Bit of a fiddle but not hard just means filing down ridge from outer plastic dust shield/washer, there are cheaper 7805 bearings on
Ebay (about half the price) tbh they are perhaps better than the shimano bearings but only my view i just want to fit and forget about them so pay the extra get a good quality bearing fit it once.
Know lots of cyclists that buy hope etc, big outlay for a BB, but again only my view i use shimano tiagra find they last ok (again only my view but last as long as 105 or ultegra) but when worn fit a 7805 good thing about these bearings is they have a rubber seal that shimano dont so perhaps keeps water etc out?
Whilst I admire the ingenuity that has gone into this I don't think it's a great idea, but I'm willing to be proved otherwise.
You say in your post that Shimano don't have a rubber seal, which for a start is not true - they do have a seal.
The seal in a Shimano HT2 bottom bracket seems to work really well, and is something that I don't think can be replicated using generic off the shelf bearings. It's not a bearing seal in the traditional sense, which seals between the inner and outer races (indeed, looking at the bearing/seal on its own, there is a great big gap between the seal and the inner race), but between the face of the bearing and the rear side of the plastic/'top hat' washer.
This has a big advantage over a conventional sealed bearing in that not only does it keep water out of the bearing, but also out of the interface between the spindle and inner race of the bearing.
I would think that in your set up, there exists a risk (small in fresh water, much bigger if ridden in winter salt) that the spindle of the crankset may become corroded, potentially resulting in the scenario outlined in Rogerzilla's post above or a need to regularly replace the often very expensive crankset.
Rogerzilla said in his post that the spindles are aluminium. Some definitely are (I have one that is!), but most are steel.
Also, why the need for angular contact bearings? I can't imagine that these bearings see significant thrust loads unless they are installed (incorrectly in my opinion) with significant preload applied using the plastic tool.
