Starship to Mars: faster than Hohmann transfer trajectories

[waveney]

This summary https://phys.org/news/2025-06-missions-mars-starship-months.html - Missions to Mars with the Starship could only take three months

Was posted on Phys.org yesterday which is a summary of this detailed paper - https://www.nature.com/articles/s41598-025-00565-7

It looks at shorter transit times using higher speeds than those for the traditional Hohmann transfers.

Read then Discuss...

[InterestedEngineer]

"15 full Starship launches per crew Starship".

That's 3000t of fuel, they only need 1500t of fuel, so they are assuming only 100t of fuel to LEO per launch.

Maybe in 2026 window, but after that?  That's too much pessimism.

1500t of fuel with 100t of cargo is a MR of 7, or by odd coincidence 7km/sec of delta V.

They are using 4.6km/sec on the TMI.   So where is the other 2.4km/sec going?  I'm pretty sure the landing on Mars is < 1km/sec of deltaV.

I also don't think 4.6km/sec is a 90 day transfer.  I seem to recall that was on the order of 120 days, but it's been a while since I dug through those calculations.

[InterestedEngineer]

General comments from the paper

All burns are assumed to be done with only the Raptor vacuum engines as the extra thrust from the sea level engines is not necessary during orbital maneuvers.

LOL, how's engine failure going to be handled with no gimbaling.

The setup I typically envision is 2 of the sea level raptors on minimum throttle, and I end up with a net Isp of 367 when I do that, or 3600m/sec exhaust velocity.

We took the dry mass of an empty tanker Starship to be 100t

That's strangely optimistic in the face of their pessimism about 100t cargo to LEO..   I typically use the figure of 150t, which give a MR of 6.0 and deltaV of 6.5km/sec for 100t of cargo.  (I include the landing fuel in the 150t).

DeltaV for a Starship with 200t to LEO (100t dry, 100t fuel) for their scenario with a 1500t tank is 7.2km/sec.  That seems in excess of what is needed to get to LEO from MECO.

DeltaV of booster to MECO is about 2.8-3.0km/sec depending on your assumptions, so the MR of 6 is what gets you the 9.2km/sec typically cited to get to LEO.

By their numbers the cargo to LEO should be 150t, not 100t.  They didn't self-check their numbers.

There are no trajectories with similarly low ejection DV until 2048-04-08 with 4.60 km/s. Each trajectory leaves the Starship with 3.5 km/s of DV, of which 0.5 km/s25 (Appendix B) is used for the landing burn and the other is used for a deceleration burn near Mars entry to reduce aeroloads during aerocapture (Fig. 6).

Oh, that's where all that extra DeltaV is going.

For instance, on the 2035 trajectory with 380s ISP, the arrival speed without a deceleration burn would be 9.73 km/s, but is 6.87 km/s with it. This leaves 1.85 km/s to be shaved off by Mars aerocapture.

So they are assuming that atmosphere arrival speed for Mars needs to be south of 7km/sec.   I can't recall the numbers, but others here have calculated max arrival speeds, and I seem to recall it was on the order of 9km/sec before you couldn't brake without multiple passes.  Anyone remember?

SpaceX will have plenty of opportunities before 2033 to figure out what the max approach speed to Mars is.

At any event, they don't propose a multi-pass capture, so they are using up a lot of engine deltaV as a result.

Looking at their porkchop plots, I can see why they can get 90 days from only 4.6km/sec TMI burn, it'll only happen once in the next dozen transit windows, that's why I thought it's typically more like 120 days for that kind of deltaV.

They didn't do a sensitivity analysis - for example what would 5km/sec TMI burn do?   What would increasing the fuel to the proposed eventual 2300t do?

I"m skeptical of Lambert solver results for deeply (high deltaV) hyperbolic orbits, they always seem to behave funny when I try them.

[Greg Hullender]

"15 full Starship launches per crew Starship".

That's 3000t of fuel, they only need 1500t of fuel, so they are assuming only 100t of fuel to LEO per launch.

Maybe in 2026 window, but after that?  That's too much pessimism.

1500t of fuel with 100t of cargo is a MR of 7, or by odd coincidence 7km/sec of delta V.

