user guide
utilities
Go to Website For Partners
Product Line-up

home > learning resource > white papers > overall reliability : MTBF & APR

overall reliability : MTBF & APR

Introduction

HDDs are, perhaps, the most significant component of a PC. A faulty HDD, therefore, most likely insures a faulty and unreliable PC. Here, we will show you how SAMSUNG assesses HDD reliability and also how SAMSUNG creates reliable and quality HDDs.

How to Measure HDD Reliability

In the PC industry, there is no universal standard in measuring HDD reliability-each company employs its own methodolgy in assesing HDDs. The most commonly used procedures are MTBF (Mean Time Between Failure) and ARP (Annual Return Rate). At the product development stage, SAMSUNG uses MTBF and continues to improve HDD design until the MTBF value meets the satisfaction of SAMSUNG designers and engineers. After the fnishing touches are added to the protocol, mass-production of the product begins. During the mass-production period, ORTs (On-Going Reliability Test) are continuously conducted to ensure product reliability. Even after product delivery, ARP tests are continuously conducted to identify any potential problems. SAMSUNG's MTBF for HDDs is 500,000 hours. That means that if you use your PC for 9 hours every day, your HDD should operate for 152 years. In imperfect, non-test conditions, however, please note that the real life span of an HDD varies because of fluctuating operating environments. Now, let us show you have the MTBF value is calculated.

MTBF Values are Derived From Arithmetic Calculation MTBF valuation by arithmetic calculation is used in the early stages of product development-it is the sum of all of an HDD's components. After the initial test phase, the MTBF value is reconfigured to account for the following factors: design faults, manufacturing faults, software bugs, and other environmental problems.

SAMSUNG conducts extensive MTBF tests to insure high product quality.

MTBF Based on Sample Tests

This method of MTBF valuation is statistically based. When the MTBF goal is not met during the test phase, problems are identified and corrected until the optimal MTBF level is achieved. Because this MTBF valuation method takes into consideration design faults, manufacturing faults, and software bugs, the MTBF value is usually lower than those that are arithmetically derived. Statistically based MTBF valuations are also more accurate. However, this is still a probability-based prediction, and cannot precisely indicate the life span of the product in non-test environments. A sample of MTBF valution is shown below:

  • A test is conducted on 480 sample units, which have passed the baseline test: Under the same temperature and voltage conditions for 4 hours as a unit and with On/Off power alternation.
  • When the test is completed, the MTBF is calculated using the following equation:
    MTBF=2×Drive Test Time × The Amount of Drive Tested ÷ X (n)
  • The sample unit, which experienced failure during the test, is sent to the Failure Analysis Group to improve design flaws.
  • MTBF ranges of acceptability: Table 1 PRST (Probability Ratio Sequential Test) Chart - If the MTBF is acceptable, the test is completed. If the MTBF is unacceptable, design flaws are corrected and tests are continued until the MTBF falls into the range of acceptability.
Table 1.PRST (Probability Ratio Sequential Test) Chart
MTBF ranges of acceptability
ORT (On-Going Reliability Test)

Even after the mass production stage, ORT tests are conducted to endsure product reliability and longevity. In ORT tests, a total of 120 sample units are tested non-stop for 72 hours and the MTBF in the ORT phase is calculated using the following equation.

Test Time × The number of Sample Units ÷ The number of Sample Units with Failure

When the MTBF fails acceptability at the development stage, all methods are used to bring the product up to acceptable standards. Failure analyses are conducted, manufacturing processes are improved, and sometimes, component suppliers are asked to take action to address the problem. In some cases, even, the the problems lies in PC assembly process.

ARR (Annualized Return Rate)

Even after the mass production stage, ORT tests are conducted to endsure product reliability and longevity. In ORT tests, a total of 120 sample units are tested non-stop for 72 hours and the MTBF in the ORT phase is calculated using the following equation:

The number of units returned for the year ÷ The total number of units Shopped for the year