Classifications

• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03K—PULSE TECHNIQUE
  • H03K17/00—Electronic switching or gating, i.e. not by contact-making and –breaking
  • H03K17/002—Switching arrangements with several input- or output terminals
• • H—ELECTRICITY
  • H01—ELECTRIC ELEMENTS
  • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
  • H01L23/00—Details of semiconductor or other solid state devices
  • H01L23/52—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames
  • H01L23/522—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames including external interconnections consisting of a multilayer structure of conductive and insulating layers inseparably formed on the semiconductor body
  • H01L23/525—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames including external interconnections consisting of a multilayer structure of conductive and insulating layers inseparably formed on the semiconductor body with adaptable interconnections
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03K—PULSE TECHNIQUE
  • H03K17/00—Electronic switching or gating, i.e. not by contact-making and –breaking
  • H03K17/51—Electronic switching or gating, i.e. not by contact-making and –breaking characterised by the components used
  • H03K17/56—Electronic switching or gating, i.e. not by contact-making and –breaking characterised by the components used by the use, as active elements, of semiconductor devices
  • H03K17/687—Electronic switching or gating, i.e. not by contact-making and –breaking characterised by the components used by the use, as active elements, of semiconductor devices the devices being field-effect transistors
  • H03K17/693—Switching arrangements with several input- or output-terminals, e.g. multiplexers, distributors
• • H—ELECTRICITY
  • H01—ELECTRIC ELEMENTS
  • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
  • H01L2924/00—Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
  • H01L2924/0001—Technical content checked by a classifier
  • H01L2924/0002—Not covered by any one of groups H01L24/00, H01L24/00 and H01L2224/00

