In the late 1970s, StorageTek, an industry leader in the production and integration of large-scale tape and disk drive systems, decided that its products were experiencing an unacceptable incidence of shipping-related failure and damage. The company resolved to correct the situation by implementing a rigorous vibration test program involving early and frequent interaction between test and design engineers.
Over the years, StorageTek’s program has ensured that its complex products reach the customer damage-free and ready to perform as specified. Here are the details about their program, focusing on the use of vibration controllers to evaluate sensitive subassemblies of disk and tape drives.
Inherent Susceptibility to Shock and Vibration
Manufacturers of tape and disk drives know that the vast majority of their product returns and warranty claims involve systems damaged by shock or vibration. Vulnerability to such damage is related to basic factors inherent in the way disk and tape drives function.
Disk drives, for example, operate with extremely tight tolerances, less than 3 millionths of one inch between the read/write head and the disk that rotates at speeds up to 15,000 rpm. The combination of high speed and tight tolerances in disk drives has been compared to flying a 747 at full speed about 1 ft above the runway.
On the disk surface, the recording tracks are so close together that vibration during operation can all too easily cause loss of or damage to data. These are common types of shock and vibration damage to disk drives:
· Head slaps, which occur when the recording head slams down with destructive force on the disk surface.
· Overstressing sensitive components such as springs and cantilevered parts of the mechanism. This damage can occur during a nonoperating condition when the drive is being shipped to the end user or during operation at the customer’s site.
Disk drives are more sensitive to damage when they are operating, but they are more likely to experience strong shock and vibration during the rigors of shipping, handling, and installation. Tape drives tend to be more rugged than disk drives. However, when slammed around in a truck or on a loading dock, tape drives can suffer head misalignment which prevents proper read/write operation or sustain cosmetic damage to the frame and cover.
The sensitivity to shock and vibration is magnified when multiple disk or tape drives are installed in the same unit. StorageTek specializes in multiple-drive units. Its products range from desktop or deskside cabinets standing about 30 in. high and integrating up to 128 disk drives to 8-ft tall cabinets used at banks or other large institutions and containing up to 6,000 tapes and eight tape drives. The diversity, size, and sensitivity of these systems give the company many reasons to be concerned about shock and vibration damage.
Integrated Approach to Testing
The test programs implemented at StorageTek are not radically different from those in place at other quality-conscious companies. However, they may be more comprehensive and more tightly integrated with the entire process of product development, delivery, and usage at the customer’s site.
All products and subassemblies designed and produced by StorageTek undergo a rigorous set of tests including:
· Horizontal and vertical vibration and horizontal and vertical shock.
· Low-level shock and vibration.
· Highly accelerated life tests (HALT) performed under hot/cold cycling conditions.
· Environmental testing for sensitivity to temperature, humidity, and in some cases altitude and dust.
· A broad range of handling tests such as tilt and loose-cargo tests, rolling on rough surfaces such as diamond decking, bumping, tipping, and falling from forklifts.
· Tests of systems in their shipping packages including stacking, compression, edge-on dropping, and package-material burst testing.
Role of the Shock and Vibration Lab
Believing that a 10-cent fix at the design stage can prevent a $10 problem at the customer level, StorageTek’s shock and vibration lab gets involved in the early stages of product prototyping and design. A test engineer is an integral part of each design team. After testing early mechanical prototypes for fragility, the lab feeds its data back to the design engineers, who use the data to develop more robust products and correct problems at the earliest possible time.
All StorageTek products pass through Mechanical Accelerated Stress Testing (MAST) at various stages of development including:
· Mechanical evaluation of machines to assist mechanical design.
· Prototype evaluation to assist in package design.
· Preproduction for operational testing and package evaluation.
· Field Repair Unit (FRU) and subassembly.
With this program, the packaging lab reports to the manager in charge of shock and vibration testing. It develops protective packaging for each system, then tests packaged prototypes for survivability in the shipping environment (most often by truck or rail, occasionally by air). The goal is to detect inherent product weaknesses and fix them at the prototype stage rather than to band-aid the problem by using a great deal of expensive packaging to protect the product during shipment.
Role of Electrodynamic Shakers
The shock and vibration lab uses electrodynamic shakers primarily to test subassemblies such as power supplies and logic chassis for disk drives, robot-arm and tape-handling assemblies for tape drives, and operator control consoles for both types of drives. Some of StorageTek’s smaller systems, such as desktop or deskside units, also are tested fully assembled and in their cabinets on electrodynamic shakers.
The tests involved are classical shock, random, and sine vibration profiles. Searches also are conducted to detect resonant frequencies within a product or subassembly. Testing in the lab is based on real-world data, gathered by mounting shock recorders on packages as they undergo shipment by rail, truck, or air to customer sites in the United States or Europe.
