robgraham wrote:

The reason they don't light is that the user is ignorant of the why
they light up and is standing on a rubber mat wearing rubber soled
boots. If I've got any doubt I usually find some way of earthing my
other hand to give a definite earth path.

That may help a little but you don't really need a conductive path to
earth. The stray capacitance of the body to earth (~150 - 1000 pF) is
enough to complete the circuit.

I've never heard of one failing - has anyone?

There are plenty of recorded cases of shock injuries resulting from neon
screwdrivers. The are potentially dangerous because:

- the series resistor can fail short circuit, particularly the old type
of solid cracked-carbon resistors they used to use in the things;

- in damp conditions the resistor can be bridged by condensation, etc.

- immunity to transient overvoltages (spikes and surges) is poor, and
again the failure mode is to lower the impedance of the device,
increasing body current;

- the neon tube itself can fracture if handled roughly;

- visibility of the live indication is at best poor in outdoor conditions.

OK, the last two points can apply to other voltage indicators, any type
can fail to indicate, but (a) the neon screwdriver is far more likely to
fail than an approved GS38-compliant indicator, and (b) test lamps and
meters tend to fail high-impedance. As has already been said, the
tester should be tested both before and after the test - this is the
basic safe procedure for 'proving dead'.

--
Andy