Definitions

• This invention relates to a universal interconnect matrix array suitable for programmably interconnecting circuit components and terminals. More particularly, this invention relates to a crosspoint switch containing programmable elements, such an antifuses or transistor switches, which allow one or more selected input or output pads of the switch to be connected to one or more of any other input or output pads of the switch.
• Crosspoint switches are well known.
• An article by Watson, published in EDN, February 16, 1989, pages 76-86, entitled “Crosspoint-Switch IC's Enter Digital Domain” describes the crosspoint switch art and the development of digital crosspoint switches.
• programmable crosspoint switches connect one of many input leads (also called “pads” or “terminals”) to one or more output leads.
• input leads also called “pads” or “terminals”
• crosspoint switches While initially crosspoint integrated circuit switches were developed to replace electro ⁇ mechanical switches such as relays in electrical telephone-switching systems, crosspoint switches are now used in parallel-processing, industrial-control-routing and data-communications systems.
• Crosspoint switches are of two types: analog and digital.
• An analog crosspoint switch is nothing more than a bi-directional transmission path which passes a selected signal substantially unchanged from an input terminal to an output terminal.
• a digital crosspoint switch is a system which will take a digital signal on one input terminal and provide a replica of that digital signal on one or more output terminals.
• all analog crosspoint switches have a number of basic elements: the switch array, an address decoder and a control memory or latches.
• N the number of input terminals
• M the number of output terminals
• the integrated circuit consists of NM distinct switches.
• a typical prior art analog switch i.e., a switch which transmits passively a signal on one input or output pad to another one or more input or output pads
• control signals must be applied to analog switch 0 and analog switch 1 to connect input pad X0 through analog switch 0 to the output pad Y0 and then connect the output pad Y0 through analog switch 1 to input pad XI. Accordingly, output pad Y0 is lost for other purposes. This inefficiency greatly restricts the usefulness of the analog switch of the type shown in Figure 2.
• U.S. patent 4,949,084 also discloses in Figure 1 a cross point circuit wherein each of the input pads in circuit "A" can be connected to one or more of the output pads in circuit "B". However, should an input pad in circuit A be desired to be connected to one or more other input pads in circuit A then an output pad must be dedicated to this function.
• crosspoint switches use decoding elements and/or multiplexers to form the desired connections. As the number of input terminals and output terminals increase, the complexity of these switches goes up extremely rapidly. Accordingly, these crosspoint switches are limited to a relatively small number of input terminals and output terminals. The total number of input and output terminals for these prior art crosspoint switches typically ranges from 8 to 128.
• a crosspoint switch of unique configuration which enables any one or more input or output pads on the switch to be connected to one or more of the remaining input or output pads on the switch without requiring the dedication of an additional input or output pad to this function.
• the programmable cross point switch of this invention is suitable for use in a variety of applications such as computers (including parallel processors) , industrial control systems, switching networks and programmable circuits, whether formed, for example, on printed circuit boards, insulating substrates or as an integrated circuit.
• the crosspoint switch (also called a "programmable interconnect chip” or a “universal interconnect matrix array”) is provided with horizontal and vertical conductive segments, programmable elements for interconnecting the horizontal and vertical conductive segments at any one or more of the plurality of intersections of said segments and conductive pads for connecting to circuitry external to the crosspoint switch.
• connection between selected pads is formed using horizontal and/or vertical conductive segments contrary to the prior art where decoders and multiplexers are used for this purpose.
• each track being capable of being broken into segments, in a channel associated with each row and column of input/output pads (i.e. by use of routing architecture to interconnect the selected pads) the complexity of the crosspoint switch is substantially reduced while the flexibility of the crosspoint switch is substantially increased.
• the programmable element in the crosspoint switch can be one time programmable antifuses and/or fuses or, alternatively, transistor switches capable of being reprogrammed as required either during or between operation of the circuitry of which the crosspoint switch is a part, should it be desired to reconfigure the circuitry interconnected by the crosspoint switch of this invention.
• Figures la and lb illustrate one embodiment of the crosspoint switch of this invention
• Figure 2a illustrates the architecture of one embodiment of the crosspoint switch of this invention claimed on a single integrated circuit
