E.S.D. Protection Basics
... part two ...
2.0 ESD Effects on Electronics
**** I'm still working on this... nearly done! ****
In this section we'll start looking at a simple example circuit and apply some
of the principles outlined in the introduction section.
I'm going to try and stick to an 'intuitive' approach, attempting to give a feel
for what is going on rather than pages of numbers, oscilloscope plots and calculus that
students use before they meet the real world.
This means that the language may stray from the textbook and I'll probably start using
descriptions that may seem strange to some junior electronics engineers (e.g. "capacitors
are like low value resistors for a while") - but this is an attempt to give a workable
methodology to people who havn't had 100's of lecture courses in electronics.
I think that good engineers are those with a sense for design, people who
can look at a circuit and say: "Hmmm, I'd shove another 0.1uF in there..." without having
to run up a simulator. It is some of this quality you need to deal with ESD since it is
extremely difficult to measure what is going on in 'zap' events. You need some imagination.
(Oh, and a bit of theory, and some experience and... )
2.1 An Example Microprocessor Circuit
Ok, here's an example... the diagram below is a piece of equipment whose safe operation
is critical - a firing box for big fireworks. The equipment will be used in all sorts of
environments and will have hundreds of meters of wire attached to it. It will be in situations
where ESD events will happen as people wire it up and operate it. It must withstand
everything that's thrown at it. If it fails, it could accidentally ignite some firework -
which could be lethal.
The diagram shows a circuit board with a ground (Green) track running around the edge,
a couple of connectors at the top for an input and output and a nice little microprocessor inside.
The whole lot is run from a 12v battery. We're assuming that this unit is grounded (by being sat on the
ground or whatever) which is shown by the dotty arrow at bottom right.
2.2 Current Paths and Jumping Grounds
We'll start looking at the effect of a static Zap on the input wires (note that this argument also
applies if this "input" is actually a button on the top of the case!).
On the above diagram, the blue line A shows the obvious "ground path" for the current, however, (remembering all
that stuff from the introduction) there are many other paths - in fact current will flow along
every possible path connecting the input point to the ground point. Also remember that the
voltage will be "divided" across these possible paths.
Now look at the blue line B - this could indicate the obvious path down the Black input wire
straight into our unprotected microprocessor (but we'll deal with this later).
It can also indicate a less obvious path...
Take a closer look at this small section of the above diagram:
The ground track (in green) around the edge of the circuit board will carry a high
current but the thinner, internal ground track will also carry some current (just like
the example in the introduction with different value resistors). If a reasonable current
flows in this small track, a voltage (according to Ohms Law) will be developed between
points X, Y and Z.
Wait a minute... if there is a voltage between points Y and Z then the
CPU ground pin is not at ground any more!
Yes, thats right. This is sometimes referred to as "ground jump" or "ground elevation" because the "ground" on a
device jumps or elevates away from zero volts when a Zap hits. This is a sure way to really worry a
microprocessor, if not kill it off completely.
A common approach to fix problems like this is to "beef up" the ground
connections within the circuit board, oddly enough, this can make things a lot WORSE!.
Have a look at this revised diagram, I've added another component (a memory chip) to
emphasise the point:
Here we have a modified circuit board with lots of cross-linked ground tracks (the thin
Green lines), however, the Red lines indicate paths for current flow from the input point
to the boards ground connection. This layout would cause the current flowing in the bit
of track from Y to Z to increase!.
The memory chip, having different connections to ground will not be elevated as much, so,
for the duration of the Zap event, there may be a significant voltage difference between
the ground (0v) of the CPU and the memory chip.
In most cases, if the differential is greater than half a volt, then the CPU might not "see"
the memory chip properly (or vice versa). Imagine the CPU ground elevates to 1.5 volts and
the memory chip's 0v pin elevates to 0.5 volts - in this case the lowest logic level the CPU
can present to the memory is 0 volts referenced to its own ground pin- but this is seen
by the memory chip as 1 volt, not what was intended at all!
