17th edition design given 18th edition certifcate

I am still somewhat unhappy with the SPD thinking, because the test specification is far too vague, and assumes a number of parameters that are in reality undefined. The place for protection is ideally next to or within the device to be protected. The fundamental problem is that people keep saying “clamped to some voltage”. This is fundamentally untrue unless the SPD has zero impedance when operating, and also there are no “transmission line” effects mentioned by Mike above. If the pulse under consideration strikes a short circuit (the SPD) in RF terms it is reflected not removed, and thus may damage something else!

As real VDRs have a significant series resistance when conducting, the remaining pulse amplitude depends on the original pulse energy, some of which is dissipated and some reflected. A pulse still continues on the original circuit, although the amplitude may be reduced. There are so many unknowns in this description that what actually happens is undefined, although in the ultimate case is probably some improvement in damage levels. To work better it is necessary that the pulse source impedance is defined, and for in-product protection, this is usually some series inductance in the supply lead. The VDR can then be chosen to have a suitable clamping voltage and current (operating series resistance) to reduce the pulse of some width to a known potential, even if the supply impedance is zero, and even less if it is significant.

From the above one can see that the standard SPDs in BS7671 may help in some circumstances, but appliances also have specifications for resistance (sic) to spike damage, although cheap stuff usually ignores these. It is not worth spending £1 to protect a led light bulb whose manufacturing cost is 50p.

There are ways to protect or remove spikes at a known voltage but these also have some problems, related to the possible pulse energy to be clamped. Again controlling the supply impedance is essential, and excess pulse energy may cause thermal failure. They are also much more expensive than the simple VDR (voltage dependant resistor) types, using power electronics.

I am still somewhat unhappy with the SPD thinking, because the test specification is far too vague, and assumes a number of parameters that are in reality undefined. The place for protection is ideally next to or within the device to be protected. The fundamental problem is that people keep saying “clamped to some voltage”. This is fundamentally untrue unless the SPD has zero impedance when operating, and also there are no “transmission line” effects mentioned by Mike above. If the pulse under consideration strikes a short circuit (the SPD) in RF terms it is reflected not removed, and thus may damage something else!

As real VDRs have a significant series resistance when conducting, the remaining pulse amplitude depends on the original pulse energy, some of which is dissipated and some reflected. A pulse still continues on the original circuit, although the amplitude may be reduced. There are so many unknowns in this description that what actually happens is undefined, although in the ultimate case is probably some improvement in damage levels. To work better it is necessary that the pulse source impedance is defined, and for in-product protection, this is usually some series inductance in the supply lead. The VDR can then be chosen to have a suitable clamping voltage and current (operating series resistance) to reduce the pulse of some width to a known potential, even if the supply impedance is zero, and even less if it is significant.

From the above one can see that the standard SPDs in BS7671 may help in some circumstances, but appliances also have specifications for resistance (sic) to spike damage, although cheap stuff usually ignores these. It is not worth spending £1 to protect a led light bulb whose manufacturing cost is 50p.

There are ways to protect or remove spikes at a known voltage but these also have some problems, related to the possible pulse energy to be clamped. Again controlling the supply impedance is essential, and excess pulse energy may cause thermal failure. They are also much more expensive than the simple VDR (voltage dependant resistor) types, using power electronics.