WO1997040579A1 - Radiation resistant logic circuit - Google Patents

Radiation resistant logic circuit Download PDF

Info

Publication number
WO1997040579A1
WO1997040579A1 PCT/US1997/006642 US9706642W WO9740579A1 WO 1997040579 A1 WO1997040579 A1 WO 1997040579A1 US 9706642 W US9706642 W US 9706642W WO 9740579 A1 WO9740579 A1 WO 9740579A1
Authority
WO
WIPO (PCT)
Prior art keywords
logic
output signals
signals
section
output
Prior art date
Application number
PCT/US1997/006642
Other languages
French (fr)
Inventor
Michael D. Lahey
Debra S. Harris
Harry N. Gardner
Michael J. Barry
Original Assignee
United Technologies Corporation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by United Technologies Corporation filed Critical United Technologies Corporation
Publication of WO1997040579A1 publication Critical patent/WO1997040579A1/en

Links

Classifications

    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03KPULSE TECHNIQUE
    • H03K19/00Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits
    • H03K19/003Modifications for increasing the reliability for protection
    • H03K19/0033Radiation hardening
    • H03K19/00338In field effect transistor circuits

Definitions

  • This invention relates to semiconductor electronic circuits and more particularly to semiconductor logic circuits.
  • a single event upset is the result of an ion transitioning through a semiconductor structure and depositing charge on a critical circuit node within that structure.
  • SEU single event upset
  • CMOS logic circuit In a CMOS logic circuit, this can cause an unintended switch in the logic state, creating potentially catastrophic consequences for the system.
  • the primary SEU problem lies in the feedback path, where amplification and feedback of noise on a critical node can permanently change the cell's logic state.
  • CMOS logic include the use of redundant circuit paths. Dual circuit paths provide redundancy and allow implementation of voting schemes to reduce the effect of SEUs. However, the typical voting scheme uses digital logic to recombine the redundant paths, and thereby actually provides amplification of the SEU.
  • An object of the present invention is to provide a logic circuit architecture which has higher reliability in a radiation environment. Still another object of the present invention is to provide circuitry for a family of logic products having higher reliabihty in a radiation environment.
  • apparatus for performing a selected logic function includes a first logic portion which receives one or more logic state input signals and performs the selected logic function on the logic state input signals to provide one or more output signals, and further includes a second logic portion which receives the one or more logic state input signals and performs the selected logic function on the logic state input signals to provide one or more output signals, and further combines each output signal from the first logic portion with the corresponding output signal from the second logic portion to provide one or more resultant logic output signals.
  • apparatus for performing a selected logic function includes an input section which receives one or more logic state input signals and provides a pair of buffered logic state signals for each, and further includes, a logic section having a first and second portion, each receiving one logic state signal from each of the pairs of buffered logic state signals, and each perforrning the selected logic function on the logic state signals received thereby to each provide one or more output signals, and each of said output signals from the first portion combined in analog fashion with a corresponding one of the output signals from the second portion to provide one or more logic section output signals, and further includes an output section receiving each of the one or more logic section output signals and providing one or more resultant logic output signals.
  • the present invention provides a combinatorial logic circuit having higher reliability in a radiation environment
  • the present invention further provides circuitry for a family of products having higher reUability in a radiation environment with little impact to the die size and no decrease in functionality.
  • FIG. 1 is a schematic illustration of a NAND logic function, which is used in the description of the present invention
  • FIG. 2 is a schematic illustration of the best mode embodiment of the present invention as used in the NAND logic function of FIG. 1;
  • FIG. 3 is a schematic illustration of one element used in the embodiment of Fig. 2;
  • FIG. 4 is a schematic illustration of another element used in the embodiment of Fig. 2. Best Mode for Carrying out the Invention
  • a two input NAND circuit 20 receives logic signal A on line 22 and logic signal B on line 24, and produces a NAND logic output signal Y on line 26.
  • logic signal A on line 22 and logic signal B on line 24 are each received by an input section 28.
  • Each of the logic signals on lines 22, 24 are presented to an associated one of a pair of signal conditioning circuits 29a, 29b within the input section 28.
  • the signal conditioning circuits 29a, 29b are identical.
  • the circuit 29a has a diode 30 which provides protection against logic signal A magnitudes less than the lower power supply voltage magnitude V S s, and a diode 32 which provides protection against logic signal A magnitudes greater than the upper power supply voltage magnitude VQ D -
  • V S s is approximately zero (0) volts and V D D is in the range of from 4.5 to 5.5 volts.
  • the diodes 30, 32 are physically located as close to where the input signal on line 22 is received, as is practical. In situations where the input signal A on line 22 cannot exceed the magnitude range of the power supplies, such as where the source of logic signal A resides on the same integrated circuit as the NAND circuit 20, the two diodes 30, 32 may not be necessary.
  • the circuit 29a further comprises a series resistor 33 that is electrically connected between the diodes 30, 32 and the signal inputs a pair of buffer stages 34, 36.
  • the resistor 33 provides electrostatic discharge protection (ESD) on line 22.
  • ESD electrostatic discharge protection
  • the buffer stages 34, 36 each receive the ESD protected signal and provide a pair of buffered, redundant A signals on lines 38, 40 respectively.
  • the pair of buffer stages 34, 36 represent independent signal paths, such that, a failure in one of the pair of buffer stages 34, 36 will not disturb the logic signal out of the other.
  • the pair of buffer stages 34, 36 are physically spaced further apart than normal so that the effect of a single ion transitioning through the semiconductor structure is less likely to cause single event upset (SEU) in both of the pair of buffer stages 34, 36.
  • SEU single event upset
