Relay leakage question

[julian1]

Consider a relay (rated 10^10 Ohm insulation resistance) with a coil switched by an open-collector opto-coupler with RIO=10^12 isolation resistance, and an isolated power supply/winding. Will current leakage from relay coil pin to pole pin be limited to 10^12 ?

When off, some dark current flows through the collector/emitter of the bjt, and extra darlington if present. But because the supply is isolated, I don't see a possible path for any voltages present on the relay pole pins.

[TimFox]

Draw the equivalent circuit:
The relay spec means > 10¹⁰ resistance from either relay pole to the relay coil.
The coupler spec means > 10¹² resistance between the output side of the isolator and the input side.
One more resistance:  what is the resistance from the isolated power supply to ground, and to any power supply voltage? 
Besides ohmic resistance across the transformer, there may be substantial capacitive coupling of the mains voltage.
That will contribute to the resistance from the relay coil to everything else (including the isolator input).
The off current of the bjt should be too small to affect the position of the relay poles.

[julian1]

Yes, I was thinking a very high-isolation transformer / or separate transformer would be needed. But maybe possible with a secondary coil wound on a separate split bobbin section of an EI core former, or unused section of a toroid. 

What I am not sure about is transformer capacitive coupling for the supply. And if that would make for a DC path. 

It does seem like a lot of complication.

[TimFox]

The winding-to-winding capacitance of the transformer may be important if you worry about 50/60 Hz, but it will not give you DC leakage.
Since power supplies are macroscopic, the leakage resistance to ground, compared with 10¹² , may not be negligibly high.

[free_electron]

And then you put them on a pcb contaminated with greasy finger goop, flux remnants , drool and whatnot. after 5 years the pcb has collected so much dust it's ready to catch fire.