I noticed on Boeing's viewgraph of future Delta IV upgrades a mention
of "cross-feed", which I took to mean propellant transfer between
common core boosters in a multi-core configuration. This reminded me of
the occasional topic on this group of Siamese and Triamese booster
configurations, and I wondered if they would be applicable to the Delta
IV core. Using the numbers on astronatica.com for the core (MR of ~8.5,
SL isp 365, vac isp 420), a rough calculation suggests that a Delta IV
Siamese would have a maximum ideal delta-V of almost 11km/sec and could
put a larger payload in LEO than a Delta IV small. If the truisms that
upper stages are more expensive than lower stages and that cost scales
more with number of stages than with stage size are true, would it be
reasonable to try something like this with the Delta IV core (which has
the great advantage of actually exisiting and having flown)? I realize
the Dela IV is no bargain in any configuration, but maybe if one was
just making cores, there could be some economies of scale. What I have
no idea of is how hard technically it is to do the propellant transfer
needed for siamese configurations (given that cost does scale strongly
with complexity). Presumably a Delta IV Triamese could sit on the same
pad as used by the recently launched Delta IV Heavy, and this
configuration could put a serious payload in LEO using the same number
of cores as the Delta IV heavy. This is probably sub-optimal compared
to possible alternatives, but having seen so many good ideas never get
off the ground, I am ever more interested in actual flying hardware.
Has anyone with better calculations looked into this?

Propellant cross-feed (from outboard CBCs to core CBC) is just one
of several methods Boeing could use to increase payload. They're
considering clustering up to six CBCs as well, but that would
require a new launch pad. Cross-feed looks 'interesting' as it
would be necessary to drain the outboards CBCs symmetrically, plus
close off four cryogenic propellant lines in flight while keeping
the core engine happy.

The upper stage will have to grow considerably as well; the
Atlas V upper stage is planned to grow substantially in volume with
up to four engines; both Atlas and Delta upper stages are under-
powered right now.

It should be possible to hit 50 tons to LEO with existing
configurations and pads; above that may force a redesign
to a larger multi-engine core as Lock-Mart is considering
for Atlas "6".

--Damon