David, also consider the effect of an Earth swingby. Look up the Oberth effect and hyperbolic trajectories on Wikipedia. The swingby amplifies the effect of your TMI. There's also a section on this in the documentation (in the "Deep Space Manual") of the freeware Orbiter flight simulator.
In any event I believe he expects more efficiency from using SEL-2 instead of L1.
At Mars, instead of rendez-vousing with Deimos, you could also brake into a much higher orbit and end up at Sun Mars L1/L2. This should be a lot cheaper. I need to update my spreadsheet for this, but in the mean-time I'd be really interested in what numbers David comes up with.
The depot doesn't have to be so big it cannot be launched without an HLV, especially if you put it at L1.
Sun Mars L1/L2 would be a good place to dock with a landing craft or even a transfer craft to Deimos or low Mars orbit. You could also refuel there. Propellant could be prepositioned by SEP and the Interplanetary Transport Network. It could come from Earth or with ISRU it could come from the moon, Phobos/Deimos or Mars itself.
Earth departure remains the same: .7 to drop from EML1 and a .5 burn at perigee for TMI for about 1.2
Totalling all these, my guess would be 1.84 km/sec for EML1 to SunMarsL1
Even though the Sun Mars L1 takes less delta V, I still favor Deimos and Phobos.
From the Marsden et al paper we're studying, it doesn't take much to get from the earth-moon Lagrange 1&2 to the sun-earth Lagrange 1&2 regions. So it might be possible to get stuff from EML1&2 to Sun-Mars L1&2 with little delta V. But since the Interplanetary Transport Network is still mysterious to me, I have to admit I'm guessing. If there are low delta V routes to Sun Mars L1&2, I'd expect them to take a long time.
As I mentioned earlier, I think it'd take about 1 km/sec to get cargo from Deimos to SML1.I don't like the idea of Mars supplying propellant. Mars has an atmosphere comparable to earth's Mesosphere where most meteors burn up. So I believe an ascent burn is needed, like on earth. I've been told Martian gravity/drag penalty is about .8 km/sec. So it would take 6 or 7 km/sec to get stuff from Mars to SML1.
A reasonably functional depot (with sun-shade, solar power, active-cooling, safe haven etc) will have reasonably large mass and volume. Think something SkyLab size here. While it would be possible to launch one in several pieces and then launch crew to assemble it, a single 50t+ HLV with a big payload fairing would make the whole process (& depot design) a lot simpler.
At Mars, instead of rendez-vousing with Deimos, you could also brake into a much higher orbit and end up at Sun Mars L1/L2. This should be a lot cheaper. I need to update my spreadsheet for this, but in the mean-time I'd be really interested in what numbers David comes up with. Sun Mars L1/L2 would be a good place to dock with a landing craft or even a transfer craft to Deimos or low Mars orbit. You could also refuel there. Propellant could be prepositioned by SEP and the Interplanetary Transport Network. It could come from Earth or with ISRU it could come from the moon, Phobos/Deimos or Mars itself.It's fun to see how small the burns are if you stage at high-energy orbits, well within reach of hypergolics. No HLVs needed.
However then i had an idea. Remeber Farquhar powered lunar swingby ? Well, what about a powered Mars swingby, to Sun-Mars L2 ? If that works for the Moon, it should work for Mars isn't it ?
A FLEM mission boosted toward Mars using a nuclear rocket during the favorable 1971 transfer opportunity could have a mass as low as 260,000 pounds at launch from Earth orbit, Titus estimated, perhaps permitting a piloted Mars stopover with only a single Saturn V launch.
Current manned Mars mission designs are looking at around 400t of cryo propellant in LEO. Probably 200+ at EML-1?