They are using 4.6km/sec on the TMI.   So where is the other 2.4km/sec going?  I'm pretty sure the landing on Mars is < 1km/sec of deltaV.

I also don't think 4.6km/sec is a 90 day transfer.  I seem to recall that was on the order of 120 days, but it's been a while since I dug through those calculations.

The full paper is free online. I'm just starting it now. Kingdon, J. 3 months transit time to Mars for human missions using SpaceX Starship. Sci Rep 15, 17764 (2025). https://doi.org/10.1038/s41598-025-00565-7

[InterestedEngineer]

This paper even with its flaws definitely kills DRACO, much as i love nuclear rockets.

Glad to see that done.

[InterestedEngineer]

Starship V3 is going to be 1550t of fuel, w/ Raptor 3s, probably 130t of "really dry" mass (add 20t of landing fuel and ullage for LEO missions).

Given 3km/sec of deltaV from Booster2, that requires 6.2km/sec of deltaV to LEO, so given 3600m/sec exhaust velocity on average yields a mass ratio of 5.6.

1550/4.6 = 336t to LEO which is 336-150 = 185t of excess fuel.  Let's round down to 155t. of excess fuel.

So that's now 10 refueling flights to fully load a crew starship.

The crew starship's dry mass is 130+63 = 193t.  Round to 200t.  The crew gets extra food or something.

The deltaV available is 3600 * ln(1750/200) = 7.8km/sec

TMI by their numbers is 3.2km/sec leftover at arrival to Mars.  their approach speed is 9.87km/sec, and with 1km/sec needed for the landing at Mars that gives them 2.2km/sec of braking, so their atmosphere interface velocity will be 9.87-2.2 = 7.67km/sec (same as LEO), which seems reasonable, and SpaceX has 3 attempts before 2033 to try higher approach velocities.

Basically, the paper has a lot of canceling offsets and assumptions that even with their incorrect/out of date info the essence is still reasonably correct - 90-120 day transits are doable.

Now, one can go even further and do near-GTO refuels and get another 3.2km/sec of deltaV out of the TMI (but divide by 2 because it has to be engine-braked at Mars), so sub 90 day transfers are possible, albeit at double the fuel to LEO, and double the crossing of the Van Allen belts for the crew.

[Vultur]

Yeah, I'm surprised by the deceleration burns. I thought you'd basically go up to the limit of your delta v or the g forces of aerobraking you can stand, whichever is stricter.

[Twark_Main]

Yeah, I'm surprised by the deceleration burns. I thought you'd basically go up to the limit of your delta v or the g forces of aerobraking you can stand, whichever is stricter.

The braking burn at Mars makes sense, because of the short transit.  The arrival speed at Mars rapidly becomes large with these short transits.

However the surprising part is that despite assuming this trajectory, that's not what they optimized for!  Instead, they optimized their trajectory for just the departure delta-v instead. 

From the paper:

We iterated the poliastro Lambert-arc solver over 1-day timesteps in the 2030 s to look for the lowest ejection DV solutions in the Sun reference frame.

Surely the solver should look for the solution with the lowest sum of the ejection DV and the arrival DV.

[AmigaClone]

What type of dV would be needed to enter Mars orbit - at least for a short amount of time? Note that I would expect some combination of aerobraking and burns used to enter and possibly circularize the Mars orbit.

This might be needed for at least part of the cargo Starships once the number of Starships going to Mars at every launch window passes a certain threshold.

[Twark_Main]

What type of dV would be needed to enter Mars orbit - at least for a short amount of time? Note that I would expect some combination of aerobraking and burns used to enter and possibly circularize the Mars orbit.

This might be needed for at least part of the cargo Starships once the number of Starships going to Mars at every launch window passes a certain threshold.

That's easy: it's just the entry velocity minus the vehicle's maximum aerocapture velocity, which is influenced by the payload mass and distribution.

If you want to get fancy, account for Oberth (hint: do it numerically while running time backwards).

[sdsds]

https://forum.nasaspaceflight.com/index.php?topic=25742
An old "Fast times to Mars" thread.

