Asymmetrical Payload Boost

Paul451 made a suggestion on the last post that triggered a thought on boosting payloads on many current vehicles at a reasonable cost. I claim 50% credit for this concept with Paul getting the rest.

He suggested that the tanks on the last post should be thin walled balloon tanks with a structural truss from the thrust structure of the first stage to the second stage. This would allow modular stacking and swap outs with minimum pain. It would also allow greater flexibility in modifying vehicles for different missions. The thin wall tanks were to compensate for the mass of the structural truss as well as for cost reasons.

My quibble was with the truss mass, though the ability to go with cheaper and lighter tanks appealed greatly. My thought based on his thought started with using an existing LV for the structural truss with the thin wall tanks attached to the sides much like in the previous post. Then SRBs are attached to the outsides of the strap on tanks. My squeamishness about SRBs is not shared by several of the experienced companies that actually launch rockets.


Several companies use strap on SRBs as a matter of course. The Atlas sometimes uses a single SRB, which answers many of the concerns expressed in comments about asymmetrical loads or thrust vectors expressed in the previous post. A variety of existing and well known SRBs are available to launch companies right now, as opposed to the liquid engines that I would prefer. SRBs are frequently mentioned as relatively inexpensive engines. My qualms have to do with the unfortunate failure modes.

If two tanks are strapped onto an LV with direct feed to the core engines, the core will be far too overloaded to leave the ground. Add enough SRBs to the outside of the strap on tanks and the vehicle will have the TW to launch. At SRB burn out, the core vehicle’s engines will be at full vacuum thrust. A lofted trajectory will allow the total vehicle to have a TW of less than one including the strap on tanks. The vehicle first stage mass ratio from Mach 3-4 and 25 miles could be as much as double that of a standard core vehicle doing a ground launch. Payload could triple or so.

An Atlas 5 with this modification could match the theoretical payload of an Atlas 5 heavy with the expenditure of one RD170 instead of three.  Other vehicles run by companies not squeamish about SRBs could get similar results.

As pointed out by Peter in the previous post, it is not required to stop at two strap on tanks as the Saturn I for instance had multiple parallel tanks.

In the event of SRB cato, the strap on tanks act as shrapnel catchers. This might make it possible for standard LVs to use SRB boosters with less concern about losing the core vehicle and payload. The payload might make it into LEO instead of GEO after losing the extra boost and propellant.  A tug could be designed and tasked to rescue these stranded vehicles and take them to their intended destination.


The following two tabs change content below.


I do construction for a living and aerospace as an occasional hobby. I am an inventor and a bit of an entrepreneur. I've been self employed since the 1980s and working in concrete since the 1970s. When I grow up, I want to work with rockets and spacecraft. I did a stupid rocket trick a few decades back and decided not to try another hot fire without adult supervision. Haven't located much of that as we are all big kids when working with our passions.

Latest posts by johnhare (see all)


About johnhare

I do construction for a living and aerospace as an occasional hobby. I am an inventor and a bit of an entrepreneur. I've been self employed since the 1980s and working in concrete since the 1970s. When I grow up, I want to work with rockets and spacecraft. I did a stupid rocket trick a few decades back and decided not to try another hot fire without adult supervision. Haven't located much of that as we are all big kids when working with our passions.
This entry was posted in Uncategorized. Bookmark the permalink.

11 Responses to Asymmetrical Payload Boost

  1. Bob Steinke says:

    Drop tanks plus SRBs does seem like a neat idea to add payload capability to an existing launcher especially if you can get the SRBs to burn out before maxQ when you will probably be throttling down anyway.

  2. Christian S. says:

    The whole concept of “parallel” LV sounds a lot like the first stage of Saturn I. In that case, tanks and tooling from existing rockets were used, albeit without the asymmetry factor. The Shuttle stack would count as an asymmetric parallel LV too, I guess.

    On the other hand, I am not sure if using existing tooling and tank structures would result in much of cost saving. Space X manages to make the Falcon at lower cost than Atlas, and they built it from scratch. In case of Saturn I there were cost savings because the Redstone and Jupiter were mass manufactured missiles. None of the launchers today is truly mass manufactured. They are really artisan creations with hours and hours of manual labor and testing on them. Which is really easy to understand given the tyranny of the rocket equation.

    To make space access cheaper we have to work around the rocket equation, because the engineering constraints it imposes are simply driving up the cost too much. One way would be to use maglev like launchers or cannons, but then we run into large capital costs. But, even reducing the required deltaV from 10 km/s to 7 km/s and exhaust velocity of 4.5 km/s (LOX/LH) would enable mass ratios below 5, which would enable building a cheaper vehicle because problems can be solved by throwing aluminium at them, instead of a nanoengineered structure from an alloy of unobtanium and sueprexpensivium (it is obviously a colorful exaggeration). Reducing the deltaV to LEO from 10 km/s to 7 km/s is basically getting rid of air drag and gravity drag.

  3. George Turner says:

    If you used four external tanks then the two types could be mounted on opposite sides of the core, so you’d stay symmetric. If you mounted the solids in the corners (of a tic-tac-toe grid) then each solid could directly attach to two of the external tanks, while each external tank is attached to two solids (one on each side of it), to distribute the loads. However, you have to drop all the external tanks at once or you get a really off-center mass distribution.

    For a solution with two external tanks with LOX RP-1 at a 2.56:1 O/F ratio, let me suggest that you keep the propellant masses symmetrical. Given a total propellant loading of 3.56 * X, where 2.56 * X is oxidizer and 1.00 * X is fuel, you’d want a mass of 1.78 * X in each tank. So one tank is loaded with 1.78 * X of LO2 and the other tank is loaded with 0.78 * X of oxidizer and 1.00 * X of RP-1, which requires that tank to have a bulkhead, but only that tank.

