OK, So I answered my own question. I thought I'd leave it here anyway in case it interests anyone else.
I measured the voltage into the aquaero at 11.3v, so I tested at the computer end of the run, and found 11.9v - I am sure I already did this a couple of weeks ago, but it got forgotten about until I got new fans and was fooling around looking at speeds etc. Incidentally, my reported voltage now on the pwm output I use to control my mcp35x is 11.3v, so that is reporting the voltage without any drop through the aquaero itself, since the pump actually draws no power on this connection, just gets a pwm signal. The actual voltage drop through the aquaero itself to my fans is therefore 0.3v which seems entirely reasonable.
0.6v seems like quite a lot of vdroop for my run - 5m of cable to the radbox, using household electrical cable. However something like
http://www.tlc-direct.co.uk/Technical/Ch…oltageDrop.html indicates that I should expect a drop of around 0.3v in the cable. Add into that, the MOLEX connectors and splitters, and my in-line plugs and sockets for detaching the cabling at either the PC or radbox ends, and the losses are not at all unreasonable.
Anyway, 11v is fine, I nearly get rated fan speed and it doesn't look like I can expect much better. At least I found the bad molex splitter and got a little more max fan speed.