Reginald Neale wrote:

>  In my application C1 was omitted. R2 was between C2 and D5, i.e.,
> C2 would charge up to peak line voltage and R2 limited current through
> D5.  Reflects the reality that if you only need a few mA in a very
> price-driven application, and you don't have a user interface to
> isolate, power resistors are cheaper than transformers.

  OTOH, the argument *for* using C1 is that it runs so much cooler,
making packaging easier and increasing overall reliability, if that is
of any significance in your product.

  I do grant you that the extra mains-rated capacitor costs a bit, but
it's not that large and this is offset by the far more conservative
rating of C2 and R2 (which being a fuze, is deliberately not over-rated.
Maybe there are considerations of cost of fuze resistors I suppose).
--
  Cheers,
        Paul B.