chip interface circuitry.
  • the present invention provides a process and structure for an impedance matching network.
  • the structure includes an adjustable-length transmission line located between the driver circuit and the receiver circuit (load) on the system printed circuit board (card).
  • the adjustable-length transmission line is comprised of two conductive stripes connected in parallel with an input-output driver circuit at one end, electrically isolated at the opposite end, and short circuited by a movable stub intermediate of the ends.
  • the adjustable length transmission line is located on the system printed circuit board between the driver circuit and the receiver circuit.
  • the impedance matching network may further comprise an adjustment mechanism and a control circuit.
  • the impedances at the driver circuit output are matched by adjusting the length of the adjustable-length transmission line such that the composite reactance of the load, the transmission line between the driver circuit chip and the load, the connections, and the adjustable-length transmission line equals the converse of the driver reactance.
  • the length of the adjustable-length transmission line can be set or adjusted manually at system set-up, such as by a systems engineer or operator.
  • the systems engineer or operator observes the waveform using electronic measuring equipment such as an oscilloscope.
  • the position of the moveable stub is manually adjusted until the magnitude of undershoots and overshoots as viewed on the electronic measuring equipment is minimized.
  • the length of the adjustable-length transmission line is adjusted by the adjustment mechanism.
  • the adjustment mechanism comprises an elecromagnetic coil, a moveable magnetic core, and a return spring.
  • the electromagnetic coil creates a magnetic field proportional to a control current from the control circuit.
  • the magnetic field provides a force to the moveable magnetic core acting in a first direction.
  • the return spring provides a force to the moveable magnetic core acting in a second direction opposite to the first direction.
  • the moveable magnetic core is attached to the moveable stub whose position determines the length of the adjustable-length transmission line.
  • the control circuit generates a control current which varies in response to high and low transient voltages on the driver circuit output.
  • the control circuit comprises a positive detect circuitry and a negative detect circuitry, each of which provides a high (up) pulse to a counter when triggered by a transient, and a CLK down generator which provides a low (down) pulse to the counter as a function of the system clock. Both the positive detect circuitry and the negative detect circuitry are self-resetting.
  • control circuit further comprises a quantity of “n” N-type field effect transistors connected in parallel between the adjustment mechanism of the adjustable length transmission line and the system ground.
  • the transistors have control gates connected to an n-bit counter which is incrementally increased by the positive detect circuitry when an overshoot occurs on the driver circuit output and by the negative detect circuitry when an undershoot occurs on the driver circuit output.
  • n-bit counter which is incrementally increased by the positive detect circuitry when an overshoot occurs on the driver circuit output and by the negative detect circuitry when an undershoot occurs on the driver circuit output.
  • the present invention provides considerable improvement over the prior art. Whereas the prior art has been directed toward compensating for transient voltages using approaches such as isolation rings, forward biasing, excess dielectric, and decoupling capacitors on the system card, the present invention is directed to preventing voltage transients on the driver circuit output by impedance matching.
  • the present invention can be economically fabricated and implemented using current process technologies and existing input-output semiconductor devices.
  • the present invention can operate in systems having a wide range of operating conditions and in systems having varying operating conditions. For instance, when operating conditions change dynamically during operation, the present invention can automatically adjust the impedance to provide matching.
  • the present invention achieves these benefits in the context of known process technology.
  • FIG. 1 illustrates a conventional circuit model for a driver circuit, an off-chip receiver circuit, and a transmission line
  • FIG. 2A illustrates a circuit model for the driver circuit, the impedance matching network of the present invention, the transmission line, the off-chip receiver circuit, and connections—showing the impedance matching network as a black box;
  • FIG. 2B illustrates the conductive lines of the transmission line
  • FIG. 3A illustrates the adjustable-length transmission line of the matching network according to the first embodiment of the present invention
  • FIG. 3B illustrates the adjustable-length transmission line of the matching network and the adjustment mechanism of the matching network (shown in the relaxed state in which the length of the adjustable-length transmission line is at its maximum) according to a second embodiment of the present invention, in which the magnetic coil is shown cut open to highlight the magnetic core;
  • FIG. 4 illustrates an alternate circuit model showing the impedance matching network as an adjustable length transmission line, according to the present invention
  • FIG. 5 illustrates a circuit model showing a composite impedance as derived for the receiver side of the matching network
  • FIG. 6A shows a block diagram of the control circuit according to the second embodiment of the present invention.
  • FIG. 6B shows a schematic of the positive detect circuit according to the second embodiment of the present invention.
  • FIG. 6C shows a schematic of the negative detect circuit according to the second embodiment of the present invention.
  • FIG. 7 shows the calculat e d result for impedance matching for the cases in which selected values of load resistance (termination resistance) range from 50 ohms to 500 mega ohms, with the results provided in terms of the length for the adjustable length transmission line required to achieve impedance matching, as a function of frequency ranging from 0.5 to 100 giga hertz (GHz);
  • FIG. 8 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the receiver inductance ranging from 1 to 8 nano henry (nH);
  • FIG. 9 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the driver inductance ranging from 1 to 16 nH;
  • FIG. 10 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the receiver capacitance ranging from 1 pF to 12 pF;
  • FIG. 11 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the driver-capacitance ranging from 1 pF to 12 pF;
  • FIG. 12 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the dielectric constant for both the transmission line and the adjustable-length transmission line ranging from 2 to 6;
  • FIG. 13 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the length of the transmission line ranging from 51 to 127 mm (2 to 5 inches);
  • FIG. 14 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the characteristic impedance of both the transmission line and the adjustable-length transmission line in the range from 20 ohms to 200 ohms.
  • FIG. 1 shows a circuit model for a driver circuit ( 20 ), an off-chip receiver circuit ( 10 ) and a transmission line ( 40 ) according to the prior art.
  • the net reactance at the output of the driver circuit ( 20 ) is not zero (impedance mis-match)
  • reflections occur on the signal line. These reflections cause overshoots and undershoots.
  • FIG. 2A shows a circuit model of the most significant resistive and reactive (e.g., inductive and capacitive) components involved in a system where a driver circuit ( 20 ) on one chip is used to supply a signal (voltage) to a receiver circuit ( 10 ) on a different chip from the driver circuit. The signal is transmitted from the driver circuit (20) to the receiver circuit (10) on the transmission line ( 40 ).
  • An impedance matching network ( 50 ) which is the focus of the present invention, is located between the driver circuit ( 20 ) and the transmission line ( 40 ).
  • the on-chip driver circuit ( 20 ) has a driver resistance (R 20 ) and a driver capacitance (C 20 ).
  • the driver resistance (R 20 ) is typically about 50 ohms.
  • the driver capacitance (C 20 ) is typically about 4.7 pF. As will be shown later, however, the present invention can function with values that differ from the given typical values.
  • the receiver resistance (R 10 ) can vary in value depending on the particular system with a typical value being about 50 ohms. It will be demonstrated that the present invention will work independently of the termination resistance of the receiver circuit ( 10 ) from a receiver resistance (R 10 ) of 50 ohms to a receiver resistance (R 10 ) approaching the “infinite” resistance of an open circuit.
  • the receiver capacitance (C 10 ) is typically about 3.7 pF.
  • the connections at the receiver circuit ( 10 ) provide a receiver inductance (L 31 ).
  • Chip packaging connections at the driver circuit ( 20 ) provide a driver inductance (L 32 ).
  • the receiver inductance (L 31 ) and the driver inductance (L 32 ) are typically each about 2 nH.
  • the transmission line ( 40 ) is composed of printed, conductive stripes ( 42 ) of conductive material on a system circuit card ( 30 ) with a length (D 40 ) of about 76 mm (3 inches).
  • the stripes ( 42 ) are separated by an insulator ( 44 ).
  • the dielectric constant ( ⁇ 40 ) of the insulator ( 44 ) between the stripes ( 42 ) is typically about 4.5.
  • the relative permeability ( ⁇ 40 ) of the insulator ( 44 ) is typically about 1.
  • the characteristic impedance (Z 40 ) of the transmission line ( 40 ) is typically about 50 ohms.
  • driver resistance R 20
  • driver capacitance C 20
  • receiver resistance R 10
  • receiver capacitance C 10
  • driver inductance L 32
  • receiver inductance L 31
  • transmission line length D 40
  • transmission line dielectric constant ⁇ 40
  • ⁇ 40 transmission line relative permeability
  • Z 40 transmission line characteristic impedance
  • the matching network ( 50 ) comprises an adjustable-length transmission line ( 52 ) having a variable length (D 52 ).
  • the adjustable-length transmission line ( 52 ) comprises two parallel conductive lines ( 54 ) (e.g., stripes or circuit traces) located on the system card ( 30 ), connected in parallel with an on-chip driver circuit ( 20 ) at one end, electrically isolated at the opposite end, and shorted by a movable stub ( 56 ) intermediate of the ends of the parallel conductive lines ( 54 ).
  • the conductive lines ( 54 ) can be composed of a variety of conductive materials, such as aluminum, gold, or most preferably copper.
  • the conductive lines ( 54 ) preferably have a width of between about 0.5 inches and one inch, a thickness of between about 400 ⁇ m and 800 ⁇ m, and are preferably spaced apart a distance of between about 0.5 inches and 2 inches.
  • the movable stub ( 56 ) can be composed of a variety of conductive materials, including aluminum, copper, or most preferably stainless steel.
  • the movable stub ( 56 ) is preferably formed using conventional micro-machining techniques.
  • the movable stub ( 56 ) simultaneously contacts both of the parallel conductive lines ( 54 ), and has a thickness of between about 3 mm and 10 mm.
  • the movable stub ( 56 ) contacts the parallel conductive lines ( 54 ) with a pressure of between about 5 grams and 20 grams.
  • the movable stub ( 56 ) further comprises a mechanism to adjust the position along the length of the parallel conductive lines ( 54 ) at which the movable stub ( 56 ) contacts the parallel conductive lines ( 54 ), thereby changing the length (D 52 ) of the adjustable-length transmission line ( 52 ).
