I'm a bit confused here - the top photo shows 3 wires - red brown and blue, and an un-cancelled current of a touch under 70mA ? Are these all on the same RCBO ? If it is a 30mA instant, yes I'd expect it to trip. But as with RCDs, you can get funny effects with NE shorts load side that both bias the device to trip on normal loads by imbalancing L and N, and yet by making a shorted turn shunt for the N part of the RCD also take the edge off its response to small errors. At the same time you do have some odd DC floating about in the meter tails, but the zero on the Dilog is a bit iffy - oddly I have one of those and it has the same problem since I over amped it looking at welding waveforms.
I'd look for an NE low resistance or short load side. But if I have misunderstood the story in pictures please correct me.
Looks tricky. What about current coming IN to neutral through the earth? If there is a fault next door for example your earth might actually be a higher voltage than the incoming neutral, assuming TN supply. And of course this could be changing all the time.
Other than equipment issues I can't think of anything else at the moment.
