An MD&DI December 1998 Column
EMI FIELD NOTESWhen electrostatic discharge cannot be prevented, effective shielding is key to protecting devices.
Electrostatic discharge (ESD), although troublesome for portable electronic medical devices, is controllable with effective shielding. Lack of complete shielding leads to compromises that leave devices open to several paths for ESD to follow. A recent article discussed avoiding ESD problems by implementing preventive measures.1 Preventing discharge to the enclosure eliminates the major ESD problem of direct discharge, leaving only indirect discharge, a much less severe problem.
It is not always possible, however, to prevent discharge. Metallic members may need to be exposed, or a metallic shield may be needed to control other electromagnetic interference. For cases in which discharge cannot readily be prevented, a means must be found to protect the circuits within the enclosure. It is important to keep in mind that if ESD cannot be prevented, then shielding problems multiply.
ESD has a rise time of less than 1 nanosecond. In that brief period, the voltage and associated electric field fall from the initial voltage (815 kV) to near zero, and the current and attendant magnetic field rise to 10 A or more. Both electric and magnetic field effects can be troublesome, as shown in Figure 1. In Figure 1a, the discharge is to a plane, with the charge capacitively coupled to internal circuits. In Figure 1b, the discharge is to a wire or trace, which may cause a voltage drop along the wire itself, and the current may couple magnetically to adjacent wires. Electric field coupling is dominant in larger devices, but in smaller devices, the current is the problem. This article concentrates on current.
Figure 1. Effects from discharge. I represents current; H is the magnetic field, which arises from current.
The current can follow several paths. The first, and simplest, path is when discharge occurs to a grounded metal member. The current density at the point of contact is very high, and hence the magnetic field immediately surrounding the actual arc is also high. Once the charge reaches the metal member, it spreads out to the farthest extent possible, depending on the amount of metal available. Discharge directly to a plane spreads out very quickly, confining the intense magnetic field to the immediate vicinity of the discharge point and to the current's constrictions downstream.
Discharge to a wire can't spread out very much because it must follow the wire. This is bad news for two reasons. First, the inductance of the wire causes a substantial back voltage along the wire path. Second, the magnetic field remains intense along the entire run of the wire.
The second possible current path involves a floating piece of metal. Metal that is completely isolated and fairly small presents little concern: such metal accepts a modest charge, but the problem ends there. Unfortunately, metal is not usually very well isolated, and there is always the other end to worry about. For example, the end of a screw securing plastic pieces may be near an internal circuit (Figure 2). In this case, the discharge actually produces two arcs, with the floating member connecting the two. The problem in this scenario is that the metal member can carry the arc to places it would otherwise never be able to reach. Accordingly, it is imperative to ground all metal members.
Figure 2. Discharge via screw to metallic member.
An ungrounded enclosure, such as a portable device, provides a third current path. In this case, the device accepts as much charge as allowed by the capacitance between it and the ground. If the device is small, little charge transfer occurs. However, if the device is large or in close proximity to a ground plane, then most of the charge transfers to the device. Thus, a device can accept ESD even if it is not grounded.
The fourth type of current path can occur when an enclosure is grounded solely via a power-line ground wire. Following discharge, the current eventually finds its way out of the ground wire, but the leading edge of the current is through capacitive or planar paths. Depending on the wire length, stray capacitance, and resistive losses, poorly dampened oscillations will follow. This is a key point when addressing ESD: an earth ground via a green wire ground is not particularly desirable.
The key concern in coping with ESD is to keep the current density as low as possible. Because the current density reaches its maximum at the discharge point, an engineer can take several preventive actions. The first is to arrange the conductors so that the current densities fall off as quickly as possible. When this is not possibleas with a wire entry pointone should keep some distance from internal circuits and wires. It is also important to avoid subsequent current constrictions.
Shielding is the next step in protecting unavoidable exposed metal surfaces. A well-shielded enclosure embodies the following features:
- A metal or metallized enclosure providing essentially 100% coverage.
- Mating seams that make frequent conductive contact (at least every 2 in. or 5 cm). Conductive gasketing is preferred.
- Cable shields terminated directly at the metal enclosure; no pigtails allowed.
- Nonshielded cables (including power lines) with filter capacitors grounded directly to the case. (Recognizing the leakage current limitations doesn't change the laws of ESD.)
- All conductors (such as switches) bonded to the enclosure.
- No internal cables and wires routed close to enclosure seams.
Often, compromises need to be made, but it is crucial to be aware that any deviation from the above recommendations involves a significant risk. It will take considerable time and money to make a poor shield adequate, if it can be done at all. Nevertheless, many designers insist on trying. The following guidelines are useful when good shielding practices cannot be implemented.
First, if full shielding is not possible, then providing a ground plane to which cable shields and filters can be referenced is mandatory. The ground plane should have a footprint at least as large as the internal electronics and wiring. It is not necessary to connect the ground plane to the earth ground and, in fact, the ground can be completely contained within the enclosure as an internal ground reference or as an additional layer of the printed circuit board (PCB) itself.
Second, if the cable or power-line filter capacitor cannot be connected to an enclosure ground, then it should be tied to a circuit ground and supplemented with a series ferrite between the capacitor and the connector pin. Multilayer circuit boards are required if filter capacitors are connected to the circuit ground; two-sided boards are not sufficient.
The first line of defense in fighting ESD problems is to prevent discharge from occurring directly to the equipment. When discharge cannot be prevented, a significant burden is placed on the shielding. Although ESD is controllable when recognized shielding practices are employed, deviation requires much trial and error to effect an acceptable solution. Lack of complete shielding inevitably involves a major trade-off, leaving no room for other compromises. Most notably, multilayer boards become mandatory.
The bottom line is that a ground plane must be present for ESD control. That ground plane may be the conductive enclosure, a plane under the PCB, or a plane in the PCB.
1. Kimmel WD, and Gerke DD, "Blocking ESD at the Enclosure," Med Dev Diag Indust, 20(5):102106, 1998.
William D. Kimmel, PE, and Daryl D. Gerke, PE, are principals in Kimmel Gerke Associates, Ltd., an electrical engineering consulting firm specializing in EMI/EMC issues, with offices in Phoenix, AZ, and St. Paul, MN.