The present invention relates generally to power systems, and, more particularly, to indicating status of an on-chip power supply system.
On-chip power systems for integrated circuits have become increasingly popular as they eliminate the need for providing several different voltage levels from external sources. It is common to have several power systems serving several memory sub-systems.
As one example, in a chip including memory macros, such as an embedded dynamic random memory (eDRAM) chip, it is necessary to provide many different voltage levels to effectively achieve performance. For example, there needs to be one voltage level (VPP) to provide a voltage greater than supply voltage Vdd, for a boosted wordline operation to enhance cell signal and access time. A negative voltage, (VBB), is needed to bias the body of the transfer devices in the memory array so that sub-threshold leakage of the cell is minimized. In some memory chip designs there may also be an additional negative voltage (VWL) to hold unselected wordlines to a level below ground so that the sub-threshold leakage of the array devices can be suppressed to improve cell retention. Additionally, a bandgap reference voltage (VBGR) is needed to provide a constant and stable reference voltage insensitive to variations in the supply voltage Vdd, and temperature. Another voltage reference (VREFDC), which varies with the supply voltage, is needed to generate voltage levels which are designed to track the supply voltage. Another bandgap circuit output is constant current reference (VCMN), which may be used to bias differential amplifiers, etc.
Diagnostics become difficult in an Application Specific Integrated Circuit (ASIC) chip having several power systems serving several memory sub-systems because the power system voltages are not observable at test. Internal probing of a failed chip is not practical under several metal levels, and consequently, field failures are presently very difficult to diagnose.
Thus, there exists a need for a technique for simply and effectively diagnosing failures in on-chip power supply systems.
According to exemplary embodiments, systems and a method are provided for indicating status of an on-chip power supply system for improving test flow and diagnosing of chip failure.
According to one embodiment, a system is provided for indicating status of an on-chip power supply system including multiple power supplies. The system includes a power system status register for receiving digital compliance signals. Each compliance signal is associated with one of the multiple power supplies. Each power supply has an associated compliance level, and each compliance signal indicates whether its associated power supply is operating at the associated compliance level. The power system status register also generates a power supply status signal based on the digital compliance signals. The power system status signal indicates status of the compliance signals associated with the power supplies. The system further includes an output for outputting the power supply status signal. If a power supply is operating at its associated compliance level, the power supply status signal indicates that the power supply is passing, and if a power supply is not operating at its associated compliance level, the power supply status signal indicates that the power supply is failing.
According to another embodiment, a system for indicating status of an on-chip power supply system including multiple power supplies also includes a sequencer circuit for activating the multiple power supplies in a predetermined order. The sequencer circuit provides a digital compliance signal for each power supply indicating whether the power supply is operating at an associated compliance level. The system further includes an analog-to-digital converter for converting analog signals indicating operating levels of power supplies into digital compliance signals indicating whether the power supplies are operating at associated compliance levels. The power system status register is coupled to the sequencer circuit and the analog-to-digital converter for receiving the digital compliance signals. The power system status register generates a power supply status signal based on the digital compliance signals. The power supply status signal indicates status of the compliance signals associated with the power supplies. If a power supply is operating at its associated compliance level, the power supply status signal indicates that the power supply is passing, and if the power supply is not operating at its associated compliance level, the power supply status signal indicates that the power supply is failing. A Built-in Self Test (BIST) receives the power supply status signal via a latched interface or a scan chain and performs tests on memories on the chip based on the power supply status signal.
According to another embodiment, a method is provided for indicating status of an on-chip power supply system including multiple power supplies. The method includes receiving digital compliance signals, each compliance signal associated with one of the multiple power supplies, and each power supply having an associated compliance level. Each compliance signal indicates whether its associated power supply is operating at the associated compliance level. A power supply status signal is generated based on the digital compliance signals, the power supply status signal indicating status of the compliance signals associated with the power supplies. The power supply status signal is output. If a power supply is operating at its associated compliance level, the power supply status signal output indicates that the power supply is passing, and if the power supply is not operating at its associated compliance level, the power supply status signal output indicates that the power supply is failing.
Referring to the exemplary drawings, wherein like elements are numbered alike in the several Figures:
According to exemplary embodiments, the health of the boosted supplies and analog voltage levels present in an on-chip power system may be easily tested and diagnosed based on a status signal indicating the status of on-chip power supplies. The systems and method described herein eliminate the need for internal probing to diagnose a field failure. A memory sparing technique is also made possible based on the status signal indicating the status of the power supply system.
The description below is directed toward embedded memory circuits and, more specifically, to embedded DRAM (eDRAM) circuits. However, it should be understood that the invention is not limited to eDRAM circuits but is applicable to all circuits making use of on-chip power systems.
Described below with reference to
Referring to
VREFXGOP is fed to VREFX_DELAY element 160, along with VCMN. VPPGOP is an active high output that activates the VPP pump system. VPPGOP is fed to VPP_DELAY element 170, along with VCMN. PWRON is an active high output that indicates that activation of the multiple power supply system is complete. PWRON is generated as an output of a delay element 180, and is fed back as an input to delay elements 150, 160 and 170.