• Figure 2b illustrates one structure for programming the one-time programming element at the intersection of two conductive leads formed on the crosspoint switch of this invention.
• Figure 3 illustrates another embodiment of the crosspoint switch of this invention in one embodiment
• Figure 4 illustrates a prior art crosspoint switch wherein the programmable elements increase as a quadratic function of the total number of input pads and output pads to be interconnected
• Figures 5 and 6 illustrate the logic block diagram and circuit diagram respectively for the transistor programming elements used to interconnect a vertical and a horizontal conductive segment or two intersecting conductive segments in accordance with this invention.
• Figure 3 illustrates one embodiment of this invention.
• the crosspoint switch of this invention as formed on an integrated circuit chip, contains a plurality of conductive pads (1,1) (1,2) through (R,C) where R represents the maximum number of rows of pads and C represents the maximum number of columns of pads.
• These conductive pads are connected by conductive traces, such as conductive trace Tl,2 interconnecting pad 1,2 to vertical conductive segment VA.
• a plurality of horizontal conductive segments 1H1, 1H2, 1H3, 2H1, 2H2, for example run horizontally across the surface of the chip.
• These horizontal conductive segments are insulated from and intersected by a multiplicity of vertical conductive segments 1V1, 1V2...1VC.
• Programmable elements such as antifuses, fuses or programmable transistors, are depicted at the intersection of each vertical and horizontal conductive segment by circles.
• a circle with an "X" in it indicates a programmed interconnection between the overlying vertical lead and the underlying horizontal lead.
• An empty circle indicates an unprogrammed element.
• the pads (1,1) through (R,C) are intended to be connected to external circuitry.
• pad (1,2) is connected to either horizontal or vertical segments HA and VA or both. These segments allow pad (1,2) (and other pads capable of being connected to these segments) to be connected to any of the other routing tracks in the array.
• pad 1,3 can be connected by conductive traces to conductive segments HB and VB, for example.
• the conductive segments between the rows of pads comprise horizontal channels. While in accordance with this invention one horizontal channel is associated with each row of pads and one vertical channel is associated with each column of pads, each channel generally contains a plurality of conductive tracks.
• the horizontal conductive tracks typically contain segmented tracks. That is, one or more of these tracks is broken into two or more conductive segments.
• the number of horizontal tracks in each horizontal channel can be set at any appropriate number "h".
• the vertical tracks are grouped in channels with each channel consisting of the tracks between two columns of conductive pads.
• one vertical channel consists of vertical conductive segments 1V1, 1V2 and VA.
• the second vertical channel consists of vertical conductive tracks 2V1, 2V2 and VB, for example.
• One or more vertical tracks in each vertical channel is also broken into two or more conductive segments. The number of vertical tracks in each vertical channel is represented as "v" for example.
• the programming elements are designated by the open circles at the intersection of the horizontal and vertical segments.
• the number of horizontal conductive tracks in each horizontal channel equals the number of vertical conductive tracks in each vertical channel.
• the total number of programming elements is proportional to the number of horizontal channels N H times the number of vertical channels N v times the number "h” of conductive tracks in each horizontal channel times the number "v” of conductive tracks in each vertical channel. Therefore the number of programming elements Q is given by the following equation: Q « (h) (v) (N H N V ) .
• the total number Y of output pins from the chip is given by:
• the programming elements can be one time programmable antifuses and/or fuses, or reprogrammable transistor switches. As described below, memory cells such as RAM cells can be used to control the setting of the switches.
• the lengths and numbers of the horizontal and vertical conductive segments are optimized to provide the best statistical coverage of the random connections required between the pads (1,1) through (R,C) to implement specific circuits.
• the segmented tracks as described above provide more efficient utilization of the tracks.
• the numbers h and v representing, respectively, the number of horizontal conductive tracks in a horizontal channel and the number of vertical conductive tracks in a vertical channel, depend weakly on the number of horizontal channels N H and the number of vertical channels N v .
• each conductive pad such as pad 1,1 can be connected to any one of the other pads in the array by programming one or more programmable elements.
• decoders In one prior art approach for implementing cross point switches which employed a hierarchy of decoders and multiplexers connecting one of a group of conductive pads to any one of another group of conductive pads, the decoders require selection bits stored in RAM cells to implement the connection of any one of the group A terminals to any one of the group B terminals (See Figure 4) .