[meekGee]

One of the key moments (IIRC) that influenced my opinion about SpaceX was when a guy called Paul Wooster published some papers on fast Mars trajectories, and almost immediately got hired (quietly) by SpaceX.

The internet says this was almost 20 years ago..

[HMXHMX]

One of the key moments (IIRC) that influenced my opinion about SpaceX was when a guy called Paul Wooster published some papers on fast Mars trajectories, and almost immediately got hired (quietly) by SpaceX.

The internet says this was almost 20 years ago..

Or maybe 37 years ago.

[meekGee]

One of the key moments (IIRC) that influenced my opinion about SpaceX was when a guy called Paul Wooster published some papers on fast Mars trajectories, and almost immediately got hired (quietly) by SpaceX.

The internet says this was almost 20 years ago..

Or maybe 37 years ago.

The hiring, not the concepts...

Back then, Starship was just a glimmer, and SpaceX Mars talk was dismissed as fake PR.  So thar hiring was note worthy.

[TheRadicalModerate]

So they are assuming that atmosphere arrival speed for Mars needs to be south of 7km/sec.   I can't recall the numbers, but others here have calculated max arrival speeds, and I seem to recall it was on the order of 9km/sec before you couldn't brake without multiple passes.  Anyone remember?

At a 9km/s, 70km periapse (slightly different from entry speed, but not by much), you need 2G of downward lift, which (I think) comes out to be 4.5G of inertial forces on the crew, assuming L/D=0.5.  Not very nice after 3 months of microgravity.

The problem with Mars is that it's light and small.  Less gravitational acceleration, shorter radius, both require the negative lift.  In contrast, to stay in level flight on Earth, even at a 12km/s periapse speed, you need a small amount of positive acceleration.

(Really, this planet is a spacefaring civilization's paradise.)

At any event, they don't propose a multi-pass capture, so they are using up a lot of engine deltaV as a result.

Multi-pass capture doesn't help you very much.  The real problem is killing enough speed on pass one to be able to capture.

I'm skeptical of Lambert solver results for deeply (high deltaV) hyperbolic orbits, they always seem to behave funny when I try them.

Me too.  The reason for funny behavior is that they do dumb things to match inclinations.  However, the things they do are smarter than my model, which assumes everything is in-plane.

Broken-plane is almost certainly the way to go, but I have no idea how to compute them.

[Vultur]

So they are assuming that atmosphere arrival speed for Mars needs to be south of 7km/sec.   I can't recall the numbers, but others here have calculated max arrival speeds, and I seem to recall it was on the order of 9km/sec before you couldn't brake without multiple passes.  Anyone remember?

At a 9km/s, 70km periapse (slightly different from entry speed, but not by much), you need 2G of downward lift, which (I think) comes out to be 4.5G of inertial forces on the crew, assuming L/D=0.5.  Not very nice after 3 months of microgravity

Probably still doable though, right?

[TheRadicalModerate]

My notes:

1) I'm just gonna accept their dry mass (100t), prop (1500t), crew module (60t), and Isp (370s) numbers.  This was clearly written before the update.  Seems really likely that a mid-2030's Starship will be the "future" variety.

2) A quick note on prop:  The SpaceX-published numbers are for subcooled prop.  When refueling in LEO, you're probably dealing with boiling prop.  Not completely certain, but a pretty good bet.  De-rate the prop mass accordingly.

3) I've simulated the flip-and-burn on Mars, and if you only have six engines, a 60t payload needs 650m/s of delta-v, starting from an altitude of 5.7km.  Terminal speed at that altitude is about 530m/s.  The model is extremely sensitive to landed mass, due to the high terminal speed, and because the header tanks get bigger very fast. 

9 engines will be better, but landing masses are radically different.  The main lesson I learned from simulating this is that there's a whole bunch of non-intuitive stuff involved.

4) I agree that an aerocapture followed by a landing is probably safer for crewed missions than direct-to-EDL, but then you have periapse-raising delta-v, and a small entry burn.  I can't tell if that's included in their 200m/s "non-optimal Oberth effect" number or not.

5) They don't seem to have any flight performance reserve added in.  That's a mistake.  I'm using 1.2% extra delta-v (about Apollo-level) across the whole budget.  Actual FPR is usually a function of your actual prop consumption vs. your predicted results.