  • the pair of the buffer stages 34, 36 provide the same logic function, non inverting buffering, and comprise logic gates of the same type, so that the pair of buffered A signals on lines 38, 40 are substantially similar to each other in logic state and timing.
  • one or both of the pair of buffer stages 34, 36 may provide a logic function other than non inverting buffering, so long as suitable modifications are made to other logic, described below, in the NAND circuit 20.
  • each of the buffer stages 34 and 36 comprise cascaded first and second inverter 42, 44 and 46, 48, respectively.
  • the inverters are shown to be identical.
  • the first inverters 42, 46 buffer and invert the logic signal .
  • the second inverters 44, 48 restore the logic states of the pair of buffered A signals on lines 38, 40 to that of the input signal A on line 22.
  • Fig. 3 is a circuit for the inverter 42, which comprises a P channel MOSFET 42a and an N channel MOSFET 42b.
  • the input pin (IN) of the first inverter 42 is electrically connected to the gate 42c of the P channel and the gate 42d of the N channel MOSFETs.
  • the source 42e of the P channel is connected to V DD and the source 42f ofthe N channel is connected to Vss-
  • the drain 42g of the P channel MOSFET is connected to the drain 42h of the N channel which then serves as the output (OUT) for the first inverter 42.
  • the other inverters 44, 46, 48 have similar transistor level circuits to that of the first inverter 42.
  • the operation of the first inverter 42 is as follows. When the signal input at IN is at a high logic state voltage, the N channel MOSFET 42b is "on", i.e. its resistance from drain to source is low, and the P channel MOSFET 42a is "off', i.e. its resistance from drain to source is high.
  • the B input logic signal on line 24 is received by signal conditioning circuitry 29b which is similar to that described with respect to circuit 29a.
  • Two diodes 50, 52 provide under voltage and over voltage protection. In situations where the input signal B on line 24 can not exceed the range of the power supplies the two diodes 50, 52 may not be necessary.
  • a series resistor 53 is positioned after the two diodes 50, 52 to provide electrostatic discharge protection (ESD) on the input signal B on line 24.
  • ESD protected input signal is then passed to a pair of buffer stages 54, 56 which provide a pair of buffered B signals on lines 58, 60 respectively.
  • a failure in one of the pair of buffer stages 54, 56 will not disturb the logic signal out of the other.
  • the pair of buffer stages 54, 56 are physically spaced further apart than normal so that the effect of a single ion is less likely to cause single event upset (SEU) in both of the pair of buffer stages 54, 56.
  • SEU single event upset
  • the pair of buffer stages 54, 56 provide the same logic function, non inverting buffering, and comprise logic gates of the same type, so that the pair of buffered B signals on lines 58, 60 are substantially similar to each other in logic state and timing.
  • one or both of the pair of buffer stages 54, 56 may provide a logic function other than non inverting buffering, so long as suitable modifications are made to other logic, described below, in the NAND circuit 20.
  • the NAND circuit 20 further comprises a logic section 70 having a first portion 72 and a second portion 74. The first and second portions 72, 74 each receives one signal from each of the buffered A and B signal pairs on lines 38, 40 and 58,60.
  • the first and second portions 72, 74 represent independent signal paths, such that, a failure in one will not disturb the logic of the other. To this end, in the best mode embodiment, the portions 72, 74 are physically spaced further apart than normal so that the effect of a single ion transitioning through the semiconductor structure is less likely to cause single event upset (SEU) in both portions 72, 74.
  • SEU single event upset
  • Each portion 72, 74 comprises the logic necessary to enable it to provide the overall logic function of the circuit, e.g. the NAND logic function of the NAND circuit 20 in the best mode embodiment.
  • the appropriate logic depends not only on the overall logic function of the circuit, but also on the logical effect of the input section 28 buffer stages 34, 36, 54, 56 and the output section, described hereinbelow.
  • the input section 28 buffer stages 34, 36, 54, 56 and the output section are non inverting buffers, and therefore have no logical effect. Consequently, the logic for the first and second portions 72, 74 of the logic section 70 each need only comprise a NAND gate 76, 78, or its equivalent, to provide the NAND function.
  • FIG. 4 A schematic representation of the transistor level circuit for the NAND gate 76 of the first portion 72 is shown in Fig. 4.
  • the NAND gate 76 has a first input 76a which is electrically connected to the gates 76b, 76c of a first pair 76d of P channel MOSFETs.
  • the first pair 76d of P channel MOSFETs have their sources 76e electrically connected to V DD and their drains 76f electrically connected to the drains 76g of a second pair 76h of P channel MOSFETs.
  • the sources 76i of the second pair 76h of P channel MOSFETs are connected to V DD and the gates 76j, 76k are connected to the second input signal 761.
  • the drains 76f, 76g of both pairs 76d, 76h of P channel MOSFETs are further connected to the drains 76m of a first pair 76n of N channel MOSFETs and the output 76p of the NAND gate 76.
  • the first pair 76n of N channel MOSFETs have their gates 76q, 76r connected to the second input 761.
  • Each of the sources 76s of the first pair 76n of N channel MOSFETs is connected to a different one of the drains 76t of a second pair 76u of N channel MOSFETs.
  • the second pair 76u of N channel MOSFETs have their gates 76v, 76 w 6642
  • the operation of the NAND gate 76 is as follows. When both of the inputs are at a high logic state voltage, both pairs 76n, 76s of N channel MOSFETs are "on", i.e. low resistance, and both pairs 76d, 76h of P channel MOSFETs are "off, i.e. high resistance. At the output 76p, this produces a high resistance to the V DD supply and a low resistance to the Vss supply. The result is an output voltage of approximately V S s, representing a low logic state.
  • the NAND gate 78 of the second portion 74 has a similar transistor level circuit schematic (not shown) and operation to that of the NAND gate 76 of the first portion 72.
  • the first and second portions 72, 74 each operate on their respective input signals and produce first and second portion output signals which are summed, or combined, in analog fashion to produce a logic section output signal on line 84.
  • the logic section output signal on line 84 takes on the logic state corresponding to the voltage of the analog combination. This is in contrast to a combination that combines the signals by perfoirriing a logical operation on each of their logic states.