  • the position of the movable stub ( 56 ) can be adjusted in a number of ways, including but not limited to direct manual manipulation of the movable stub ( 56 ) and indirect manipulation of the movable stub ( 56 ) through a vernier device ( 58 ) as is known in the art.
  • a vernier device is a short, graduated scale that slides along a longer, graduated instrument and is used to indicate fractional parts of divisions, as in a micrometer.
  • the movable stub ( 56 ) is initially positioned along the length of the parallel conductive lines ( 54 ) such that the adjustable-length transmission line ( 52 ) is set to its maximum length (shorted as far as possible from the connection of the parallel conductive lines ( 54 ) to the output of the driver circuit ( 20 )).
  • the output of the driver circuit ( 20 ) is monitored using measuring equipment such as an oscilloscope to determine the magnitude of overshoots and undershoots. If the magnitude of overshoots and undershoots exceeds an acceptable threshold for the system, the position of the movable stub ( 56 ) is adjusted incrementally to increase the length (D 52 ) of the adjustable-length transmission line ( 52 ).
  • the output of the driver circuit ( 20 ) is monitored while the length (D 52 ) of the adjustable-length transmission line ( 52 ) is simultaneously adjusted until the magnitude of overshoots and undershoots is within the acceptable threshold. Care must be taken in choosing the monitoring equipment to minimize the impedance-loading effect introduced by the monitoring equipment.
  • the acceptable threshold for overshoot and undershoot magnitude depends upon the particular system based on performance, system set-up, and reliability requirements for the particular system.
  • the matching network ( 50 ) comprises an adjustable-length transmission line ( 52 ) having a variable length (D 52 ), an adjustment mechanism ( 60 ), and a control circuit ( 70 ).
  • the matching network is preferably located on the system card ( 30 ). This allows for use of existing I/O chips and manufacturing methods for conductive lines and a moveable stub on a system card which are known in the art.
  • the adjustable-length transmission line ( 52 ) comprises two parallel conductive lines ( 54 ) located on the system card ( 30 ), connected in parallel with an on-chip driver circuit ( 20 ) at one end, electrically isolated at the opposite end, and shorted by a movable stub ( 56 ) intermediate of the two ends of the parallel conductive lines ( 54 ).
  • the conductive lines ( 54 ) can be composed of a variety of conductive materials, such as aluminum, gold, or most preferably copper.
  • the conductive lines ( 54 ) preferably have a width of between about 0.1 and 0.25 inches, a thickness of between about 400 and 800 ⁇ m, and are preferably spaced apart a distance of between about 0.1 and 0.25 inches.
  • the movable stub ( 56 ) can be composed of a variety of conductive materials, including aluminum, copper, or most preferably stainless steel.
  • the movable stub ( 56 ) is preferably formed using conventional micro-machining techniques.
  • the movable stub ( 56 ) simultaneously contacts both of the parallel conductive lines ( 54 ) and has a thickness of between about 3 and 10 mm.
  • the movable stub ( 56 ) contacts the parallel conductive lines ( 54 ) with a pressure of between about 20 and 40 grams/cm 2 .
  • the moveable stub ( 56 ) is attached to a magnetic core ( 68 ) which is part of the adjustment mechanism ( 60 ).
  • the adjustment mechanism ( 60 ) receives a control current from the control current node ( 71 ) of the control circuit ( 70 ) causing it to transfer a linear motion to the moveable stub ( 56 ).
  • the linear movement of the moveable stub ( 56 ) adjusts the length (D 52 ) of the adjustable-length transmission line ( 52 ).
  • the adjustment mechanism ( 60 ) is located adjacent to the adjustable-length transmission line ( 52 ) on the system card ( 30 ).
  • the adjustment mechanism ( 60 ) preferably comprises an electromagnetic coil ( 62 ) having between 50 and 300 turns and preferably 100 turns of gold, aluminum, or most preferably copper wire.
  • the copper wire preferably has a diameter of between about 0.2 and 0.5 mm.
  • the electromagnetic coil ( 62 ) is connected at one end to a power supply ( 64 ) and at the other end to the control current node ( 71 ) of the control circuit ( 70 ).
  • the electromagnetic coil ( 62 ) creates a magnetic field that provides a force acting in a first direction on the magnetic core ( 68 ) located inside of the electromagnetic coil ( 62 ).
  • the magnetic core ( 68 ) is attached such that it has only one degree of freedom parallel to the axis of the electromagnetic coil ( 62 ). Most preferably, the magnetic force created by the electromagnetic coil ( 62 ) acting on the magnetic core ( 68 ) causes the length (D 52 ) of the adjustable-length transmission line ( 52 ) to decrease.
  • a return spring ( 66 ) fixedly attached to the system card ( 30 ) on one end and attached to the magnetic core ( 68 ) on the opposite end provides a force on the magnetic core ( 68 ) acting in a second direction opposite to the first direction.
  • the control circuit ( 70 ) detects overshoots or (positive) high voltage transients and undershoots or (negative) low voltage transients on the output of the driver circuit ( 20 ) and provides a control current in the control current node ( 71 ) to the adjustment mechanism ( 60 ).
  • the control current in control current node ( 71 ) causes the adjustment mechanism ( 60 ) to adjust the length (D 52 ) of the adjustable-length transmission line ( 52 ), thereby minimizing overshoots and undershoots.
  • the control circuit ( 70 ), as shown in FIG. 6A, comprises positive detect circuitry ( 72 ) and negative detect circuitry ( 74 ) connected to the output of the driver circuit ( 20 ).
  • the positive detect circuitry ( 72 ) detects the overshoot and generates a high pulse at node UP 1 ( 77 ).
  • the negative detect circuitry ( 74 ) detects the undershoot and generates a high pulse at node UP 2 ( 79 ).
  • Nodes UP 1 ( 77 ) and UP 2 ( 79 ) are connected to the inputs of an OR gate ( 76 ), which is in turn connected at its output to an n-bit counter ( 78 ).
  • the n-bit counter ( 78 ) is connected at its outputs to the control gates of a quantity “n” N-type field effect transistors or NFETs ( 80 ) which are connected in parallel.
  • the quantity “n” is large enough to provide adjustment increments that will meet the overshoot-undershoot threshold established for the system design, preferably between about 4 and 16.
  • a quantity of these NFETs ( 80 ) are made active (energized or turned on) equal to the count of the n-bit counter ( 78 ).
  • a current is drawn through the NFETs ( 80 ) which is proportional to the number of NFETs ( 80 ) which are energized.
  • the NFETs ( 80 ) are connected to ground at their sources and to the node CONTROL CURRENT ( 71 ) at their drains.
  • the node CONTROL CURRENT ( 71 ) is connected to the adjustment mechanism.
  • the control current at the node CONTROL CURRENT ( 71 ) is determined by the total currents in the “n” NFETs which are energized.
  • the node CONTROL CURRENT ( 71 ) is connected to the adjustment mechanism ( 60 ), providing the control current to the adjustment mechanism ( 60 ).
  • a down signal is periodically generated by the CLK down generator ( 83 ) on the node DOWN ( 85 ) as a function of the system clock ( 88 ).
  • the down signal causes the n-bit counter ( 78 ) to count down, thereby reducing by one the quantity of NFETs ( 80 ) which are energized.
  • the periodic down signal prevents the impedance matching network from being over-adjusted because of impedance changes due to dynamic environmental conditions such as temperature.
  • the positive detect circuitry ( 72 ) When a positive transient or overshoot occurs on the output of the driver circuit ( 20 ), the positive detect circuitry ( 72 ) generates a high pulse at node UP 1 ( 77 ).
  • the positive detect circuitry ( 72 ) is preferably implemented as shown in FIG. 6B.
  • a pulse up transistor (T 100 ) which is an N-well PFET, is connected to the driver circuit power supply ( 101 ) at its body and to a high reference potential, typically the driver circuit power supply ( 101 ), at its control gate (VGATE).
  • the pulse up transistor (T 100 ) is connected to the output of the driver circuit ( 20 ) at its source and to node PULSE ( 103 ) at its drain.
  • the pulse up transistor (T 100 ) switches the node PULSE ( 103 ) from low to high.
  • An up interrogate transistor (T 104 ) which is an NFET, is connected to the node PULSE ( 103 ) at its control gate, to ground at its source, and to node UPC 1 ( 105 ) at its drain.
  • the up interrogate transistor (T 104 ) switches node UPC 1 ( 105 ) from high to low.
  • Nodes UPC 1 ( 105 ) and DOWN ( 85 ) are connected to the inputs of a NOR gate (G 106 ) which generates an up (high) pulse on node UP 1 ( 77 ) when node UPC 1 ( 105 ) is low due to a high transient, provided node DOWN ( 85 ) has not been switched to a high level by the CLK down generator ( 83 ).
  • Node UPC 1 ( 105 ) is also connected to a voltage keeper comprising an inverter (I 108 ) connected to the control gate of a PFET keeper transistor (T 110 ).
  • the source of the keeper transistor (T 110 ) is connected to internal power and the drain is connected to node UPC 1 ( 105 ) such that node UPC 1 ( 105 ) is held at a high state until interrogation by the up interrogate transistor (T 104 ) can occur, thereby preventing accidental tripping by system noise.
  • Node UPC 1 ( 105 ) is also connected to a reset loop comprising an odd number of reset delay inverters (I 112 ) connected in series.
  • the last reset delay inverter (I 112 ) is connected to node RESET ( 109 ).
  • the number of reset delay inverters (I 112 ) is sufficient to cause a delay longer than the duration of the transient. Experimental modeling has shown that seven reset delay inverters (I 112 ) are typically sufficient.
  • the node RESET ( 109 ) is connected to the control gate of a pulse reset transistor (T 114 ), which is an NFET connected to ground at its source and to node PULSE ( 103 ) at its drain.
  • T 114 When node RESET ( 109 ) switches to high, the pulse reset transistor (T 114 ) switches node PULSE ( 103 ) back to its original state of low.
  • Node RESET ( 109 ) is also connected to a UPC 1 reset inverter (I 116 ), which is in turn connected to the control gate of a UPC 1 reset transistor (T 118 ).
  • the UPCl reset transistor (T 118 ) is a PFET connected to internal power at its source and to node UPC 1 ( 105 ) at its drain. When node RESET ( 109 ) switches to high, the UPC 1 reset transistor (T 118 ) switches node UPC 1 ( 105 ) back to its original state of high.
  • a first optional isolating resistive device (R 120 ) can be connected between internal power and node RESET ( 109 ) to prevent accidental switching due to coupling to external signals or system noise.
  • a second optional isolating resistive device (R 122 ) can be connected between ground and node PULSE ( 103 ).
  • the isolating resistive devices (R 120 , R 122 ) can be resistors or resistively connected FETs.
  • the negative detect circuitry ( 74 ) is a mirror image of the positive detect circuitry ( 72 ). When a negative transient or undershoot occurs on the output of the driver circuit ( 20 ), the negative detect circuitry ( 72 ) generates a high pulse at node UP 2 ( 79 ).
  • the negative detect circuitry ( 72 ) is preferably implemented as shown in FIG. 6C.
  • the pulse down transistor (T 200 ) is connected to the output of the driver circuit ( 20 ) at its source and to node PULSE DOWN ( 203 ) at its drain.
  • the pulse down transistor (T 200 ) switches the node PULSE DOWN ( 203 ) from high to low.
  • the node PULSE DOWN ( 203 ) switches to low
  • the down interrogate transistor (T 204 ) switches node UPC 2 ( 205 ) from low to high.
  • Down ( 85 ) is connected to the input of an inverter (I 240 ) whose output is connected to an input of a NAND gate (G 206 ).
  • Node UPC 2 ( 205 ) is connected to the other input of NAND gate (G 206 ).
  • NAND gate (G 206 ) is connected in series with a UP 2 inverter (I 230 ) then a node UP 2 ( 79 ).
