This application claims priority under 35 USC §119 to German Application No. DE 10 2006 019075.0, filed on Apr. 25, 2006, and titled “Integrated Circuit to Store a Datum,” the entire contents of which are hereby incorporated by reference.
Fuse circuits can be used in combination with storage circuits in integrated semiconductor memory devices, for example in a DRAM (Dynamic Random Access Memory) semiconductor memory devices (for example, for the activation of redundant word and bit lines of a memory cell array). The fuse circuit includes a programmable element, such as a fusible link, that registers a programmable state of “0” when the fusible link is not severed and a programmable state of “1” when the fusible link is severed.
A fuse circuit can be arranged, for example, on a memory chip and surrounded by a housing composed of a plastic material. Due to alpha-particles that impinge on the memory chip proceeding from the plastic material of the housing, charge carriers in the material of the memory chip can be torn from their bonds. This gives rise, on the chip, to low-impedance connections between a conductor track and a substrate of the chip, which is generally charged to a ground potential. A high potential on the conductor track is conducted away through the resultant conductor track to the substrate. A storage state that was buffer-stored in the storage circuit can be corrupted by such a discharge process.
In addition, the state of the storage circuit can be influenced by neutrons that likewise generate charge carriers that establish a conductive connection between a conductor track and the substrate. Consequently, the influence of neutrons can also have the effect that the output terminal of the storage circuit drives an inaccurate datum value (e.g., a “1” state instead of a “0” state in the case of a non-blown fuse of the fuse circuit, or drives a “0” state instead of a “1” state in the case of a blown fuse of the fuse circuit). The change in state at the output terminal of the storage circuit can lead to a malfunction of the semiconductor memory device that persists until the voltage supply is switched off and switched on again, since the state of the fuse is evaluated anew as a result of the switching on of the voltage supply.
An integrated circuit and corresponding method of operating an integrated circuit are described herein. The integrated circuit comprises a programmable circuit configured to be programmed into a selected programmable state and to generate a programming state signal that is dependent upon the selected programmable state, and a storage circuit configured to receive the programming state signal from the programmable circuit, to store a first storage state or a second storage state depending upon the programming state signal received from the programmable circuit, and to generate an output signal that is dependent upon the stored storage state. The storage circuit comprises a first inverter circuit and a second inverter circuit, each of the first and second inverter circuits being connected between a first supply voltage terminal and a second supply voltage terminal, where at least one of the first inverter circuit and the second inverter circuit includes a first controllable switch connected between the first supply voltage terminal and an output terminal of the first inverter circuit and a second controllable switch connected between the output terminal of the inverter circuit and the second supply voltage terminal. The first and the second controllable switches have different conductivities in a conductive state.
The above description and still further features and advantages will become apparent upon consideration of the following detailed description of specific embodiments thereof, particularly when taken in conjunction with the accompanying drawings wherein like reference numerals in the various figures are utilized to designate like components.
The fuse circuit is programmed with a state “1” or “0” depending on the state of the programmable element 101. If the programmable element 101, which is embodied as a fusible link, for example, is severed, the fuse circuit is programmed with the state “1”. In the non-severed state of the programmable element 101, the state “0” is programmed in the fuse circuit. In order to read out the programming state of the fuse circuit, an activation signal AS is applied to a control terminal A3. A high level of the activation signal AS is converted into a low level by the inverter 104, which low level controls the p-channel transistor 102 into the on state and the n-channel transistor 103 into the off state. As a result, firstly the output terminal M is charged to a high potential.
In order to read out the programming state of the programmable element 101, a low level of the activation signal AS is applied to the control terminal A3, by means of which the transistor 102 is controlled into the off state and the transistor 103 is controlled into the on state. If the fusible link, as illustrated in
The high or low potential state of the output terminal M, which identifies the programming state of the fuse circuit, is buffer-stored by two inverters 105 and 106 and also the feedback of the output side of the inverter 106 to the input side of the inverter 105 in the storage circuit 107.