  • the transistors in the NAND gates 76, 78 have "on" resistances in the range of from about 500 ohms to 5,000 ohms. Lower resistances are not required because the NAND gates do not to provide high magnitude output current. Accordingly, the connection should have an electrical resistance of about less than 500 ohms, preferably less than 10 ohms. In the best mode embodiment, a direct electrical connection having a resistance below / ohm, specifically about .01 ohms, is used. However, those skilled in the art will recognize that other suitable transistors having other "on" resistances , and other connections having other resistances may also be used.
  • the transistors of both of the NAND gates 76, 78 are in similar "on" and “off” states. Consequently, the voltage contributions from the first and second portions to the logic section output signal on line 84 are both approximately V S s, or both approximately V DD , and the combined signal, the logic section output signal on line 84, is also approximately Vss, corresponding to a logic low, or V DD , corresponding to a logic high. Statistically, only one transistor will experience SEU at a given time. Thus when an ion transitions through the semiconductor one portion 72, 74 of the logic section 70, may experience SEU in one of the transistors in its NAND gate 76, 78.
  • That transistor may change state from “on” to “off” or “off” to “on”.
  • any change in logic section output signal voltage on line 84 is small, since the output from the affected portion is combined, in analog fashion, as described above, with the output from the unaffected portion. n Consequently, the voltage effect of the SEU is reduced, or attenuated.
  • At or below the designed ion energy level of 80 MeVcmVmg there is insufficient energy in the ion to cause the logic section output signal on line 84 to transition to an incorrect logic state.
  • the logic section output signal on line 84 remains at the correct logic state throughout the SEU.
  • the typical SEU lasts for about 20 nano seconds after which the transistor experiencing SEU returns to its prior state. Capacitance in the output portion, described hereinbelow, provides additional filtering for the small voltage changes in the logic section output signal on line 84.
  • the summation approach of the present invention does not operate as a standard logic function.
  • the summation here should not be used for combining logic signals which are not substantially similar in state and tuning for the very reason that it does not provide a normal logical combination. Instead, it provides a means for combming substantially similar signals wherein a SEU failure in one of the signals does not cause failure of the combined signal.
  • the logic section output signal on line 84 is passed to an output section 86. Nodal capacitance, not shown, on line 84, provides some additional filtering for the logic section output signal.
  • the output section 86 further comprises a pair of inverters 88, 90 for buffering of the signal.
  • the pair of inverters 88, 90 are similar to that of the first inverter 42 of the A input signal buffer stage 34 with the exception that the transistors sizes are larger to reduce the likelihood that they themselves will experience SEU. Larger and stronger transistors are better able to resist SEU by counting the collection of charge through the substrate that is brought on by the entering ion.
  • the pair of inverters 88, 90 are also larger to provide the desired capability for driving additional logic circuits that may be connected to the output logic signal Y on line 26.
  • the output section further includes a pair of diodes 92, 94 to protect against under and over voltage damage. As with all circuitry, appropriate decoupling should be used when employing the present invention.
  • the NAND circuit 20 is implemented in CMOS technology, preferably 1.2 micro meters (um) or smaller, however, any other suitable technology may also be used. At 1.2 um the circuit is small enough in size, has relatively low power, and provides high speed operation.
  • each of the logic signals on lines 22, 24, 26 are of the type having CMOS logic levels, however, as will be obvious to those skilled in the art, the NAND logic circuit 20 may also be designed to accommodate TTL logic signals. In the preferred embodiment, this only requires threshold modifications in the input section.
  • the best mode of the present invention creates the logic section output signal on line 84 by combuiing the outputs from NAND gates 76, 78, those skilled in the art will recognize that other transistor configurations may also be suitable. For example, transistor configurations to be connected may also be of the type used in inverters. While the best mode of the present invention includes an input section, the logic section of the present invention may also be used without the input section or portions of the input section. For example, each input logic signal may be routed directly to both portions of the logic section, without passing through the input section, so long as the logic section is suitably adapted for operation with the logic levels of the input logic signals. The present invention may also be used without the output section of the best mode embodiment.
  • the logic section output signal may be used to drive other logic directly so long as the logic section is suitably adapted for driving such other logic.
  • the present invention may be used in logic circuits of any type, including but not limited to that of OR, NOR, XOR, XNOR, inverter, AND, buffer, flip flop, line drivers and transceivers, with or without three-state outputs.
  • the NAND circuit of the present invention has 2 inputs and 1 output, the present invention ;may be used in logic circuits having any number of inputs, e.g. N, where N is one or more, and any number of outputs, e.g. K, where K is one or more.
  • the input section still preferably receives each of the N input signals and produces a pair of buffered signals for each.
  • One signal from each pair of buffered signals is received by the first portion of the logic section and the other signal from each pair is received by the second portion.
  • Each of the two logic section portions has all of the necessary logic gates for implementing the overall function.
  • the first portion produces K first portion output signals and the second portion produces K second portion output signals.
  • Each of the K first portion output signals is combined with the counterpart signal from the K second portion output signals to produce K logic section output signals.
  • the K logic section output signals are passed to the output section where they are buffered to produce the K outputs for the circuit.
  • the present invention may also be used in logic circuits of the type including but not limited to comparators, flip flops, decoders/demultiplexers, selectors/multiplexers, counters, shift registers, parity generators, adders, and memories.