• decoder l,C is configured by bits stored in RAM cell 2.
• decoder 1,1 is also configured by bits stored in a RAM cell as are decoders R,l and R,C.
• decoder 1,C To connect, for example, input pad Al in the group A terminals to pad Bl in the group B terminals, decoder 1,C must produce a high output signal thereby to gate the signal on terminal Al to terminal Bl through gate G. All other decoders will have a low input signal, thereby preventing these decoders from enabling the gates to which the output signals from these decoders are transmitted.
• each gate G shown in Figure 4 can be a simple transistor switch (such as transistor Nl shown, for example, in Figure 5.)
• the gate can also provide a digital connection using equivalent input buffers for A terminals and output buffers for B terminals.
• the number of transistors/circuits required to implement the decoding architecture shown in Figure 4 is proportional to the number of input pads multiplied by the number of output pads. If there are M group A terminals and N group B terminals, the total number of terminals is N+M and the number of transistor circuits T required to implement this decoding structure is:
• k represents the number of transistors/circuits used in the gates G
• I represents the number of transistors/circuits of the RAM cells and the decoders.
• N the total number of transistors required to implement this scheme is: T « N 2 [k + I x 2lnM] .
• the terminals in group A cannot be connected to each other.
• the terminals in group B cannot be connected to each other. Only connections between one or more terminals in group A and one or more terminals in group B are possible.
• the number of transistors/circuits increases almost quadradically with the number of pins to be connected. Therefore the architecture is limited to a small number of pins such as 64 input pins and 64 output pins, for a total of 128 pins or some lesser number. A larger number of pins would require a very large increase in the array area and the chip area.
• Figures la and lb illustrate a crosspoint switch formed on an integrated circuit chip 605 in accordance with this invention.
• chip 605 contains a plurality of cells 606-1,1 through 606-S,T where S represents the number of rows of cells in the chip 605 and T represents the number of columns of cells in chip 605.
• Each cell has an array of electrically conductive pads 607-1,1 through 607-M,N where M represents the number of rows of pads in the cell and N represents the number of columns of pads in the cell.
• each cell is identical in configuration, only the conductive pads 607 associated with cell 606-1,1 will be described in detail with the understanding that the conductive pads associated with each of the other cells 606-s,t (where s is an integer given by 1 ⁇ s ⁇ S and t is an integer given by l ⁇ t ⁇ T) in chip 605 function identically.
• Figure lb illustrates the configuration of cell 606-1,1 and also of each of the other cells 606-s,t in Figure la.
• horizontal conductive tracks 608-1 through 608-J (where J is an integer representing the maximum number of horizontal conductive tracks formed on chip 605) are shown.
• vertical conductive tracks 609-1 through 609-K are shown where K is an integer representing the maximum number of columns of conductive tracks formed on chip 605.
• the horizontal conductive tracks 608-1 through 608-J are formed on one level of interconnections on chip 605 while the vertical conductive tracks 609-1 through 609-K are formed on a second level of interconnections on chip 605.
• the horizontal conductive leads 608-1 through 608-J have differing lengths across the chip.
• the cell 606-1,1 shown in the upper left hand corner of both Figure la and Figure lb has a plurality of horizontal conductive leads 608 originating in and extending from cell 606-1,1 to each of the other cells 606-1,2 through 606-1,T in the same row.
• cell 606-1,1 has a plurality of vertical conductive leads 609 extending from cell 606-1,1 to each of the other cells 606-2,1 to 606-S,l in the same vertical column.
• the horizontal and vertical traces 608 and 609 have at each of their intersections a programmable connective structure (i.e. programmable elements) such as for example, an antifuse and/or a fuse, or a programmable and reprogrammable transistor.
• a programmable connective structure i.e. programmable elements
• an antifuse comprises a capacitive structure with a dielectric capable of being broken down by the application of a selected voltage to provide a conductive path between the two plates of the capacitor.
• Antifuses are well known in the art and thus will not be described in detail.
• the substrate of the chip 605 may have in it selected circuitry to enable the programming of the programmable elements at selected intersections in accordance with design requirements.
• vertical leads 609-1 through 609-K are formed on the programmable interconnect chip 605 so as to extend at a minimum across one cell 606 and at a maximum across all cells.
• a plurality of vertical leads 609 cross each cell with the length of leads varying from being such as to extend across just that cell to being such as to extend across all cells in a column.
• Horizontal conductive leads 608-1 through 608-J likewise extend across the chip 605.