As noted above, the use circuits of this type in combination with a storage circuit connected downstream can be used in an integrated semiconductor memory, for example in a DRAM (Dynamic Random Access Memory) semiconductor memory, for the activation of redundant word and bit lines of a memory cell array. For this purpose, the integrated circuit shown in
In addition, the state of the storage circuit can be influenced by neutrons that likewise generate charge carriers that establish a conductive connection between a conductor track and the substrate. Consequently, the influence of neutrons can also have the effect that the output terminal A5 of the storage circuit 107 drives a “1” state instead of a “0” state in the case of a non-blown fuse 101, or drives a “0” state instead of a “1” state in the case of a blown fuse 101. The change in state at the output terminal A5 of the storage circuit 107 can lead to a malfunction of the semiconductor memory that persists until the voltage supply is switched off and switched on again, since the state of the fuse is evaluated anew as a result of the switching on of the voltage supply.
The output terminal A10 of the programmable circuit unit is connected to an input terminal E20 of the storage circuit 20. The storage circuit 20 comprises an inverter circuit 21 and an inverter circuit 22, which are connected between the input terminal E20 of the storage circuit 20 and an output terminal A20 of the storage circuit. The output terminal A20 is coupled with feedback to the input terminal E20.
The inverter circuit 21 comprises a controllable switch P2, which is embodied as a p-channel transistor, and a controllable switch N2, which is embodied as an n-channel transistor. The controllable switch P2 is connected between a supply voltage terminal V1 for application of a supply voltage VDD and an output terminal A21 of the inverter circuit 21. The controllable switch N2 is connected between the output terminal A21 of the inverter circuit 21 and a supply voltage terminal V2 for application of the supply voltage VSS. The control terminals SP2 of the controllable switch P2 and SN2 of the controllable switch N2 are connected to the input terminal E20 of the storage circuit 20. The output terminal A21 of the inverter circuit 21 is connected to an input side of the inverter circuit 22.
The inverter circuit 22 contains an activatable inverter 23. The activatable inverter 23 comprises a controllable switch P3, which is embodied as a p-channel transistor, a controllable switch N3, which is embodied as an n-channel transistor, and a controllable switch N4, which is embodied as an n-channel transistor. The controllable switch P3 is connected between a supply voltage terminal V1 for application of a supply voltage VDD and the output terminal A20 of the storage circuit 20. The controllable switches N3 and N4 are connected in series between the output terminal A20 of the storage circuit 20 and a supply voltage terminal V2 for application of a supply voltage VSS. The control terminals SP3 of the controllable switch P3 and SN3 of the controllable switch N3 are connected to the output terminal A21 of the inverter circuit 21.
The storage circuit 20 further comprises a controllable switch P4, which is embodied as a p-channel transistor, and a controllable switch P5, which is likewise embodied as a p-channel transistor. The two controllable switches P4 and P5 are connected in series between a supply voltage terminal V1 for application of a supply voltage VDD and the output terminal A20 of the storage circuit 20. A control terminal SP4 of the controllable switch P4 is connected to the output terminal A21 of the inverter circuit 21.
The functioning of the circuit arrangement shown in
In order to read out the programmed-in state of the programmable circuit unit and in order to buffer-store the programming state in the storage circuit 20, the programmable circuit unit must first be initialized. During a time phase T0, control terminal SP1 of the controllable switch P1 is driven with a low level of an activation signal PCH. An activation signal SET likewise drives control terminal SN1 of the controllable switch N1 with a low level. As a result, the controllable switch P1 is in a conducting state and the controllable switch N1 is in a blocking state. The output terminal A10 is therefore charged to a high potential (“1” state) (initialization state). A programming state signal PZS, which occurs at the output terminal A10, therefore has the programming state “1”.
The input terminal E20 of the storage circuit 20 is driven by the programming state signal PZS. The programming state “1” is inverted by the inverter circuit 21, whereby the controllable switch P4 is controlled into the conducting state. The controllable switch P5 is likewise controlled into the conducting state by the low level of the activation signal SET, with the result that a storage state “1” occurs at the output terminal A20 of the storage circuit 20. The integrated circuit is now initialized for the actual read-out process of the programmable circuit unit 10.
In order to read out the programming state of the programmable element F of the programmable circuit unit 10, the activation signal PCH is subsequently applied with a high level to the control terminals SP1 of the controllable switch P1 and SN4 of the controllable switch N4. Furthermore, the activation signal SET is still present with a low level at the control terminals SN1 of the controllable switch N1 and SP5 of the controllable switch P5. The controllable switch N4 is switched into the conducting state by the high level of the activation signal PCH. The activatable inverter 23 is thus activated. At the time phase T1, therefore, the state of the programming state signal PZS generated at the output terminal A10 is buffer-stored in the storage circuit 20.