Abstract

Apparatus for performing a selected logic function includes a first logic portion which receives one or more logic state input signals and performs the selected logic function on the input signals to provide one or more output signals, and further includes a second logic portion which receives the one or more input signals and performs the selected logic function on the input signals to provide one or more output signals, and combines each output signal from the first logic portion with the corresponding output signal from the second logic portion to provide one or more resultant logic output signals which have improved reliability in a radiation environment.

Description

Description
RADIATION RESISTANT LOGIC CIRCUIT
Technical Field
This invention relates to semiconductor electronic circuits and more particularly to semiconductor logic circuits.
Background Art
A single event upset (SEU) is the result of an ion transitioning through a semiconductor structure and depositing charge on a critical circuit node within that structure. In a CMOS logic circuit, this can cause an unintended switch in the logic state, creating potentially catastrophic consequences for the system. In the case of storage cells, the primary SEU problem lies in the feedback path, where amplification and feedback of noise on a critical node can permanently change the cell's logic state. Known SEU hardening techniques for CMOS logic include the use of redundant circuit paths. Dual circuit paths provide redundancy and allow implementation of voting schemes to reduce the effect of SEUs. However, the typical voting scheme uses digital logic to recombine the redundant paths, and thereby actually provides amplification of the SEU. T/US97/06642
Disclosure of Invention
An object of the present invention is to provide a logic circuit architecture which has higher reliability in a radiation environment. Still another object of the present invention is to provide circuitry for a family of logic products having higher reliabihty in a radiation environment.
According to the present invention, apparatus for performing a selected logic function includes a first logic portion which receives one or more logic state input signals and performs the selected logic function on the logic state input signals to provide one or more output signals, and further includes a second logic portion which receives the one or more logic state input signals and performs the selected logic function on the logic state input signals to provide one or more output signals, and further combines each output signal from the first logic portion with the corresponding output signal from the second logic portion to provide one or more resultant logic output signals.
In fijrther accord with the present invention, apparatus for performing a selected logic function includes an input section which receives one or more logic state input signals and provides a pair of buffered logic state signals for each, and further includes, a logic section having a first and second portion, each receiving one logic state signal from each of the pairs of buffered logic state signals, and each perforrning the selected logic function on the logic state signals received thereby to each provide one or more output signals, and each of said output signals from the first portion combined in analog fashion with a corresponding one of the output signals from the second portion to provide one or more logic section output signals, and further includes an output section receiving each of the one or more logic section output signals and providing one or more resultant logic output signals. The present invention provides a combinatorial logic circuit having higher reliability in a radiation environment The present invention further provides circuitry for a family of products having higher reUability in a radiation environment with little impact to the die size and no decrease in functionality. These and other objects, features, and advantages of the present invention will become apparent in light of the following detailed description of a best mode embodiment, accompanying Drawings, and claims.
Brief Description of the Drawings FIG. 1 is a schematic illustration of a NAND logic function, which is used in the description of the present invention;
FIG. 2 is a schematic illustration of the best mode embodiment of the present invention as used in the NAND logic function of FIG. 1;
FIG. 3 is a schematic illustration of one element used in the embodiment of Fig. 2; and
FIG. 4 is a schematic illustration of another element used in the embodiment of Fig. 2. Best Mode for Carrying out the Invention
The apparatus of the present invention is disclosed with respect to a best mode embodiment for use in a logic circuit providing the NAND logic function represented in FIG. 1. As should be understood by those skilled in the art, the drawing figures are only schematic representations, and are not intended to portray physical details.
Referring to FIG. 1, in a prior art schematic of a selected logic device, a two input NAND circuit 20 receives logic signal A on line 22 and logic signal B on line 24, and produces a NAND logic output signal Y on line 26. As known, the logical operation of the NAND circuit 20 can be described by the equation Y = NOT(A*B).
Referring now to FIG. 2, in the present invention's best mode embodiment of the NAND circuit 20 of FIG. 1, logic signal A on line 22 and logic signal B on line 24 are each received by an input section 28.
Each of the logic signals on lines 22, 24 are presented to an associated one of a pair of signal conditioning circuits 29a, 29b within the input section 28. As described hereinafter, the signal conditioning circuits 29a, 29b are identical. The circuit 29a has a diode 30 which provides protection against logic signal A magnitudes less than the lower power supply voltage magnitude VSs, and a diode 32 which provides protection against logic signal A magnitudes greater than the upper power supply voltage magnitude VQD- In the best mode embodiment VSs is approximately zero (0) volts and VDD is in the range of from 4.5 to 5.5 volts. To afford the best protection, the diodes 30, 32 are physically located as close to where the input signal on line 22 is received, as is practical. In situations where the input signal A on line 22 cannot exceed the magnitude range of the power supplies, such as where the source of logic signal A resides on the same integrated circuit as the NAND circuit 20, the two diodes 30, 32 may not be necessary.
The circuit 29a further comprises a series resistor 33 that is electrically connected between the diodes 30, 32 and the signal inputs a pair of buffer stages 34, 36. The resistor 33 provides electrostatic discharge protection (ESD) on line 22. The buffer stages 34, 36 each receive the ESD protected signal and provide a pair of buffered, redundant A signals on lines 38, 40 respectively. The pair of buffer stages 34, 36 represent independent signal paths, such that, a failure in one of the pair of buffer stages 34, 36 will not disturb the logic signal out of the other. To further this end, in the best mode embodiment, the pair of buffer stages 34, 36 are physically spaced further apart than normal so that the effect of a single ion transitioning through the semiconductor structure is less likely to cause single event upset (SEU) in both of the pair of buffer stages 34, 36.
In the best mode embodiment, the pair of the buffer stages 34, 36 provide the same logic function, non inverting buffering, and comprise logic gates of the same type, so that the pair of buffered A signals on lines 38, 40 are substantially similar to each other in logic state and timing. However, as will be obvious to those skilled in the art, one or both of the pair of buffer stages 34, 36 may provide a logic function other than non inverting buffering, so long as suitable modifications are made to other logic, described below, in the NAND circuit 20.