  • An up (high) pulse is generated on node UP 2 ( 79 ) when node UPC 2 ( 205 ) is high due to a low transient, provided node DOWN ( 85 ) has not been switched to a high level by the CLK down generator ( 83 ).
  • Node UPC 2 ( 205 ) is also connected to a voltage keeper comprising an inverter (I 208 ) connected to the control gate of an NFET keeper transistor (T 210 ).
  • the source of the NFET keeper transistor (T 210 ) is connected to ground and the drain is connected to node UPC 2 ( 205 ) such that node UPC 2 ( 205 ) is held at a low state until interrogation by the down interrogate transistor (T 204 ) can occur, thereby preventing accidental tripping by system noise.
  • Node UPC 2 ( 205 ) is also connected to a second reset loop comprising an odd number of reset delay inverters (I 212 ) connected in series.
  • the last reset delay inverter (I 212 ) is connected to node RESET 2 ( 209 ).
  • the number of reset delay inverters (I 212 ) is sufficient to cause a delay longer than the duration of the transient. Experimental modeling has shown that seven reset delay inverters (I 212 ) are typically sufficient.
  • the node RESET 2 ( 209 ) is connected to the control gate of a pulse down reset transistor (T 214 ), which is a PFET connected to internal power at its source and to node PULSE DOWN ( 203 ) at its drain.
  • T 214 When node RESET 2 ( 209 ) switches to low, the pulse down reset transistor (T 214 ) switches node PULSE DOWN ( 203 ) back to its original state of high.
  • Node RESET 2 ( 209 ) is also connected to a UPC 2 reset inverter (I 216 ), which is in turn connected to the control gate of a UPC 2 reset transistor (T 218 ).
  • the UPC 2 reset transistor (T 218 ) is an NFET connected to ground at its source and to node UPC 2 ( 205 ) at its drain. When node RESET 2 ( 209 ) switches to low, the UPC 2 reset transistor (T 218 ) switches node UPC 2 ( 205 ) back to its original state of low.
  • a third optional isolating resistive device (R 220 ) can be connected between ground and node RESET 2 ( 209 ) to prevent accidental switching due to coupling to external signals or system noise.
  • a fourth optional isolating resistive device (R 222 ) can be connected between internal power and node PULSE DOWN ( 203 ).
  • the isolating resistive devices (R 220 , R 222 ) can be resistors or resistively connected FETs.
  • the terms “low pulse,” “low state,” and “low” and, similarly, the terms “high pulse,” “high state,” and “high” refer to relative voltage states and do not imply a specific magnitude.
  • the purpose of the adjustable-length transmission line ( 52 ) is to reduce transient voltages by providing a matching impedance for the reactive component of the impedance of the receiver circuit ( 10 ) to the driver circuit ( 20 ).
  • the adjustable-length transmission line ( 52 ) When the adjustable-length transmission line ( 52 ) is short circuited at one end, the input impedance at the other end will have no resistive component.
  • the reactive component (i.e., capacitive or inductive) of the input impedance will have a magnitude dependent on the length (D 52 ) of the adjustable-length transmission line ( 52 ) and the frequency (f) of the signal from the driver circuit ( 20 ).
  • impedance matching between the on-chip driver circuit ( 20 ) and the off-chip receiver circuit ( 10 ) can be achieved for a wide range of configurations.
  • Frequency (f) is dependant upon system design, where performance demands increasingly require operation at ultra-high frequencies (0.5 GHz and above).
  • the required length (D 52 ) of the adjustable-length transmission line ( 52 ) can be calculated using electromagnetic principles as described in chapter 12 of Engineering Electromagnetics, by William H. Hayt, Jr., McGraw-Hill, 1981, incorporated herein by reference.
  • the resistive component (R1) and the reactive component (X1) of the input impedance (Z 1 ) on the receiver side of the transmission line ( 40 ) are given by Equations 1 and 2, respectively, in terms of frequency (f), receiver resistance (R10), receiver capacitance (C10), and receiver inductance (L31).
  • Equations 1 and 2 can be derived for an inductor (such as receiver inductance (L31)) in series with a parallel capacitor (such as receiver circuit capacitance (C10)) and resistor (such as receiver circuit resistance (R10)), as shown in FIG. 2A.
  • Angular frequency ( ⁇ ) is equal to 2 ⁇ f.
  • R1 R10 1 + ( ⁇ ⁇ R10 ⁇ C10 ) 2 ⁇ ( Equation ⁇ ⁇ 1 )
  • X1 ⁇ ⁇ ⁇ L31 - ( ⁇ ⁇ R10 2 ⁇ C10 ) + ( ⁇ 3 ⁇ R10 2 ⁇ L31 ⁇ C10 2 ) 1 + ( ⁇ ⁇ R10 ⁇ C10 ) 2 ( Equation ⁇ ⁇ 2 )
  • the input impedance (Z4) at the driver side of the transmission line ( 40 ) can be expressed in terms of the characteristic impedance (Z40) of the transmission line ( 40 ), the load or input impedance (Z 1 ), the transmission line length (D40), and a phase factor ( ⁇ 40) for the transmission line ( 40 ).
  • Z4 Z40 ⁇ Z1 + [ j ⁇ Z40 ⁇ tan ⁇ ( ⁇ ⁇ 40 ⁇ D40 ) ]
  • the resistive component (R4) and the reactive component (X4) of the input impedance (Z4) on the driver side of the transmission line ( 40 ) can be determined by substituting Equation 0 into Equation 3 and separating the real (resistive) and imaginary (reactive) components.
  • the resistive component (R4) of the input impedance (Z4) at the driver side of the transmission line ( 40 ) can be expressed as Equation 5
  • the reactive component (X4) of the input impedance (Z4) at driver side of the transmission line ( 40 ) can be expressed as Equation 6.
  • R4 ( R1 ⁇ Z40 ) 2 ⁇ 1 + tan 2 ( ⁇ ⁇ 40 ⁇ D40 ) [ Z40 - ( X1 ⁇ tan ( ⁇ ⁇ 40 ⁇ D40 ) ) ] 2 + [ R1 ⁇ tan ( ⁇ ⁇ 40 ⁇ D40 ) ] 2 ( Equation ⁇ ⁇ 5 )
  • X4 Z40 ⁇ ⁇ [ X1 + ( Z40 ⁇ tan ( ⁇ ⁇ 40 ⁇ D40 ) ) ] ⁇ [ Z40 - ( X1 ⁇ tan ( ⁇ ⁇ 40 ⁇ D40 ) ] ⁇ - [ R1 2 ⁇ tan ( ⁇ ⁇ 40 ⁇ D40 ) ] [ Z40 - ( X1 ⁇ tan ( ⁇ ⁇ 40 ⁇ D40 ) ] 2 + [ R1 ⁇ tan ( ⁇ ⁇ 40 ⁇ D40 ) ] 2 ( Equation ⁇ ⁇
  • R 4 and X 4 are also the resistive and reactive components of the impedance at the receiver side of the matching network ( 50 ).
  • the driver circuit ( 20 ), matching network ( 50 ), transmission line ( 40 ) and receiver circuit ( 10 ) of FIG. 2A can now be reduced to the circuit shown in FIG. 4, where the impedance (Z50) of the matching network ( 50 ) and the input impedance (Z4) on the receiver side of the matching network ( 50 ) are in parallel.
  • the equivalent impedance (Z50) of the matching network ( 50 ) has no resistive component because the adjustable-length transmission line ( 52 ) is short circuited.
  • the equivalent impedance (Z50) of the matching network ( 50 ), as given by Equation 7, can be expressed as a function of the characteristic impedance (Z52) of the adjustable-length transmission line ( 52 ), the phase factor ( ⁇ 52) of the adjustable-length transmission line ( 52 ), and the length (D 52 ) to which the adjustable-length transmission line ( 52 ) is set.
  • the phase factor ( ⁇ 52) for the adjustable-length transmission line ( 52 ) is dependent upon the frequency (f) of the signal from the driver circuit ( 20 ) and the dielectric constant ( ⁇ 52) and the relative permeability ( ⁇ 52) for the adjustable-length transmission line ( 52 ).
  • the phase factor ( ⁇ 52) for the adjustable-length transmission line ( 52 ) can be determined using equations 8A and 8B, where “c” is the speed of light (3 ⁇ 10 8 m/s).
  • ⁇ ⁇ 52 2 ⁇ ⁇ ⁇ ⁇ f v52 (Equation 8A)
  • v52 c ( ⁇ 52 ⁇ ⁇ ⁇ 52 ) - 0.5 (Equation 8B)
  • Equation 9 the net input impedance (Z3) at the driver side of the matching network ( 50 ) is given by Equation 9.
  • the resistive component (R 3 ) of the net input impedance (Z3) is given by Equation 10.
  • the reactive component (X 3 ) of the net input impedance (Z3) is given by Equation 11.
  • the equivalent impedance (ZI) for the driver circuit ( 20 ) can be derived from the circuit model of FIG. 2A, and is given by Equation 12.
  • the resistive component (RI) of the input impedance (ZI) to the driver side of the matching network ( 50 ) is given by Equation 13
  • the reactive component (XI) of the input impedance (ZI) to the driver side of the matching network ( 50 ) is given by Equation 14.
  • Equations 13 and 14 can be derived for a parallel driver circuit capacitance (C20) and driver circuit resistance (R 20 ) in series with a driver inductance (L32), as shown in FIG. 2A.
  • the reactive components of the net input impedance (Z3) at the driver side of the matching network ( 50 ) and of the equivalent impedance (ZI) for the driver circuit ( 20 ) must cancel out (i.e., the matching network ( 50 ) must provide a reactance, which in combination with the driver and receiver circuits, will give no net reactance).
  • the reactive component (X 3 ) must be the conjugate (equal in magnitude with the opposite sign) of the reactive component (XI), as shown in Equation 15. The matching of the reactive components will result in no overshoots or undershoots.
  • Equation 11 and 14 can be substituted into Equation 15.
  • Selected values for frequency (f) and length (D 52 ) of the adjustable-length transmission line ( 52 ) can be substituted into Equation 15 to determine, by the process of iteration, the values at which the net reactance (XI+X 3 ) is equal to zero (0), thereby achieving impedance matching.
  • FIG. 7 shows the values for length (D52) of the adjustable-length transmission line ( 52 ) as a function of frequency (f) which provide impedance matching over a range of selected values for the receiver circuit resistance (termination load) (R10) ranging from 50 ohms to 500 mega ohms. The values shown in FIG.
  • FIG. 8 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line ( 52 ) for a receiver inductance (L31) in the range between 1 nH and 8 nH. The values shown in FIG.
  • FIG. 9 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line ( 52 ) for a driver inductance (L32) in the range between 1 nH and 16 nH.
  • the values shown in FIG. 9 are determined using typical values as previously provided for C10 (3.7 pF), R 10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), ⁇ 40 (4.5), L31 (2 nH), C20 (4.7 pF), Z52 (50 ohms), and ⁇ 52 (4.5).
  • FIG. 10 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line ( 52 ) for a receiver capacitance (C10) in the range between 1 pF and 12 pF.
  • the values shown in FIG. 10 are determined using typical values as previously provided for C20 (4.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), ⁇ 40 (4.5), L31 (2 nH), L32 (2 nH), Z52 (50 ohms), and ⁇ 52 (4.5).
  • FIG. 11 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line ( 52 ) for a driver capacitance (C20) in the range between 1 pF and 12 pF.
  • the values shown in FIG. 11 are determined using typical values as previously provided for C10 (3.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), ⁇ 40 (4.5), L31 (2 nH), L32 (2 nH), Z52 (50 ohms), and ⁇ 52 (4.5).
  • FIG. 12 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line ( 52 ) for a dielectric constant ( ⁇ 40 , ⁇ 52) for the transmission line ( 40 ) and the adjustable-length transmission line ( 52 ) in the range between 2 and 6.
  • the values shown in FIG. 12 are determined using typical values as previously provided for C10 (3.7 pF), C20 (4.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), L31 (2 nH), L32 (2 nH), and Z52 (50 ohms).
  • FIG. 13 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line ( 52 ) for a length (D40) of the transmission line ( 40 ) in the range between 51 mm (2 inches) and 127 mm (5 inches).
  • the values shown in FIG. 13 are determined using typical values as previously provided for C10 (3.7 pF), C20 (4.7 pF), R10 (50 ohms), Z40 (50 ohms), Z52 (50 ohms), ⁇ 40 (4.5), L31 (2 nH), L32 (2 nH), and ⁇ 52 (4.5).
  • FIG. 14 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line ( 52 ) for a characteristic impedance (Z40, Z52) of the transmission line ( 40 ) and of the adjustable-length transmission line ( 52 ) in the range between 20 ohms and 200 ohms.
  • the values shown in FIG. 14 are determined using typical values as previously provided for C10 (3.7 pF), C20 (4.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), ⁇ 40 (4.5), L 31 (2 nH), L32 (2 nH), and ⁇ 52 (4.5).

