Conversion Table
Technical Articles:

Comparison Testing of Shock vs. Vibration ESS Systems

Principal Investigator: Dr. Hong-sun Liu, Quanta Laboratories

Contributors: Larry Foshee, Motion Engineering, Ron Weglinski, Susan Mercovich, Greg Wilterdink, Transistor Devices Inc.

It is well known that Environmental Stress Screening (ESS) is a very effective tool to precipitate the weaknesses of a product. Currently however, the most available equipment for this process is the pneumatic hammer (shock) system. The author has developed a new ESS system that utilizes a skewed fixture together with an electromagnetic shaker system (or a hydraulic shaker system) in conjunction with a fast ramping chamber (Ref 1). The advantages of this system are:
  • Frequency Control - The random vibration profile can be precisely controlled, i.e. it is able to cut off inappropriate high frequency energy while at the same time maintaining the low frequency energy needed to screen out damage that could be caused by shipping and operations in the real world.

  • Table Uniformity - The vibration intensity is uniform across the whole table because it has only one actuation device - the shaker.

  • Less Time and Money - All three axes are screened simultaneously on the skewed fixture, which saves test time and money.
These ESS systems apply very different technologies, and since the response of products to vibration or to shock are very different; but to date, the effectiveness and cost of screening by each system have not been compared. In order to evaluate the merits of these systems, such a comparison study has been performed by Quanta Laboratories. Quanta, in conjunction with three reliability conscious companies - Motion Engineering, Transistor Devices Inc., and a large network device company (who has asked to remain anonymous) established a joint project to perform HALT on three products of different sizes, a printed circuit board, a power supply and a large system (chassis). All three commercial products have previously gone through Highly Accelerated Life Test (HALT) on the pneumatic hammer system. Quanta’s responsible engineer and the technical team from each participating company worked together to ensure that the identical diagnostic system and vibration test conditions were used for the HALT comparison on each test unit. In order to make fair comparisons, the exact same model of each product was used in HALT testing on the new system as was tested in the previous HALT tests on the pneumatic hammer system. The major difference between the tests is that the vibration profile on the new system is controlled over the 5 to 500 Hz range, (even though this system is capable of frequencies up to 3000 Hz). The reason we concentrated on these lower level frequencies is that the test units were commercial products; it is unlikely that these units will ever see vibration levels much higher than 500Hz in their operating or transportation lives. However, for certain products for aircraft, missiles, fiber optic devices, etc., we would choose a spectrum between 5~20 to 2000 Hz.

The two main differences between the skewed fixture system and the pneumatic hammer system are:
  1. On the skewed fixture system, the spectrum can be precisely controlled, such that inappropriate high frequency energy can be removed and real-life low frequency energy can be correctly imposed in direct contrast to the pneumatic hammer systems

  2. The vibration intensity across the skewed fixture is very uniform as opposed to the pneumatic hammer system, which has large variations across the table (Ref 2).
Product #1 – SQID/Motion Printed Circuit Boards from Motion Engineering

Fig. 1. Unit on the skewed ESS system Fig. 2. Accelerometer mounted on the capacitor

The HALT test on the new system is designed to duplicate the original test on the pneumatic hammer system, except the frequency range for vibration is controlled over the range 5 to 500 Hz with a flat profile. The comparison unit did not go through thermal cycling with vibration when tested on the pneumatic hammer system, so for this comparison test, we also did not include temperature cycling as part of the test. The response accelerometer model(s) used and their locations were exactly the same in both tests. The results of the tests are listed in the following table:

ESS Testing System ED Shaker with Skewed Fixture Pneumatic Hammer System
Profile Flat Spectrum from 5-500 Hz Uncontrolled
Vibration Intensity Level (Grms) 5 5, 10, 15, 20, 25, 30, 35
Duration of Vibe at Each Level 5 min. 10 min.
Failed at 5 Grms 35 Grms
Total Time to Failure (min.) 5 70

During both tests, the principal failure mode was that the same capacitor broke off. However, the pneumatic hammer system went to 35 Grms before this failure was found. Significantly, the electromagnetic shaker ESS system uncovered the problem at 5 Grms. Note that the Grms values reported for the pneumatic hammer system are the mathematical average of the combined three-axis Grms values, filtered to a low frequency, usually to 2000 Hz, even though most of the energy for a pneumatic hammer table is between 2,000 to 25,000 Hz. Therefore, the Grms values seen by the product on the three axes are actually much higher than the calculated Grms values for the pneumatic hammer shock method. Thus, the calculated 35 Grms value is much lower than the actual G-level input to the product. It should also be noted that the broken capacitor had an accelerometer mounted on it (see Fig. 2), but another identical capacitor, mounted immediately beside the broken one, did not break loose. Another test was performed on the new ESS system without mounting the accelerometer on the capacitor, and the capacitor did not break off. Clearly, it was mass loading by the accelerometer that caused the failure of the capacitor. In a later test, the new system found a real problem, which was a break in the Ethernet ring; however, after re-plugging the connecter, the unit recovered.