• the horizontal leads 608 extending across one cell vary from a length such that they extend across only that one cell up to a length which will extend across all cells.
• breaklines are included to indicate that the semiconductor chip 605 is only partially shown with interior portions of the chip having been removed for clarity.
• a horizontal conductive lead thus might comprise one conductive segment extending across the whole chip 605 or a plurality of conductive segments extending across a section of the chip.
• the vertical conductive leads likewise vary from one conductive lead which will extend across the entire height of the chip or two or more conductive segments each extending across a selected portion of the chip.
• the particular configuration of the conductive leads extending across one cell and from that cell to adjacent cells depends upon an analysis of the electrical functions to be carried out by the programmable interconnect matrix array and is selected using the most probable types of system requirements to be imposed on programmable interconnect chip 605. This selection depends upon an analysis of the circuit functions to be performed by the circuit using the crosspoint switch of this invention and thus the actual configuration of the crosspoint switch is determined in light of the proposed uses.
• an interconnection between the appropriate vertical conductor 609 and the appropriate horizontal conductor 608 is formed.
• pad A to pad B both in cell 606-1,1
• the intersection of vertical lead 609-1 and horizontal lead 608-1 is programmed.
• antifuses are used as the programmable elements a high voltage is applied to this intersection in the circuit so as to break down the dielectric between these two points and form a conductive path therebetween.
• Figure lb also illustrates the particular connections which must be formed to connect pad A to pad D, pad A to pad C or pad A to pad E. Should all of these connections be made then pads B, D, C and E will also be connected to each other through pad A.
• FIG. 2a shows in block diagram form one architecture of the crosspoint switch 605 of this invention.
• the interior 605A of chip 605 contains the cells 606 (as described in conjunction with Figures la and lb) and the horizontal and vertical tracks 608 and 609 respectively.
• peripheral area 605B which forms an annular square around interior 605A are placed control and programming circuits including shift registers for selecting particular horizontal and vertical tracks the intersections of which are to be programmed.
• buffer circuitry for the testport bus and the control port bus are provided in this region of chip 605.
• Annular region 605C surrounds annular region 605B and contains additional circuitry essential to the operation of the chip such as mode selection circuitry which will determine whether the programmable interconnect chip is in the test mode, the operating mode or the programming mode. Additional special circuitry as required will also be placed in peripheral region 605C.
• Figure 2b illustrates a programming structure, in the case antifuses are used as the programmable elements, and particularly programming transistors and circuits to select the intersection to be programmed of horizontal and vertical conductive leads on the chip using only two transistors in the programming circuit path. Utilization of the structure shown in Figure 2b allows the programming current to reach into the hundreds of milliamps to amperes range necessary to break down the dielectric between the vertical and horizontal conductive leads to form an interconnection therebetween with sufficiently low resistance.
• transistors Ql and Q2 are provided to program the intersection of vertical conductive track V and horizontal conductive track H 1 .
• Transistor Ql has its gate connected to voltage source VGP and transistor Q2 has its gate connected to a voltage source HGP...
• the source of transistor Ql is connected to vertical conductive track VI while the drain of transistor Ql is connected to conductive lead VDP .
• the source of transistor Q2 is connected to horizontal lead H. and the drain of transistor Q2 is connected to conductive lead HDP .
• VGP. is applied to take the gate of Ql to a high voltage V GH , the gates of other transistors in the array such as transistor Q3 are held at zero volts and the drain voltage VDP1 on transistor Ql is taken to V pp .
• the gate voltage of Q4 is taken high because HGP is taken to a high voltage to turn on transistor Q2.
• the voltage on the drain of Q2 is taken to zero volts by driving HDP to zero and HDP 2 applied to the drain of Q4, is taken to zero or to V pp /2 (which voltage is selected so as not to program the programming element at the intersection of V 2 and H_.
• V pp the programming voltage, is typically 15 to 100 volts.
• V GH which is applied to lead VGP- , is larger than V pp by the transistor threshold voltage and thus is approximately 18 to 103 volts. Because the devices Ql to Q4 operate under high voltage, the threshold voltage of these transistors is made approximately three (3) volts. As a result of the above- described voltages, only the programming element at the intersection of conductive lead segments HI and VI will receive the full programming
• the programming element be an antifuse and/or a fuse, or a programmable transistor.
