Accendo Reliability

Your Reliability Engineering Professional Development Site

  • Home
  • About
    • Contributors
  • Reliability.fm
    • Speaking Of Reliability
    • Rooted in Reliability: The Plant Performance Podcast
    • Quality during Design
    • Critical Talks
    • Dare to Know
    • Maintenance Disrupted
    • Metal Conversations
    • The Leadership Connection
    • Practical Reliability Podcast
    • Reliability Matters
    • Reliability it Matters
    • Maintenance Mavericks Podcast
    • Women in Maintenance
    • Accendo Reliability Webinar Series
    • Asset Reliability @ Work
  • Articles
    • CRE Preparation Notes
    • on Leadership & Career
      • Advanced Engineering Culture
      • Engineering Leadership
      • Managing in the 2000s
      • Product Development and Process Improvement
    • on Maintenance Reliability
      • Aasan Asset Management
      • CMMS and Reliability
      • Conscious Asset
      • EAM & CMMS
      • Everyday RCM
      • History of Maintenance Management
      • Life Cycle Asset Management
      • Maintenance and Reliability
      • Maintenance Management
      • Plant Maintenance
      • Process Plant Reliability Engineering
      • ReliabilityXperience
      • RCM Blitz®
      • Rob’s Reliability Project
      • The Intelligent Transformer Blog
    • on Product Reliability
      • Accelerated Reliability
      • Achieving the Benefits of Reliability
      • Apex Ridge
      • Metals Engineering and Product Reliability
      • Musings on Reliability and Maintenance Topics
      • Product Validation
      • Reliability Engineering Insights
      • Reliability in Emerging Technology
    • on Risk & Safety
      • CERM® Risk Insights
      • Equipment Risk and Reliability in Downhole Applications
      • Operational Risk Process Safety
    • on Systems Thinking
      • Communicating with FINESSE
      • The RCA
    • on Tools & Techniques
      • Big Data & Analytics
      • Experimental Design for NPD
      • Innovative Thinking in Reliability and Durability
      • Inside and Beyond HALT
      • Inside FMEA
      • Integral Concepts
      • Learning from Failures
      • Progress in Field Reliability?
      • Reliability Engineering Using Python
      • Reliability Reflections
      • Testing 1 2 3
      • The Manufacturing Academy
  • eBooks
  • Resources
    • Accendo Authors
    • FMEA Resources
    • Feed Forward Publications
    • Openings
    • Books
    • Webinars
    • Journals
    • Higher Education
    • Podcasts
  • Courses
    • 14 Ways to Acquire Reliability Engineering Knowledge
    • Reliability Analysis Methods online course
    • Measurement System Assessment
    • SPC-Process Capability Course
    • Design of Experiments
    • Foundations of RCM online course
    • Quality during Design Journey
    • Reliability Engineering Statistics
    • Quality Engineering Statistics
    • An Introduction to Reliability Engineering
    • An Introduction to Quality Engineering
    • Process Capability Analysis course
    • Root Cause Analysis and the 8D Corrective Action Process course
    • Return on Investment online course
    • CRE Preparation Online Course
    • Quondam Courses
  • Webinars
    • Upcoming Live Events
  • Calendar
    • Call for Papers Listing
    • Upcoming Webinars
    • Webinar Calendar
  • Login
    • Member Home

by Anne Meixner 1 Comment

What Makes Memory Test Hard

What Makes Memory Test Hard

Just One More Thing

Memory circuits store 1’s and 0’s for you so that you can retrieve them later. The 1/0’s can represent data or instructions.

Today, no computing device operates without some memory. Memory devices store pictures, music, documents. How many USB (Unified Serial Bus) memory sticks do you have scattered on your desk?

Semiconductor memory comes in many flavors: Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Read Only Memory (ROM), Erasable Programmable Memory (EPROM), NAND/NOR Memory (a.k.a. Flash), Solid State Drive (SSD). Without power volatile memories like SRAM and DRAM lose their stored values. Non-Volatile memories retain their stored value when power is absence.

Stuck at Fault Model Not Sufficient for Memory Test

As memory just stores 1’s and 0’s what could be so hard about it testing it?  You apply the stuck at fault model and done! With this introductory article I want you to leave with an appreciation of the nuanced nature of memory test. I will use CMOS (Complementary Metal Oxide Silicon) SRAM to illustrate.

Remember the stuck at fault model was developed for testing logic circuits. Slightly removed from the physical implementation of the circuitry—tis technology agnostic. This abstraction quickly breaks for memory circuits. As a starting point let’s apply the S@ model. The test would consist of:

Write a 1, then read a 1, if you get a 0 you have a fail.

Write a 0, then read a 0, if you get a 1 you have a fail.

Voila, Fini!

Oh one more thing, there’s that write and read operation. Specific circuitry performs that function. Also, read/write operations are not performed 1 bit at a time, but a row of bits at time. Address circuitry connects the read/write circuitry to that row in memory. How many bits are there?  Well that depends if it is a stand-alone memory embedded. Sizes are typically on order of 512 megabytes or more.

Tally the Memory Parts to Test

First let’s look at memory from top view at architecture and drill down to the transistor level description. Along the way you can tally the circuitry that requires to be tested.

