Hence the question “Is it me?” because I am not getting a grip on what the thinking is behind all of this.
Total lack of care, I would have said. If it's Chinese, then earthing seems to be optional, regardless of whether the components actually have earth connectors.
There are a number of countries that some companies send their staff to [well, used to, back when we used to travel] accompanied with plug-in “socket-checkers” for this reason. I don't think it's fair to single out a specific one … although I'm not arguing too strongly.
That is a very curious list Graham because none of those things requires an Earth connection, and if the design does it does not properly meet the SELV requirements. Let's actually examine what you have written:
EMC is not in any way affected by an Earth wire of unknown length, routing, or anything else, in fact, it may well increase EMC levels. Whilst effective RF Earth connections may reduce EMC under certain circumstances, in the generalised case they cannot, because there are far too many unknown conditions. The normal way to reduce conducted EMC is to use common-mode transformers (often called chokes) so that the signal on each live wire is equal and opposite in polarity, so that they largely cancel in the connecting cables, therefore radiate very little. Capacitors between live conductors may be effective for some frequencies but tend to make the signal on both live conductors equal, which cancels out the differential effectiveness of method 1.
As the objects are in plastic cases and the output is completely isolated, what might need Earthing as nothing is accessible?
Why does a static charge on the plastic matter, it will be tiny?
Surge protection inside the equipment only needs to deal with voltages between the live conductors unless Earthed!
That is a very curious list Graham because none of those things requires an Earth connection, and if the design does it does not properly meet the SELV requirements.
We're talking about the mains side, which will definitely NOT meet SELV requirements …
Let's actually examine what you have written:
EMC is not in any way affected by an Earth wire of unknown length, routing, or anything else, in fact, it may well increase EMC levels. Whilst effective RF Earth connections may reduce EMC under certain circumstances, in the generalised case they cannot, because there are far too many unknown conditions. The normal way to reduce conducted EMC is to use common-mode transformers (often called chokes) so that the signal on each live wire is equal and opposite in polarity, so that they largely cancel in the connecting cables, therefore radiate very little. Capacitors between live conductors may be effective for some frequencies but tend to make the signal on both live conductors equal, which cancels out the differential effectiveness of method 1.
The way in which the device meets the EMC criteria is one of design. Many designs use some form of capacitance to PE. I agree it's not the only solution, but it's the one used here.
As the objects are in plastic cases and the output is completely isolated, what might need Earthing as nothing is accessible?
I agree that someone might have made a luminaire with Class I components that they have since encased … but that doesn't change the design assumption in the first place, if earthing is required for another reason and fault protection is being used.
Does the isolation transformer have a screen which requires earthing? Some do, in which case the original driver design is absolutely fine to consider Class I - doesn't mean the isolation transformer will now actually meet its requirements (and hence the output no longer SELV)
We can go round in circles, but without further info all we can go off is what the original driver manufacturer intended, and we already know what that is because they went to the trouble of marking it on their product.
Why does a static charge on the plastic matter, it will be tiny?
Static charge on the SELV output matters, though - again I agree it could be addressed by alternative design solutions, but some of those lead to “tingles” from the SELV side (I won't mention the one that's often discussed here, which uses a type of fruit for branding and for which it had a court battle with a band named after silver insects, but hit it big a little later when they dropped the silver and just became insects).
Surge protection inside the equipment only needs to deal with voltages between the live conductors unless Earthed!
See above. Your assumption it's not necessary, I'm happy to say I don't know so I'll go with what's marked on the driver.
They in turn belong to a Japanese company, but have significant autonomy.
Other companies with similar names are based in Peru and the US, but not the logo..
The Ansell website is strangely silent on ohmtronic so I assume it is an ‘off brand ’ aimed at a different market segment, perhaps only for sale to other makers of fittings and not direct to public.
they have a very liberal approach to declaring CE conformity - basically ‘all our products comply, not going to bother to list them '
Most likely highlighting my ignorance here, but would a more obvious answer be that the earth terminal is just there to ‘park’ the cpc where a light fitting is changed but is using the existing wiring from a fitting which previously utilized it?
Most likely highlighting my ignorance here, but would a more obvious answer be that the earth terminal is just there to ‘park’ the cpc where a light fitting is changed but is using the existing wiring from a fitting which previously utilized it?
Reasonable assumption, but the IEC 60417-5017 symbol on the bottom-left of the driver is used where the equipment must be earthed.