Abstract

An impedance matching system and a network for impedance matching at a driver circuit output for high frequency input-output devices. The impedance matching network comprises an adjustable-length transmission line having a length adjusted in proportion to the magnitude of transients on the driver circuit output and an input impedance, which is purely reactive, and is a function of its length. The purpose of the adjustable-length transmission line is to reduce transient voltages by providing a matching impedance for the reactive component of the impedance of the receiver circuit to the driver circuit. In the preferred embodiment, the impedance matching network comprises two parallel conductive lines formed on the system card, shorted by a movable stub, and connected in parallel to the driver circuit. Optionally, the impedance matching network further comprises a control circuit which detects overshoots and undershoots on the driver circuit output and provides a control current proportional to the magnitude of overshoots and undershoots to an electromagnetic adjustment mechanism which provides a linear adjustment to the moveable stub proportional to the control current.

Description

    CROSS REFERENCE TO RELATED APPLICATIONS
  • 1. The present application is a continuation-in-part of U.S. patent application Ser. No. 09/255,997, filed on Feb. 23, 1999, which is pending and is assigned to the assignee of the subject application.
  • TECHNICAL FIELD
  • 2. The present invention relates generally to fabrication of electronic devices and, more particularly, to a system and structure for impedance matching in very high speed input-output (I/O) chip interfacing to improve performance and reliability of I/O chips and systems that contain I/O chips.
  • BACKGROUND OF INVENTION
  • 3. One of the most important aspects in integrated chip system design is the interface considerations for very short channel Complementary Metal-Oxide-Semiconductor (CMOS) devices in high-speed applications. Beginning with the driving circuit, an on-chip voltage swing of ground to Vdd (internal power supply) must be communicated off-chip to external devices. The path from the driver output to the output pad involves capacitive coupling effects to other on-chip signals. Moreover, the chip packaging system adds inductances to the circuit. The external signals must then traverse some interconnect (or transmission line), whether card traces or multi-chip module (MCM) connections. The signals are subject to additional deformations at this point due to transmission line effects introduced by the interconnects. The use of external decoupling capacitors to stabilize the card power supply provided to the chip is known in the art. For low-impedance card power plane connections, these external capacitors do not reduce the on-chip simultaneous switching noise of the driver circuit.
  • 4. Transmission line effects become significant when the round trip propagation delay from the sending chip to the receiving chip is greater than the rise time of the transmitted signal. This condition is almost always met for modern CMOS-based digital systems. When this is the case, reflections occur on the signal line due to impedance mis-matches between the source, the transmission line, and the load. These reflections are superimposed on the transmitted signal, causing significant overshoot, undershoot, and system-wide noise. These conditions can cause both performance problems and reliability problems.
  • 5. The performance problem is a reduction in the valid-data window of transmitted pulses with respect to the system clock. The reliability problem involves devices with ultra-thin gate dielectric layers, less than 5 nm thickness, for which overshoots and undershoots can have a significant effect in reducing the effective useful life of the thin gate dielectric. This can lead to catastrophic breakdown. The voltage overshoots and undershoots can also cause latch-up in CMOS devices leading to serious reliability problems. For high speed CMOS applications where the data rates are in the range of 1 GHz to 10 GHz and above, these performance and reliability problems can cause severe design limitations and prove costly.
  • 6. The deficiencies of conventional, high-speed input-output interfaces show that a need exists for improvement. To overcome the shortcomings of conventional systems, a new system and structure for impedance matching in high-speed input-output chip interfacing is provided. It is an object of the present invention to provide a system and structure to achieve impedance matching at a driver circuit output, thereby preventing and not merely compensating for high voltage transients (overshoots) and low voltage transients (undershoots). It is another object of the present invention to provide a system and structure which can be manually adjusted to provide impedance matching at a driver circuit output during system set-up. It is yet another object of the present invention to provide an economical and manufacturable system and structure which can be manually adjusted to provide impedance matching at a driver circuit output for actual operating conditions of a specific system.
  • 7. It is still another object of the present invention to provide a system and structure capable of automatically adjusting to provide impedance matching at a driver circuit output when operating conditions cause changes to the system. Another object of the present invention is to provide a control circuit for an impedance matching system which detects overshoots and undershoots on the driver circuit output and automatically provides a control voltage to an adjustment mechanism to adjust the length of an adjustable-length transmission line providing impedance matching based on actual overshoots and undershoots. Still another object of the present invention is to provide an impedance matching system capable of performing its functions under a wide range of values for the various components comprising both the on-chip driver circuit and the off-chip interface circuitry.
  • SUMMARY OF THE INVENTION
  • 8. To achieve these and other objects, and in view of its purposes, the present invention provides a process and structure for an impedance matching network. The structure includes an adjustable-length transmission line located between the driver circuit and the receiver circuit (load) on the system printed circuit board (card). The adjustable-length transmission line is comprised of two conductive stripes connected in parallel with an input-output driver circuit at one end, electrically isolated at the opposite end, and short circuited by a movable stub intermediate of the ends. The adjustable length transmission line is located on the system printed circuit board between the driver circuit and the receiver circuit. Optionally, the impedance matching network may further comprise an adjustment mechanism and a control circuit.
  • 9. The impedances at the driver circuit output are matched by adjusting the length of the adjustable-length transmission line such that the composite reactance of the load, the transmission line between the driver circuit chip and the load, the connections, and the adjustable-length transmission line equals the converse of the driver reactance. In a first embodiment, the length of the adjustable-length transmission line can be set or adjusted manually at system set-up, such as by a systems engineer or operator. In order to adjust the adjustable-length transmission line, the systems engineer or operator observes the waveform using electronic measuring equipment such as an oscilloscope. The position of the moveable stub is manually adjusted until the magnitude of undershoots and overshoots as viewed on the electronic measuring equipment is minimized.
  • 10. In a second embodiment, the length of the adjustable-length transmission line is adjusted by the adjustment mechanism. The adjustment mechanism comprises an elecromagnetic coil, a moveable magnetic core, and a return spring. The electromagnetic coil creates a magnetic field proportional to a control current from the control circuit. The magnetic field provides a force to the moveable magnetic core acting in a first direction. The return spring provides a force to the moveable magnetic core acting in a second direction opposite to the first direction. The moveable magnetic core is attached to the moveable stub whose position determines the length of the adjustable-length transmission line.
  • 11. The control circuit generates a control current which varies in response to high and low transient voltages on the driver circuit output. In the second embodiment, the control circuit comprises a positive detect circuitry and a negative detect circuitry, each of which provides a high (up) pulse to a counter when triggered by a transient, and a CLK down generator which provides a low (down) pulse to the counter as a function of the system clock. Both the positive detect circuitry and the negative detect circuitry are self-resetting.
  • 12. In the second embodiment, the control circuit further comprises a quantity of “n” N-type field effect transistors connected in parallel between the adjustment mechanism of the adjustable length transmission line and the system ground. The transistors have control gates connected to an n-bit counter which is incrementally increased by the positive detect circuitry when an overshoot occurs on the driver circuit output and by the negative detect circuitry when an undershoot occurs on the driver circuit output. Thus, the number of transistors drawing current and the resulting control current are adjusted until no transients are occurring on the driver circuit output.
  • 13. The present invention provides considerable improvement over the prior art. Whereas the prior art has been directed toward compensating for transient voltages using approaches such as isolation rings, forward biasing, excess dielectric, and decoupling capacitors on the system card, the present invention is directed to preventing voltage transients on the driver circuit output by impedance matching. The present invention can be economically fabricated and implemented using current process technologies and existing input-output semiconductor devices. In addition, by adjusting the length of the adjustable-length transmission line based on the magnitude of the transients in a particular system, the present invention can operate in systems having a wide range of operating conditions and in systems having varying operating conditions. For instance, when operating conditions change dynamically during operation, the present invention can automatically adjust the impedance to provide matching. The present invention achieves these benefits in the context of known process technology.
  • 14. It is to be understood that both the foregoing general description and the following detailed description are exemplary, but are not restrictive, of the invention.
  • BRIEF DESCRIPTION OF THE DRAWING
  • 15. The features and advantages of the impedance matching network according to the present invention and further details of the process and structure for such an impedance matching network are best understood from the following detailed description when read in connection with the accompanying drawing. It is emphasized that, according to common practice, the various features of the drawing are not to scale. On the contrary, the dimensions of the various features are arbitrarily expanded or reduced for clarity. Included in the drawing are the following figures:
  • 16.FIG. 1 illustrates a conventional circuit model for a driver circuit, an off-chip receiver circuit, and a transmission line;
  • 17.FIG. 2A illustrates a circuit model for the driver circuit, the impedance matching network of the present invention, the transmission line, the off-chip receiver circuit, and connections—showing the impedance matching network as a black box;
  • 18.FIG. 2B illustrates the conductive lines of the transmission line;
  • 19.FIG. 3A illustrates the adjustable-length transmission line of the matching network according to the first embodiment of the present invention;
  • 20.FIG. 3B illustrates the adjustable-length transmission line of the matching network and the adjustment mechanism of the matching network (shown in the relaxed state in which the length of the adjustable-length transmission line is at its maximum) according to a second embodiment of the present invention, in which the magnetic coil is shown cut open to highlight the magnetic core;
  • 21.FIG. 4 illustrates an alternate circuit model showing the impedance matching network as an adjustable length transmission line, according to the present invention;
  • 22.FIG. 5 illustrates a circuit model showing a composite impedance as derived for the receiver side of the matching network;
  • 23.FIG. 6A shows a block diagram of the control circuit according to the second embodiment of the present invention;
  • 24.FIG. 6B shows a schematic of the positive detect circuit according to the second embodiment of the present invention;
  • 25.FIG. 6C shows a schematic of the negative detect circuit according to the second embodiment of the present invention;
  • 26.FIG. 7 shows the calculat e d result for impedance matching for the cases in which selected values of load resistance (termination resistance) range from 50 ohms to 500 mega ohms, with the results provided in terms of the length for the adjustable length transmission line required to achieve impedance matching, as a function of frequency ranging from 0.5 to 100 giga hertz (GHz);
  • 27.FIG. 8 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the receiver inductance ranging from 1 to 8 nano henry (nH);
  • 28.FIG. 9 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the driver inductance ranging from 1 to 16 nH;
  • 29.FIG. 10 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the receiver capacitance ranging from 1 pF to 12 pF;
  • 30.FIG. 11 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the driver-capacitance ranging from 1 pF to 12 pF;
  • 31.FIG. 12 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the dielectric constant for both the transmission line and the adjustable-length transmission line ranging from 2 to 6;
  • 32.FIG. 13 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the length of the transmission line ranging from 51 to 127 mm (2 to 5 inches); and
  • 33.FIG. 14 shows the length of the adjustable-length transmission line required for impedance matching, as a function of frequency, for selected values of the characteristic impedance of both the transmission line and the adjustable-length transmission line in the range from 20 ohms to 200 ohms.
  • DETAILED DESCRIPTION OF THE INVENTION
  • 34. The present invention will now be described in detail with reference to the accompanying drawing, in which like reference numbers designate similar or corresponding elements, regions, and portions. Several processes of compensating for the overshoots and undershoots caused by impedance mis-match are known in the art. Guard rings, excess dielectric, and forward biasing are some examples of the techniques used to compensate for the transient voltages caused by impedance mis-match. The present invention provides a process for matching the reactive component of the impedance of the load and the transmission line to the impedance of the driver circuit, thereby preventing overshoots and undershoots and their associated performance and reliability problems.
  • 35.FIG. 1 shows a circuit model for a driver circuit (20), an off-chip receiver circuit (10) and a transmission line (40) according to the prior art. When the net reactance at the output of the driver circuit (20) is not zero (impedance mis-match), reflections occur on the signal line. These reflections cause overshoots and undershoots.
  • 36.FIG. 2A shows a circuit model of the most significant resistive and reactive (e.g., inductive and capacitive) components involved in a system where a driver circuit (20) on one chip is used to supply a signal (voltage) to a receiver circuit (10) on a different chip from the driver circuit. The signal is transmitted from the driver circuit (20) to the receiver circuit (10) on the transmission line (40). An impedance matching network (50), which is the focus of the present invention, is located between the driver circuit (20) and the transmission line (40).