In the best mode embodiment, each of the buffer stages 34 and 36, comprise cascaded first and second inverter 42, 44 and 46, 48, respectively. The inverters are shown to be identical. The first inverters 42, 46 buffer and invert the logic signal . The second inverters 44, 48 restore the logic states of the pair of buffered A signals on lines 38, 40 to that of the input signal A on line 22. Fig. 3 is a circuit for the inverter 42, which comprises a P channel MOSFET 42a and an N channel MOSFET 42b. The input pin (IN) of the first inverter 42 is electrically connected to the gate 42c of the P channel and the gate 42d of the N channel MOSFETs. The source 42e of the P channel is connected to VDD and the source 42f ofthe N channel is connected to Vss- The drain 42g of the P channel MOSFET is connected to the drain 42h of the N channel which then serves as the output (OUT) for the first inverter 42. The other inverters 44, 46, 48 have similar transistor level circuits to that of the first inverter 42. The operation of the first inverter 42 is as follows. When the signal input at IN is at a high logic state voltage, the N channel MOSFET 42b is "on", i.e. its resistance from drain to source is low, and the P channel MOSFET 42a is "off', i.e. its resistance from drain to source is high. At the output, this produces a low resistance to the Vss supply and a high resistance to the VDD supply. The result is a voltage on the output (OUT), approximately equal to VSs, which represents a low logic state. When IN is at a low logic state voltage, the P channel MOSFET 42a is "on" and the N channel MOSFET 42b is "off'. At the output, this produces a low resistance to the VDD supply and a high resistance to the VSs supply. The result is a voltage on the output (OUT), approximately equal to VDD, which represents a high logic state.
Referring again to FIG. 2, the B input logic signal on line 24 is received by signal conditioning circuitry 29b which is similar to that described with respect to circuit 29a. Two diodes 50, 52 provide under voltage and over voltage protection. In situations where the input signal B on line 24 can not exceed the range of the power supplies the two diodes 50, 52 may not be necessary. A series resistor 53 is positioned after the two diodes 50, 52 to provide electrostatic discharge protection (ESD) on the input signal B on line 24. The ESD protected input signal is then passed to a pair of buffer stages 54, 56 which provide a pair of buffered B signals on lines 58, 60 respectively. A failure in one of the pair of buffer stages 54, 56 will not disturb the logic signal out of the other. In the best mode embodiment, the pair of buffer stages 54, 56 are physically spaced further apart than normal so that the effect of a single ion is less likely to cause single event upset (SEU) in both of the pair of buffer stages 54, 56.
In the best mode embodiment, the pair of buffer stages 54, 56 provide the same logic function, non inverting buffering, and comprise logic gates of the same type, so that the pair of buffered B signals on lines 58, 60 are substantially similar to each other in logic state and timing. As will be obvious to those skilled in the art, one or both of the pair of buffer stages 54, 56 may provide a logic function other than non inverting buffering, so long as suitable modifications are made to other logic, described below, in the NAND circuit 20. In the best mode embodiment, each of the buffer stages 54 and
56, comprise a first and second inverter 62, 64 and 66, 68, respectively. The first inverters 62, 66 buffer and invert the input signal B on line 24. The second inverters 64, 68 restore the logic states of the pair of buffered B signals on lines 58, 60 to that of the input signal B on line 24. All four ofthese inverters 62, 64, 66, 68 have similar transistor level circuits to that of the first inverter 42. The NAND circuit 20 further comprises a logic section 70 having a first portion 72 and a second portion 74. The first and second portions 72, 74 each receives one signal from each of the buffered A and B signal pairs on lines 38, 40 and 58,60. The first and second portions 72, 74 represent independent signal paths, such that, a failure in one will not disturb the logic of the other. To this end, in the best mode embodiment, the portions 72, 74 are physically spaced further apart than normal so that the effect of a single ion transitioning through the semiconductor structure is less likely to cause single event upset (SEU) in both portions 72, 74.
Each portion 72, 74 comprises the logic necessary to enable it to provide the overall logic function of the circuit, e.g. the NAND logic function of the NAND circuit 20 in the best mode embodiment. The appropriate logic depends not only on the overall logic function of the circuit, but also on the logical effect of the input section 28 buffer stages 34, 36, 54, 56 and the output section, described hereinbelow. In the best mode embodiment, the input section 28 buffer stages 34, 36, 54, 56 and the output section are non inverting buffers, and therefore have no logical effect. Consequently, the logic for the first and second portions 72, 74 of the logic section 70 each need only comprise a NAND gate 76, 78, or its equivalent, to provide the NAND function.
A schematic representation of the transistor level circuit for the NAND gate 76 of the first portion 72 is shown in Fig. 4. Referring now to Fig. 4, the NAND gate 76 has a first input 76a which is electrically connected to the gates 76b, 76c of a first pair 76d of P channel MOSFETs. The first pair 76d of P channel MOSFETs have their sources 76e electrically connected to VDD and their drains 76f electrically connected to the drains 76g of a second pair 76h of P channel MOSFETs. The sources 76i of the second pair 76h of P channel MOSFETs are connected to VDD and the gates 76j, 76k are connected to the second input signal 761.
The drains 76f, 76g of both pairs 76d, 76h of P channel MOSFETs are further connected to the drains 76m of a first pair 76n of N channel MOSFETs and the output 76p of the NAND gate 76. The first pair 76n of N channel MOSFETs have their gates 76q, 76r connected to the second input 761. Each of the sources 76s of the first pair 76n of N channel MOSFETs is connected to a different one of the drains 76t of a second pair 76u of N channel MOSFETs. The second pair 76u of N channel MOSFETs have their gates 76v, 76 w 6642
connected to the first input 76a and their sources 76x connected to Vss.
The operation of the NAND gate 76 is as follows. When both of the inputs are at a high logic state voltage, both pairs 76n, 76s of N channel MOSFETs are "on", i.e. low resistance, and both pairs 76d, 76h of P channel MOSFETs are "off, i.e. high resistance. At the output 76p, this produces a high resistance to the VDD supply and a low resistance to the Vss supply. The result is an output voltage of approximately VSs, representing a low logic state. When one or more of the inputs 76a, 761 are at a low logic state voltage, the pair 76d, 76h of P channel MOSFETs having that input connected to its gate 76b, 76c are "on" and the pair 76n, 76s of N channel MOSFETs having that input connected to its gate are "off". At the output 76p, this produces a low resistance to the VDD supply and a high resistance to the VSs supply. The result is an output voltage of approximately VDD, representing a high logic state.
Referring again to FIG. 2, the NAND gate 78 of the second portion 74 has a similar transistor level circuit schematic (not shown) and operation to that of the NAND gate 76 of the first portion 72. The first and second portions 72, 74 each operate on their respective input signals and produce first and second portion output signals which are summed, or combined, in analog fashion to produce a logic section output signal on line 84.. The logic section output signal on line 84 takes on the logic state corresponding to the voltage of the analog combination. This is in contrast to a combination that combines the signals by perfoirriing a logical operation on each of their logic states.
In the best mode embodiment, the transistors in the NAND gates 76, 78 have "on" resistances in the range of from about 500 ohms to 5,000 ohms. Lower resistances are not required because the NAND gates do not to provide high magnitude output current. Accordingly, the connection should have an electrical resistance of about less than 500 ohms, preferably less than 10 ohms. In the best mode embodiment, a direct electrical connection having a resistance below / ohm, specifically about .01 ohms, is used. However, those skilled in the art will recognize that other suitable transistors having other "on" resistances , and other connections having other resistances may also be used.
