Angotti Product Development

 Electronic Product Development Questions answered for a LOW COST, FIXED FEE -- > Click here

Back to Other Newsletters --> Click Here

Angotti Product Deveopment Newsletter

Volume 6, No. 3 (Click Here for pdf Version)

"Helping You Accelerate Your High-Tech Development Projects"

Welcome to the ANGOTTI PRODUCT DEVELOPMENT e-mail newsletter!

The goal of this  newsletter is to help you accelerate your development projects by sharing many of the tips, techniques, and strategies we've honed during three decades of providing high-tech consulting services.

Optimizing the Cost of Production Test Equipment

Excutive Summary

Testing of new designs is best thought out at the very beginning. Managers must decide what to do about testing the new product in the early and later versions of a Product's Life Cycle.
This article describes the tradeoffs for creating electronic test equipment for the Final Test of a PCB Assembly or System.

It describes the economics of testing at all product development phases, from Engineering Design to Manufacturing Final Test, QA Returns, and others. The least expensive place to consider testing is at the Early Design Phase. The costliest phases start at the beginning of the Final Manufacturing Test, especially after a series of field failures.

The considerations for each NPD phase are as follows:

1. During Pre-Design Phases

During this phase, a decision should be made on the choice of the test system architecture. This choice significantly impacts the overall cost of tester development and production test costs. The architecture depends on several factors. These include:
  • Experience with Building Such Testers
  • Final Cost of Device to be Tested
  • Volume Level of Nominal Build
  • Volume Level of the Estimated Life Cycle Build Total
  • Funds Available for Test System and Software or Test Procedures
  • Cost to Build the Tester and Write the Software and Test Procedures
  • Skill Level of Test Operator
  • Risk of Field Failures for Cost and Reputation Considerations
  • Cost of Field Failure Detection, Return, and Replacement
To assist in the design process, Angotti Product Development has created a spreadsheet. It covers the tradeoffs between the above factors to make manual, semi-automatic, or automatic testers in practice. It is available in Appendix 1 of the pdf version of this document at this page: (PDF version of Appendix 1).
    2. During Engineering Product Development

    At this point, a good test plan can significantly increase test fault coverage and lower the overall cost for the final functional test (Design for Testability or DFT). This plan can include design-in of good test points and access to them, and the Built-in Initial System Test performed at power-up (BIST or Power-On Self-Test (POST), or Continuous Built-In Test (CBIT).

    Of considerable advantage is the plan for using high-level commands in the embedded code to perform System Test. These System-level tests can be used during the Development Debug and utilized in the final Manufacturing Functional Test. For board-level testing, the test designer makes use of the unique board-level JTAG port or ports. These tests can be used for lower-level debug or to expand some of the system-level tests available outside the firmware.

3.  Tester Development Process

Much like a typical New Product Development (NPD) process, a tester development process follows a plan with milestones that track the Tester Development process. This process would look something like this:

Link to Flow Diagram

    4. Pre- Production Functional Testing

    Pre-production testing is testing at the PCB Fab and Assembly level. Such testing would include testing the fabricated PCB for trace opens and shorts. Fab testing allows for the assessment of PCB manufacturing quality to catch problems before the PCB assembly.

    After the Assembly process is completed, PCBs can be visually inspected to a high degree using sophisticated optical comparison methods and electrical flying probe testing to potentially pick up problems such as bad parts and solder joints, faulty parts installation, among others.
    Testing at this level can achieve test coverage accuracy from 90 to 95%. Such testing today can be performed at a relatively low cost per PCB assembly.

    After the Assembly process, PCBs can be visually inspected to a high degree using sophisticated optical comparison methods and electrical flying probe testing to potentially pick up problems such as bad parts and solder joints, faulty parts installed, etc.
    Testing at this level can achieve test accuracy from 90 to 95%. Such testing today can be performed at a relatively low cost per PCB assembly.

4. During Low Volume Production, including batches of 10 to 100 builds
    This testing is performed on PCBs and Systems at the final Functional Test level. The intention is to catch about 99% or more of the electronic and potential mechanical functional failures. These tests can reduce returns significantly and impact the reputation of the fabrication house and the product design system company.

    a. There have been several approaches to produce Final Testing in the past. These include:
    For larger systems or multi-card assemblies - use a "gold standard" reference system and then substituted for a previously tested reference PCBs with the new UUT.

    b. This approach still requires generating the written final test procedures for final tests but can be one of the least expensive methods for producing hardware test facilities. This approach often has the requirement to have testing done by more highly skilled technicians or even the design engineers and can be very costly on a per-unit test cost basis.

    c. This testing also does not provide any automated capturing of test results data. It also doesn't allow for any margin testing of signal Input/Outputs.

    d. For Low production volumes of 10 to 100 units per month

    e. For this production level, custom manual or semi-automated testers for the Device under test (DUT) can be designed. This approach can work for a PCB Sub-Assembly or the testing of small systems or single PCB systems.

