# US 2004/0205433 A1 · High reliability memory module with a fault tolerant address and command bus

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2004/0205433 A1
- Document ID: US20040205433A1
- Status: Published application
- Kind: A1
- Application number: 10/413,605
- Filed: 2003-04-14
- Published: 2004-10-14
- Earliest claimed: 2003-04-14
- Assignees: International Business Machines Corporation
- Inventors: Kevin C. Gower; Bruce Hazelzet; Mark W. Kellogg; David J. Perlman
- Web page: https://patentgrep.com/patent/US20040205433A1

## CPC

- H05K 1/117: {Pads along the edge of rigid circuit boards, e.g. for pluggable connectors}
- G06F 11/1044: {with specific ECC/EDC distribution}
- G11C 5/04: Supports for storage elements {, e.g. memory modules}; Mounting or fixing of storage elements on such supports
- G11C 2029/0409: {Online test}
- H05K 2201/09145: Edge details
- H05K 2203/167: Using mechanical means for positioning, alignment or registration, e.g. using rod-in-hole alignment
- H05K: PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G11C: STATIC STORES (semiconductor memory devices H10B)

## Abstract

A high reliability dual inline memory module with a fault tolerant address and command bus for use in a server. The memory module is a card approximately 151.35 mm or 5.97 inches long provided with about a plurality of contacts of which some are redundant, a plurality of DRAMs, a phase lock loop, a 2 or 32K bit serial EE PROM and a 28 bit and a 1 to 2 register having error correction code (ECC), parity checking, a multi-byte fault reporting circuitry for reading via an independent bus, and real time error lines for determining and reporting both correctable errors and uncorrectable error conditions coupled to the server's memory interface chip and memory controller or processor such that the memory controller sends address and command information to the register via address/command lines together with check bits for error correction purposes to the ECC/Parity register. By providing the module with a fault tolerant address and command bus fault-tolerance and self-healing aspects necessary for autonomic computing systems compatible with industry-standards is realized. The memory module corrects single bit errors on the command or address bus and permits continuous memory operation independent of the existence of these errors and can determine any double bit error condition. The redundant contacts on the module prevents what would otherwise be single points of failure.

## Claims

### Claim 1

A dual in-line memory module (DIMM) comprising:
a rectangular printed circuit board having a first side and a second side, a length of between 149 and 153 millimeters and first and second ends having a width smaller than said length;
a first plurality of connector locations on said first side extending along a first edge of said board that extends the length of the board;
a second plurality of connector locations on said second side extending on said first edge of said board;
a locating key having its center positioned on said first edge and located between 82 mm and 86 mm from said first end of said card and located between 66 and 70 mm from said second end of said card.

### Claim 2 (depends on 1)

The DIMM of claim 1 wherein there is further provided on said first side multiple dynamic random access memories (DRAMs);
a phase lock loop circuit; and
a 28 bit 1 to 2 register circuit having error correction code (ECC) across data inputs, parity checking, and a real time error line for reporting both correctable errors and uncorrectable error conditions mounted on said first side of said board.

### Claim 3 (depends on 1)

The DIMM of claim 1 wherein there is further provided on said first side
multiple dynamic random access memories (DRAMs);
a phase lock loop circuit; and
a 28 bit 1 to 2 register circuit having error correction code (ECC) across data inputs and a real time error line for reporting both correctable errors and uncorrectable error conditions mounted on said first side of said board.

### Claim 4 (depends on 2)

The DIMM of claim 2 wherein there are 22 data inputs to the register ECC circuitry.

### Claim 5 (depends on 1)

The DIMM of claim 1 wherein said multiple (DRAMs) are mounted on said first side of said printed circuit board.

### Claim 6 (depends on 1)

The DIMM of claim 1 wherein said multiple (DRAMs) are mounted on said second side of said printed circuit board.

### Claim 7 (depends on 1)

The memory module of claim 1 wherein the output pins of said multiple DRAMS are connected to selected ones of said connector locations on said board.

### Claim 8 (depends on 1)

The memory module of claim 1 wherein the first edge of the first side of said board has one hundred and thirty-eight (138) connector locations and an identical plurality of connector locations on the said first edge of said second side of said board for a total of two hundred and seventy-six (276) contacts on said board.

### Claim 9 (depends on 1)

The memory module of claim 1 wherein selected contacts on said first side surface are coupled to selected contacts on said second side to provide redundant contacts for selected signals sent to and received from said DRAMs, said register chip, said EE PROM and/or said phase locked loop.