    The standalone LOX tank’s mass flow is set to 1.78 * Y, while the other tank feeds LOX at a mass flow of 0.78 * Y and RP-1 at a rate of 1.00 * Y, giving you a 2.56:1 O/F ratio while keeping both tanks at the same weight through the whole ascent.

  4. john hare says:


    Using existing tooling is likely to be much cheaper that building more new tooling. There is also ground transport to consider when suggesting larger tanks that don’t fit on highways and trains. One of the points of the post was that relatively cheap light balloon tanks with SRBs could boost payloads for an existing LV at fairly low cost.

    There are many methods of reducing DeltaV for a given vehicle. Staging, airlaunch, orbital rotovator, Lunar pellet stream propulsion, RTLS stages, and so on. We don’t have to work around the rocket equation, we have to work with and understand it, along with various other fields.


    Thrust vectoring is cheap and simple.

  5. George Turner says:

    I was also thinking of commonality between the two tanks, but another thought occurs to me given that the volume ratio of oxidizer and fuel is very close to 2 to 1. Make 3 virtually identical tanks and fill one with RP-1 and the other two about 96% full of LOX. If you space in a triangle so the angle between the RP-1 tank and either LOX tank is 113 degrees (and the angle between the two LOX tanks is 134 degrees), the center of mass stays at the center axis.

  6. john hare says:

    This still doesn’t match masses as the LOX will mass 2.5 times the Kero, which has the Kero tank carrying 80% of the mass of either LOX tank. Also LOX has a density 1.42 times that of Kero which makes those tanks more like 87% full each. These are not show stoppers if one decided to do this anyway, but it still leaves an asymmetrical vehicle after all that effort.

    Another problem with the triangle is that it could more easily interfere with normal pad access to the payload. Inconvenience can get expensive in the aerospace field.

  7. George Turner says:

    Pad access might be a problem, but possibly not that much worse than having large clusters of solids all around, but it could end up like launching a Soyuz. My thinking on the balance is that by changing the top angle on the ‘Y’, where the tops are the oxidizer tanks, you can balance out almost any mixture ratio, even hydrogen, by bending the top LOX arms downward. Of course, with LH the tank volumes wouldn’t work at all.

    Or you could use a big tank and a little tank, possibly extending the concept into the second stage, where the long second stage tank was above the short first stage tank and vice-versa. That idea is inspired by the Space Shuttle carrying a 66,000 lb tank almost all the way to orbit, instead of splitting the fuel between two tanks, a big and a little one, allowing it to dump more dead mass on the way up.

  8. johnhare johnhare says:

    Sounds complicated. I can’t quite figure out what your idea or your point is. Thrust vectoring through the center of a shifting mass is simple, cheap, and well understood. What benefit is there from making things complicated?

  9. George Turner says:

    Well, having to tip and yaw at lift (maybe 2 degrees) off might look a little alarming at first, but I don’t think it’s an issue. In any event you could prevent by having the solids apply a canceling torque, either by having different thrust levels or different mounting offsets.

    I’m mostly just exploring the design space of separating the fuel tanks from the oxidizer tanks to see if there’s other local optimums where you can get some other serendipitous benefits.

    As an example, last night I ran some simplified numbers (based on cylinders) to look at a single RP-1 tank of diameter d and length L, offset by a LOX tank of length 1.8 L (similar to your example drawing), keeping the external tank diameters the same for tooling commonality. Then I went with a core first stage of diameter 1.66d and length L, to match the length of the RP-1 ET. It’s kind of odd looking but lets the first stage ET’s and the first stage core have about the same fuel mass.

    Then, on top of the first stage LOX ET of length 1.8 L you have a stubby second stage LOX ET of length 0.47 L. On top of the 1st stage RP-1 ET of length L you have a second stage LH2 tank of length 1.27 L, so the stacked ET’s are the same length at launch (though not nearly the same weight). That took a little algebra on relative fuel/oxidizer densities so the ET stacks came out the same length with proper O/F ratios.

    That configuration gives you a second stage with the LH2 ET the same length as the second stage core, with an stubby LOX ET attached on the other side near its top. The relative propellant masses work out to give you about 100 X of RP-1/LOX propellant in the first stage ET’s, 100 X in the first stage core, 22 X of LH2/LOX in the second stage ET’s, and 44 X in the second stage core, so the second stage total fuel mass is about a third of the first stage fuel mass.

    It looks symmetric on the pad (most of the mass on the LOX side), becomes very asymmetric after the first stage ETs drop off (one is long and one is short), stays asymmetric after the first stage core separates (a long core and LH2 tank and a stubby LOX tank), and stays that way until the second stage ETs drop off. But the relative stage weights look pretty good for an RP-1 first stage and a cryogenic upper stage, so maybe it could be some serendipity.

  10. john hare says:

    By George I think you might have a shooting solution here.
    Compact, simple, strong, and mostly common tooling in all tanks.

  11. George Turner says:

    Well, here’s another serendipitous thing. LOX is 2.7 times as dense as liquid methane (at the boiling point), and 2.7 is a fine mixture ratio for a methane engine like Raptor, so your LOX tanks and your fuel tanks would be the same dimensions.

    You could use a variation of the above double stack idea by having the LOX ET on the second stage sitting on top the first stage’s methane tank, and vice versa on the other side. That way the vehicle is balanced at liftoff (no yawing needed), grows gradually more unbalanced toward first stage ET separation, gradually shifts (due to first stage core fuel consumption), then gradually and completely rebalances as the second stage ET’s propellant is consumed (since both empty tanks are the same size and weight).

Leave a Reply

Your email address will not be published. Required fields are marked *