Ben Mack wrote:

We found some 50Hz toroids from Maplins that work a treat over the same
wiring, proving the wiring was OK and frequency was the issue.

A very sensible solution I must say, using a transformer rather than a
long-wave transmitter to power your lighting :~)

A further thought is that at such a high frequency (125 kHz) the
inductance of the bulb filaments could also be significant, meaning that
(a) the inductive voltage drop in the wiring is not necessarily in phase
quadrature with the load voltage, thus making its effect worse, and (b)
the voltage across the lamp terminals might need to be significantly
higher than 12V to get the lamp running at the correct power.

--
Andy