In article , Rick Hughes
writes

Box will be all fully insulated nylon ...

Never use neons ... what would cct look like with these .... although I
would prefer to use what I just bought .... (if it is possible)

The whole reason for going to capacitor dropper was to avoid dissipating
heat in resistors which many of the simple ccts do.

Just so annoying I followed the design as per here ....
http://www.marcspages.co.uk/tech/6103.htm

(I have had contact from page owner he insists with 220R it works fine,
but for additional reliability it could be increased to 680R )


If you want to test the reliability of the circuit with the original
values then set up a circuit to switch the mains to it every second or
so and if it lasts more than 24hrs then stick with those values to
monitor your occasionally switched heating zones, if they don't last
then you have your answer.

220R will result in 1.5A peak LED current when switched on at the peak
of the ac waveform, 500mA for 680R. I'd say that the former will pop all
but the toughest pulse capable LEDs in short order and that the latter
will be a 50-50 risk.

You still risk popping the LEDs for excessive forward current when lit,
things last longer when derated, which is why they have a nominal and a
max rating, they'll last 10-100 times longer when operated at nominal
over what they will when operated at max.

--
fred
it's a ba-na-na . . . .