Hi Torquetalk,
It's very brave of you to set yourself up as a target.
(Most of that which is said below applies when more than nine passengers are carried.)
As Mars posted, the absolute performance requirement is set out in Part 29.1; however, caution is advised because, I am reliably informed, no helicopter has been submitted solely for certification in accordance with 29.1(e) - i.e. all the Category A requirements except for Subpart B - Flight (with the exception of those three stipulated). Manufacturers submit their helicopters for Category A certification when equipped with more than nine seats - i.e. they ignore the break at 9,072 kg.
When compliance is shown with 29.1(e) (specifically, in compliance with 29.25 and 29.67(a)(2)) the resulting performance graph represents the Category A WAT - which can be used for establishing the PC2 take-off mass.
The difference from certification in 29.1(c) and 29.1(e) is the requirement to provide
the Category A 'procedures/profiles'; as Mars said, the three rules mentioned are those which equate to operations in PC2 - i.e. the provision of an HV diagram, its stipulation as a
limitation and second segment performance.
The main reason this is being penned (apart from providing additional information), is because of the statement in your post:
"because they don't meet other design criteria, or because they are pretty rubbish at staying aloft following an engine failure at their certificated max weight"
That can only be true for Part 27 twins (which have not been submitted for Category A certification), or Part 29 twins that are not intended for passenger transport (I don't know of any).
The whole amendment of Part 29.1 in 'FAA NPRM 80-25' (and the introduction of a discriminant of more than 9 passengers) was to permit a lower certification basis for utility helicopters only (or utility operations for those submitted for dual certification).
Jim