Normally, the transistors of both of the NAND gates 76, 78 are in similar "on" and "off" states. Consequently, the voltage contributions from the first and second portions to the logic section output signal on line 84 are both approximately VSs, or both approximately VDD, and the combined signal, the logic section output signal on line 84, is also approximately Vss, corresponding to a logic low, or VDD, corresponding to a logic high. Statistically, only one transistor will experience SEU at a given time. Thus when an ion transitions through the semiconductor one portion 72, 74 of the logic section 70, may experience SEU in one of the transistors in its NAND gate 76, 78. That transistor may change state from "on" to "off" or "off" to "on". However, any change in logic section output signal voltage on line 84, is small, since the output from the affected portion is combined, in analog fashion, as described above, with the output from the unaffected portion. n Consequently, the voltage effect of the SEU is reduced, or attenuated. At or below the designed ion energy level of 80 MeVcmVmg, there is insufficient energy in the ion to cause the logic section output signal on line 84 to transition to an incorrect logic state. Thus the logic section output signal on line 84 remains at the correct logic state throughout the SEU. The typical SEU lasts for about 20 nano seconds after which the transistor experiencing SEU returns to its prior state. Capacitance in the output portion, described hereinbelow, provides additional filtering for the small voltage changes in the logic section output signal on line 84.
In comparison to the present invention, other approaches, which may use a voting stage, tend to amplify the effect of the SEU because they first translate the affected signal to its own logic state prior to the voting stage, and then in the voting stage, they attempt to eliminate the effect of the SEU. These other approaches also require more gates, since they require a voting stage and often require three voting signals rather than just two, as in the present invention.
It should be recognized that this approach is not a typical method for the combination of logic signals from logic gates having active pull ups and pull downs in the output stage. For example, while open collector TTL gates may be wired-ANDed, the wired- ANDing of active pull up TTL gates is prohibitive due to the high power dissipation that occurs if one gate output is high and the other gate output is low. Similarly, in ECL logic, gates may be wired- ORed but the ECL output stage only has an active pull up and does not have an active pull down. However, in the present invention, the output of each CMOS gate does have both active pull up and active pull down.
Furthermore, unlike the TTL and ECL techniques above which seek to conserve logic gates while implementing standard logic functions, the summation approach of the present invention does not operate as a standard logic function. In fact, the summation here should not be used for combining logic signals which are not substantially similar in state and tuning for the very reason that it does not provide a normal logical combination. Instead, it provides a means for combming substantially similar signals wherein a SEU failure in one of the signals does not cause failure of the combined signal.
The logic section output signal on line 84 is passed to an output section 86. Nodal capacitance, not shown, on line 84, provides some additional filtering for the logic section output signal. The output section 86 further comprises a pair of inverters 88, 90 for buffering of the signal. The pair of inverters 88, 90 are similar to that of the first inverter 42 of the A input signal buffer stage 34 with the exception that the transistors sizes are larger to reduce the likelihood that they themselves will experience SEU. Larger and stronger transistors are better able to resist SEU by counting the collection of charge through the substrate that is brought on by the entering ion. The pair of inverters 88, 90 are also larger to provide the desired capability for driving additional logic circuits that may be connected to the output logic signal Y on line 26. The output section further includes a pair of diodes 92, 94 to protect against under and over voltage damage. As with all circuitry, appropriate decoupling should be used when employing the present invention.
In the best mode embodiment, the NAND circuit 20 is implemented in CMOS technology, preferably 1.2 micro meters (um) or smaller, however, any other suitable technology may also be used. At 1.2 um the circuit is small enough in size, has relatively low power, and provides high speed operation.
In the best mode embodiment, each of the logic signals on lines 22, 24, 26 are of the type having CMOS logic levels, however, as will be obvious to those skilled in the art, the NAND logic circuit 20 may also be designed to accommodate TTL logic signals. In the preferred embodiment, this only requires threshold modifications in the input section.
Although the best mode of the present invention creates the logic section output signal on line 84 by combuiing the outputs from NAND gates 76, 78, those skilled in the art will recognize that other transistor configurations may also be suitable. For example, transistor configurations to be connected may also be of the type used in inverters. While the best mode of the present invention includes an input section, the logic section of the present invention may also be used without the input section or portions of the input section. For example, each input logic signal may be routed directly to both portions of the logic section, without passing through the input section, so long as the logic section is suitably adapted for operation with the logic levels of the input logic signals. The present invention may also be used without the output section of the best mode embodiment. Instead, the logic section output signal may be used to drive other logic directly so long as the logic section is suitably adapted for driving such other logic. Although disclosed with respect to a best mode embodiment for use in a NAND circuit, the present invention may be used in logic circuits of any type, including but not limited to that of OR, NOR, XOR, XNOR, inverter, AND, buffer, flip flop, line drivers and transceivers, with or without three-state outputs. Furthermore, while the NAND circuit of the present invention has 2 inputs and 1 output, the present invention ;may be used in logic circuits having any number of inputs, e.g. N, where N is one or more, and any number of outputs, e.g. K, where K is one or more. The input section still preferably receives each of the N input signals and produces a pair of buffered signals for each. One signal from each pair of buffered signals is received by the first portion of the logic section and the other signal from each pair is received by the second portion. Each of the two logic section portions has all of the necessary logic gates for implementing the overall function. The first portion produces K first portion output signals and the second portion produces K second portion output signals. Each of the K first portion output signals is combined with the counterpart signal from the K second portion output signals to produce K logic section output signals. The K logic section output signals are passed to the output section where they are buffered to produce the K outputs for the circuit. As such, the present invention may also be used in logic circuits of the type including but not limited to comparators, flip flops, decoders/demultiplexers, selectors/multiplexers, counters, shift registers, parity generators, adders, and memories.
Still further, although the best mode embodiment uses signal conditioning circuits that provide a pair of buffered signals and logic sections having two portions, signal conditioning circuits providing more than a pair of buffered signals and logic sections having more than two portions may also be used and similarly combined to provide even greater immunity to SEU. While the particular invention has been described with reference to illustrative embodiments, this description is not meant to be construed in a limiting sense. It is understood that, various modifications of the illustrative embodiments, as well as additional embodiments of the invention, will be apparent to persons skilled in the art upon reference to this description without departing from the spirit of the invention, as recited in the claims appended hereto. Thus, upon understanding the present invention, one of ordinary skill in the art could employ the present invention in a variety of logic circuit applications. Those skilled in the art will know of the forms which are suitable for each application. It is therefore contemplated that the appended claims will cover any such modifications or embodiments as fall within the true scope of the invention. What is claimed is:

