CT1 and CT2

In the example you give, you are correct.

Consider as the pulse comes along, the leading edge of it causes the L voltage to rise, but in this case for whatever reason not the N.  Once  voltage between L-N is > than the breakdown of the first clamp, current will flow and the L-N voltage will start to be be limited, and as LE voltage limits, the NE voltage will start to rise. Note that even on a TNC supply, unless the surge arrestor  is very close to the N-E bond, the neutral voltage will also start to rise relative to earth once the LN device has broken down and started to pass current.

An observer at the SPD cannot see the short circuit, distance 'd' away, until a time 2* d/c later - this is the time it takes the current to get down the wires from the SPD, to the NE bond and back again - this delay is measured in nano-seconds per foot so does not affect normal 50Hz thinking, which is quite pedestrian, but is important for fast edged pulses  -it is a short circuit at a distance....

The NE device may never break down if the NE bond is near enough to take effect before the voltage on the second arrestor has risen to the breakdown voltage.

M

If this is not helping come back and I can try another tack.

Bear in mind that such a device may well start conducting at around a 500V difference, but due to inductive effects, lead lengths etc, it may only clamp the effective voltage between the two lines to 1.5kV. So a 2kV difference across 2 SPDs in series may well cause them both to start conducting. But I defer completely to Mike over the details.

also true, I was trying to gloss over the rather non-linear turn on curve.

M.

The reason I ask about the series arrangement

Just to add into the mix that both CT1 and CT2 have a series arrangement - where you might get 2kV instead of 1kV between L & PE with CT2 you could equally get 2kV instead of 1kV between L & N with CT1.  In terms of equipment connected between L & N only - CT2 might actually be better. For Class 1 equipment, something's going to get stressed anyway - either live conductors to chassis or between live conductors. There's insulation and often components between both.

I think the N-PE SPD in CT2 arrangements is usually different technology (spark gap rather than MOV perhaps?) - which may have an effect on relative clamping voltages (maybe someone with more knowledge than me could illuminate).

   - Andy.