5. During Low Volume Production, including batches of 10 to 100 builds

This level of testing involves the use of more highly skilled persons to perform the testing. The test person performs tests according to a written procedure, and manually sets up inputs and assesses outputs. Test technicians or engineers often do this level of testing. If done intelligently, it can capture data reports for the DUTs tested and provide for margin testing, but is usually costly and slow on a per unit UUT.



6. Manual Testers

This level of testing involves the use of more highly skilled persons to perform the testing. The test person performs tests according to a written procedure and manually sets up inputs and assesses outputs. Test technicians or engineers often do this level of testing. If done intelligently, it can capture data reports for the DUTs tested and provide for margin testing, but it is usually costly and slow on a per unit UUT.

During high volume of production, much more expensive approaches are used to test to take advantage of high volume fully automated testers.

7. Semi-Automated Testing
In this approach, the Manual Testing can be extended by having manual inputs and outputs manually selected. Then, if required, system-level test commands are sent to the DUT. In this case, the operator still assesses the PASS/FAIL of the test results.

Often these tests can be performed by trained, lower-level test persons. Using this approach can reduce production test costs and lower the system input errors by sending repetitive commands to the DUT. The approach is still prone to errors in input and output switching and interpretation of test results.

8. During High Volume Production

During high volume of production much more expensive approaches to test to take advantage of high volume fully automated testers.

Automated Testers

High Production Volume testing is where automated testers are genuinely needed, whether the DUT cost is low or high. At very high test volumes, sample testing instead of 100% screening is utilized. Such testers can often be quite expensive for both hardware and especially software.

A clear case can be made for automatic testers in the full system test except for lower production volumes. In this case, subassemblies usually use automated testing for sub-system DUTs. In such cases, a full-fledged test plan is a must.

At the system level, high-level commands are often sent to the full system, and responses are measured. These can be done in a semi-automatic or automatic testing approach, and measurements are made either automatically or manually.
In automated testing, all system inputs and outputs are automatically switched, commands are sent to the DUT, and measurements are made with automatic test equipment via a control program. The test results are then usually read by the test program and compared with the test program limits. Testing proceeds in an automated manner one after another until testing is complete. The measurements are often stored in an output file generated to keep track of the test results for later reference.

Even during automated testing, occasionally hand probe tests must be made. The program stops execution, the probe is placed appropriately, and then a command is issued to continue the program.

9. Testing a Very Expensive DUT

For an very expensive or even an expensive PCB DUT, test probe mechanical fixtures are generated that insert test probe pins at various internal points in the circuitry to permit the measurement of signals. These test probes can be created semi-automatically from the formal fabrication drawings and can add significant additional expense to the Tester. These allow isolation of failures to a much smaller section of the board circuitry than just utilizing system Inputs and Outputs for measurements.

For this approach to be effective preplanned test points must be placed throughout the PCB layout.

Of course, this approach must be coupled with the Fully Automated Test approach to be effective.

    FOR FUTURE NEWSLETTERS

    Future newsletters will discuss in more detail each of the testers described in this newsletter. Get on our mailing list by sending an email to carl@angotti.com and requesting to be included. You can learn much more about the test engineering services we provide by going to the Angotti Product Development Test Website.

    FEEL FREE TO FORWARD THIS NEWSLETTER TO YOUR ASSOCIATES

    If you think this information would be valuable to others, please feel free to forward this newsletter to your associates. I would appreciate it if you would not alter its contents.

    If you were forwarded this e-mail, and wish to subscribe, please send an e-mail to mailto:carl@angotti.com with SUBSCRIBE in the Subject Heading, and I will add you to the list of subscribers.

    MORE FREE MATERIAL ON PROJECT MANAGEMENT

     For more FREE Project Management tips, techniques and strategies, and to learn more about the services available to my clients, be sure to visit the ANGOTTI PRODUCT DEVELOPMENT website at www.angotti.com.