Claims

Claims
1. Apparatus for perfoπnirig a selected logic function on one or more logic state input signals presented thereto to provide one or more resultant logic output signals, comprising:
a first logic portion, receiving each logic state input signal and perfoπning the selected logic function on the logic state input signals to provide one or more output signals; and
a second logic portion, receiving each logic state input signal and performing the selected logic function on the logic state input signals to provide one or more output signals, and each of said output signals from said second logic portion combined in analog fashion with a corresponding one of said output signals from said first logic portion to provide the one or more resultant logic output signals.
2. The apparatus of claim 1 wherein said output signals of said first and second logic portions are provided by logic gates having outputs with active pull ups and active pull downs.
3. The apparatus of claim 1 wherein said first and second logic portions comprise CMOS technology.
4. The apparatus of claim 1 wherein said combination in analog fashion of said first logic portion output signals and said second logic portion output signals comprises an electrical connection having a resistance of about less than 500 ohms.
5. The apparatus of claim 1 wherein said first logic portion and said second logic portion comprise identical logic circuitry.
6. The apparatus of claim 1 wherein said first logic portion and said second logic portion are physically separated to prevent an ion from causing SEU in both portions.
7. The apparatus of claim 1 wherein said combination in analog fashion of said first logic portion output signals and said second logic portion output signals to produce the one or more resultant output signals is immune to SEU at ion energy levels of up to and including 80 MeVcmVmg.
8. The apparatus of claim 1 wherein said combination in analog fashion of said first logic portion output signals and said second logic portion output signals comprises a direct electrical connection having a resistance of about less than 1 ohm.
9. Apparatus for performing a selected logic function on one or more logic state input signals presented thereto to provide one or more resultant logic output signals, comprising:
ig an input section having a signal conditioning circuit for each of the logic state input signals, for providing for each of the logic state input signals, a pair of buffered logic state signals;
a logic section having a first and second portion, each receiving one logic state signal from each of said pairs of buffered logic state signals, and each performing the selected logic function on said logic state signals received thereby to each provide one or more output signals, and each of said output signals from said first portion combined in analog fashion with a corresponding one of said output signals from said second portion to provide one or more logic section output signals; and
an output section receiving each of said one or more logic section output signals and providing the one or more resultant logic output signals.
10. The apparatus of claim 9 wherein said output signals of said first and second portions of said logic section are provided by logic gates having outputs with active pull ups and active pull downs.
11. The apparatus of claim 9 wherein said first and said second portions of said logic section comprise CMOS technology.
12. The apparatus of claim 9 wherein said combination in analog fashion of said first portion output signals and said second portion S97/06642
output signals comprises an electrical connection having a resistance of about less than 500 ohms.
13. The apparatus of claim 9 wherein said first portion and said second portion of said logic section comprise identical logic circuitry.
14. The apparatus of claim 9 wherein said first portion of said logic section and said second portion of said logic section are physically separated to prevent the an ion from causing SEU in both portions.
15. The apparatus of claim 9 wherein said output section comprises circuitry that is immune to SEU at ion energy levels of up to and including 80 MeVcm2/mg.
16. The apparatus of claim 9 wherein said one or more logic section output signals are immune to SEU at ion energy levels of up to and including 80 MeVcπrVmg.
17. The apparatus of claim 9 wherein said input section, said logic section, and said output section are on the same integrated circuit.
18. The apparatus of claim 9 wherein said input section comprises, for each of the N input signals, a pair of identical buffer stages that produce said pair of buffered signals.
19. The apparatus of claim 9 wherein for each of said pair of buffer stages, one of said pair of buffer stages is physically separated from the other of said pair of buffer stages to prevent an ion from causing SEU in both buffer stages.
20. The apparatus of claim 9 wherein said combination in analog fashion of said first portion output signals and said second portion output signals comprises a direct electrical connection having a resistance of about less than 1 ohm.
21. The method for combining first and second logic signals, each having both active pull and active pull down, to produce an output logic signal comprising the step of:
summing the first and second logic signals in analog fashion to produce an output logic signal.
PCT/US1997/006642 1996-04-22 1997-04-18 Radiation resistant logic circuit WO1997040579A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US63585396A 1996-04-22 1996-04-22
US635,853 1996-04-22