At the time phase T2, the activation signal SET is applied with a high level to the control terminal SN1 and the control terminal SP5, while the activation signal PCH retains the high level. As a result, the controllable switch N1 is controlled into the conducting state and the controllable switch P5 is controlled into the turned off state. In the case of a non-blown (non-severed) programmable element F, the charge to which the output terminal A10 was charged during the initialization process flows away to the supply voltage terminal V2 via the controllable switch N1 that has been controlled into the conducting state and the intact fusible wire. In the case of a blown (severed) programmable element F, the output terminal A10 continues to remain at the high potential to which it was charged during the initialization phase. Since the activatable inverter 23 is still active during the time phase T2, the state of the programming state signal PZS that is present at the input terminal E20 is read into the storage circuit 20 and buffer-stored there as the storage state. The output signal FLAT occurs with a high or low level at the output terminal A20 depending on the buffer-stored storage state.
In the case of a p-channel transistor, the doped regions NG1 and NG2 are embodied as p-doped regions and the substrate PS is embodied as an n-doped substrate. In the case of an n-channel transistor, the doped regions are in each case embodied as n-doped regions and the substrate is embodied as a p-doped substrate. The resistance of the channel K is dependent on the channel length LK and the channel width WK. Thus, when a channel for a first transistor is shorter or wider than a channel for a second transistor, the impedance exhibited by the first transistor in the conducting state is correspondingly lower than the impedance exhibited by the second transistor.
The transistor N2 in the conducting state is provided with lower impedance (strengthening of the transistor N2) than the transistor P2 in the conducting state (weakening of the transistor P2). Furthermore, in the conducting state of the transistors P4 and P5, the series circuit comprising the transistors P4 and P5 is provided with lower impedance (strengthening of the transistors P4 and P5) than that with which the series circuit comprising the transistors N3 and N4 (weakening of the transistors N3 and N4) is embodied in the conducting state of the transistors N3 and N4.
A strengthening and weakening of transistors can be obtained, for example, by changing the channel lengths and channel widths of the transistors (i.e., as described above with reference to
As a result of the strengthening of the transistors N2, P4 and P5 and the weakening of the transistors P2, N3 and N4, the state of the output signal FLAT=1, corresponding to the initialization state at the time phase T0 and the state in the case of a severed programmable element F at the time phase T2, becomes resistant to an undesired state change due to alpha-particles or neutrons. As a result of the strengthening of the transistors N2, P4 and P5 and the weakening of the transistors P2, N3 and N4, the state of the output signal FLAT=0, corresponding to the state in the case of a non-severed programmable element F at the time phase T2, becomes more susceptible to an undesired state change on account of alpha-particles or neutrons. A corrupted state of the output signal can be corrected again, however, by repeated evaluation of the programmed-in state of the programmable element F. For this purpose, it suffices for a pulse to be applied to the activation signal SET (time phase Tn) in order to reverse the undesired state change on account of alpha-particles or neutrons that altered the state of the output signal FLAT from the state “0” to the state “1”, and thus to reestablish the state of the output signal FLAT=0 for the programmable circuit unit with a non-severed programmable element F.
The susceptibility to an undesired state change on account of alpha-particles or neutrons can be lowered overall by the strengthening of the transistors N2, P4 and P5 and the weakening of the transistors P2, N3 and N4 and the repeated evaluation of the programmable state of the programmable element F.
If the output signal FLAT at the output terminal A20 has the state “0” and the programmable element F is not severed, the output signal FLAT has the correct programming or storage state. No error has occurred in this case. The state of the storage circuit has not been corrupted on account of alpha-particle and neutron influence. If the activation signal SET is fed in with a high level at specific time intervals Δt onto the control terminals SN1 and SP5 of the transistors N1 and P5, and the control terminals SP1 and SN4 are permanently driven with a high level of the activation signal PCH, the output signal FLAT continues to remain at the state “0”.
If the output signal FLAT has the state “0” and the programmable element F is severed (blown), the storage state of the storage circuit 20 has been corrupted. In this case, the programming state of the programmable element F cannot be read out merely by driving the control terminals SN1 and SP5 with a high pulse of the activation signal SET, since a low potential is likewise present at the output terminal A10 on account of the feedback. Consequently, the initialization state would not be present at the output terminal A10. However, since the transistors N2, P4 and P5 have been strengthened relative to the transistors P2, N3 and N4, it is possible to prevent the situation in which a state change of the output signal FLAT=“1” to the state FLAT=“0” occurs as a result of alpha-particles and neutrons while the fusible wire is severed. By contrast, the state “1” of the output signal FLAT at the output terminal A20 is reliably held at the state “1” as a result of the strengthening of the transistors N2, P4 and P5 relative to the transistors P2, N3 and N4.