We'll come back to current paths around our circuit after considering the other
little problem we side-stepped earlier - that of the obvious path for a Zap to
get at our electronics: an input.
2.3 Input Protection
We considered inputs a little bit in the Introduction section with this diagram:
Unfortunatly, there is some bad news: it is not usually as simple as this!
The good news is that the same broad principles apply.
In this diagram we saw that a low "shunt" resistance (100 ohm) could lower the voltage
present at the junction with a higher (10k) input series resistor. The series resistor
reduced the current which could otherwise cook our chip.
The problem in practice is that a simple low value resistor also stops the normal
input signal we want to detect. What we need is something (or somethings)
that will act on the static discharge in a different way to a normal input signal.
We need to look at the differences between normal, expected signals and static zaps:
Static Discharge Normal Signal
---------------- ---------------------------------
Voltage 1000s of volts low voltage (e.g. 5 volts)
Speed very fast (uS) relatively slow (10mS to seconds)
So, we need to get rid of very fast, high voltage spikes and leave slow, low-voltage
stuff alone.
Time to dig out some more components...
2.4 Input Protection components
B = parasitic ferrite bead
L = inductor
R1 = protection shunt limiting resistor
TZ = TransZorb
C = capacitor
Z = Zener diode (mention failure modes)
R2 = input current limiter
R3 = gate pull-up
What do all these do then ?
You can think of these components acting like different resistors during the very
short time that a discharge event lasts. Ferrite beads B are useful for damping down
very high frequencies, when threaded over a wire they act like an inductor. Inductors L
are coils of wire that resist changes in current flow - in this case the act like
a resistor, but they have no effect on slow signals. Ferrites, inductors and the
ordinary resistor R1 have a combined effect of increasing the source resistance of
a fast spike.
The TransZorb and Zener Diode (TZ, Z)are similar devices in that they are semiconductors
and they act like high value resistors at low voltages and change into low resistances
at a higher voltage, depending on their type. The big difference between them is that
a Transzorb is designed for protection, it can withstand very high currents for
short periods whereas Zener diodes are designed for less arduous tasks and have a
tendancy to be damaged by high current pulses (they often go short). In either case a
series resistor R1 can be used to keep the peak current within limits for the device.
The other component in this little cluster is a Capacitor C- its use here is to look like
a very low resistance as it quickly "soaks up" the current from the Zap (like a flat
battery takes high current from a charger).
The combination of the series components (B, L, R1) increasing resistance and the "shunt"
components acting as low resistance can reduce the voltage at the input to R2 to
only a few tens of volts.
The remainder of the circuit is fairly standard for protecting logic inputs. R2
limits the current into a CMOS Schmitt-input device. This has a high input resistance
so R2/R3 can be high values. R3 is there to ensure that the gate is in a known state
when no external input is connected.
Next we'll put our input protection components together with what we learned about
grounding earlier...
2.5 Going to Ground
Design layout on a PCB (printed circuit board) is most important. You need to think
of two main things:
1. is there a current path through any devices to ground?
2. will ESD current path to ground pass near any sensitive devices?
...more discussion about these points and the current path in BLUE
NOTE: location of input protection components and their ground connections - as close
as possible to the input line
FIXINGS!!!
On the above diagram there are circles marked in yellow at the corners of our imaginary
circuit board. These circles indicate where the board might be bolted down inside a
metal case. The important point to note is that the top mounting points will be
connected to the green ground track on the board whereas the bottom holes have space
around them. This is because, if all four corners were bolted to the ground track, there
would be a current path through the case to the bottom of the board.
2.6 Problems with outputs
...especially high-current drivers (in the case of a pyro firing rig for example) where
adding series resistance is not possible.
Note: relays vs. semiconductor in these situations
Maybe add... (in next section???)
protection devices:
TransZorb and other 'breakdown' devices
resistive devices
Zener diodes (and failure modes)
diodes
good ol' spark gaps
R's and C's (discuss: 'spreading out' the spike)
inductors, coils, parasitics (and problems of 'tuning')
opto-isolation (also mention supply isolation, layout, etc.)
Back to: TonyWilk Home Page