Hello Dani,
the RD15HVF1 was tried in an early prototype of the H-L PA; if I recall correctly it did not bring any big advantage w.r.t the current PA with the AFT05MS003N and could actually be not as stable and could oscillate in some conditions.
As you saw the bias circuit can be modified to provide a bias voltage in the range needed by the RD15HVF1 gates; while this bias circuit configuration has been simulated I don't think it has actually been verified in practice.
You can, and actually should, use a 12 V (or up to 13.6 V) drain supply for the RD15HVF1 to get more output power in the lower HF; if you increase the PA driver gain, at 1.8 MHz with 13.6 V of drain supply you can get more than 20 W at 1 dB compression but at 30 MHz you'll get only about 10 W at 1 dB compression.
For reference here below is a graph of the output power measured for an H-Lv2b2 a RD15HVF1 PA (I think this was at 12 V drain supply)
and for comparison what was obtained with the current PA configuration at 7.5 V drain supply. Note that this is not exactly the current PA on the H-Lv2, as its drain supply there is 8 V and the driver gain was lowered a bit but should not bee too different.

In case you want to do some experiments with a RD15HVF1 PA I can try to dig out more measured data and build info.
73 de Claudio, IN3OTD / DK1CG