|
Equipment
Reliability Institute |
|
Environmental Test -
Understanding Vibration of Electronic Systems - by John Starr
and Ed Abner How does one "prove" that a product will function properly throughout its anticipated life? Dave Douthit points to some serious weaknesses in climatic testing in Reliability and Dependability - How to prove them. Now is the time for
training - by Wayne Tustin Readers, please ask questions. One reader's question led to a discussion of physical demonstrations, using a shaker, during vibration training. Another reader asks about shock transmissibility. Readers - please respond. Some graduates of Wayne's basic vibration and shock course have asked about advanced vibration training. See "Electronic Circuit Cards - Applied Vibration" by John Starr. And don't forget to check the right-hand column for other reliability news. Climatics test training has new dates. ERI's vibration and shock glossary has just been updated. Check it out! If you are involved with vibration or shock testing, then you should be involved with fixtures - see "Fixture Group". Best wishes, |
| ******************************* |
|
Environmental
Test All electronic systems experience vibration as they get transported on land, by air or rocket, some time during their life cycles. In addition, vibration is very efficient for environmental stress screening (ESS) of electronics because of the large number of cycles of stress that can be applied in a short time. When setting up a vibration testing program for electronic systems, using either in-house testing or the services of an environmental test lab, the most important goal is understanding of your product. This applies for accelerated life testing and for ESS. For life testing, what life is expected (of each of the component parts of system? For ESS, what areas of your system are adequately screened? What areas are questionable? In order to develop and test electronic systems with confidence, someone (and preferably everyone - design, development, production, test - in the development chain, should understand vibration capabilities. This requires understanding at the component level. Understanding vibration means being able to identify those few components at risk of failure during a vibration test. Also to list the many components that would survive testing for 10 years. Vibration has always been presented as very complex, necessitating oversimplification. Empirical formulas predicted capabilities. Empirical methods used in the often served to define a conservative limit for design of electronic card systems. 1970’s methods fail to account for the complex configurations of electronic systems; and often result in expensive failures. Life capability of electronics is usually dominated by the modal response of circuit cards. Failure results from high cyclic bending of a few critical components. Failures often result local effects - stresses caused by "stress risers". These are structural discontinuities (such as supports, card cutouts, or adjacent components). Fatigue failures within any structural system vary widely, even within apparently identical structures. Survival is further complicated by unavoidable variations in geometric features and material properties of electronic parts.
The empirical formula approach served well when electronics companies could not afford the needed computer power and expertise to evaluate their products. No equation is capable of covering all the configuration variations and modal response combinations found in modern electronics. But today, desktop computer power and computer software programs make that capability ... to evaluate vibration life ... available to all. The CirVibe software package is an example of a PC Windows tool built for one purpose - vibration life analysis of electronics. CirVibe software analyzes electronics devices and structures, recognizing the complexity of non-precise mechanical construction and configuration. When interfaced with Electronic CAD (e.g. Mentor Graphics, Cadence Allegro - which supply geometric detail) and coupled with CirVibe’s Life Usage / Expectations method, users quickly learn to create fully detailed models in a few minutes. The computer's detailed analysis may take several unattended hours, but it provides detailed information. It rates life capabilities at component level. It identifies the modes which dominate failure of each. With this understanding, all steps in the vibration development process become more efficient. This is especially true for testing. Testing programs, which tend to be very expensive, become far more efficient because they are now fully understood. Some think of vibration ESS as black magic, mystically causing flaws to turn to failures. Do we understand that flaws become failures by application of stress cycles? Understand your system at stress level and you understand the process. The electronics industry has a name for failures for which no cause can be determined. They are called “Cannot Duplicate Failures” (CND’s) or “no fault identified” or “re-test OK”. Let us cease to need or use these terms. Let us support our tests with proper, effective analyses of our electronic systems. ESS is called a process (unfortunately a trial and error process). ESS supported by proper analysis is an informed process. Ed Abner at CPI Satcom designs, develops and tests (HALT/HASS/ESS/QUAL) satellite communications equipment. He uses CirVibe in his development processes. His responsibilities include ruggedization of equipment and upscreening of COTS to meet the MIL Specs, Bellcore specs, European/Asian/Russian specs, IEC/EMC/RFI specs and others. These specs are often more severe than MIL-Specs. The demands on his schedule require Ed to understand his products with minimum time investment. Ed incorporates CirVibe analysis early and throughout the development process. This has helped him to meet tough development schedules and gives him confidence in the capability of his products. An example: the Outdoor LPA Hub Mount TWT Low Power Amplifier. Defining a vibration test plan for electronic systems with confidence in the outcome of the test program creates the most efficient and cost effective use of the test facility. It also allows quick recovery from unexpected results. It allows virtual evaluation of many alternate solutions to problems that do occur. Virtual evaluations are performed in a small fraction of the time otherwise required to create design revisions and test. J. Starr is a registered professional engineer in the state of Pennsylvania with 33 years of continuous and varied experience in structural capabilities in Nuclear, Chemical, and Defense industries. He works as a Consulting Engineer at CirVibe Inc. To contact John send an e-mail to starr@equipment-reliability.com To contact Ed Abner: |