  • 37. The on-chip driver circuit (20) has a driver resistance (R20) and a driver capacitance (C20). The driver resistance (R20) is typically about 50 ohms. The driver capacitance (C20) is typically about 4.7 pF. As will be shown later, however, the present invention can function with values that differ from the given typical values.
  • 38. The receiver circuit (10), which is typically not but might be on the same chip as the driver circuit (20), has a receiver resistance (R10) and a receiver capacitance (C10). The receiver resistance (R10) can vary in value depending on the particular system with a typical value being about 50 ohms. It will be demonstrated that the present invention will work independently of the termination resistance of the receiver circuit (10) from a receiver resistance (R10) of 50 ohms to a receiver resistance (R10) approaching the “infinite” resistance of an open circuit. The receiver capacitance (C10) is typically about 3.7 pF.
  • 39. The connections at the receiver circuit (10) provide a receiver inductance (L31). Chip packaging connections at the driver circuit (20) provide a driver inductance (L32). The receiver inductance (L31) and the driver inductance (L32) are typically each about 2 nH.
  • 40. As shown in FIG. 2B, the transmission line (40) is composed of printed, conductive stripes (42) of conductive material on a system circuit card (30) with a length (D40) of about 76 mm (3 inches). The stripes (42) are separated by an insulator (44). The dielectric constant (∈40) of the insulator (44) between the stripes (42) is typically about 4.5. The relative permeability (μ40) of the insulator (44) is typically about 1. The characteristic impedance (Z40) of the transmission line (40) is typically about 50 ohms.
  • 41. Unless otherwise specified, all calculations are performed and all graphs are made using the typical values provided above for driver resistance (R20), driver capacitance (C20), receiver resistance (R10), receiver capacitance (C10), driver inductance (L32), receiver inductance (L31), transmission line length (D40), transmission line dielectric constant (∈40), transmission line relative permeability (μ40), and transmission line characteristic impedance (Z40). As will be demonstrated later, however, the present invention can provide impedance matching over a wide range for each of these characteristics.
  • 42. First Embodiment of the Present Invention
  • 43. In a first embodiment, the matching network (50) comprises an adjustable-length transmission line (52) having a variable length (D52). The adjustable-length transmission line (52), as shown in FIG. 3A, comprises two parallel conductive lines (54) (e.g., stripes or circuit traces) located on the system card (30), connected in parallel with an on-chip driver circuit (20) at one end, electrically isolated at the opposite end, and shorted by a movable stub (56) intermediate of the ends of the parallel conductive lines (54). The conductive lines (54) can be composed of a variety of conductive materials, such as aluminum, gold, or most preferably copper. The conductive lines (54) preferably have a width of between about 0.5 inches and one inch, a thickness of between about 400 μm and 800 μm, and are preferably spaced apart a distance of between about 0.5 inches and 2 inches.
  • 44. The movable stub (56) can be composed of a variety of conductive materials, including aluminum, copper, or most preferably stainless steel. The movable stub (56) is preferably formed using conventional micro-machining techniques. The movable stub (56) simultaneously contacts both of the parallel conductive lines (54), and has a thickness of between about 3 mm and 10 mm. The movable stub (56) contacts the parallel conductive lines (54) with a pressure of between about 5 grams and 20 grams.
  • 45. The movable stub (56) further comprises a mechanism to adjust the position along the length of the parallel conductive lines (54) at which the movable stub (56) contacts the parallel conductive lines (54), thereby changing the length (D52) of the adjustable-length transmission line (52). The position of the movable stub (56) can be adjusted in a number of ways, including but not limited to direct manual manipulation of the movable stub (56) and indirect manipulation of the movable stub (56) through a vernier device (58) as is known in the art. A vernier device is a short, graduated scale that slides along a longer, graduated instrument and is used to indicate fractional parts of divisions, as in a micrometer.
  • 46. In the preferred process, the movable stub (56) is initially positioned along the length of the parallel conductive lines (54) such that the adjustable-length transmission line (52) is set to its maximum length (shorted as far as possible from the connection of the parallel conductive lines (54) to the output of the driver circuit (20)). The output of the driver circuit (20) is monitored using measuring equipment such as an oscilloscope to determine the magnitude of overshoots and undershoots. If the magnitude of overshoots and undershoots exceeds an acceptable threshold for the system, the position of the movable stub (56) is adjusted incrementally to increase the length (D52) of the adjustable-length transmission line (52). Most preferably, the output of the driver circuit (20) is monitored while the length (D52) of the adjustable-length transmission line (52) is simultaneously adjusted until the magnitude of overshoots and undershoots is within the acceptable threshold. Care must be taken in choosing the monitoring equipment to minimize the impedance-loading effect introduced by the monitoring equipment. The acceptable threshold for overshoot and undershoot magnitude depends upon the particular system based on performance, system set-up, and reliability requirements for the particular system.
  • 47. Second Embodiment of the Present Invention
  • 48. In a second embodiment, the matching network (50) comprises an adjustable-length transmission line (52) having a variable length (D52), an adjustment mechanism (60), and a control circuit (70). For ease of manufacturing and economic implementation, the matching network is preferably located on the system card (30). This allows for use of existing I/O chips and manufacturing methods for conductive lines and a moveable stub on a system card which are known in the art.
  • 49. As shown in FIG. 3B, the adjustable-length transmission line (52) comprises two parallel conductive lines (54) located on the system card (30), connected in parallel with an on-chip driver circuit (20) at one end, electrically isolated at the opposite end, and shorted by a movable stub (56) intermediate of the two ends of the parallel conductive lines (54). The conductive lines (54) can be composed of a variety of conductive materials, such as aluminum, gold, or most preferably copper. The conductive lines (54) preferably have a width of between about 0.1 and 0.25 inches, a thickness of between about 400 and 800 μm, and are preferably spaced apart a distance of between about 0.1 and 0.25 inches.
  • 50. The movable stub (56) can be composed of a variety of conductive materials, including aluminum, copper, or most preferably stainless steel. The movable stub (56) is preferably formed using conventional micro-machining techniques. The movable stub (56) simultaneously contacts both of the parallel conductive lines (54) and has a thickness of between about 3 and 10 mm. The movable stub (56) contacts the parallel conductive lines (54) with a pressure of between about 20 and 40 grams/cm2. The moveable stub (56) is attached to a magnetic core (68) which is part of the adjustment mechanism (60).
  • 51. The adjustment mechanism (60) receives a control current from the control current node (71) of the control circuit (70) causing it to transfer a linear motion to the moveable stub (56). The linear movement of the moveable stub (56) adjusts the length (D52) of the adjustable-length transmission line (52). The adjustment mechanism (60) is located adjacent to the adjustable-length transmission line (52) on the system card (30).
  • 52. As shown in FIG. 3B, the adjustment mechanism (60) preferably comprises an electromagnetic coil (62) having between 50 and 300 turns and preferably 100 turns of gold, aluminum, or most preferably copper wire. The copper wire preferably has a diameter of between about 0.2 and 0.5 mm. The electromagnetic coil (62) is connected at one end to a power supply (64) and at the other end to the control current node (71) of the control circuit (70). The electromagnetic coil (62) creates a magnetic field that provides a force acting in a first direction on the magnetic core (68) located inside of the electromagnetic coil (62). The magnetic core (68) is attached such that it has only one degree of freedom parallel to the axis of the electromagnetic coil (62). Most preferably, the magnetic force created by the electromagnetic coil (62) acting on the magnetic core (68) causes the length (D52) of the adjustable-length transmission line (52) to decrease. A return spring (66) fixedly attached to the system card (30) on one end and attached to the magnetic core (68) on the opposite end provides a force on the magnetic core (68) acting in a second direction opposite to the first direction.
  • 53. When the number of energized NFETs (80) increases due to transients on the output of the driver circuit (20), the control current through the electromagnetic coil (62) increases. As the control current through the electromagnetic coil (62) increases, the electromagnetic force acting on the magnetic core (68) increases, causing the magnetic core (68) to move in the first direction, decreasing the length (D52) of the adjustable-length transmission line (52).
  • 54. When the number of energized NFETs (80) decreases due to the clock down generator (83), the control current through the electromagnetic coil (62) decreases. As the control current through the electromagnetic coil (62) decreases, the return spring (66), which is in tension, causes the magnetic core (68) to move in the second direction, increasing the length (D52) of the adjustable-length transmission line (52).
  • 55. Referring to FIG. 6A, the control circuit (70) detects overshoots or (positive) high voltage transients and undershoots or (negative) low voltage transients on the output of the driver circuit (20) and provides a control current in the control current node (71) to the adjustment mechanism (60). The control current in control current node (71) causes the adjustment mechanism (60) to adjust the length (D52) of the adjustable-length transmission line (52), thereby minimizing overshoots and undershoots.
  • 56. The control circuit (70), as shown in FIG. 6A, comprises positive detect circuitry (72) and negative detect circuitry (74) connected to the output of the driver circuit (20). When an overshoot occurs on the output of the driver circuit (20), the positive detect circuitry (72) detects the overshoot and generates a high pulse at node UP1 (77). When an undershoot occurs on the output of the driver circuit (20), the negative detect circuitry (74) detects the undershoot and generates a high pulse at node UP2 (79).
  • 57. Nodes UP1 (77) and UP2 (79) are connected to the inputs of an OR gate (76), which is in turn connected at its output to an n-bit counter (78). The n-bit counter (78) is connected at its outputs to the control gates of a quantity “n” N-type field effect transistors or NFETs (80) which are connected in parallel. The quantity “n” is large enough to provide adjustment increments that will meet the overshoot-undershoot threshold established for the system design, preferably between about 4 and 16. A quantity of these NFETs (80) are made active (energized or turned on) equal to the count of the n-bit counter (78). A current is drawn through the NFETs (80) which is proportional to the number of NFETs (80) which are energized. The NFETs (80) are connected to ground at their sources and to the node CONTROL CURRENT (71) at their drains. The node CONTROL CURRENT (71) is connected to the adjustment mechanism.
  • 58. The control current at the node CONTROL CURRENT (71) is determined by the total currents in the “n” NFETs which are energized. The node CONTROL CURRENT (71) is connected to the adjustment mechanism (60), providing the control current to the adjustment mechanism (60).
  • 59. A down signal is periodically generated by the CLK down generator (83) on the node DOWN (85) as a function of the system clock (88). The down signal causes the n-bit counter (78) to count down, thereby reducing by one the quantity of NFETs (80) which are energized. The periodic down signal prevents the impedance matching network from being over-adjusted because of impedance changes due to dynamic environmental conditions such as temperature.
  • 60. When a positive transient or overshoot occurs on the output of the driver circuit (20), the positive detect circuitry (72) generates a high pulse at node UP1 (77). The positive detect circuitry (72) is preferably implemented as shown in FIG. 6B. A pulse up transistor (T100), which is an N-well PFET, is connected to the driver circuit power supply (101) at its body and to a high reference potential, typically the driver circuit power supply (101), at its control gate (VGATE). The pulse up transistor (T100) is connected to the output of the driver circuit (20) at its source and to node PULSE (103) at its drain. When a positive transient occurs on the output of the driver circuit (20), the pulse up transistor (T100) switches the node PULSE (103) from low to high. An up interrogate transistor (T104), which is an NFET, is connected to the node PULSE (103) at its control gate, to ground at its source, and to node UPC1 (105) at its drain. When node PULSE (103) switches to high, the up interrogate transistor (T104) switches node UPC1 (105) from high to low.
  • 61. Nodes UPC1 (105) and DOWN (85) are connected to the inputs of a NOR gate (G106) which generates an up (high) pulse on node UP1 (77) when node UPC1 (105) is low due to a high transient, provided node DOWN (85) has not been switched to a high level by the CLK down generator (83). Node UPC1 (105) is also connected to a voltage keeper comprising an inverter (I108) connected to the control gate of a PFET keeper transistor (T110). The source of the keeper transistor (T110) is connected to internal power and the drain is connected to node UPC1 (105) such that node UPC1 (105) is held at a high state until interrogation by the up interrogate transistor (T104) can occur, thereby preventing accidental tripping by system noise.
  • 62. Node UPC1 (105) is also connected to a reset loop comprising an odd number of reset delay inverters (I112) connected in series. The last reset delay inverter (I112) is connected to node RESET (109). The number of reset delay inverters (I112) is sufficient to cause a delay longer than the duration of the transient. Experimental modeling has shown that seven reset delay inverters (I112) are typically sufficient. When node UPC1 (105) switches to low, the reset loop switches node RESET (109) from low to high. The node RESET (109) is connected to the control gate of a pulse reset transistor (T114), which is an NFET connected to ground at its source and to node PULSE (103) at its drain. When node RESET (109) switches to high, the pulse reset transistor (T114) switches node PULSE (103) back to its original state of low.
  • 63. Node RESET (109) is also connected to a UPC1 reset inverter (I116), which is in turn connected to the control gate of a UPC1 reset transistor (T118). The UPCl reset transistor (T118) is a PFET connected to internal power at its source and to node UPC1 (105) at its drain. When node RESET (109) switches to high, the UPC1 reset transistor (T118) switches node UPC1 (105) back to its original state of high.
  • 64. A first optional isolating resistive device (R120) can be connected between internal power and node RESET (109) to prevent accidental switching due to coupling to external signals or system noise. Likewise, a second optional isolating resistive device (R122) can be connected between ground and node PULSE (103). The isolating resistive devices (R120, R122) can be resistors or resistively connected FETs.
