Environmental
Last Updated: Apr 9th, 2008 - 15:00:00
|
See this article in our Digital Edition
Download a pdf of this article (right click to save the pdf)
Finding and correcting the weaknesses in a product during the design phase is Engineering 101, and is adopted by most design engineers. Environmental stress screening (ESS) is a very effective tool for finding mechanical/electrical weaknesses in a product; however it is underutilized. Maintaining the reliability of a product throughout its life, however, has proven to be more of a challenge, though it is widely known that doing so can significantly reduce costs and increase your company’s reputation as a quality conscious supplier who produces a superior product.
Because of the long screening time currently required for ESS during manufacturing, and the high cost associated with it, most companies choose not to include this very beneficial process. After over 13 years of development, a new chamber/vibration table system that can reduce the ESS process time is now being manufactured to help alleviate this problem.
The Evolution of the ESS Process
The evolution of the ESS process came about because companies were seeing products arrive at their customer sites “dead on arrival,” even after putting their products through design verification and/or production testing. Although they went through burn-in, temperature cycling, vibration and shock testing individually, they could not uncover all the design weaknesses and weed out manufacturing defects.
A new approach in which temperature cycling and vibration testing is conducted concurrently has been proven to be more effective. This process is called environmental stress screening (ESS).
How the System Works
The ESS system discussed in this article has a chamber with doors on each side that open and close, and an electromechanical shaker with a skewed table in the center of the chamber. The system enables the operator to conduct thermal cycling evaluations on a given product while imposing vectored vibration input on all 3 axes, providing an effective and efficient method of screening.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5: Exploded view of the accordion assembly when it is pulled apart
Figure 5 shows the exploded view of the accordion assembly when it is pulled apart.
The system uses a temperature chamber with openings on each side. Printed circuit boards (PCBs) are loaded in between support plates and are pushed together on standoffs, similar to their mounting in a finished product. Compressed air shaft collar locks are employed to hold the PCBs in position, and these specialized collar locks hold the shaft, and lock in the PCBs tightly on the standoffs.
Movement of the PCBs is done with a hovercraft transport system. An accordion fixture holds the PCBs in place, and the hovercraft moves them in and out of the sides of the chamber. The PCBs are easily loaded on the accordion fixture from either side of the chamber. The hovercraft transport system on the accordion fixture makes it easy to move the fixture from the loading table to the chamber.
Once the units are in the chamber, the doors close and the units are quickly fastened by vacuum onto the shaker table and are then vibrated at cold and hot temperatures. While screening is being performed on one group of boards, new boards are being unloaded and reloaded outside the chamber. Once the screened PCBs are done, the boards are unloaded and the new PCBs are moved into the chamber for screening. The process is then repeated.
Figure 6
There are 2 main ESS systems available today; the combined temperature chamber with an electromagnetic, skewed shaker table, and the pneumatic hammer shock table system. These ESS systems apply very different technologies, but to date, the effectiveness and cost of screening by each system have not been compared.
The two main differences between the skewed fixture chamber system and the pneumatic hammer system are:
- On the skewed fixture chamber 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. The reason why low frequency is more important to product testing is because high frequency displacement is much lower than low frequency displacement. In order to precipitate defects, larger displacement for high frequency displacement is very low. The displacement at high frequency is much lower than low frequency (just like bending a wire).
- The vibration intensity across the skewed fixture is uniform, compared with the pneumatic hammer system, which can evidence large variations across the table [2].
Because of these advantages, the skewed fixture chamber system should be able to precipitate defects at a much lower overall vibration level in a much shorter time, thereby increasing product throughput and reducing the cost of screening.
In order to evaluate the merits of these very different systems, a comparison study was performed, executing highly accelerated life testing (HALT) using the skewed fixture chamber system on three products of different sizes: a printed circuit board, a power supply and a large system (chassis). All three commercial products had previously undergone HALT on a pneumatic hammer ESS system. However, since the temperature component was not utilized during the pneumatic hammer system testing, we duplicated the screening process using only the vibration table portion of the HALT chamber system. Efforts were made to ensure that conditions during the HALT evaluation using the skewed fixture chamber were identical to the HALT evaluation using the pneumatic hammer system.
The Results
The skewed fixture chamber ESS system was able to precipitate defects at a much lower overall vibration level in a much shorter time, thus increasing product throughput and reducing the cost of testing.
Product # 1 – Printed Circuit Boards
The HALT test using the skewed fixture chamber system was designed to duplicate the original test performed with the pneumatic hammer system, except the frequency range for vibration was 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, and was not included as part of this test.) The response accelerometer model(s) used and their locations were exactly the same in both tests.
Figure 7
Figure 8
The results of the tests are listed in the Table 1.
| 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 |
Table 1
During both tests, the principal failure mode was that the same capacitor broke off. However, the pneumatic hammer system went to 35 Grms and took 70 minutes before this failure was found. Significantly, the electromagnetic shaker ESS system uncovered the problem at 5 Grms, in just 5 minutes.
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 Figure 8), but another identical capacitor, mounted immediately beside the broken one, did not break loose. Another test was performed with the skewed fixture chamber 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 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 system, the power supply tested was subjected only to random vibration stress.
Figure 9
Figure 10
The data from these two tests are shown in the Table 2.
| 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 |
Table 2
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 skewed fixture ESS system precipitated the same weakness at 7 Grms, and took only 32 minutes.
Product #3 – Network Device
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 skewed fixture system in order to minimize the number of variables.
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, under testing 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.
Figure 11
Figure 12
| 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 | ? |
Table 3
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 skewed fixture ESS system only about 25 minutes of vibration time to find the backed-out compact flash memory problem, and 28 minutes to fail the power supplies. The testing time required with the skewed fixture system was considerably shorter than the pneumatic hammer system, and failures were identified at a much lower vibration level.
Conclusions
The test results for the three different products demonstrated that the skewed fixture chamber system precipitated the weaknesses in a product in much shorter time and at lower vibration levels than the pneumatic hammer shock system. In addition, the skewed fixture chamber system also found additional defects that were not uncovered by the pneumatic hammer system.
The reason is that, for the same G level, the displacement is much larger at low frequencies than at high frequencies. For example, bending a wire (or any structure) thousands of times at a small displacement will not break it, but bending it at a large displacement a few times will break it.
A shorter time to precipitate hidden defects can mean lower cost for the ESS process and a shorter product development cycle, as well as greater throughput during manufacturing. In addition, effective lower random vibration levels would use less of the product’s life during the HALT process, leaving more useful life in the product for the customer. n
Dr. Hong-sun Liu is with Quanta Laboratories, and can be reached at hliu@quantalabs.com.
References
- New approach for production line Environmental Stress Screening by Dr. Hong-sun Liu, Test Engineering and Management, August/September 2005
- Environmental Stress Screening Equipment: search, evaluation, design, experimentation by Dr. Hong-sun Liu, August/September 1994
- Comparison Testing of Shock v/s Vibration ESS System by Dr. Hong-sun Liu, October 2007
© 2007 Conformity
Top of Page
|