| ******************************* |
|
Reliability
and Dependability Most accelerated testing methods currently in use are neither adequate nor accurate. Tests must be designed to reflect the synergistic/combined effects of environmental (such as vibration, temperature, moisture and contaminants) conditions, which are likely to cause failures and shorten the life of electronic systems. Performance Based Specifications have traditionally been and are presently negotiated between the military or other end user and the manufacturer. Military or other end users demand "proof" that the proposed system will perform as designed for a specified length of time. These contractual arrangements plus the rapidly expanding use of COTS are creating serious conflicts including the inability to trace and/or track the manufacturing history of COTS components and confusion i.e. how to validate the reliability of COTS components. Reliability engineering, as currently practiced, is heavily based on statistics. The accuracy of reliability-predicting statistical analysis which uses field failure data to predict reliability has long and vigorously been questioned. Certainly there are various formulas and software programs in use, which claim to "predict" the reliability of electronic products. Such predictions cannot be any better than the numbers programmed into them. Inaccurate and inadequate tests and very limited field failure data (if any) will create highly optimistic reports that can not accurately predict reliability. The only value of such predictions is to obtain contracts, to meet contractual milestone requirements for payments, and for advancing an existing project (SAE ARP 5890, Aug '01 - Guidelines for Preparing Reliability Assessment Plans for Electronic Engine Controls). This SAE document has been adopted by the IEC (International Electro technical Commission), to replace the widely discredited MIL-HDBK-217. Reliability engineering has historically been based upon calculations and upon predictions. These bases have compromised too many tests as well as the reporting (if they get reported at all) of test and field failures. Another historically based (unfortunately it still exists) engineering practice also limits the development of adequate reliability tests. Reported test failures determine the amount of money spent and type of engineering effort devoted to correcting problems. This process is sometimes called "Tombstone Engineering", which has a built in "Catch-22". Lacking adequate tests there will be no failures to report, if there are no reported failures there are no problems, if there are no problems we don't need different tests! The electronics industry is developing new technology far too rapidly to wait for reports of field failures. Once a product is "in the field", it is always difficult to get accurate failure data, even during warranty periods. Later, beyond warranty periods, there is no incentive at all to actively pursue or report failures. Further, the reporting of field failures by field personnel can lead to accusations (and worse) of "airing dirty laundry". Gaining an understanding of and taking steps to correct life limiting failures can reduce a manufacturers profits. Correcting problems (such as test and field failures) reduces the need for redundancy and for replacement assemblies for "worn out "systems, thus depriving the manufacturer of follow-on orders that can be more lucrative than the original contract. Dave Douthit has over 30 years experience in troubleshooting and repair of electronic circuitry and systems. He is an electronics reliability expert. To contact him, send an e-mail to douthit@equipment-reliability.com Dave is preparing a course about Reliability and Dependability to teach at ERI next year. But, for now, he has a quiz for you. Try to answer the 14 questions below and then click on "Answers" to confirm if you are an expert in Chemistry. If you would like to let us know your score, click here to send us an e-mail.
|
| ******************************* |
|
Now
is the time for training This is written for "the boss". Are things a bit slow at your place of business? Congratulations! Now you have time to accomplish goals for which a year ago you were too busy. Goals such as training your people. Beat them to the punch. Don't wait for your people to remind you of the excuse you gave them a year ago: "Sorry, we just can't spare the time." Demonstrate your own confidence. There's nothing like a modest investment in training to demonstrate your own confidence in the firm and in your people. You want to keep those valued people, the ones you didn't lay off. Right now, some of them are thinking about other jobs or even changing careers. Move quickly, before you lose them. Hire fewer consultants. Are a few of your retirees, whose specialties a year ago you needed, still coming in, perhaps as consultants? Now is your chance to dispense with them and save some money. Train your present staff in those specialties. Things to say. Here are a few ideas for you to rephrase to fit your own situation and temperament. Possibly say them to one or more groups. Possibly say them to individuals. Hopefully, you can make points like these to your present staffs:
A major benefit: your team will be ready to move fast when business picks up. It will pick up, you know. It always has picked up. It always will pick up. A reminder: immediately, when individuals attain that training, reward them with at least a token increase in take-home pay. You might say something like "I'm sorry we can't afford more right now, but we want to show that we appreciate you." Wayne Tustin, ERI's president, can be reached at tustin@equipment-reliability.com or at 805/564-1260. |