  • 65. The negative detect circuitry (74) is a mirror image of the positive detect circuitry (72). When a negative transient or undershoot occurs on the output of the driver circuit (20), the negative detect circuitry (72) generates a high pulse at node UP2 (79). The negative detect circuitry (72) is preferably implemented as shown in FIG. 6C.
  • 66. A pulse down transistor (T200), which is an NFET, is connected to a low reference potential, typically ground, at its control gate (VGATE2). The pulse down transistor (T200) is connected to the output of the driver circuit (20) at its source and to node PULSE DOWN (203) at its drain. When a negative transient occurs on the output of the driver circuit (20), the pulse down transistor (T200) switches the node PULSE DOWN (203) from high to low. A down interrogate transistor (T204), which is a PFET, is connected to the node PULSE DOWN (203) at its control gate, to internal power at its source, and to node UPC2 (205) at its drain. When the node PULSE DOWN (203) switches to low, the down interrogate transistor (T204) switches node UPC2 (205) from low to high.
  • 67. Down (85) is connected to the input of an inverter (I240) whose output is connected to an input of a NAND gate (G206). Node UPC2 (205) is connected to the other input of NAND gate (G206). NAND gate (G206) is connected in series with a UP2 inverter (I230) then a node UP2 (79). An up (high) pulse is generated on node UP2 (79) when node UPC2 (205) is high due to a low transient, provided node DOWN (85) has not been switched to a high level by the CLK down generator (83). Node UPC2 (205) is also connected to a voltage keeper comprising an inverter (I208) connected to the control gate of an NFET keeper transistor (T210). The source of the NFET keeper transistor (T210) is connected to ground and the drain is connected to node UPC2 (205) such that node UPC2 (205) is held at a low state until interrogation by the down interrogate transistor (T204) can occur, thereby preventing accidental tripping by system noise.
  • 68. Node UPC2 (205) is also connected to a second reset loop comprising an odd number of reset delay inverters (I212) connected in series. The last reset delay inverter (I212) is connected to node RESET2 (209). The number of reset delay inverters (I212) is sufficient to cause a delay longer than the duration of the transient. Experimental modeling has shown that seven reset delay inverters (I212) are typically sufficient. When node UPC2 (205) switches to high, the reset loop switches node RESET2 (209) from high to low. The node RESET2 (209) is connected to the control gate of a pulse down reset transistor (T214), which is a PFET connected to internal power at its source and to node PULSE DOWN (203) at its drain. When node RESET2 (209) switches to low, the pulse down reset transistor (T214) switches node PULSE DOWN (203) back to its original state of high.
  • 69. Node RESET2 (209) is also connected to a UPC2 reset inverter (I216), which is in turn connected to the control gate of a UPC2 reset transistor (T218). The UPC2 reset transistor (T218) is an NFET connected to ground at its source and to node UPC2 (205) at its drain. When node RESET2 (209) switches to low, the UPC2 reset transistor (T218) switches node UPC2 (205) back to its original state of low.
  • 70. A third optional isolating resistive device (R220) can be connected between ground and node RESET2 (209) to prevent accidental switching due to coupling to external signals or system noise. Likewise, a fourth optional isolating resistive device (R222) can be connected between internal power and node PULSE DOWN (203). The isolating resistive devices (R220, R222) can be resistors or resistively connected FETs.
  • 71. In the above description, the terms “low pulse,” “low state,” and “low” and, similarly, the terms “high pulse,” “high state,” and “high” refer to relative voltage states and do not imply a specific magnitude.
  • 72. Impedance Matching Calculations
  • 73. The purpose of the adjustable-length transmission line (52) is to reduce transient voltages by providing a matching impedance for the reactive component of the impedance of the receiver circuit (10) to the driver circuit (20). When the adjustable-length transmission line (52) is short circuited at one end, the input impedance at the other end will have no resistive component. The reactive component (i.e., capacitive or inductive) of the input impedance will have a magnitude dependent on the length (D52) of the adjustable-length transmission line (52) and the frequency (f) of the signal from the driver circuit (20).
  • 74. Thus, with a reactive impedance that can be adjusted, impedance matching between the on-chip driver circuit (20) and the off-chip receiver circuit (10) can be achieved for a wide range of configurations. Frequency (f) is dependant upon system design, where performance demands increasingly require operation at ultra-high frequencies (0.5 GHz and above). The required length (D52) of the adjustable-length transmission line (52) can be calculated using electromagnetic principles as described in chapter 12 of Engineering Electromagnetics, by William H. Hayt, Jr., McGraw-Hill, 1981, incorporated herein by reference.
  • 75. Referring again to FIG. 2A, the resistive component (R1) and the reactive component (X1) of the input impedance (Z1) on the receiver side of the transmission line (40) are given by Equations 1 and 2, respectively, in terms of frequency (f), receiver resistance (R10), receiver capacitance (C10), and receiver inductance (L31). As is known in the art, impedance can generally be expressed as Z=R+j·X and, therefore, the input impedance (Z1) can be expressed in the form of Equation 0.
  • Z1=R1+j·X 1  (Equation 0)
  • 76. Equations 1 and 2 can be derived for an inductor (such as receiver inductance (L31)) in series with a parallel capacitor (such as receiver circuit capacitance (C10)) and resistor (such as receiver circuit resistance (R10)), as shown in FIG. 2A. Angular frequency (ω) is equal to 2πf. R1 = R10 1 + ( ω · R10 · C10 ) 2 ( Equation 1 ) X1 = ω L31 - ( ω · R10 2 · C10 ) + ( ω 3 · R10 2 · L31 · C10 2 ) 1 + ( ω · R10 · C10 ) 2 ( Equation 2 )
    Figure US20010000428A1-20010426-M00001
  • 77. From Engineering Electromagnetics, the input impedance (Z4) at the driver side of the transmission line (40) can be expressed in terms of the characteristic impedance (Z40) of the transmission line (40), the load or input impedance (Z1), the transmission line length (D40), and a phase factor (β40) for the transmission line (40). Z4 = Z40 · Z1 + [ j · Z40 · tan ( β 40 · D40 ) ] Z40 + [ j · Z1 · tan ( β 40 · D40 ) ] ( Equation 3 )
    Figure US20010000428A1-20010426-M00002
  • 78. The phase factor (β40) is dependant upon the frequency (f) of the signal from the driver circuit (20) and the dielectric constant (∈40) and the relative permeability (μ40) for the transmission line (40), and can be determined using equations 4A and 4B, where “c” is the speed of light (3×10−8 m/s) β 40 = 2 π f v40 (Equation 4A) v40 = c ( μ 40 · ɛ 40 ) - 0.5 (Equation 4B)
    Figure US20010000428A1-20010426-M00003
  • 79. The resistive component (R4) and the reactive component (X4) of the input impedance (Z4) on the driver side of the transmission line (40) can be determined by substituting Equation 0 into Equation 3 and separating the real (resistive) and imaginary (reactive) components. The resistive component (R4) of the input impedance (Z4) at the driver side of the transmission line (40) can be expressed as Equation 5, and the reactive component (X4) of the input impedance (Z4) at driver side of the transmission line (40) can be expressed as Equation 6. R4 = ( R1 · Z40 ) 2 1 + tan 2 ( β 40 · D40 ) [ Z40 - ( X1 · tan ( β 40 · D40 ) ) ] 2 + [ R1 · tan ( β 40 · D40 ) ] 2 ( Equation 5 ) X4 = Z40 · { [ X1 + ( Z40 · tan ( β 40 · D40 ) ) ] · [ Z40 - ( X1 · tan ( β 40 · D40 ) ) ] } - [ R1 2 · tan ( β 40 · D40 ) ] [ Z40 - ( X1 · tan ( β 40 · D40 ) ) ] 2 + [ R1 · tan ( β 40 · D40 ) ] 2 ( Equation 6 )
    Figure US20010000428A1-20010426-M00004
  • 80. As shown in FIG. 2A, R4 and X4 are also the resistive and reactive components of the impedance at the receiver side of the matching network (50). The driver circuit (20), matching network (50), transmission line (40) and receiver circuit (10) of FIG. 2A can now be reduced to the circuit shown in FIG. 4, where the impedance (Z50) of the matching network (50) and the input impedance (Z4) on the receiver side of the matching network (50) are in parallel.
  • 81. The equivalent impedance (Z50) of the matching network (50) has no resistive component because the adjustable-length transmission line (52) is short circuited. The equivalent impedance (Z50) of the matching network (50), as given by Equation 7, can be expressed as a function of the characteristic impedance (Z52) of the adjustable-length transmission line (52), the phase factor (β52) of the adjustable-length transmission line (52), and the length (D52) to which the adjustable-length transmission line (52) is set.
  • Z50=j·Z52·tan (β52·D52)  (Equation 7)
  • 82. The phase factor (β52) for the adjustable-length transmission line (52) is dependent upon the frequency (f) of the signal from the driver circuit (20) and the dielectric constant (∈52) and the relative permeability (μ52) for the adjustable-length transmission line (52). The phase factor (β52) for the adjustable-length transmission line (52) can be determined using equations 8A and 8B, where “c” is the speed of light (3×108 m/s). β 52 = 2 π f v52 (Equation 8A) v52 = c ( μ52 · ɛ 52 ) - 0.5 (Equation 8B)
    Figure US20010000428A1-20010426-M00005
  • 83. Referring to FIG. 5, the net input impedance (Z3) at the driver side of the matching network (50) is given by Equation 9. The resistive component (R3) of the net input impedance (Z3) is given by Equation 10. The reactive component (X3) of the net input impedance (Z3) is given by Equation 11. Z3 = [ j · Z52 · tan ( β 52 · D52 ) ] ( R4 + j · X4 ) R4 + j · X4 + [ j · Z52 · tan ( β 52 · D52 ) ] ( Equation 9 ) R3 = [ X4 + Z52 · tan ( β 52 · D52 ) ] [ R4 · Z52 · tan ( β 52 · D52 ) ] - [ X4 · R4 · Z52 · tan ( β 52 · D52 ) ] ( R4 2 + [ X4 + Z52 · tan ( β 52 · D52 ) ] 2 ) ( Equation 10 )
    Figure US20010000428A1-20010426-M00006
    X3 = [ X4 · Z52 · tan ( β 52 · D52 ) ] [ X4 + Z52 · tan ( β 52 · D52 ) ] + [ R4 2 · Z52 · tan ( β 52 · D52 ) ] R4 2 + [ X4 + Z52 · tan ( β 52 · D52 ) ] 2 ( Equation 11 )
    Figure US20010000428A1-20010426-M00007
  • 84. The equivalent impedance (ZI) for the driver circuit (20) can be derived from the circuit model of FIG. 2A, and is given by Equation 12. The resistive component (RI) of the input impedance (ZI) to the driver side of the matching network (50) is given by Equation 13, and the reactive component (XI) of the input impedance (ZI) to the driver side of the matching network (50) is given by Equation 14. Equations 13 and 14 can be derived for a parallel driver circuit capacitance (C20) and driver circuit resistance (R20) in series with a driver inductance (L32), as shown in FIG. 2A. ZI = R20 · ( 1 / j · ω · C20 ) R20 + ( 1 / j · ω · C20 ) + j · ω · L32 ( Equation 12 )
    Figure US20010000428A1-20010426-M00008
    RI = R20 1 + ( ω · R20 · C20 ) 2 ( Equation 13 )
    Figure US20010000428A1-20010426-M00009
    XI = - ( ω · R20 2 · C20 ) 1 + ( ω · R20 · C20 ) 2 + ω · L32 ( Equation 14 )
    Figure US20010000428A1-20010426-M00010
  • 85. Referring again to FIG. 5, for impedance matching, the reactive components of the net input impedance (Z3) at the driver side of the matching network (50) and of the equivalent impedance (ZI) for the driver circuit (20) must cancel out (i.e., the matching network (50) must provide a reactance, which in combination with the driver and receiver circuits, will give no net reactance). To do this, the reactive component (X3) must be the conjugate (equal in magnitude with the opposite sign) of the reactive component (XI), as shown in Equation 15. The matching of the reactive components will result in no overshoots or undershoots.
  • XI=−X3  (Equation 15)
  • 86. Equations 11 and 14 can be substituted into Equation 15. Selected values for frequency (f) and length (D52) of the adjustable-length transmission line (52) can be substituted into Equation 15 to determine, by the process of iteration, the values at which the net reactance (XI+X3) is equal to zero (0), thereby achieving impedance matching. FIG. 7 shows the values for length (D52) of the adjustable-length transmission line (52) as a function of frequency (f) which provide impedance matching over a range of selected values for the receiver circuit resistance (termination load) (R10) ranging from 50 ohms to 500 mega ohms. The values shown in FIG. 7 are determined using typical values as previously provided for C10 (3.7 pF), L31 (2 nH), D40 (76 mm or 3 inches), Z40 (50 ohms), ∈40 (4.5), L32 (2 nH), C20 (4.7 pF), Z52 (50 ohms), and ∈52 (4.5).
  • 87. Additional iterative solutions were plotted for the length (D52) of the adjustable-length transmission line (52) as a function of frequency (f) over a range of selected values for L31, L32, C10, C20, ∈40, ∈52, D40, Z40, and Z52, successively. FIG. 8 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line (52) for a receiver inductance (L31) in the range between 1 nH and 8 nH. The values shown in FIG. 8 are determined using typical values as previously provided for C10 (3.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), ∈40 (4.5), L32 (2 nH), C20 (4.7 pF), Z52 (50 ohms), and ∈52 (4.5).
  • 88.FIG. 9 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line (52) for a driver inductance (L32) in the range between 1 nH and 16 nH. The values shown in FIG. 9 are determined using typical values as previously provided for C10 (3.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), ∈40 (4.5), L31 (2 nH), C20 (4.7 pF), Z52 (50 ohms), and ∈52 (4.5).
  • 89.FIG. 10 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line (52) for a receiver capacitance (C10) in the range between 1 pF and 12 pF. The values shown in FIG. 10 are determined using typical values as previously provided for C20 (4.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), ∈40 (4.5), L31 (2 nH), L32 (2 nH), Z52 (50 ohms), and ∈52 (4.5).
  • 90.FIG. 11 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line (52) for a driver capacitance (C20) in the range between 1 pF and 12 pF. The values shown in FIG. 11 are determined using typical values as previously provided for C10 (3.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), ∈40 (4.5), L31 (2 nH), L32 (2 nH), Z52 (50 ohms), and ∈52 (4.5).
  • 91.FIG. 12 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line (52) for a dielectric constant (∈40, ∈52) for the transmission line (40) and the adjustable-length transmission line (52) in the range between 2 and 6. The values shown in FIG. 12 are determined using typical values as previously provided for C10 (3.7 pF), C20 (4.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), Z40 (50 ohms), L31 (2 nH), L32 (2 nH), and Z52 (50 ohms).
  • 92.FIG. 13 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line (52) for a length (D40) of the transmission line (40) in the range between 51 mm (2 inches) and 127 mm (5 inches). The values shown in FIG. 13 are determined using typical values as previously provided for C10 (3.7 pF), C20 (4.7 pF), R10 (50 ohms), Z40 (50 ohms), Z52 (50 ohms), ∈40 (4.5), L31 (2 nH), L32 (2 nH), and ∈52 (4.5).
  • 93.FIG. 14 shows that the present invention can provide impedance matching using a comparable length (D52) adjustable-length transmission line (52) for a characteristic impedance (Z40, Z52) of the transmission line (40) and of the adjustable-length transmission line (52) in the range between 20 ohms and 200 ohms. The values shown in FIG. 14 are determined using typical values as previously provided for C10 (3.7 pF), C20 (4.7 pF), R10 (50 ohms), D40 (76 mm or 3 inches), ∈40 (4.5), L31 (2 nH), L32 (2 nH), and ∈52 (4.5).
  • 94. Although illustrated and described above with reference to certain specific embodiments, the present invention is nevertheless not intended to be limited to the details shown. Rather, various modifications may be made in the details within the scope and range of equivalents of the claims and without departing from the spirit of the invention.

