It charges at about 10 miles per hour, so, if I drive it 20 miles, it takes 2 hours to top up, which is typical for what I do.

This is a really helpful answer, but it doesn't answer the original question which is why my granny lead switches itself on and off at about 1 cycle per second. This seems far too fast for a thermal cycling issue.

Also, it seems that it only does this when I plug in a car that has just been driven. If I wait an hour before plugging in, it is happy.

Also, it seems that it only does this when I plug in a car that has just been driven. If I wait an hour before plugging in, it is happy.

Sounds like an issue inside the car then (or possibly with the connecting lead). As above, the granny lead (just like a mode 3 box on the wall) will disconnect if any of the basic correct "all-OK" signals no longer arrive).

   - Andy.

Then I wonder whether the battery is getting a little warm. So you plug it in, it warms just a little and then it cuts out. Then is cools just a little and it connects again, and so on.

When you wait an hour, the battery may have cooled down enough not to affect charging.

Or maybe the temperature sensor in the battery is defective.

Isn't it really as simple as the EV being charged communicating with EVSE (charger) to determine the current rating of the available electricity supply?

There can be additional protection built into the EVSE, but not always. 

A similar question would be, why do electricians spend four hundred pounds or more on an EVSE testing adapter to fool the EVSE into thinking the electrician is plugging in an EV to charge rather than their installation tester to carry out tests?

I was starting to wonder why I wasted my hard-earned money on the course, the COP, the EV socket adaptor and a new megger !

Sparkingchip said:

A similar question would be, why do electricians spend four hundred pounds or more on an EVSE testing adapter to fool the EVSE into thinking the electrician is plugging in an EV to charge rather than their installation tester to carry out tests?

That's a good question for the most part ... although protective devices such as RCDs can be built into EV charging equipment ... having said that, in designs like the "cabinet" type approach with DIN-rail mounted components, the RCD could possibly  be tested using Test Method 2 (upstream'/downstream) from GN3 9th Ed (2022) ...

I think, however, fault-finding and commissioning are the answers to why you really need an EV  simulator/adaptor.

There must be more than this, even in a simple granny lead, as the thing was clicking on and off and I can't find out what could have been making it do this.

The granny lead will contain relays to disconnect the power from the car until it is satisfied that it has established a data connection over the pilot wire.  The same relays are probably also used to implement the RCD function.  The whole thing will be controlled by a microporcessor, because everything is these days.

If the microprocessor is faulty, or the lead is damaged, then anything could be happrning inside.

www.electricalcounter.co.uk/.../bge_evcp2135sl.pdf