### Claim 10 (depends on 1)

The Dual Inline Memory Module of claim 1 wherein;
a selected number of said contacts on said first side are commonly connected to a selected number of contacts on said second side to provide said card with a selected number of redundant contacts;
a plurality of dynamic random access memory chips on said first and second sides;
a 2 or 32K bit serial EE PROM and a 28 bit, 1 to 2 register circuit chip having error correction code (ECC), a multi-byte fault reporting register, read via an independent bus, and real time error lines for both correctable errors and uncorrectable error conditions on said first side; and
a phase lock loop circuit coupled to said plurality of dynamic random access memory chips on said first and second sides and to said register chip on said first side.

### Claim 11 (depends on 1)

The Dual Inline Memory Module of claim 1 wherein;
a selected number of said contacts on said first side are commonly connected to a selected number of contacts on said second side to provide said card with a selected number of redundant contacts;
a plurality of dynamic random access memory chips on said first and second sides;
a 2 or 32K bit serial EE PROM and a 28 bit, 1 to 2 register circuit chip having parity checking, a multi-byte fault reporting register, read via an independent bus, and real time error lines for both correctable errors and uncorrectable error conditions on said first side; and
a phase lock loop circuit coupled to said plurality of dynamic random access memory chips on said first and second sides and to said register chip on said first side.

### Claim 12 (depends on 1)

The memory module of claim 1 wherein selected contacts CS1, CKE0, CKE1, RAS, CAS, WE, CK0, and CK0B are positioned at selected nominal distances from the key on said first edge of said DIMM as follows:
6 NOMINAL DIRECTION DISTANCE FROM CONTACT or SIDE of FROM CENTER CENTER SIGNAL PIN# DIMM OF KEY OF KEY CS0 86 FRONT 11.495 mm RIGHT CS0 224 BACK 11.495 mm LEFT CS1 91 FRONT 16.495 mm RIGHT CS1 229 BACK 16.495 mm LEFT CKE0 65 FRONT 13.505 mm LEFT CKE0 203 BACK 13.505 mm RIGHT CKE1 62 FRONT 16.505 mm LEFT CKE1 200 BACK 16.505 mm RIGHT RAS 222 BACK 9.495 mm LEFT CAS 87 FRONT 12.495 mm RIGHT WE 84 FRONT 9.495 mm RIGHT CK0 77 FRONT 2.495 mm RIGHT CK0 215 BACK 2.495 mm LEFT CK0B 78 FRONT 3.495 mm RIGHT CK0B 216 BACK 3.495 mm LEFT

### Claim 13

A high-reliability memory arrangement with a fault tolerant address and command bus for use as a main memory that will achieve the degree of fault-tolerance and self-healing necessary for autonomic computing systems comprising;
a memory interface chip;
a memory controller; and
a dual inline memory module having a register coupled to the memory interface chip and the memory controller via address/command lines together with check bits for error correction such that the memory controller sends address and command information to the register via the address/command lines together with check bits for error correction purposes to the register.

### Claim 14 (depends on 13)

The memory of claim 13 wherein said register is an ECC register.

### Claim 15 (depends on 13)

The memory of claim 13 wherein said register is an parity register.

### Claim 16

A server comprising:
a memory interface chip;
a memory controller; and
a plurality of dual inline memory modules (DIMMs), each having a phase lock loop circuit chip, a error correction code/parity register chip and a plurality of dynamic random access memory (DRAM) chips, thereon; and
said memory controller and said memory interface chip being coupled to said plurality of dual inline memory modules by data lines, clock lines, an address bus and a command bus.

### Claim 17 (depends on 16)

The server of claim 16 wherein said register chip is a 28 bit 1:2 register containing error correction code circuitry (ECC) to correct for single bit errors on the command bus or address bus, and permit continuous memory operation independent of the existence of these errors.

### Claim 18 (depends on 16)

The server of claim 16 wherein said register chip is a parity register.

### Claim 19 (depends on 16)

The server of claim 16 where said register further includes means for parity checking 22 data bit lines, error latches and an error reporting mode whereby the system may interrogate the device to determine the error condition thereby allowing accurate fault determination and preventive maintenance thereby reducing unplanned system outages.

### Claim 20

A server memory structure comprising
a dual inline memory module provided with selective redundant contacts;
a phase locked loop;
said dual inline memory module is provided with a 2 to 32K bit serial EE PROM; and
a multi-byte fault reporting register coupled to the memory interface and memory controller via an independent bus;
means for reading said multi-byte fault reporting register bus; and
real time error lines for reporting both correctable errors and uncorrectable error conditions.

