Wylex SPD

[JohnW2]

The ones I've seen teardown videos of use a spring-loaded fusible element if the surge protection element gets too hot the element melts and the spring pulls the connection apart.

Again, that all makes complete sense, but I'm rather amazed that I've yet to see any mention of such a mechanism (let alone any details of the characteristics thereof) in any technical documentation, and that one therefore has to resort to teardown videos on the Internet to find such information!

This also causes a failure indication to appear on the SPD.

Yes, that was actually going to be one of my questions - i.e. what triggers the 'failure indication'. I think I assumed that it indicated that the 'surge protection element' had 'burned out'.

SPDs are becoming increasing common in CUs. Has anyone yet seen one which has developed a 'failure indication' ... or, indeed, one which has caused a DNO fuse to operate?

In passing, I note that, although many/most SPDs have 'replaceable cartridges, per TLC prices a replacement cartridge for the Wylex SPD costs appreciably more than a whole new SPD

Any SPD without such protection would be dangerous as heck in the event of a sustained overvoltage that was sufficient to cause significant current flow in the surge protection element, but not so high as to cause a current flow that would trip the overcurrent protective device.

Indeed - but it again begs the (so far unanswered, at least for me) question as to what the operating characteristics of this internal 'protection mechanism' actually are. Nearly all SPDs say that they require ('back-up') protection of a maximum 100 A or 125 A fuse, so the implication presumably is that the internal mechanism would disconnect before the upstream fuse operated?

As for the differences between ones that are ok with 100A upstream protection, and ones where the manufacturer suggests smaller I don't know ...

This seems to be the most important question. As above, they all seem to say that they require at least 100A (or 125A) upstream protection, but some say that additional 32/40 A protection is also required.

, but I would speculate that it comes down to the details like how big the contact gaps are, how good the case is at tolerating arcs until the upstream overcurrent protection blows and so-on.

Quite so - which is why I said that it's very strange that none of the manufacturers seem to provide any information on the operating characteristics of whatever 'internal protection' they have.

I suppose that an alternative explanation would be that all the devices are essentially the same, but with some manufacturers being prepared to trust the internal protection, but others being more cautious ('protecting their backsides'?) by indicating the need for a 32/40A external MCB.