• the programming element is a transistor (which is used as a switch) the state of this transistor is stored in memory such as, for example, a static RAM cell.
• the RAM cell comprising cross-coupled inverter 530A and 530B is programmed to be in a high or low state by data in transmitted through AND gate 510 enabled by write signal going high.
• the state of the RAM cell can be read through AND gate 520 by applying a high level read signal to enable input of AND gate 520.
• the output signal from AND gate 510 is connected to one node of the RAM cell 530 while one input lead of AND gate 520 is also connected to this same node.
• the other node of RAM cell 530 is connected to one input lead of AND gate 540, the other input lead to which is connected to an enable signal.
• AND gate 540 will pass the state of RAM cell 530 to the gate of programming element comprising NMOS transistor Nl. If the output signal from AND gate 540 is high, transistor Nl turns on and provides a conductive path from the pad to the track segment. If the output signal of AND gate 540 is low, transistor Nl remains off, thereby providing an open circuit between the pad and the track segment.
• Figure 6 illustrates the specific circuitry of the structure shown in logic block diagram form in Figure 5.
• AND gate 510 of Figure 5 is shown as N-type pass transistor N2 with a write enable signal applied to the gate of transistor N2.
• a signal data in is passed through transistor N2 to node A of RAM cell 530.
• RAM cell 530 consists of P-type transistor PI, the drain of which is connected to a power supply and the source of which is connected to the drain of N-type transistor N4.
• the source of N4 is connected to a reference voltage, shown as ground.
• the gates of transistors PI and N4 are connected together and also connected to the node B connected to the source of P-type transistor P2 in the drain of N-type transistor N5.
• the RAM cell 530 consists of P-type transistors PI, P2 and N- type transistors N4 and N5.
• P-type transistor PI When the voltage on node B is high, P-type transistor PI is off, N-type transistor N4 is on, thereby providing a low voltage on node A.
• the voltage on node A is connected to the gates of P-type transistor P2 and N-type transistor N5, thereby turning on P-type transistor P2 and turning off N-type transistor N5, thus insuring that the voltage on node B is high.
• the high voltage on node B is transmitted to one terminal of pass transistor N6, the gate of which is enabled by a high level enable signal.
• transistor N6 When transistor N6 is enabled, the high voltage on node B is transmitted to the gate of pass transistor N7, thereby turning on transistor N7 to connect the pad to the conductive trace.
• the gate of transistor N7 can be bootstrapped by the voltage swing on the track segment connected to the source of N-type transistor N7, provided the enable gate N6 is off, thereby to isolate the gate of N7.
• RPIC RPIC
• RPIC RPIC
• the programming data to program the crosspoint switch is retained in memory, typically a conventional SRAM.
• the reprogrammable interconnect component of this invention is programmed.
• the reprogrammable interconnect component of this invention can be used with interfaces where the ability to dynamically reassign interconnect paths is highly useful. These applications include custom control panels, displays, data and voice communications, data acquisition systems, test equipment, test interfaces, parallel processors and systolic arrays.
• ASICs application specific integrated circuits
• the universal crosspoint switches of this invention may also be used to create products where a portion of the circuitry may be customized to the needs of a particular user or class of users.
• the universal crosspoint switch of this invention is particularly suitable for implementation in the form of an integrated circuit packaged a pin grid array package or a QFP package.
• Critical paths may be defined prior to device configuration to provide a lower resistance and capacitance for the resulting circuitry. Incremental changes can then be made in the circuit and connections can be programmed to be either "make-before-break” or "break-before-make”. This provides flexibility in communications and other applications where conductive links may optionally be maintained while new connections are being formed or may be broken before new connections are formed.

Landscapes

• Physics & Mathematics (AREA)
• Condensed Matter Physics & Semiconductors (AREA)
• General Physics & Mathematics (AREA)
• Engineering & Computer Science (AREA)
• Computer Hardware Design (AREA)
• Microelectronics & Electronic Packaging (AREA)
• Power Engineering (AREA)
• Design And Manufacture Of Integrated Circuits (AREA)

Applications Claiming Priority (3)

Publications (2)

Family

ID=25070187

Family Applications (1)

Country Status (4)

Families Citing this family (4)

Citations (3)

Family Cites Families (11)

• 1992
  • 1992-09-23 EP EP92920784A patent/EP0609264A4/en not_active Withdrawn
  • 1992-09-23 JP JP5506365A patent/JPH09506481A/ja active Pending
  • 1992-09-23 WO PCT/US1992/008115 patent/WO1993006559A1/fr not_active Application Discontinuation
  • 1992-09-23 AU AU26793/92A patent/AU2679392A/en not_active Abandoned

Patent Citations (3)

Non-Patent Citations (2)

Also Published As

Similar Documents

Legal Events