Hi Indy-Trier, I think you're all right. As far as the "two word" thing goes, I have the same problem. The preamble says that the components can be deduced from three pairs of unclued entries (two of two words). I think it's poorly worded, but have taken the "two of two words" to refer to the that to mean two of the descriptors of the components have two words, rather than that two of the unclued entries have two words, as I agree only one does.
So 1d + lowest of the three unclued across entries is a descriptor of the first component, and the right hand unclued down entry is a descriptor of the second component. My full breakdown was here:
http://www.theanswerb.../Question1119455.html
I hope that makes some kind of sense!