Our company, Quality of Supply Technical Services, has a close working relationship with the South African Bureau of Standards SABS National Testing Facility (NETFA). Since 1995 we have had our recommendations for SPD specifications adopted with success widely all over Africa, and particularly by Mobile Telephone Networks in Africa. I therefore believe that we are appropriately qualified and experienced to assist with the interpretation, selection, testing and installation recommendations of SPDs in harsh electrical environments.

We have carefully studied the theory and practical application of test requirements for SPDs in various locations and based on that study can make the following conclusions and recommendations.

The South African Bureau of Standards (SABS) is affiliated to the International Electrotechnical Commission, (IEC). SABS official policy is to adopt the IEC specifications either unchanged or where deemed necessary to adapt them to suit South African conditions.

In the case of IEC 61643-1, this specification was adopted unchanged, and is officially referred to in this country as SABS/IEC 61643-1. The standard is referred to as a compulsory requirement in the newly revised SABS standard – SABS 0142 – the electrical wiring of premises. This standard makes the installation of an SPD at the electrical service entrance compulsory for all new buildings. As the specification was released only in November 2001, there has been much discussion and debate in South Africa of which I have been involved.

I have held a meeting recently with a representative of SABS NETFA, Mr. H Kroninger, to discuss their viewpoint on the purpose and application of the test classes. Mr. Kroninger has over 20 years of testing experiences in high voltage laboratories and has studied in Switzerland and had a paper on lightning characteristics published by CIGRE.

I think that it is important to note that due to the wiring code being compulsory in South Africa, all SPDs sold or allowed to be sold in South Africa have had to undergo testing by SABS NETFA, which means that they have a considerable knowledge of the specification.

Due to SABS policy it is rare for them to make recommendations for one technology versus another. As a national standards body, this request falls outside of their scope and they must remain totally objective. In this case in their capacity as a testing laboratory they cannot decide on application matters. They were however agreeable to discuss the issue with my company and allow us to interpret the IEC61643-1 based on our discussions.

It is easy to confuse the issues of application recommendations and test standards. The IEC 61643-1 is a test standard, and whilst it makes references to test classes and states current impulse test values, it in no way recommends a particular technology or particular type device, for a zone or location. The appropriate document for application is a final draft international standard (FDIS) IEC61643-12. This will not be revised for a further four years.

The SABS stated that the manufacturer must decide on the ratings of their equipment and the SABS will test the devices in accordance with these ratings to 61643-1 and report the results.

From our conversation it became apparent that there is a perception amongst end users that a test class relates to a location of the SPD. This is not so. It is evident from the marketing material that I have studied that certain German SPD manufacturers infer or state that a service entrance requires a Class 1 tested device. Broadly however a Class 1 test would be performed on a crowbar type of device, an arrestor, as opposed to a suppressor, due to the extremely high coulomb ratings specified in 61643-1. The figures are the source of intense debate.

To expand further on the test requirements, 61643-1 defines an additional Class 1 test which states that a value of 10 coulombs must be dissipated into the device within a period of 10 milli -seconds, (a lifetime in fast transient terms). I imp is specified but no waveshape. Only the charge "Q coulombs" and I imp. There are any number of waveshapes which could satisfy these requirements. A number of vested interests are promoting the adoption of the 10 x 350 micro second waveshape. Others promote a 10 x 1000 and so on. IEC 61000-4 quotes a 10 x 700 micro second waveshape.

The figure of 10 coulombs is extremely large. In nature, this level of charge can only be found in a direct lightning strike. As the characteristics of a particular lightning strike i.e. its waveshape(s), it is therefore not possible to specify a wave shape for a simulation test, so there is a reason why the IEC 61643-1 does not specify a waveshape. A lightning strike usually consists of a first strike followed by a number of smaller strikes that vary depending on the environmental conditions and routes to earth etc.

Therefore a device which would be tested to Class 1 would typically be located on a powerline and would be a first line defense to discharge high levels of energy to earth as effectively as possible. Their purpose is to protect equipment such as transformers and switchgear from damage; the voltage protection level would be of less importance than its ability to dissipate the charge to earth.

However the situation at a service entrance to a building is entirely different. Unless the building was remote and there was only a single route to earth via the service entrance, which would be extremely rare, a service entrance will never be exposed to 10 coulombs. What is important for a building SPD is to dissipate (absorb) the residual current pulse, whilst controlling the let through voltage ( sometimes referred to as clamping voltage or voltage protection level VPL).

These devices would need to be tested to a peak current whilst simultaneously measuring the VPL. The Class 2 test accomplishes this requirement. The waveshape is defined, not the charge Q.

Over the years the accumulated knowledge in the industry has resulted in the adoption of the 8 x 20 microsecond current waveshape, as being representative or a lightning transient entering a building. There have been numerous technical papers on the subject.

The IEEE C62.41 section 9 specifies the 8 x 20 waveshape. The IEC have subsequently adopted it in 61643-1, which is proof that the waveshape is adequate for SPD tests where simulation of surges likely to be present at a service entrance of a building.

It is up to the manufacturer to specify the ratings of the SPD and what test class is appropriate. IEC 61643-1 does not state this.

Based on my discussions with the SABS NETFA, it is clear that the reason that a Class 1 test does not specify a waveshape and that a Class 2 test does, is that a direct lightning strike does not exhibit a particular simple waveshape, but is rather a series of impulses, the energy of which is best determined by its charge (Q). A single waveshape cannot be representative of a lightning strike and so one is not specified.

However, by the time a lightning strike has traveled through the distribution network, surge arrestors will have dissipated a proportion of the charge, network impedance will have shaped the impulse and the residual energy can be represented in terms of its waveshape, and its peak current which is what will be present at a service entrance.

This is the waveshape specified in IEC 61643-1 Class 2. Based on these facts, it is the Class 2 test which is the correct class of test for an SPD recommended for installation in the service entrance of a building. Class 1 is appropriate for arrestors.

The Class 3 test is the combination waveform test which I recommend as the appropriate test for an SPD situated inside a building or as an SPD integrated into equipment.

It should be mentioned in conclusion there are certain parties intent on distorting the highly technical issues of SPD testing for their own benefit, and one needs to be aware of these subtle influences. They attempt to use specifications as a means to "lock out" competitors products and increase their sales unfairly.

Richard R. Goodland