That does seem to be the core problem with cooling cells - a 60k coolant averages out to a regular heat vent, so they only shine in a Mark III or Mark V. Without RedPower, Mark III & V are far less appealing, especially given how slim the gains are now. I'd disagree about the "ridiculous cooldown time" part, however - the incredible heat output of adjacent quads tends to mean you've got an impressive cooling system, so once the reactor shuts off, all that can be applied to the cooling cells.
----
For Copper Plates, it looks like they each cost 400k EU (or more accurately, ~430k EU including the Recycler - still need to factor in macerator and compressor to be truly accurate):
(166666 EU/ UU matter, 3 UU matter / 5 copper ore, 2 bars / ore, 8 bars / plate)
So a Quad costs 2m extra (5 plates * 400k EU/each). A 2x2 of single-cells is 3 eff, as is a quad, meaning you lose 2m with the Quad. Adjacent Quads is 4eff, gaining you 4m over using 8 cells (3.5 eff) ... which is exactly as much as you spent on Copper Plates (excluding processing). So to actually get a *gain* from Quads, you either need more of them, or a more complex arrangement (e.g. a quad with adjacent single-cells). That's much worse than I'd expected, so thanks for getting me to look more closely at the numbers >< It does still provides a relatively efficient copper->EU conversion, though, which can be circumstantially useful, and of course it does create a more compact reactor layout, allowing for more cooling components.
Ignoring processing costs (macerating, compressing, recycling), a Reflector costs 4.2 UU, or ~700k EU. Each Reflector seems like it would yields a raw 1m EU extra, translating into a ~300k EU gain. While the processing costs will eat into that, I suspect the Reflectors are still a gain overall, but probably not a large enough gain to be worth the hassle. I expect the Thick reflector is a net loss or negligible gain, as the 400k EU copper plate would probably consume any actual gain from the four reflectors.
All of this, of course, presumes you need to account for the materials consumed. If you're just burning what you dug out of the ground and would otherwise sit unused in a stockpile somewhere, then most of these 'losses' could evaporate. That's largely a playstyle preference, but useful to keep in mind with respect to the value of the feature to players overall.
----
Something to make reactor design more 'interesting' would certainly be welcome. Personally, I think the Component Vents are the primary source of the problem, because they provide a unique and incredibly powerful capability (moreso, imo, than the Overclocked Vents, if only one of the two were present).
You essentially seem to be suggesting a world which heavily favors Mk II, III, and V reactors over Mk I, which I agree would be more appealing. How you'd actually balance that, I have no idea, and figuring out how to make such a system *work* is a large obstacle.
Thats all I use.