Claims (28)

What is claimed is:
1. An impedance matching network comprising an adjustable-length transmission line connected in parallel with an input-output driver circuit having an output, wherein the length of the adjustable-length transmission line can be manually adjusted to provide impedance matching at the output of the driver circuit.
2. The impedance matching network of
claim 1
wherein the adjustable-length transmission line comprises two parallel conductive lines connected at one end to the driver circuit and a movable stub, the adjustable-length transmission line electrically shorted by the movable stub.
3. The impedance matching network of
claim 2
further comprising a system card and a semiconductor chip located on the system card, the driver circuit located on the semiconductor chip and the parallel conductive lines comprise stripes of conductive material formed directly on the system card.
4. The impedance matching network of
claim 2
further comprising a receiver circuit having a resistance and being connected to the output of the driver circuit.
5. The impedance matching network of
claim 2
wherein the length of the adjustable-length transmission line is adapted to be adjusted to precisely set the distance between the point where the parallel conductive lines are shorted by the moveable stub and the point where the parallel conductive lines are connected to the driver circuit to provide impedance matching at the driver circuit output.
6. The impedance matching network of
claim 5
wherein the length of the adjustable-length transmission line is adapted to be adjusted using a vernier.
7. An impedance matching system, comprising:
a system card;
a semiconductor chip located on the system card;
a driver circuit located on the semiconductor chip and having an output and a capacitance;
a driver connection having an inductance;
an impedance matching network located on the system card and connected in parallel to the driver circuit;
a receiver circuit having a resistance and a capacitance;
a receiver connection having an inductance; and
a transmission line connected at one end to both the impedance matching network and, through the driver connection, to the output of the driver circuit and connected on the opposite end through the receiver connection to the receiver circuit, the transmission line having a characteristic impedance and a dielectric constant.
8. The impedance matching system of
claim 7
wherein the adjustable-length transmission line is located off-chip from the driver circuit.
9. The impedance matching system of
claim 7
wherein impedance matching is provided for a receiver circuit resistance between about 50 and 100 mega ohms.
10. The impedance matching system of
claim 7
wherein impedance matching is provided for a receiver connection inductance between about 1 and 8 nH.
11. The impedance matching system of
claim 7
wherein impedance matching is provided for a driver connection inductance between about 1 and 16 nH.
12. The impedance matching system of
claim 7
wherein impedance matching is provided for a receiver circuit capacitance in the range between about 1 and 12 pF.
13. The impedance matching system of
claim 7
wherein impedance matching is provided for a driver circuit capacitance between about 1 and 12 pF.
14. The impedance matching system of
claim 7
wherein impedance matching is provided for a dielectric constant for the transmission line in the range between about 2 and 6.
15. The impedance matching system of
claim 7
wherein impedance matching is provided for a length of the transmission line in the range between about 50 and 125 mm.
16. The impedance matching system of
claim 7
wherein impedance matching is provided for a characteristic impedance of the transmission line in the range between about 20 and 200 ohms.
17. A process for reducing overshoots and undershoots on a driver circuit by impedance matching, comprising the steps of:
detecting overshoots and undershoots on the output of the driver circuit using electronic monitoring equipment;
locating a movable stub which is connected across two parallel conductive lines a distance determined by the magnitude of overshoots and undershoots detected, the parallel conductive lines being connected in parallel to the driver circuit at one end thereby forming an adjustable- length transmission line comprising the parallel conductive lines and the movable stub; and
repeating the first and second steps until the adjustable-length transmission line is adjusted to a length at which the impedance at the driver circuit is matched and overshoots and undershoots are reduced to an acceptable magnitude.
18. The process of
claim 17
wherein the movable stub is initially positioned such that the adjustable-length transmission line is at its maximum length.
19. The process of
claim 17
wherein the position of the movable stub is changed using a vernier device.
20. The process of
claim 17
wherein the electronic monitoring equipment is connected to the output of the driver circuit throughout the adjustment of the adjustable-length transmission line and the first step and the second step are performed simultaneously.
21. An automatic impedance matching network providing impedance matching for a driver circuit located on a semiconductor chip on a system card and providing a voltage signal to a receiver circuit, the driver circuit having an output and the network comprising:
an adjustable-length transmission line located on the system card and connected in parallel with the driver circuit;
a control circuit located on the system card and generating a control current varied in response to high and low transient voltages on the driver circuit output; and
adjustment means located on the system card for adjusting the length of the adjustable-length transmission line proportionally to the control current generated by the control circuit.
22. The impedance matching network of
claim 21
wherein the adjustable-length transmission line comprises two parallel conductive lines connected at one end to the driver circuit and a movable stub, the adjustable-length transmission line electrically shorted by the movable stub.
23. The impedance matching network of
claim 22
wherein the adjustment means comprises:
an electromagnetic coil creating a force in a first direction proportional to the control current;
a return spring creating a force in a second direction opposite to the first direction; and
a magnetic core, attached to the moveable stub, having a position determined by the equilibrium of the force created by the electromagnetic coil and the force created by the return spring.
24. The impedance matching network of
claim 21
wherein the control circuit comprises:
an n-bit counter;
a source of internal power;
a CONTROL CURRENT node;
a quantity “n” N-type field effect transistors connected in parallel between the source of internal power and the CONTROL CURRENT node, said N-type field effect transistors having control gates connected to the n-bit counter;
a positive detect circuitry incrementally increasing the n-bit counter when an overshoot occurs on the driver circuit output; and
a negative detect circuitry incrementally increasing the n-bit counter when an undershoot occurs on the driver circuit output.
25. A process for reducing overshoots and undershoots on the output of a driver circuit by impedance matching, comprising the steps of:
detecting overshoots on the output of the driver circuit using a positive detect circuitry and transmitting an incremental pulse to a counter;
detecting undershoots on the output of the driver circuit using a negative detect circuitry and transmitting an incremental pulse to a counter;
transmitting an activation voltage to the control gates of a quantity of N-type field effect transistors connected in parallel between an internal power source and a CONTROL CURRENT node, wherein the quantity of N-type field effect transistors so activated is proportional to the count of the counter;
transmitting a control current from the CONTROL CURRENT node to an electromagnetic adjustment means, wherein the control current is proportional to the quantity of the N-type field effect transistors that are activated; and
moving a moveable stub which is connected across two parallel conductive lines a distance proportional to the magnitude of the control current, the parallel conductive lines being connected in parallel to the driver circuit at one end thereby forming an adjustable-length transmission line comprising the parallel conductive lines and the moveable stub, whereby the adjustable-length transmission line is adjusted to a length at which the impedances at the output of the driver circuit are matched and overshoots and undershoots are reduced.
26. An automatic, self-resetting control circuit for an impedance matching network that detects overshoots and undershoots on a driver circuit output and provides, to an adjustment mechanism for the impedance matching network having an adjustable-length transmission line, a control current whose magnitude is proportional to the magnitude of overshoots and undershoots on the output of the driver circuit, the control circuit comprising:
positive detect circuitry detecting overshoots on the output of the driver circuit and providing an incremental pulse;
negative detect circuitry detecting undershoots on the output of the driver circuit and providing an incremental pulse;
a counter receiving the incremental pulses from the positive detect circuitry and the negative detect circuitry, counting the incremental pulses, calculating a count, and providing an activation voltage;
a source of power;
a system ground;
a CONTROL CURRENT node;
a quantity of N-type field effect transistors connected in parallel and having control gates connected to and receiving the activation voltage from the counter, wherein the quantity of N-type field effect transistors being activated is equal to the count of the counter; the N-type transistors drawing current, through the CONTROL CURRENT node, to system ground, thereby providing a control current at the CONTROL CURRENT node for adjusting the adjustable-length transmission line for impedance matching; and
a CLK down generator providing a down pulse to the counter to prevent over-adjustment of the adjustable-length transmission line.
27. The automatic, self-resetting control circuit of
claim 26
wherein:
(a) the positive detect circuitry comprises:
a node PULSE having a low state and a higher voltage state,
a pulse up transistor activated by an overshoot on the output of the driver circuit, the pulse up transistor switching the node PULSE from the low state to the higher voltage state, the magnitude of the higher voltage state being proportional to the magnitude of the transient overshoot at the output of the driver circuit,
a node UPC1 having a low and a high state,
an up interrogate transistor activated when the node PULSE switches from the low state to the higher voltage state, the up interrogate transistor switching the node UPC1 from the high to the low state,
a voltage keeper holding node UPC1 high until activation of the up interrogate transistor,
a reset delay loop delaying a reset signal for a length of time greater than the duration of a voltage transient at the driver circuit output,
a pulse reset transistor resetting node PULSE to its original low state,
a UPC1 reset transistor resetting node UPC1 to its original high state, and
a NOR gate having inputs connected to the node UPC1 and to the output of the CLK down generator and having an output connected to the counter, whereby the counter is incremented when an overshoot occurs unless the CLK down generator receives a down pulse; and
(b) the negative detect circuitry comprises:
a node PULSE DOWN having a high state and a lower voltage state,
a pulse down transistor activated by an undershoot on the output of the driver circuit, the pulse down transistor switching the node PULSE DOWN from the high state to the lower voltage state, the magnitude of the lower voltage state being proportional to the magnitude of the undershoot at the driver circuit output,
a node UPC2 having a low and a high state,
a down interrogate transistor activated when the node PULSE DOWN switches from the high state to the lower voltage state, the down interrogate transistor switching the node UPC2 from the low to the high state,
a voltage keeper holding node UPC2 low until activation of the down interrogate transistor,
a pulse down reset transistor resetting node PULSE DOWN to its original high state,
a UPC2 reset transistor resetting node UPC2 to its original low state, and
a NAND gate having inputs connected to the node UPC2 and to the inverted output of the CLK down generator and having an input connected to the counter, whereby the counter is incremented when an undershoot occurs unless the CLK down generator receives a down pulse.
28. The automatic, self-resetting control circuit of
claim 27
further comprising:
a RESET node;
a first isolating resistive device connected between the source of internal power and the RESET node;
a second isolating resistive device connected between the system ground and the node PULSE;
a RESET2 node;
a third isolating resistive device connected between the system ground and the RESET2 node; and
a fourth isolating resistive device connected between the internal power source and the node PULSE DOWN, thereby reducing the susceptibility of the control circuit to accidental switching due to system noise.
US09/735,679 1999-02-23 2000-12-13 Termination resistance independent system for impedance matching in high speed input-output chip interfacing Expired - Lifetime US6278339B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US09/735,679 US6278339B2 (en) 1999-02-23 2000-12-13 Termination resistance independent system for impedance matching in high speed input-output chip interfacing

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US09/255,997 US6140885A (en) 1999-02-23 1999-02-23 On-chip automatic system for impedance matching in very high speed input-output chip interfacing
US09/320,902 US6249193B1 (en) 1999-02-23 1999-05-27 Termination impedance independent system for impedance matching in high speed input-output chip interfacing
US09/735,679 US6278339B2 (en) 1999-02-23 2000-12-13 Termination resistance independent system for impedance matching in high speed input-output chip interfacing

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
US09/320,902 Division US6249193B1 (en) 1999-02-23 1999-05-27 Termination impedance independent system for impedance matching in high speed input-output chip interfacing

Publications (2)

Publication Number Publication Date
US20010000428A1 true US20010000428A1 (en) 2001-04-26
US6278339B2 US6278339B2 (en) 2001-08-21

Family

ID=26945096

Family Applications (2)

Application Number Title Priority Date Filing Date
US09/320,902 Expired - Lifetime US6249193B1 (en) 1999-02-23 1999-05-27 Termination impedance independent system for impedance matching in high speed input-output chip interfacing
US09/735,679 Expired - Lifetime US6278339B2 (en) 1999-02-23 2000-12-13 Termination resistance independent system for impedance matching in high speed input-output chip interfacing

Family Applications Before (1)

Application Number Title Priority Date Filing Date
US09/320,902 Expired - Lifetime US6249193B1 (en) 1999-02-23 1999-05-27 Termination impedance independent system for impedance matching in high speed input-output chip interfacing

Country Status (1)

Country Link
US (2) US6249193B1 (en)