6) My model is circular, with both planets in-plane, but that should make my model more optimistic, not less.  In order to get down to a speed of 7km/s at a 55km periapse altitude, I need an Earth departure angle (vs. Earth's orbital direction) of 42º, and time of flight is 3.7 months (112 days).  That's still terrific, but it's not as terrific as they're getting.

7) Their aerocapture model is much better than mine.  I'm just assuming a particular periapse altitude and assuming that you have to use (negative) lift to maintain that altitude until you're through peak heating.  I agree that L/D=0.5 is correct.  At the periapse altitude, lift and drag should generate about 2.4G of inertial acceleration on the crew.  I don't have a heating model.

8 ) I lost the will to live on the Mars-to-Earth leg.  Maybe later.

[TheRadicalModerate]

So they are assuming that atmosphere arrival speed for Mars needs to be south of 7km/sec.   I can't recall the numbers, but others here have calculated max arrival speeds, and I seem to recall it was on the order of 9km/sec before you couldn't brake without multiple passes.  Anyone remember?

At a 9km/s, 70km periapse (slightly different from entry speed, but not by much), you need 2G of downward lift, which (I think) comes out to be 4.5G of inertial forces on the crew, assuming L/D=0.5.  Not very nice after 3 months of microgravity

Probably still doable though, right?

Depends on the heating model.  As lift goes up, so does drag.  As drag goes up, peak heating goes up non-linearly.

I've never seen anybody discussing 9km/s entry speeds at Mars.  7500m/s, yes, but that only shaves off a couple of days from the 7000m/s speed.

[Twark_Main]

4) I agree that an aerocapture followed by a landing is probably safer for crewed missions than direct-to-EDL, but then you have periapse-raising delta-v, and a small entry burn.  I can't tell if that's included in their 200m/s "non-optimal Oberth effect" number or not.

You don't need these burns.

6) My model is circular, with both planets in-plane, but that should make my model more optimistic, not less.  In order to get down to a speed of 7km/s at a 55km periapse altitude, I need an Earth departure angle (vs. Earth's orbital direction) of 42º, and time of flight is 3.7 months (112 days).  That's still terrific, but it's not as terrific as they're getting.

Technically a circular coplanar model isn't strictly more pessimistic (especially when it comes to arrival velocity, and especially at the better-than-average windows where Mars's radial motion is helping you), so I'm inclined (no pun intended) to trust the real orbits and poliastro.

[TheRadicalModerate]

4) I agree that an aerocapture followed by a landing is probably safer for crewed missions than direct-to-EDL, but then you have periapse-raising delta-v, and a small entry burn.  I can't tell if that's included in their 200m/s "non-optimal Oberth effect" number or not.

You don't need these burns.

If you’re not coming straight in, then your periapse is down in the atmosphere. In theory, you could finish landing at the next periapse. In practice, the post-aerocapture orbital parameters aren’t known well enough to do that without a correction maneuver.

6) My model is circular, with both planets in-plane, but that should make my model more optimistic, not less.  In order to get down to a speed of 7km/s at a 55km periapse altitude, I need an Earth departure angle (vs. Earth's orbital direction) of 42º, and time of flight is 3.7 months (112 days).  That's still terrific, but it's not as terrific as they're getting.

Technically a circular coplanar model isn't strictly more pessimistic (especially when it comes to arrival velocity, and especially at the better-than-average windows where Mars's radial motion is helping you), so I'm inclined (no pun intended) to trust the real orbits and poliastro.

Circular coplanar is usually more optimistic, not pessimistic. So if it doesn’t have a shorter time of flight, that’s weird.

My model works by letting you allocate the delta-v budget how you like, letting you plug in a departure angle, using that to compute your angular momentum, which gives you your perihelion and your departure true anomaly. Now you can compute your arrival true anomaly by setting r = Mars, which gives you ToF, arrival angle, and arrival vInfinity, thereby avoiding needing a Lambert solver.

I believe their solver is producing correct results. I think my results come from the more pessimistic landing delta-v and the addition of FPR. BTW, it’s a straight-in model—no post-aerocapture fiddling.