Device Sensitivity and Testing

Two of the key elements in any successful static-control program are the identification of those items—whether components, assemblies, or finished products—that are sensitive to ESD, and the determination of the level of their sensitivity. The damage done to an electrostatic-discharge-sensitive (ESDS) device by an ESD event will depend on the device's ability either to dissipate the energy of the discharge or to withstand the current levels involved. This is known as device ESD sensitivity or ESD susceptibility.

Certain devices may be more readily damaged by discharges occurring within automated equipment, while others may be more prone to damage from handling by personnel. In this article, we will review the models and test procedures used to characterize, determine, and classify the sensitivity of components to ESD. These test procedures are based on the three primary models of ESD events: human body model (HBM), machine model (MM), and charged device model (CDM). While the models employed to perform component testing cannot replicate the full spectrum of all possible ESD events, they have proven to be successful in reproducing over 95% of all ESD field-failure signatures. The use of standardized test procedures has allowed the industry to:

One of the common causes of electrostatic damage is the direct transfer of electrostatic charge through a significant series resistor (±1.5k ohms) from either the human body or a charged material to the ESDS device. When a person walks across a floor, an electrostatic charge accumulates on his or her body. Simple contact of a finger to the leads of an ESDS device or assembly permits the body to discharge, possibly causing device damage. The model used to simulate this event is called the Human Body Model, or HBM.

HBM is the oldest and most commonly used method of classifying device sensitivity to ESD. The testing model, which represents the discharge delivered to the device from the fingertip of a standing individual, comprises a 100-pF capacitor discharged through a switching component and a 1.5k-ohm series resistor into the component. Dating from the nineteenth century, this model was originally developed for the purpose of investigating explosive gas mixtures in mines. It was adopted by the military in MIL-STD-883 Method 3015 and is also used in ESD Association standard ESD-STM5.1-1998—Device Testing: Human Body Model. The HBM circuit is illustrated in Figure 1.

Figure 1. Typical human body model (HBM) circuit.

Testing for HBM sensitivity is generally performed using automated test systems, with the devices being placed in the test system and contacted through a relay matrix. After ESD zaps are applied, the post-stress I-V current traces are reviewed to see if the device has failed. The ESD Association's HBM test standard has been revised and includes several new technical changes. First, the number of zaps per stress level and polarity has been reduced, from three to one. The minimum time interval between zaps has also been reduced, from one second to 300 milliseconds. Taken together, these modifications serve to cut the total HBM qualification test time.

The second major technical change is a revision in the HBM tester specifications: the maximum rise time for an HBM waveform measured through a 500-ohm load has been increased from 20 to 25 nanoseconds. This will allow manufacturers of HBM test equipment to build high-pin-count testers, which typically have a higher parasitic test-board capacitance that slows down the 500-ohm waveform.

Another kind of discharge, similar to the HBM event, can occur from a charged conductive object such as a metallic tool or fixture. Originating in Japan as a result of attempts to create a worst-case HBM event, this ESD model, known as the machine model (MM), consists of a 200-pF capacitor discharged directly into a component, with no series resistor (see Figure 2).

Figure 2. Typical machine model (MM) circuit.

As a worst-case HBM, the machine model may be overly severe. However, there are certain real-world situations that this model represents—for example, the rapid discharge from a charged board assembly or from the charged cables of an automatic tester.

The testing of devices for MM sensitivity using ESD Association standard ANSI/ESD-S5.2-1994—Device Testing: Machine Model is much like HBM testing. But while the test equipment used for the two models is the same, the test head is slightly different, in that the MM version does not have a 1.5k ohm resistor. (The test board and the socket replicate those employed for HBM testing.)

The transfer of charge from an ESDS device is also an ESD event. A device may, for instance, become charged when sliding down the feeder in an automated assembler. If it then contacts the insertion head or some other conductive surface, a rapid discharge may occur from the device to the metal object. This so-called charged device model (CDM) event can be even more destructive than the HBM event for some devices. Although the duration of the discharge is very short—often less than one nanosecond—the peak current can reach several tens of amperes.

Several test methods have been explored to duplicate the real-world CDM event and replicate the conditions that have been observed in CDM-caused field failures. Efforts in this area are currently focusing on two separate test methods. The first, known as CDM, better simulates an actual charged-device event, while the second addresses devices that are inserted into a socket and then charged and discharged in the same socket. This second method is termed the socketed discharge model, or SDM.

A draft standard for CDM, designated ESD-DS5.3.1-1996—Device Testing: Charged Device Model, was released in 1996. (Work is ongoing to release a full standard in the near future.) The test procedure involves placing the device on a field plate with its leads pointing up, then charging and discharging it. Figure 3 illustrates a typical CDM test setup.

Figure 3. Typical charged device model (CDM) test circuit.

SDM testing is similar to testing for HBM and MM sensitivity. The device is placed in a socket, charged from a high-voltage source, and then discharged. This procedure remains a work in progress and still has a number of limitations, including too great a dependence on the specific design of the SDM tester.

Each of the device-testing methods provides a classification system for defining the component's sensitivity to the specified model (see Tables I, II, and III). These classification systems have a couple of advantages: first, they allow for easy grouping and comparison of components according to their ESD sensitivity; second, they give an indication of the level of ESD protection required for a particular component.

A fully characterized component should be classified using all three models—the HBM, the MM, and the CDM. For example, a fully characterized component may have the following designations: Class 1B (500 to < 1000 V HBM), Class M1 (25 to < 100 V MM), and Class C3 (500 to < 1000 V CDM). This would alert a potential user of the component to the need for a controlled environment, whether assembly and manufacturing operations are performed by human beings or by machines.

A word of caution is in order, however. These classification systems and component-sensitivity test results should function as guides, not necessarily as absolutes. The events defined by the test procedures produce narrowly restrictive data that must be carefully considered and judiciously used. The three ESD models represent discrete points developed in an attempt to characterize ESD vulnerability. The data points are informative and useful, but their arbitrary extrapolation into a real-world scenario can be misleading. The true value of the data lies in comparing one device with another and in providing a starting place for the establishment of an effective ESD-control program.