What's important here is that it's not "leakage to earth" (in the sense that I think you're thinking about it), it's an impedance to the supply. Since the supply has some path to earth, the circuit is supply > leakage > earth > direct or leakage path back to supply.

Or to put it another way, if you connect two 1Meg resistors in series, connect one to 240V live and one to neutral, the middle will be somewhere around 115V to earth (because neutral is, you hope, somewhere close to earth). And that's probably what you're seeing.  

The driver has no earth connection, only line and neutral, hence the confusion. 

Leakage doesn't necessarily have to be to c.p.c. - it's just the potential for some current to escape as it were -  exactly what route the current can take and what limitations there are to current flow is down to individual circumstances.

In some test arrangements things are wrapped in foil and current measurements taken from that to Earth, or tests might be from the LV to ELV sides. In some circumstances the current that can actually escape may be negligible, and the part just it there at something like half mains voltage - but almost(*) any attempt to measure the voltage will allow some current to flow and so reduce the apparent voltage (which can be very confusing in some circumstances, and measured voltages then seem not to add up correctly).

(*) there are voltage measurement techniques whereby you compare the voltage you want to measure with a known but variable voltage source (with a common earth/N connection) and adjust things until no current flows - the setting of voltage source then tells you test voltage. But not many test meters work that way.

    - Andy.

That seems like a severe shock! I have seen a similar incident involving a plumber. Perhaps van tech is measuring the total leakage currents from other class 1 devices that are in contact with the grid, and the LED panel with its isolated metal frame is in contact with the grid when he conducted the measurements.

Hi, 

The panel lights are class 2, so in theory the LEDs inside should be isolated from the frame unless they are leaking through this via induction? 

Hi Andy,

The panel light, is powered via a constant current output from the driver. The LEDs inside of the panel are isolated from the metalwork of the luminaire. 

Perhaps the leakage is via induction into the bodywork of the luminaire, via the secondary side of the driver? 

So in this instance, the leakage is via the frame/chassis of the fixture? 

All components inside should be isolated, but can understand that impurities can lead to discharge. I guess here the leakage is via the secondary side of the SELV transformer? 

I’m wondering if it’s nowt to do with the fitting but the ceilings grid ? That’s why I asked if the panel was sat in the grid when you carried out the measurements 

Hi AMK, 

The panel is sat in the ceiling grid, so perhaps yes this is the issue. 

Thanks. 

Perhaps the leakage is via induction into the bodywork of the luminaire, via the secondary side of the driver? 

Probably not induction - as you'd generally need pretty hefty currents or conductors coiled up many times to generate a noticeable voltage. Capacitive coupling or even resistive (eg.insulation not quite a good as it appears) might be more likely.

   - Andy.

Any one  CE marked compliant LED driver at least working properly should not allow more that a fraction of a mA from it's Live/neutral input to come out out through the LED side low voltage terminals,  and that is the maximum that one may expect to get on to the mounting frame.

But a ceiling full of them may well add up to a lethal shock.

Most of us start to feel tingles around about 0;1 to 0,5 mA and it gets very painful and the flinch reflex becomes expected, as we increase the current  up through the whole nos of mA. By 30mA you need to be disconnected pretty pronto to save you - hence RCDs - hoping to be  thrown clear by muscle contractions is not reliable enough....

Exact sensitivities vary quite a bit from person to person, and with other factors such as skin dryness/humidity , area of skin in contact etc,  so treat the figures above as a very rough guide. (in most years there are a few more electrocutions in summer when folk are more likely to have bare arms or even bare feet, and are likely to sweat more.)

As others have said above, the problem is that within the LED driver, there are transformer windings, which as well as being windings are made of metal and spaced by insulation, so also act as the two 'plates' of a rather oddly shaped adventitious capacitor straddling between incoming mains and the ELV secondary side - for a switch mode supply this capacitance  is typically some hundreds of pF to nana-farads, depending on the detail of transformer construction - not enough current will pass to do damage, but certainly enough to be seen on a modern digital meter, and a few such 'capacitors'  in parallel between the supply and metal ceiling grids, so the currents add,  could become  dangerous..

Earthing the ceiling grid will ensure that part at least is held near 0V,  so that the effects of many sensibly spaced lamps cannot add up. but on the other hand this makes the effect of putting yourself  between the grid and  any one light more obvious.

Mike