Product #2 – Switching TDI Power Supply, Model # SPS4387

Fig. 3. Power supply on the skewed fixture Fig. 4. Failed power supply

For the HALT process on the pneumatic hammer shock system, the power supply was only subjected to vibration, so for a direct comparison using the skewed fixture HALT system, the power supply tested by Quanta was subjected only to random vibration stress. The data from these two tests are shown in the following table:

ESS Testing System ED Shaker with Skewed Fixture Pneumatic Hammer System
Profile Flat Spectrum from 5-500 Hz Uncontrolled
Vibration Intensity Level (Grms) 1, 2, 3, 4, 5, 6, 7 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
Duration of Vibe at Each Level 5 min. 20 min. at 1~5 Grms; 30 min. at 6~10 Grms
Failed at 7 Grms 10 Grms
Total Time to Failure (min.) 32 244

The same weakness was found in both tests (mechanical fatigue failure of the FET transistor); however, the pneumatic hammer ESS system required vibration steps up to 10 Grms and took 244 minutes to precipitate the weakness. The new ESS system precipitated the same weakness at 7 Grms, and only took 32 minutes.

Product #3 – Network Device

Fig. 5. Network device on skewed ESS system Fig. 6. Accelerometer locations on the unit

This chassis, when tested on the pneumatic hammer system, went through a high/low temperature step test, a vibration test and a combined vibration & temperature cycling test. Since the important difference between the skewed fixture ESS system and the pneumatic hammer system is in the spectrum and uniformity of the vibration imposed, only the vibration test was performed using the new ESS system in order to minimize the number of variables.

ESS Testing System ED Shaker with Skewed Fixture Pneumatic Hammer System
Profile Flat Spectrum from 5-500 Hz Uncontrolled
Vibration Intensity Level (Grms) 1, 2, 3, 4, 5, 6, 7, 8, 10 2, 3, 5, 10, 5, 15, 5, 10, 12.5, 15
Duration of Vibe at Each Level 3.5 min. 20 min. at 1~5 Grms; 30 min. at 6~10 Grms
Failed at 7 & 10 Grms 15 Grms
Total Time to 1st & 2nd Failure (min.) 24.5 & 28 ?

When this unit was tested on the pneumatic hammer system, the power supply failed at 15 Grms, and no other failure modes were found. However, when Quanta used an electromagnetic shaker with the skewed fixture ESS system, we found the compact flash memory came loose at 7 Grms (a weakness not uncovered by the pneumatic hammer system). Then at 8 Grms, there was a failure of one of the power supplies, but it came back when the vibration was stopped. When the random vibration level was stepped up to 10 Grms, less than 10 seconds after the start of the vibration both power supplies failed and could not be revived. Because of different vibration imposition approaches taken for each test, it was not possible to compare the total time it took to propagate the weaknesses by each of the systems. But the whole ESS process including the high/low temperature tests took the pneumatic hammer system roughly 3 days. It took the new ESS system only about 24.5 minutes of vibration time to find the backed-out compact flash memory problem and 28 minutes to fail the power supplies. The time it took is considerably shorter than the pneumatic hammer system vibration time and also the failures were identified at a much lower vibration level.

Conclusions

The test results for the three different products indicate that the electromagnetic shaker with the skewed fixture will precipitate the weaknesses in a product in much shorter time and at lower vibration levels than the pneumatic hammer shock system. It also may find additional defects that were not uncovered by the pneumatic hammer system. A shorter time to precipitate hidden defects means lower cost for the ESS process and a shorter product development cycle, as well as greater throughput during manufacturing. Effective lower random vibration levels will use less of the product’s life during the HASS process, leaving more useful life in the product for the customer.

It is hoped that with this advancement in the ESS process, more products will start to include the HALT/HASS process in their development and manufacturing cycles, so the overall reliability of the product will be significantly increased.

References
  1. New approach for production line Environmental Stress Screening by Dr. Hong-sun Liu, Test Engineering and Management, August/September 2005

  2. Environmental Stress Screening Equipment: search, evaluation, design, experimentation by Dr. Hong-sun Liu, August/September 1994



Factory Sales and Technical Support Center
Hastest Solutions Inc., 4423 Fortran Drive, Building 190, San Jose, CA 95134, USA
Tel: 408 945 7861 Fax: 408 262 6699
 Contact Form