### Claim 21 (depends on 16)

The server of claim of claim 16 wherein there is provided;
an error correction code parity register chip on each of said DIMMs; and
means for transmitting data between the DRAMs on said DIMMs, the memory interface with the memory interface chip sending address and command information to the register via an addresses command line together with check bits for error correction purposes controller via independent bus means for reading said multi-byte fault reporting register bus; and
real time error lines for reporting both correctable errors and uncorrectable error conditions.

### Claim 22

An error correction code register comprising:
a 28 bit 1:2 register segment; and
an error correction code circuit and an error logic circuit for reporting errors and for correcting single bit errors to permit continuous memory operation independent of the existence of single bit errors;
said register segment comprising a first set of differential receivers coupled to a check bit bus, a second set of differential receivers coupled to a data line bus and a plurality of input differential bit receivers and latch selection input differential receiver;
each of said differential receivers having two inputs and a single output;
a plurality of input multiplexers, a plurality of primary latches, a plurality of secondary multiplexers and a plurality of secondary latches;
the first input of each differential receiver in said first set of differential receivers , said second set of differential receivers and said plurality of input differential bit receivers being coupled to a reference voltage source and the second input of each differential receivers being coupled to a respective input from said memory interface chip;
each receiver in said first set having an input respectively coupled to a respective check bit line in said set of check bit lines and to a respective output connected through a respective check bit input multiplexer and a respective input latch to a single bit error correction/double bit error detection circuit circuit;
each receiver in said second set having an input respectively coupled to a respective data line in said set of data lines and an output connected through a respective data primary multiplexer and a respective primary latch, a single bit error correction/double bit error detection circuit, a respective secondary multiplexer and secondary latch to an output line;
a first input differential receiver having an input coupled to a check bit 0 signal line and an output coupled through a primary multiplexor and primary latch to a single bit error correction/double bit error detection circuit;
second and third input differential receivers having their second input respectively coupled to Chip select lines /CS0 and /CS1 and their outputs coupled respectively to the first inputs of a pair of primary latches, the first and second inputs of a an input select NAND gate;
the output of the NAND gate is coupled to the selection inputs of the primary multiplexors;
a fourth differential receiver having an input coupled to a first clock enable signal source (CKE0) and an output coupled to a respective primary latch;
a fifth differential receiver having an input coupled to a second clock enable signal sources (CKE1) and an output coupled to a respective primary latch;
a sixth receiver having an input coupled to a first On Die Termination line signal line (ODT0) and an output coupled to a respective primary latch;
a seventh Receiver having an input coupled to a second On Die Termination line signal line (ODT1) and an output coupled to a respective primary latch; and
a receiver having an input coupled to the reset (/RST) signal line and an output to the reset input of all the primary and secondary latches.

### Claim 23 (depends on 22)

The error correction code register of claim 22 wherein the third input of the NAND gate is coupled to a chip select (CS) gate enable circuit which can be set low to cause the primary multiplexors coupled to said first set to pass signals from the first set of differential receivers independent of the levels on inputs on said pair of differential receivers.

### Claim 24 (depends on 22)

The error correction code of claim 22 wherein there is further provided an error logic circuit comprising a 16-bit counter that increments as it receives errors inputs from the error correction code circuit, a status register block that contains a plurality of status registers, and an IIC logic block interconnected to one another through logic circuits whereby the error counter continues to count errors even while the status registers are being read out on the IIC bus.

### Claim 25

A method of operating a server for detecting the status of a DIMM inserted therein to assure the DIMM can monitor the address and control bus integrity, and can correct errors on the address and control bus and report errors and log and count errors consisting of the steps of;
accessing said DIMM using an industry standard IIC protocol and address input range sources (SA0, SA1, SA2),
activating said DIMM to produce a byte of data which includes the contents of Byte 0, i.e.,
7 Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0 RFU DIMM DIMM DIMM Mode Parity ECC error ECCerror Address Address Address 1=ECC Error 1=UE 1=CE 0 SA2 SA1 SA0 0=Pty 1=PERR
stimulating the DIMM to echo back the input range sources (SA0, SA1, SA2) on bits four, five and six of said Byte 0, and
applying a high signal level (a “1”) on bit 3 to provide a unique signature on bits four, five and six to determine if the signature matches a code sent to the DIMM.