To develop a vibration controller to facilitate the process of translating real-world data into test profiles, StorageTek worked closely with Dactron. Lab personnel were involved in developing the new controller, including definition of its feature set.
One vital feature was user-friendliness. Thirteen test engineers work at the StorageTek lab, and management wanted a controller that would enable any one of them to run a vibration test in a way that could be repeated by any of the others. Another important criterion was the capability to store test profiles and later recall them with ease.
The desire for user friendliness also pointed the lab toward a system that offered PC-based, Windows-driven control of shock and vibration testing. A common interface added continuity from one piece of test equipment to another and shortened the learning curve.
With previous generations of control systems, the lab encountered frustrating delays in compiling test data and printing it in formats useful and timely for design engineers. Dactron’s shaker controller solved this problem with special capabilities for report generation.
From within the system’s software, an operator can preselect the contents of a test report, such as setup parameter listings, data plots, and the test log. The software automatically opens a Word document and inserts the selected data and listings. This feature not only saves time, but it also allows StorageTek’s lab to quickly provide graphics and other documentation to their product design teams.
Example Case: Rotary Vibration Testing
Perhaps the most challenging test performed at StorageTek is the rotational vibration testing of hard drives. For disk drives, rotational vibration causes more problems than translational vibration. Head gimbal assemblies are very susceptible to rotational vibration since this type of vibration easily knocks the head off the track causing read/write errors. A simple bump on the corner of a computer is sufficient to cause read/write errors under the right conditions.
A specially designed electrodynamic shaker system performs the rotational vibration tests. It uses a round table mechanically linked directly to the shaker armature. A cam assembly allows up to +15° to -15° of rotation.
The rotary shaker can drive test articles at angular accelerations from 9 to 40,000 radians per second squared (rad/s2). At 40,000 rad/s2, the shaker’s rated maximum output of 200g is required.
The electrodynamic shaker has many advantages over spring-loaded test methods. Most importantly, when used with the Dactron control system, it performs closed-loop control with full compensation for the dynamics of the shaker and table. It also provides much greater flexibility since both negative and positive polarity shock pulses can be run with a mouse click, and shock, sine, and random tests can all be performed with the same test system.
This shaker system, however, puts additional demands on the control system. A high sampling rate is required for faithful shock pulse reproduction and the mechanics of the table, and the linkage assembly demands a wide dynamic range to achieve accurate control.
Industry-standard rotary vibration tests include shock with pulse durations from 0.5 to 6 ms, swept sine tests over a 5- to 500-Hz frequency range, and random vibration for a bandwidth of 5 to 500 Hz. The drives are tested in both the operational and nonoperational modes. Functional testing of the drives during operational-mode vibration verifies and records the reliability of data reads and writes for the drives.
Figure 1 shows results from a typical shock test. The pulse is a 2-ms half-sine pulse with a peak amplitude of 20k rad/s2. A triaxial accelerometer mounted on the rotating table provides the control signal and characterizes cross-axis translational acceleration. One accelerometer measures the tangential acceleration of the table and is used as the control signal.
Signals from the other two accelerometers characterize the up/down and right/left motion of the table. It is crucial to know the cross-axis translational component during rotary testing to determine failure mechanisms when read/write errors occur.
Other accelerometers often are mounted on the disk drive chassis to characterize the drive’s structural response during vibration. Swept sine and random tests are similarly instrumented for control and structural response characterization. Figures 2 and 3 show typical results for swept sine and random vibration tests, respectively.
Quantifying the Results
In a field where most players hold quality figures close to the vest, it is difficult to compare the reliability of competing storage systems. The shipping industry, however, does release figures on damage to disk and tape drives.
StorageTek estimates that its test programs have reduced shipping damage to about 10% to 15% of the level sustained by leading competitors. The trucking segment of the industry pegs shipping damage at about 2.4% of the total dollar value of shipments. The StorageTek damage figure is about 0.5% of the shipment value or one-fifth the industry average.
About the Authors
Steve Kajewski is the manager of StorageTek’s Advanced Packaging Technology Lab. He is a mechanical engineer with 35 years of engineering and management experience. StorageTek, 1 StorageTek Dr., MS 0196, Louisville, CO 80028-0196, (303) 673-6838.
Dave Galyardt is the marketing manager at Dactron. He has more than 20 years experience in vibration test, with a special focus on the design and marketing of products for vibration control and analysis. Dactron, 1629 S. Main St., Milpitas, CA 95035, (408) 934-9160.
Copyright 2000 Nelson Publishing Inc.
January 2000