Memory is organized in blocks and a grid system in which there are columns and rows. Address circuitry connects the row of bits (often organized in bytes- 8 bits) to the read and write circuitry.  Registers hold the bit values to be written or read.

SRAM ARch doc.anc70

Let’s start the Tally:

  • Address circuitry- Did you write to the correction location? Does it matter if it is the wrong location as long as it’s consistent?
  • Read circuitry- Did it read correctly within time allowed? How could you tell the difference between a faulty read circuitry and a faulty bit?
  • Write circuitry- Did it write the correct value within the time allowed?  How long does it take to write the opposite value? How could you tell the difference between a faulty write circuitry and a faulty bit?
  • What about those storage buffers?
  • Bit cells- is the bit faulty?

No BIT is an Island

No woman is an island, the same holds true for a bit cell.  They are surrounded by neighbors (except the ones on the edges- though physically they are not the last row or column). Consider a 3×3 array. Neighbors north, south and west are immediate and then neighbors on the diagonal.  Columns of Bit and Bit bars connect multiple cells in the North South direction. A row addresses connect multiple cells in the East West Direction. Bit and Bitbar lines are shared with north and south neighbors.

In the presence of defects neighbors can lead to interactions which are often described as pattern sensitive faults. Consider the following pattern in which a bit cell with a zero is surrounded by 1’s. That bit could be influenced to be a 1.

1 1 1

1 0 1

1 1 1

Why is it like this, well this relates to how densely packed the cells are physically.

Single Cell Test Perspective

Now let’s look at a single bit cell.  One way of representing it is as two inverters connected to one another.

SRAM_Cell_Inverter_Loop Wiki Commons

More often the transistor level representation provides some further insight. Six transistors compose the most often used SRAM cell; therefore, called 6-T cell.

6T sram-cell A

And as Columbo would say “Just one more thing.”  Integrated circuit memory devices pack a lot of storage in their area. For a semiconductor process the design rules for memory circuits are more aggressive than they are for logic circuits.  This is because the shapes are highly regular and can be optimized for the process node while logic circuit is much more random. Due to the higher density memory cells exhibit higher sensitivity to spot defect. The physical layout provides you with the lowest level of a memory cell. The implementation will vary according to the process node and design style. Below shows you an example.

SRAM layout_celula

Size, Density, Timing make Memory Test Hard

The S@ fault model lacks the ability to model interactions between cells.  These interactions have resulted in numerous algorithm test patterns that need to be supplied. Timing related faults occur—how long between the next read/write to the same cell. Memory size dictates the number of bits that need to be tested with all these patterns. However, the very regular nature of memory cell design and architecture can lead to some advantages when developing a test solution.

What’s the just one more thing for you when you think about memory test?  Do you want to know more about DRAM and SRAM testing? Or would you be more interested into delving into the non-volatile memory?  Have you heard about a specific test that you want to know about?  Curious about DFT solutions?  Let me know and I will follow-up with an article addressing your specific needs.

Meanwhile remember testing takes time and thoughtful application,

Anne Meixner, PhD

Additional Reading:

Wikipedia has several articles on Memory design including Memory Cell, SRAM and DRAM

Several books on memory testing can be found on Amazon

Testing Semiconductor Memories: Theory and Practice by A. J. Van De Goor. I have used this book and I met the author when he did a summer visiting professor at Intel.  It’s the classic text on testing memories.

High Performance Memory Testing: Design Principles, Fault Modeling and Self-Test by R. Dean Adams.  I met Dean over 10 years ago at a test conference when he worked for IBM. While I have not read his book, I know that he brings his many years as a designer to this subject.

Semiconductor Memories: Technology, Testing, and Reliability by Ashok K Sharma.  I have not read this book, but I thought for the reliability engineering audience this might be a good reference.

 

Filed Under: Articles, on Tools & Techniques, Testing 1 2 3 Tagged With: Mastery 1, Memory Test, Stuck at Fault Model, VLSI Test

« Listening Skills for Reliability Engineers
10+ Dropped Items from the ASQ CRE Body of Knowledge »

Comments

  1. Aol Technical Support says

    July 14, 2018 at 10:04 AM

    Memory is an integral part of the computer system. RAM required for the better performance of the system. The thing which makes memory testing hard is size density and timing.

    Reply

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Article by Anne Meixner
in the Testing 1 2 3 series

Join Accendo

Receive information and updates about articles and many other resources offered by Accendo Reliability by becoming a member.

It’s free and only takes a minute.

Join Today

Recent Posts

  • How to Reduce Maintenance Cost The Right Way
  • Significance Over Success. Innovation Over Change. Anticipation Over Agility
  • Maintenance Planning and Scheduling for World Class Reliability and Maintenance Performance
  • Self-Discipline Part 1
  • Is Safety Training Helpful?

© 2023 FMS Reliability · Privacy Policy · Terms of Service · Cookies Policy

This site uses cookies to give you a better experience, analyze site traffic, and gain insight to products or offers that may interest you. By continuing, you consent to the use of cookies. Learn how we use cookies, how they work, and how to set your browser preferences by reading our Cookies Policy.