The VINTON signal is the first input to the Power-on sequencer and when initially low, the L1/L2 latch elements (Latch_VBB 110, Latch_VREFX 120, Latch_VPP 130 and Latch_PWRON 140) are in the ‘0’ state. When VINTON goes high to indicate Vdd is on, the Latch_VBB element 110 latches a ‘1’ state, and VBBGOP output goes high to start the −0.5V VBB system. When the VBB system reaches its target level, the VBBPUMP signal goes low, and after a short delay, this sets the Latch_VREFX 120 to the high state to activate VREFXGOP.
According to one embodiment, the VREFX level is a regulated positive voltage (e.g., 0.45 V), and there is no comparator device to judge when it is at its specified level. Therefore, the “PUMP” input to the corresponding VREFX-delay element is simply grounded. Once VREFXGOP is activated, the VPPGOP signal is also activated (with the Latch_VPP 130 set to the high state), after a short delay, to enable the 1.7V VPP system. When the VPP pump attains an output level of about 1.7V, the VPPUMP signal goes low, which sets the Latch_PWRON 140 to a high state.
The latch output VPWRONGO from Latch_PWRON 140, is delayed by a delay element 180, and is output as PWRON. Signal PWRON going high indicates all power systems are up and running. This is further illustrated in
According to an exemplary embodiment, the digital power supply system status outputs (also referred to as digital compliance signals) described above may be provided as bits, with a “1” indicating operation of a power supply at its associated compliance level, and a “0” indicating the power supply is not operating at its compliance level. These signals, in turn, may be provided to a BIST, either as a single “master” value or as a set of individual signals. If all the power supply status signals (compliance signals) are “1”, this indicates that the power supplies are all passing. If any one of the power supply status signals is “0”, this indicates that one of the power supplies is failing. According to one embodiment, if one power supply is failing, all are considered to be failing for BIST memory testing purposes. In this embodiment, the BIST may be sent a single master signal indicating passing or failing of the power supplies by performing a logical AND operation on the power supply signals, or the BIST may receive multiple power supply status signals and determine passage or failure based on whether one of the power supply status signals is a “0”.
Described below and with reference to
Each BIST test status register 520 may be connected in a serial shift chain with all other BIST engines on a chip. This makes it possible to obtain status information obtained via a non-destructive shift of only the test status registers without disturbing other latches on the chip during the shift operation.
The power system status of memories tested by all BIST engines on a chip can be collected in each BIST and then serially shifted off chip. As explained above, an advantage of this over a normal scan operation is that only the status information stored in the BIST engine is serially unloaded off chip, preserving the state of all other latches (and consequently the current test state) of the chip. This allows for nondestructive monitoring of the power system status, without needing to restart testing following a scan operation.
The power system status information can be combined with the BIST memory test status information in a configuration such as that depicted in
According to an exemplary embodiment, the pass/fail indicator for all BIST engines on a chip are logically combined via, e.g., comparator 630, such that if any power supply or memory is found to be failing, the chip will output a “memory test failed” status flag. If the status flag indicates a memory test failure, then the memory test can be aborted, saving test time by not running the full test sequence on chips with known bad memories. By combining the power system status information with the BIST pass/fail indicator, a memory with a bad power system can be marked as failing as soon as a BIST test begins, removing the need to perform any memory read/write operations, and saving valuable test time.
In the flow chart shown in
Referring back to
The power system 1020 includes voltage regulators 1022, and a VINTON source 1044 inputting voltage signals into a sequencer 1025 as described above with reference to
As explained above with reference to
The pass/fail indicators are passed to the tester 1030, which diagnoses chip failure based on various factors, including the BIST memory test results. Chips may have different criteria for passing. For example, all memories must pass for a chip to pass, or only a certain number of memories must pass for a chip to pass.
Chip failure diagnosis is made easier, according to exemplary embodiments, because power status information of analog power supplies is made readily available. This makes it easier to diagnose problems with analog circuitry on the chip, as it avoids inferring from a failed memory test that there might be an analog circuit (power supply) issue. According to exemplary embodiments, rather than running extensive (and time consuming) BIST tests only to find that there appear to be issues with underlying analog circuits, the status of the analog circuits may be queried before hand. If analog circuits are failing a chip failure determination can be made quickly and testing may be terminated early.
The above embodiments describe a method for accessing the power system status in a simple manner. The health of a chip can be assessed in the field, or upon field return through simple test procedures. This can be vital in diagnosing and dealing with field or process issues that create failures in analog circuitry. Typically, these failures manifest themselves as failures in downstream logic or testing processes. The problem is in identifying the root cause or source of the fail signature to aid in debug. Often times, many different circuit failures can generate the same system level fail signature. In large systems, analog circuit failures can create memory write/read failures that are detected as system level data errors (which have many possible sources). Determining the failing circuit on a large system is difficult and expensive without good diagnostics. According to exemplary embodiments, BIST or scan diagnostics are extended to include status information of previously unobservable circuits. A check of this status information upon system failure can provide a quick check of multiple, isolated circuit blocks. This can isolate the root cause to a small subset of the total system, saving time and money in debug efforts.
While exemplary embodiments have been described above, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope thereof. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed as the best mode contemplated for carrying out this invention, but that the invention will include all embodiments falling within the scope of the appended claims.