Surge & Impulse Protection

Having just gone through the academic reasons for installing impulse protection devices, this small section concentrates on some of the two types available, and their strong points.

Now, they say if you put three lightning protection specialists in a room, you'll come out with four ways of doing things! Having been involved in many discussions on the subject, this is just so true. And when one turns to the Internet and other sources, do the really odd theories come to light!

Before I embark on solutions that have proven to work in every situation, I want to establish what is really needed to achieve successful lightning protection.


There are three requirements to achieving successful protection of devices against lightning damage.

  • Bonding (equipotential zone)
  • Zone Integrity (removing any 'punctures')
  • Voltage Clamping (limiting breakdown voltages)

The one aspect staring you in the face by the mere fact of its absence is 'grounding'. I'll cover this later, but suffice to say for the moment that it is a fallacy that grounding is paramount in protecting devices against lightning damage.


Bonding is not Grounding, and Grounding is not Bonding. Again, Bonding is not Grounding, and Grounding is not Bonding. And again, Bonding is not Grounding, and Grounding is not Bonding. Get the idea? They are not related in any way whatsoever! Regardless of what others say.

Bonding has one purpose. And that is to maintain an equal potential, or as equal as possible, between components forming the item requiring protection. This can work from the micro through to the macro scale.

On a micro scale it would be to ensure the various electronic components of a device are not placed under undue stress of undesired current flowing through them. This done by ensuring the common reference does not vary between one point and another during the high current flow.

On a macro scale, the same applies except in this instance the components are now devices, and the reference is much larger than in the micro scale. A micro scale exists at device level e.g. a computer, and macro scale exists at premises level e.g. a house or building.

Said another way; I have designed radio hi-sites and the one thing I have always made a point of is ensuring that during a lightning strike, everything was at one potential. I did not care if the installation reached a million volts during a strike, so long as the whole installation - and this includes the floor I was standing on - was at one megavolt.

As the floor was at a million volts, so was I. This meant that no current could flow through me as there was no potential difference between anything and me so as to cause current to flow. Never mind that it did not flow through anything else either (except for the paths I designed to take the lightning current).

There is a limit to how big the reference can become. Some published material states that it is possible to protect a building at the entrance only without any further protection required within the building. This is just not true.

Sure, one might be able to protect a small 2 story office block to some degree, especially if an all metal construction. But no ways to doing this on a 50 story office tower - and talking of towers, I even include the all-metal Eiffel tower!

In instances where the reference is just too big to maintain equal potential, then one needs to subdivide the larger reference into smaller, manageable chunks, each with its own protection system.

Zone Integrity:

It's all very well having bonded a complete structure together so as to create an equipotential zone, to then bring some external bit (a superb example is a vent) and connect it to some point within the protected zone. You can be rest assured the external bit will take a direct hit and bring all this energy directly into the protected zone. I've seen this sort of thing time and time again.

On a micro scale; It's like creating a beautifully protected environment for a delicate bit of electronic circuitry, to only then bring the input signal via a hole punched in the side of the case and wire this directly to the middle of the circuit board. You can be certain the wire where all the current will be induced will be the input wire, and all this current will find its way directly to the input circuitry.

In both of these above examples, this action of bypassing the bonding is referred to as "puncturing the zone" i.e. creating a potential difference within the protected zone (thus defeating the whole purpose of creating an equipotential zone in the first place).

The best method of avoiding this is based on a simple philosophy of creating what can be thought of as a "bottle-neck" approach in that the protected zone be thought of as an imaginary bottle - the interior of the bottle being the protected zone - with all cables entering and leaving the protected zone only through the neck of the bottle; The neck being the location where all extraneous currents are contained.

Voltage Clamping:

It's all very well talking about maintaining the protected zone's integrity and to not puncture it, but this is unavoidable as every signal cable is doing just that. The only way a signal cable can be made to not puncture the zone is to bond it to Earth, but it is rendered useless and no longer a signal.

In order to both maintain the usefulness of the signals as well as keeping the zone's integrity intact calls for a means to allow the signal through while ensuring any extraneous current does not find its way into the protected zone.

This is accomplished though voltage clamping i.e. any voltage deemed to be a signal is allowed to proceed unhindered, but when deemed beyond that of a normal signal is clamped and diverted to the awaiting equipotential bond (located at the "bottle's neck"). There are generally two types of voltage clamping viz. 'voltage limiting' and 'voltage switching'.

Voltage Limiting operate simply by ensuring the voltage cannot exceed a predetermined limit. They do this through starting to conduct the moment the limit is reached. The more the voltage attempts to increase, the more current is taken by the device so as to maintain the limit. Such devices are Metal Oxide Varistors and Zener based diodes (e.g. "Tranzorbs").

Voltage Switching operate by attempting to short out any signal that has exceeded the limit. They do this by becoming as close as possible a dead short while the offending current flows. As soon as the current ceases, they become an open circuit once more. Such devices are the simple Spark Gap, Gas Discharge Tubes, and Silicon Avalanche Diodes (no, Zeners are not SADs as is incorrectly published in many documents!).


No matter how carefully you read the requirements for successful protection, as laid out above, you will not read anything about grounding i.e. establishing a connection to terra firma, being needed for successful protection.

As said earlier; It is a fallacy that grounding is necessary to protect devices against lightning damage. If anything, grounding can be the very reason devices are damaged! By having a superb ground, higher currents can flow during a lightning strike. And, using what real-life has taught that it is not voltage but current that damages, means by increasing the current we have increased the likelihood of damage.

This is exactly why the actual requirements to successful lightning protection is bonding, zone integrity, and voltage clamping, and does not include grounding.

Grounding is important, however, when there is a requirement for the Rise Of Earth Potential (ROEP) to be kept below a specific limit based on the location and application of the item when subjected to a lightning strike. In other words, there are times the equipotential plane created by bonding needs to have as solid a connection with terra firma as possible so as to avoid damage to other installations or danger to life.

Said a simpler way; The reason some specs bang on about grounding (as opposed to Earthing, which is another word for 'bonding') is when the installation takes a direct hit, it is likely to be raised to many thousands of volts. It is most undesirable that the installation use other installations connected to it (such as the mains supply!) as the path for dissipating this absorbed energy.

As the lightning has only one aim and that was to discharge the cloud to ground, a good design puts in place a short (electrically speaking) path to terra firma for the energy to travel down rather than using any service conductors connected to it (power companies get a little annoyed when their transformers are also used as lightning arrestors!).

But don't get confused. Grounding does not protect an installation itself from damage. Grounding is there to limit the damage to other installations connected to an installation that takes a direct hit.


The page "A Modular Surge Protection System" concentrates on a wonderful 'top-hat' DIN-rail mountable system with impressive impulse current handling capabilities. As shown in the pages leading up to this section, effective protection is best achieved using many stages and having some distance between each. When this is not able to occur, such as radio huts where there is no distance achievable, then other plans need to be implemented. This system provides such a solution.

"Individual Appliance Surge Protection", in conjunction with the above, completes the chain of protection. The type featured here is not unique, but one where the manufacturer was prepared to put their money where their mouths are! What I like is it is repairable should it be forced to bite off a little more than it can chew.

A Modular Surge Protection System  >>

| | Ask a Question |

© 02.08.05