If the output signal FLAT has the state “1” and the programmable element F is not severed, the storage state of the storage circuit 20 has been corrupted on account of alpha-particles or neutrons. The output signal FLAT at the output terminal A20 should actually have the state “0” if the fusible wire of the programmable element F is not severed. If the control terminals SN1 and SP5 are driven by a high pulse of the activation signal SET at time intervals Δt, while the control terminals SP1 and SN4 are permanently driven by a high level of the activation signal PCH, the programming state of the programmable element F is read out again, since the output terminal A20, via the feedback, has assumed the corrupted state, in this case the high level necessary for read-out, of the output signal FLAT. As a result of driving with the high pulse of the activation signal SET, in this case the programming state of the programmable circuit unit is read out again, with the result that the output signal FLAT has the correct state “0” again after the end of the read-out process.
If the output signal FLAT has the state “1” and the programmable element F is severed, the storage state of the storage circuit 20 has not been corrupted. In this case, too, the read-out of the programming state of the programmable circuit unit 10 is only possible by driving the control terminals SN1 and SP5 with the high pulse of the activation signal SET, since the output terminal A10 has been charged, via the feedback, to a high potential state, that is to say is in the initialization state.
Providing strong transistors N2, P4 and P5 and weak transistors P2, N3 and N4 makes it possible to virtually preclude the situation in which the state of the output signal FLAT=“1” is corrupted into the state FLAT=“0” if the fusible wire of the programmable element is severed. This enables the programmable circuit unit 10 to be read merely by a high pulse on the activation signal SET, while the activation signal PCH, which drives the transistors P1 and N4, is held at a high level.
The power demand of the integrated circuit can therefore be significantly reduced by comparison with a read-out of the programmable circuit unit by the activation signal sequence applied in the time phases T0, T1 and T2. If it is assumed that a plurality of the programmable elements F are not blown, the state of the output signal FLAT would have to be subjected to charge reversal twice when carrying out the steps during the time phases T0, T1 and T2 in the case of the multiplicity of the integrated circuits. By contrast, the strengthening of the transistors N2, P4 and P5 relative to the transistors P2, N3 and N4 makes it possible for the storage state of the storage circuit to be updated merely by driving the control terminals SN1 and SP5 with a high pulse of the activation signal SET. Charge reversal of the output terminal A20 twice occurs only when the state of the storage circuit 20 has changed as a result of alpha-particles or neutrons.
The bit line decoder 200 contains a storage unit 220 comprising a plurality of the integrated circuits 210. Bit line addresses of defective bit lines BL are stored in the storage circuits 20 of the integrated circuits 210. The storage unit 220 is coupled to a comparator unit 230. A bit line address applied to the address terminal A100 is compared, in the comparator unit 230, with the bit line addresses of defective bit lines that are stored in the storage unit 220.
The word line decoder 300 comprises a storage unit 320 containing a plurality of integrated circuits 310. Addresses of defective word lines are stored in the storage circuits 20 of the integrated circuits 310. The storage unit 320 is coupled to a comparator unit 330. A word line address applied to the address terminal A100 is compared, by comparator unit 330, with the word line addresses of defective word lines that are stored in the storage circuits of the integrated circuits 310.
Upon application of a bit line address identifying a defective bit line, and upon application of a word line address identifying a defective word line, a redundant word line WLr and a redundant bit line BLr, respectively, are selected instead of the defective word line and bit line and the memory cell SZr connected to said redundant word line and redundant bit line is read. The storage content of the storage circuits 20 of the integrated circuits 210 and 310 is updated by driving the integrated circuits with a high pulse of the activation signal SET at specific time intervals. This prevents malfunctions of the integrated semiconductor memory device due to alpha-particles or neutrons.
While the invention has been described in detail and with reference to specific embodiments thereof, it will be apparent to one skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope thereof. Accordingly, it is intended that the present invention covers the modifications and variations of this invention provided they come within the scope of the appended claims and their equivalents.
Number | Date | Country | Kind |
---|---|---|---|
102006019075.0 | Apr 2006 | DE | national |