ESD damage to a GMR head can occur when a very fast (1 nanosecond) current transient (500 mA) is permitted to pass through the GMR head. Under some circumstances, such an ESD event has been shown to provide sufficient energy to produce a damaging current transient.

It has been widely written that GMR sensors can be easily damaged by an ESD event. In particular, damage has been noted by a significant change in amplitude and in the pinned layer orientation.^1,2 Head-stack assemblies (HSA) have been tested to judge the correlation of GMR condition and the severity of an ESD event.³

This article examines the condition of GMR heads that have been subjected to ESD events up to 30 kV using an ESD gun. The drives were evaluated at the final assembly level without the printed circuit board (PCB) attached. The hard-disk assemblies (HDAs) were equipped with GMR heads and single-ended preamplifiers. In a manner similar to evaluations of HSAs, the HDAs were stressed using an ESD gun at various strategic points of the drive. The PCBs were in place to direct ESD events to the base, but removed to attack the motor pins. The GMR head condition was evaluated by examining the signal amplitude and GMR resistance using a quasi-static tester (QST). The QST was able to evaluate the GMR head condition without disassembly of the drive. This article compares the condition of the GMR heads before and after ESD stress.

The drives tested in this study use single-ended preamplifiers. To judge the GMR change, peak-to-peak head amplitude and GMR resistance were reviewed using an ISI QST 2001 tester.⁴ In some cases, a kink (area of varying slope) in the amplitude versus the external magnetic field strength (B) field trace can signal a weakened GMR head, which leads to unstable performance. As a possible indicator of GMR change, the head's output-signal-amplitude (µV) response to varying magnetic fields (B [Oe]) was reviewed at each stage of the exercise.

The 3.5-in. HDAs were six-head models with 10.26 Gbyte per platter (5 Gbyte per surface or head) and 21 K tracks per inch. The ISI QST 2001 tester was used to measure the resistance and peak-to-peak amplitude before and after ESD stress testing.

A KeyTek MiniZap ESD gun (up to 17 kV) in the air discharge mode simulated hand and metal discharge according to the requirements in IEC 801 (now IEC 1004-2).⁵ A Hewlett-Packard ESD gun capable of up to 30 kV was used to discharge to the cover and base.

The HDA was placed on a surface covered with dissipative material connected to ground. The HDA was grounded using a wire from a cover screw to ground. With the PCB installed, the HDA was zapped with 30 kV to many points spread over the entire area of the cover and sides of the casting, and the edges of the underside of the PCB's mount side. The PCB itself was avoided (see Figure 1). With the PCB removed, the motor pins were also zapped. The motor pins were zapped at various voltage levels between 5 and 15 kV using the MiniZap ESD gun.

Cover and Base. Table I shows a population example of the resistance and track-averaged amplitude behavior versus a 30-kV zap and a sample of the virgin HDA population (prezap of any kind). Figure 2 is a graphical representation of the data shown in Table I. The median within the box is identified. The box ranges represent from 25 to 75% of the data (50% of the data within the box). The legs extending from the box represent the nonoutlier maximum and minimum data. Based on the sample populations, the GMR resistance and amplitude of the prezap group are considered equivalent to the sample after a 30-kV zap to the HDA body.

Motor Pins. With the PCB removed, the spindle motor pins and the head stack are exposed. The head stack pins were avoided. The HSA had been evaluated earlier for sensitivity to an ESD attack.³

The HDA was placed on a dissipative table surface and hard grounded to the cover screw. The MiniZap ESD gun was used to hit the motor pins with voltage levels of 5–15 kV. It was noted that heads in all positions exhibited a change in resistance and amplitude. A change in amplitude occurred at the lower voltage levels. Head position 0 is nearest the base. Figures 3, 4, 5, 6, 7, 8 graphically represent the change in GMR resistance and amplitude measured against the voltage level of the zap to the motor pins. The figures show head positions 0 through 5.

A very high energy ESD event directed at the cover and base of a complete HDA did not appear to alter the resistance or the amplitude of the GMR elements. Using the QST, the relationship of amplitude versus field strength was reviewed before and after the 30-kV event. The field was varied from –100 to +100 Oe. Damage levels of less than 30% were not detected by amplitude loss alone. Signal sensitivity can be lost before wholesale changes in amplitude. By using the QST to view the amplitude response to field over a wide range of field strengths, less-severe damage to the GMR element is more likely to be detected.

A simulated ESD event directed toward the motor pins of a complete HDA (with the PCB removed) appears to damage the GMR elements. The amplitude is damaged more severely than the resistance. All of the heads were damaged, although the heads nearest the base (heads 0 and 1) exhibited damage more consistently.

Quasi-static testing with consideration of amplitude, resistance, or the amplitude response to field strength (–100 Oe to +100 Oe) did not show any evidence of damage to the GMR element. A complete HDA is virtually impervious to ESD events directed to the cover and base (with the PCB removed). These samples have been subjected to a 30-kV event without apparent damage.

The motor pins, however, display some sensitivity to ESD events. Approximately 8 kV directed toward the motor pins of a complete HDA (with the PCB removed) causes a considerable reduction in amplitude, significantly damaging the GMR element. Substantial changes in resistance can be observed when ESD events of greater than 10 kV are directed at the motor pins.

The author wishes to thank Alfredo Larios for substantial assistance in taking the data; Al Wallash for inspiration and consultation; and Wen-San Lee Morgan for considerable assistance obtaining samples.