# US 2019/0095275 A1 · Error detection for training non-volatile memories

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

- Number: US 2019/0095275 A1
- Document ID: US20190095275A1
- Status: Published application
- Kind: A1
- Application number: 15/717,572
- Filed: 2017-09-27
- Published: 2019-03-28
- Earliest claimed: 2017-09-27
- Assignees: Western Digital Technologies, Inc.
- Applicants: Western Digital Technologies, Inc.
- Inventors: Gautam Ashok Dusija; Venkatesh Prasa Ramachandra; Mrinal Kochar
- Web page: https://patentgrep.com/patent/US20190095275A1

## CPC

- G06F 11/1004: {to protect a block of data words, e.g. CRC or checksum (G06F11/1076 takes precedence; security arrangements for protecting computers or computer systems against unauthorized activity G06F21/00)}
- G06F 3/0619: {in relation to data integrity, e.g. data losses, bit errors}
- G06F 3/064: {Management of blocks}
- G06F 3/0679: {Non-volatile semiconductor memory device, e.g. flash memory, one time programmable memory [OTP]}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

The present disclosure, in various embodiments, describes technologies and techniques for detecting errors in a non-volatile memory (NVM) device prior to performing re-training/recalibration. A processing device in a NVM controller detects a cyclic redundancy check (CRC) condition for detecting error in the NVM device, and a re-training condition that is based on the CRC condition. A CRC circuit generates CRC code when a CRC condition is detected, and the processing is configured to compare CRC code received from the NVM controller with the generated CRC code to detect error. A calibration circuit then re-trains the NVM device if the CRC circuit detects error and the re-training condition has been met.

## Claims

### Claim 1

A non-volatile memory (NVM) controller, comprising:
a host interface for communicating with a host device;
a memory interface for communicating with a NVM via one or more channels;
a cyclic redundancy check (CRC) circuit configured to generate a CRC code; and
a calibration circuit configured to provide re-training data to the NVM,
wherein the NVM controller is configured to:
(i) receive a CRC code from the NVM via the memory interface,
(ii) compare the received CRC code with the generated CRC code to detect error,
(iii) determine if a re-training condition has been met, based on the comparison of the received CRC code with the generated CRC code, and
(iv) activate the calibration circuit, if the re-training condition has been met, to re-train the NVM using the re-training data.

### Claim 2 (depends on 1)

The NVM controller of claim 1, wherein the NVM comprises a NAND memory, and the NVM controller is configured to receive the CRC code in response to a read command.

### Claim 3 (depends on 1)

The NVM controller of claim 1, wherein the NVM controller is further configured to:
determine if a CRC condition is present, and
receive CRC code from the NVM only if the CRC condition is present.

### Claim 4 (depends on 3)

The NVM controller of claim 3, wherein the CRC condition comprises at least one of (i) expiration of a predetermined period of time, (ii) a predetermined number of read errors detected, and/or (iii) a temperature delta detected.

### Claim 5 (depends on 1)

The NVM controller of claim 1, wherein the re-training condition comprises at least one of:
a predetermined number of CRC errors detected within a predetermined time period, and/or
a predetermined number of CRC errors detected after a temperature delta threshold has been met or exceeded.

### Claim 6 (depends on 1)

The NVM controller of claim 1, wherein the generated and received CRC codes each comprise one of (i) a fixed-length binary sequence or (ii) a data pattern.

### Claim 7 (depends on 1)

The NVM controller of claim 1, wherein the calibration circuit is configured to calibrate the NVM by matching impedance between the NVM controller and the NVM.

### Claim 8 (depends on 1)

The NVM controller of claim 1, wherein the calibration circuit is configured to calibrate the NVM based on a duty cycle of a strobe of the controller.

### Claim 9 (depends on 1)

The NVM controller of claim 1, wherein the calibration circuit is configured to calibrate the NVM by correcting at least one of a strobe and a phase relationship between data and the strobe, wherein the strobe and the data are provided via the memory interface to the NVM.