| ******************************* |
|
Questions our readers have asked... This section of our newsletter was created for you, reader! Feel free to send questions or suggestions to the webmaster. They will be either responded by Wayne Tustin or forwarded to one of our specialists. Here is the question for this issue: A visit to "the lab" is an important part of vibration and shock training. The classroom is certainly important for ease of hearing and for gaining understanding of theory. But a laboratory demonstration, using a controllable vibratory force (a shaker), can instill real understanding of some topics better than can the instructor's mouth and technical notes. For example, in the lab visit we use an electrodynamic shaker. We mount accelerometers at various locations, using one of them as "control". Then we conduct a simulated random vibration test, plotting a first graph in Tu or g²/Hz vs. Hz. To the surprise of many visitors, there are marked differences in the spectra generated by the several different accelerometers. Only the spectrum generated by the "control" accelerometer matches the desired spectrum. Why is this? To find out, we change the shaker from "random" to "sine" and do a 1g sine sweep, plotting a second graph in g vs. Hz. At the "major difference" frequency noted on the first graph, we again find major differences. We manually adjust shaker motion to that frequency and maintain it while observer fingertips feel what is happening: diaphragming or "oil canning" resonance, with maximum motion at table center. Internal portions of the shaker armature are not usually visible, so the group examines a spare armature. All agree that it looks very stiff. But now they realize that it is not so stiff after all. From that realization it is only a small step to considering their own hardware, which is much less stiff and must therefore resonate at lower forcing frequencies. They also realize that they must monitor vibration tests with several accelerometers. And they realize that when gathering vibration and shock data in the field (automobile, airplane or other vehicle or machine), they should use more than one accelerometer. Additional nearby accelerometers will probably record somewhat different spectra. For another demonstration (probably the next day), we use the same electrodynamic shaker, but now we drive a "slip plate". Again, we mount accelerometers at various locations on the plate, and use one of them as "control". Again, we conduct a simulated random vibration test. We plot a third graph in Tu or g²/Hz vs frequency in Hz. Again, we find marked differences in the spectra generated by different accelerometers. Only the spectrum generated by the "control" accelerometer matches the desired spectrum. Why is this? To find out, we change the shaker from "random" to "sine" and do a 1g sine sweep, plotting a fourth graph in g vs. Hz. At the "major difference" frequency noted on the third graph, we again find major differences. We manually adjust shaker motion to that frequency and dwell while observers feel what is happening: "stretching of the vibrating system". That system includes the shaker armature, the "slip plate" and possibly an intermediate "bull nose" structure that joins the shaker armature to the slip plate. The overall length L of that system is critical. This stretching should occur at a frequency of about 49000 divided by system length in inches or about 1250 divided by system length in metres. Then comes group realization that their hardware field (automobile, airplane or other vehicle or machine) must also stretch (and flex and twist) at certain frequencies. For some participants, lessons such as these are better learned in the lab than in the classroom. That's why we like to take every group into the lab for these (and other) demonstrations. It's also why our home study CD lessons tell students how to conduct these demonstrations in their own shaker labs. I'm especially gratified when a product design specialist admits he has never before visited the lab to which he has been sending prototypes for test. "Now I understand what the test engineer was trying to tell me. Next time I'll go see him, in our own lab, early in my projects."
|
| ******************************* |
|
Welcome
to ERI News volume 5. Here is a little "rundown" on the articles
in this issue. Way down the page is a quiz on climatics testing
of electronics assemblies; let's see how well you can do. 
|
|
|
Vibration
and Shock courses coming up
|
|
Pico
Rivera Santa Barbara, California, February 11-13, 2002 Huntsville,
Alabama, Livonia,
Michigan, Baltimore,
Maryland, |
|
Computer-based
Training
|
One way to study vibration without travel is to participate in ERI's "computer-based training" (distance learning) Vibration and Shock Program. Shortly after you register, you will receive a CD-ROM. It brings you several thousand Power Point slides divided into 31 lessons, text slides, colorful line drawings, color photographs, video clips. Also 31 sets of review questions and problems, which you e-mail to "Mr. Random Vibration", Wayne Tustin, for grading (and possible assistance). See details at http://www.vibrationand shock.com/distance.htm |
|
Training
without travel
|
|
We ask reader comments and suggestions. How do you
feel about alternative ways to obtain training, such that you need
not travel to obtain that training? Please e-mail editor@equipment- |
|
Witnessing
Environmental Tests
|
|
|
|
Announcements
|
|
COTS in San Diego Fixture group
to meet Fuel Cell Technology Seeking teacher |
| A Question for you |
"What is shock transmissibility and is there such a thing?" We would like to encourage you to answer this question and share it with us. Please e-mail Wayne and help us to better define this term. |
| Check our Glossary |
Check our Vibration and Shock Glossary. We have just added new words and their definitions. This list evolved from Wayne's 50 years of work experience and it is constantly updated. |
|
Contact
information
|
|
Wayne Tustin tustin@equipment- Webmaster webmaster@equipment Web sites Copyright © 2000-2001 Equipment Reliability Institute. All rights reserved. |
|
Free
Newsletter
|
|
Recommend Format Previous issues Unsubscribe |