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6600339B2 (en) * 1999-02-09 2003-07-29 Micron Technology, Inc. Current mode signal interconnects and CMOS amplifier
US20030173653A1 (en) * 2002-03-13 2003-09-18 Micron Technology, Inc. High permeability thin films and patterned thin films to reduce noise in high speed interconnections
US20030176028A1 (en) * 2002-03-13 2003-09-18 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20040233010A1 (en) * 2003-05-22 2004-11-25 Salman Akram Atomic layer deposition (ALD) high permeability layered magnetic films to reduce noise in high speed interconnection
US20050023650A1 (en) * 2002-01-30 2005-02-03 Micron Technology, Inc. Capacitive techniques to reduce noise in high speed interconnections
US20050030803A1 (en) * 2002-03-13 2005-02-10 Micron Technology, Inc. High permeability layered films to reduce noise in high speed interconnects
US20060131702A1 (en) * 1999-07-30 2006-06-22 Micron Technology, Inc. Novel transmission lines for CMOS integrated circuits
US20060258187A1 (en) * 2005-05-16 2006-11-16 Teradyne, Inc. Impedance controlled via structure
EP1395894A4 (en) * 2001-06-13 2007-04-04 Sony Computer Entertainment Inc Low power clock distribution methodology
DE102011075333A1 (en) * 2011-05-05 2012-11-08 Rohde & Schwarz Gmbh & Co. Kg Amplifier has end terminal with contact element which is adapted to electrically contact spacer at predetermined position, where contact element is connected to ground terminal
US20120309456A1 (en) * 2011-05-31 2012-12-06 Renesas Mobile Corporation Semiconductor integrated circuit device, electronic device, and radio communication device
US8501563B2 (en) 2005-07-20 2013-08-06 Micron Technology, Inc. Devices with nanocrystals and methods of formation
WO2015099918A1 (en) * 2013-12-27 2015-07-02 Intel Corporation High speed short reach input/output (i/o)
US20150381173A1 (en) * 2014-06-30 2015-12-31 Fujitsu Limited Impedance matching driver
US9407382B1 (en) * 2015-03-26 2016-08-02 Sprint Communications Company L.P. Positioning stub for mitigation of return loss during signal transmission
US20170054192A1 (en) * 2015-08-17 2017-02-23 Electronics And Telecommunications Research Institute High-frequency high-power terminator
US20180048094A1 (en) * 2014-05-08 2018-02-15 Apple Inc. Connector system impedance matching

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI288528B (en) * 2002-06-21 2007-10-11 Realtek Semiconductor Corp Impedance matching circuit
US7248636B2 (en) * 2004-04-20 2007-07-24 Hewlett-Packard Development Company, L.P. Systems and methods for adjusting an output driver
US7224180B2 (en) * 2004-11-18 2007-05-29 Seiko Epson Corporation Methods and systems for rise-time improvements in differential signal outputs
US20080218290A1 (en) * 2007-02-27 2008-09-11 International Business Machines Corporation Automatic Driver/Transmission Line/Receiver Impedance Matching Circuitry
US7622946B2 (en) * 2007-11-05 2009-11-24 International Business Machines Corporation Design structure for an automatic driver/transmission line/receiver impedance matching circuitry
US9899833B2 (en) 2015-02-12 2018-02-20 Aria Controls Corporation Electronic direct current transformer

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4492960A (en) * 1982-01-18 1985-01-08 The United States Of America As Represented By The Secretary Of The Navy Switching mixer
US6140885A (en) * 1999-02-23 2000-10-31 International Business Machines Corporation On-chip automatic system for impedance matching in very high speed input-output chip interfacing

Cited By (62)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6600339B2 (en) * 1999-02-09 2003-07-29 Micron Technology, Inc. Current mode signal interconnects and CMOS amplifier
US6737887B2 (en) 1999-02-09 2004-05-18 Micron Technology, Inc. Current mode signal interconnects and CMOS amplifier
US20060131702A1 (en) * 1999-07-30 2006-06-22 Micron Technology, Inc. Novel transmission lines for CMOS integrated circuits
EP1395894A4 (en) * 2001-06-13 2007-04-04 Sony Computer Entertainment Inc Low power clock distribution methodology
US7737536B2 (en) 2002-01-30 2010-06-15 Micron Technology, Inc. Capacitive techniques to reduce noise in high speed interconnections
US20060261438A1 (en) * 2002-01-30 2006-11-23 Micron Technology, Inc. Capacitive techniques to reduce noise in high speed interconnections
US20060244108A1 (en) * 2002-01-30 2006-11-02 Micron Technology, Inc. Capacitive techniques to reduce noise in high speed interconnections
US7101770B2 (en) 2002-01-30 2006-09-05 Micron Technology, Inc. Capacitive techniques to reduce noise in high speed interconnections
US20050023650A1 (en) * 2002-01-30 2005-02-03 Micron Technology, Inc. Capacitive techniques to reduce noise in high speed interconnections
US6844256B2 (en) 2002-03-13 2005-01-18 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20030176028A1 (en) * 2002-03-13 2003-09-18 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US7829979B2 (en) 2002-03-13 2010-11-09 Micron Technology, Inc. High permeability layered films to reduce noise in high speed interconnects
US6833317B2 (en) 2002-03-13 2004-12-21 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20050006727A1 (en) * 2002-03-13 2005-01-13 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20050007817A1 (en) * 2002-03-13 2005-01-13 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US6794735B2 (en) 2002-03-13 2004-09-21 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US6846738B2 (en) 2002-03-13 2005-01-25 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20050017327A1 (en) * 2002-03-13 2005-01-27 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US6787888B2 (en) 2002-03-13 2004-09-07 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20050030803A1 (en) * 2002-03-13 2005-02-10 Micron Technology, Inc. High permeability layered films to reduce noise in high speed interconnects
US6884706B2 (en) 2002-03-13 2005-04-26 Micron Technology Inc. High permeability thin films and patterned thin films to reduce noise in high speed interconnections
US6900116B2 (en) 2002-03-13 2005-05-31 Micron Technology Inc. High permeability thin films and patterned thin films to reduce noise in high speed interconnections
US6903003B2 (en) 2002-03-13 2005-06-07 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US6903444B2 (en) 2002-03-13 2005-06-07 Micron Technology Inc. High permeability thin films and patterned thin films to reduce noise in high speed interconnections
US6906402B2 (en) 2002-03-13 2005-06-14 Micron Technology Inc. High permeability thin films and patterned thin films to reduce noise in high speed interconnections
US20030173653A1 (en) * 2002-03-13 2003-09-18 Micron Technology, Inc. High permeability thin films and patterned thin films to reduce noise in high speed interconnections
US6914278B2 (en) 2002-03-13 2005-07-05 Micron Technology Inc. High permeability thin films and patterned thin films to reduce noise in high speed interconnections
US7483286B2 (en) 2002-03-13 2009-01-27 Micron Technology, Inc. Semiconductor memory device with high permeability lines interposed between adjacent transmission lines
US20030176025A1 (en) * 2002-03-13 2003-09-18 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20030176027A1 (en) * 2002-03-13 2003-09-18 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20030176053A1 (en) * 2002-03-13 2003-09-18 Micron Technology, Inc. High permeability thin films and patterned thin films to reduce noise in high speed interconnections
US7391637B2 (en) 2002-03-13 2008-06-24 Micron Technology, Inc. Semiconductor memory device with high permeability composite films to reduce noise in high speed interconnects
US20060261448A1 (en) * 2002-03-13 2006-11-23 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20030176024A1 (en) * 2002-03-13 2003-09-18 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US7335968B2 (en) 2002-03-13 2008-02-26 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US20070029645A1 (en) * 2002-03-13 2007-02-08 Micron Technology, Inc. High permeability layered films to reduce noise in high speed interconnects
US6815804B2 (en) 2002-03-13 2004-11-09 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US7235457B2 (en) 2002-03-13 2007-06-26 Micron Technology, Inc. High permeability layered films to reduce noise in high speed interconnects
US7327016B2 (en) 2002-03-13 2008-02-05 Micron Technology, Inc. High permeability composite films to reduce noise in high speed interconnects
US6970053B2 (en) 2003-05-22 2005-11-29 Micron Technology, Inc. Atomic layer deposition (ALD) high permeability layered magnetic films to reduce noise in high speed interconnection
US20050140462A1 (en) * 2003-05-22 2005-06-30 Micron Technology, Inc. High permeability layered magnetic films to reduce noise in high speed interconnection
US20040233010A1 (en) * 2003-05-22 2004-11-25 Salman Akram Atomic layer deposition (ALD) high permeability layered magnetic films to reduce noise in high speed interconnection
US7154354B2 (en) 2003-05-22 2006-12-26 Micron Technology, Inc. High permeability layered magnetic films to reduce noise in high speed interconnection
US20060258187A1 (en) * 2005-05-16 2006-11-16 Teradyne, Inc. Impedance controlled via structure
US7492146B2 (en) * 2005-05-16 2009-02-17 Teradyne, Inc. Impedance controlled via structure
US8501563B2 (en) 2005-07-20 2013-08-06 Micron Technology, Inc. Devices with nanocrystals and methods of formation
US8921914B2 (en) 2005-07-20 2014-12-30 Micron Technology, Inc. Devices with nanocrystals and methods of formation
DE102011075333B4 (en) * 2011-05-05 2019-02-21 Rohde & Schwarz Gmbh & Co. Kg Amplifier with multi-purpose high frequency line
DE102011075333A1 (en) * 2011-05-05 2012-11-08 Rohde & Schwarz Gmbh & Co. Kg Amplifier has end terminal with contact element which is adapted to electrically contact spacer at predetermined position, where contact element is connected to ground terminal
US20120309456A1 (en) * 2011-05-31 2012-12-06 Renesas Mobile Corporation Semiconductor integrated circuit device, electronic device, and radio communication device
US9136801B2 (en) * 2011-05-31 2015-09-15 Renesas Electronics Corporation Semiconductor integrated circuit device, electronic device, and radio communication device
WO2015099918A1 (en) * 2013-12-27 2015-07-02 Intel Corporation High speed short reach input/output (i/o)
US10396840B2 (en) 2013-12-27 2019-08-27 Intel Corporation High speed short reach input/output (I/O)
US20180048094A1 (en) * 2014-05-08 2018-02-15 Apple Inc. Connector system impedance matching
US10103494B2 (en) * 2014-05-08 2018-10-16 Apple Inc. Connector system impedance matching
US9768774B2 (en) * 2014-06-30 2017-09-19 Fujitsu Limited Impedance matching driver
US20150381173A1 (en) * 2014-06-30 2015-12-31 Fujitsu Limited Impedance matching driver
US9407382B1 (en) * 2015-03-26 2016-08-02 Sprint Communications Company L.P. Positioning stub for mitigation of return loss during signal transmission
US20170054192A1 (en) * 2015-08-17 2017-02-23 Electronics And Telecommunications Research Institute High-frequency high-power terminator
KR20170021062A (en) * 2015-08-17 2017-02-27 한국전자통신연구원 High frequency high power termination
US9905899B2 (en) * 2015-08-17 2018-02-27 Electronics And Telecommunications Research Institute High-frequency high-power terminator
KR102365761B1 (en) * 2015-08-17 2022-02-23 한국전자통신연구원 High frequency high power termination

Also Published As

Publication number Publication date
US6278339B2 (en) 2001-08-21
US6249193B1 (en) 2001-06-19

Similar Documents

Publication Publication Date Title
US6249193B1 (en) Termination impedance independent system for impedance matching in high speed input-output chip interfacing
US6118310A (en) Digitally controlled output driver and method for impedance matching
US6147520A (en) Integrated circuit having controlled impedance
US6140885A (en) On-chip automatic system for impedance matching in very high speed input-output chip interfacing
US5021684A (en) Process, supply, temperature compensating CMOS output buffer
US8487663B2 (en) Driver circuit with EMI immunity
US6509779B2 (en) System for providing electrostatic discharge protection for high-speed integrated circuits
US6133749A (en) Variable impedance output driver circuit using analog biases to match driver output impedance to load input impedance
DE10101066A1 (en) Driver circuit, receive circuit and signal transmission bus system
EP1720306B1 (en) On-chip source termination in communication systems
JP4562211B2 (en) Output buffer with pre-driver to compensate slew rate for process variations
US7151298B1 (en) Electrostatic discharge protection network having distributed components
EP0637842B1 (en) LC element, semiconductor device and LC element manufacturing method
US7053670B2 (en) Semiconductor integrated circuit device and semiconductor integrated circuit
EP0705008A2 (en) Load termination sensing circuit
US6359465B1 (en) CMOS small signal switchable adjustable impedence terminator network
US20040189342A1 (en) Termination circuits having pull-down and pull-up circuits and related methods
US6373276B1 (en) CMOS small signal switchable impedence and voltage adjustable terminator with hysteresis receiver network
Hanafi et al. Design and characterization of a CMOS off-chip driver/receiver with reduced power-supply disturbance
NL9201779A (en) METHOD AND APPARATUS FOR A HIGH-SPEED OUTPUT BODY.
US5406215A (en) Open drain driver circuit which eliminates overshoot caused by parasitic capacitances
US5701071A (en) Systems for controlling power consumption in integrated circuits
EP0596637A1 (en) Input buffer circuit
US6310490B1 (en) CMOS small signal switchable, impedence and voltage adjustable terminator network and receiver integration
US6472927B1 (en) Circuit having increased noise immunity and capable of generating a reference voltage or terminating a transmission line

Legal Events

Date Code Title Description
STCF Information on status: patent grant

Free format text: PATENTED CASE

FPAY Fee payment

Year of fee payment: 4

FPAY Fee payment

Year of fee payment: 8

REMI Maintenance fee reminder mailed
FPAY Fee payment

Year of fee payment: 12

SULP Surcharge for late payment

Year of fee payment: 11