Publications (1)

Publication Number Publication Date
WO1997040579A1 true WO1997040579A1 (en) 1997-10-30

Family

ID=24549388

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1997/006642 WO1997040579A1 (en) 1996-04-22 1997-04-18 Radiation resistant logic circuit

Country Status (1)

Country Link
WO (1) WO1997040579A1 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000054410A1 (en) * 1999-03-09 2000-09-14 Iroc Technologies Logic circuit protected against transitory perturbations
WO2001086397A2 (en) * 2000-04-28 2001-11-15 Bae Systems Apparatus for a radiation hardened clock splitter

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2386199A1 (en) * 1977-04-01 1978-10-27 Bailey Controle PROCESS FOR SELF-SIGNALING OF FAILURES OF A STATIC SAFETY AUTOMATION MODULE AND MODULE IMPLEMENTING SUCH PROCEDURE
JPS59108422A (en) * 1982-12-14 1984-06-22 Nec Corp Clock selecting circuit
US4709166A (en) * 1986-05-22 1987-11-24 International Business Machines Corporation Complementary cascoded logic circuit
EP0285789A2 (en) * 1987-04-10 1988-10-12 International Business Machines Corporation Fault tolerant logical circuitry
US5418473A (en) * 1992-10-28 1995-05-23 Idaho Research Foundation, Inc. Single event upset immune logic family
US5600260A (en) * 1995-06-29 1997-02-04 Motorola, Inc. Method and apparatus for hardening current steering logic to soft errors

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2386199A1 (en) * 1977-04-01 1978-10-27 Bailey Controle PROCESS FOR SELF-SIGNALING OF FAILURES OF A STATIC SAFETY AUTOMATION MODULE AND MODULE IMPLEMENTING SUCH PROCEDURE
JPS59108422A (en) * 1982-12-14 1984-06-22 Nec Corp Clock selecting circuit
US4709166A (en) * 1986-05-22 1987-11-24 International Business Machines Corporation Complementary cascoded logic circuit
EP0285789A2 (en) * 1987-04-10 1988-10-12 International Business Machines Corporation Fault tolerant logical circuitry
US5418473A (en) * 1992-10-28 1995-05-23 Idaho Research Foundation, Inc. Single event upset immune logic family
US5600260A (en) * 1995-06-29 1997-02-04 Motorola, Inc. Method and apparatus for hardening current steering logic to soft errors

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
"Circuit using CMOS that has Redundant Duality", IBM TECHNICAL DISCLOSURE BULLETIN, vol. 30, no. 9, February 1988 (1988-02-01), ARMONK, US, pages 180 - 182, XP002038940 *
BARBOUR A E ET AL: "A GENERAL, CONSTRUCTIVE APPROACH TO FAULT-TOLERANT DESIGN USING REDUNDANCY", IEEE TRANSACTIONS ON COMPUTERS, vol. 38, no. 1, January 1989 (1989-01-01), pages 15 - 29, XP000023247 *
PATENT ABSTRACTS OF JAPAN vol. 008, no. 229 (E - 273) 20 October 1984 (1984-10-20) *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000054410A1 (en) * 1999-03-09 2000-09-14 Iroc Technologies Logic circuit protected against transitory perturbations
FR2790887A1 (en) * 1999-03-09 2000-09-15 Univ Joseph Fourier LOGIC CIRCUIT PROTECTED AGAINST TRANSIENT INTERFERENCE
US7380192B1 (en) 1999-03-09 2008-05-27 Iroc Technologies Logic circuit protected against transient disturbances
WO2001086397A2 (en) * 2000-04-28 2001-11-15 Bae Systems Apparatus for a radiation hardened clock splitter
WO2001086397A3 (en) * 2000-04-28 2003-03-13 Bae Systems Apparatus for a radiation hardened clock splitter

Similar Documents

Publication Publication Date Title
EP0895638B1 (en) High reliability logic circuit for a radiation environment
US6703858B2 (en) Logic architecture for single event upset immunity
US8207753B2 (en) Method and apparatus for reducing radiation and cross-talk induced data errors
US6326809B1 (en) Apparatus for and method of eliminating single event upsets in combinational logic
US6504411B2 (en) Redundant latch circuit and associated methods
US20060267653A1 (en) Single-event-effect hardened circuitry
WO2005071837A1 (en) Radiation hardening of logic circuitry using a cross-enabled, interlocked logic system and method
US20040022005A1 (en) Radiation-hard circuit
US6906388B2 (en) SEU hard majority voter for triple redundancy
US9768757B1 (en) Register circuitry with asynchronous system reset
Naseer et al. The DF-dice storage element for immunity to soft errors
US6549443B1 (en) Single event upset resistant semiconductor circuit element
US6642802B2 (en) Ring oscillator providing single event transient immunity
US7071749B2 (en) Error correcting latch
US7688112B2 (en) Anti-SEE protection techniques for high-speed ICs with a current-switching architecture
EP1665530B1 (en) A static latch
US6825691B1 (en) Apparatus and method for a radiation resistant latch with integrated scan
US6937053B2 (en) Single event hardening of null convention logic circuits
JP2007129689A (en) Dual path redundancy with stacked transistor voting
WO1997040579A1 (en) Radiation resistant logic circuit
US20070229132A1 (en) Scannable domino latch redundancy for soft error rate protection with collision avoidance
US11791831B1 (en) Modular redundant threshold circuit and method
Zatsarinny et al. Failure-Tolerant Self-Timed Circuits

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): JP

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE

121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase

Ref country code: JP

Ref document number: 97538246

Format of ref document f/p: F

122 Ep: pct application non-entry in european phase