### Claim 10

A non-volatile memory (NVM) controller, configured to communicate with a host and a NVM, comprising:
a processing circuit configured to determine (i) a cyclic redundancy check (CRC) condition indicative of error in the NVM, and (ii) if a re-training condition, based on the CRC condition, has been met;
a CRC circuit operatively coupled to the processing circuit and configured to generate a CRC code when the CRC condition is met, wherein the processing circuit is configured to compare a CRC code received from the NVM controller with the generated CRC code to detect error; and
a calibration circuit operatively coupled to the processing circuit and configured to re-train the NVM if the processing circuit detects error and the re-training condition has been met.

### Claim 11 (depends on 10)

The NVM controller of claim 10, wherein the NVM comprises a NAND memory, and the NVM controller is configured to receive the CRC code in response to a read command.

### Claim 12 (depends on 10)

The NVM controller of claim 10, wherein the CRC condition comprises at least one of (i) expiration of a predetermined period of time, (ii) a predetermined number of read errors detected, and/or (iii) a temperature delta detected.

### Claim 13 (depends on 10)

The NVM controller of claim 10, wherein the re-training condition comprises at least one of:
a predetermined number of CRC errors detected within a predetermined time period, and/or
a predetermined number of CRC errors detected after a temperature delta threshold has been met or exceeded.

### Claim 14 (depends on 10)

The NVM controller of claim 10, wherein the generated and received CRC codes each comprise one of (i) a fixed-length binary sequence or (ii) a data pattern.

### Claim 15 (depends on 10)

The NVM controller of claim 10, wherein the calibration circuit is configured to calibrate the NVM by matching impedance between the NVM controller and the NVM.

### Claim 16 (depends on 10)

The NVM controller of claim 10, wherein the calibration circuit is configured to to calibrate the NVM based on a duty cycle of a strobe of the controller.

### Claim 17 (depends on 10)

The NVM controller of claim 10, wherein the calibration circuit is configured to calibrate the NVM by correcting at least one of a strobe and a phase relationship between data and the strobe, wherein the strobe and the data are provided via the memory interface to the NVM.

### Claim 18

A method of operating a non-volatile memory (NVM) controller, configured to communicate with a host and a NVM, comprising:
determining a cyclic redundancy check (CRC) condition for detecting error in the NVM;
generating a CRC code comprising a fixed-length binary sequence when the CRC condition is determined;
receiving a fixed-length binary sequence CRC code from the NVM in response to a read command from the NVM controller;
comparing the CRC code received from the NVM with the generated CRC code to detect error;
determining if a re-training condition has been met based on detected error in the comparing of the CRC code; and
re-training the NVM if the comparison of the CRC codes is indicative of error and the re-training condition has been met.

### Claim 19 (depends on 18)

The method of claim 18, wherein the CRC condition comprises at least one of (i) expiration of a predetermined period of time, and/or (ii) a temperature delta detected.

### Claim 20 (depends on 18)

The method of claim 18, wherein the re-training condition comprises at least one of:
a predetermined number of CRC errors detected within a predetermined time period, and/or
a predetermined number of CRC errors detected after a temperature delta threshold has been met or exceeded.

### Claim 21 (depends on 18)

The method of claim 18, wherein re-training the NVM comprises calibrating the NVM by matching impedance between the NVM controller and the NVM.

### Claim 22 (depends on 21)

The method of claim 21, wherein re-training the NVM comprises calibrating the NVM based on re-training data received from the NVM and correcting at least one of timing and phase of the re-training data received from the NVM.

### Claim 23

A non-volatile memory (NVM) controller, configured to communicate with a host and a NVM, comprising:
means for determining a cyclic redundancy check (CRC) condition for detecting error in the NVM;
means for generating a CRC code when the CRC condition is determined;
means for comparing a CRC code received from the NVM controller with the generated CRC code to detect error;
means for determining if a re-training condition has been met, based on detected error in the comparing of the CRC code; and
means for re-training the NVM if the means for comparing CRC code detects error and the re-training condition has been met.
