tavert

Thanks, glad you liked it! Setting up the fuel tanks to deplete low-Isp, high-TWR engines before high-Isp, low-TWR engines was how I got a rocket-only "SSTOMAB" to work, many months ago (before the 48-7S and its current overpowered-ness): http://imgur.com/a/uA4c5 This technique is the only way to get around the single-stage limit at TWR=1 of 6746 m/s for a 48-7S with zero payload. I'm looking now at a different way of laying out the design with no fuel lines at all, just clever timing of engine-to-fuel-tank ratios and burn times. This might look saner (build out instead of up), but the optimization problem's a very different formulation. My two designs so far were both determined by a mixed-integer nonlinear optimization solver, and the code for the SSTO was fairly similar to the asparagus version, just had to change the dry mass calculation to account for not dropping empty tanks and burnt-out engines.

Here's a 9124 m/s SSTO at under 100 tons: And in orbit, same number of parts (246) minus the 3 launch clamps: I'm pretty sure this is a record for a rocket-only SSTO that can get itself into orbit (until shown otherwise by another entrant here, I guess). The vertical design was so I didn't have to worry about symmetry and complicated fuel flow. There's a little gap below each set of 48-7S's with a couple of cubic struts, attached in a way that they don't crossfeed. Could've also used docking ports and disabled crossfeed that way, but they're heavier than struts MechJeb doesn't get the TWR right, since all of the 48-7S engines are out of fuel by the time this reaches orbit. 60 kN thrust from 1 LV-N, divided by (26.67 tons times 9.81 m/s^2) gives TWR = 0.23, at least at the point the screenshot was taken. Minimum TWR is a little lower, right at the moment the last of the 48-7S engines runs out of fuel.

I had a 29+ km/s design built to a 500 ton limit (and something like 600 parts), but it kept losing control after dropping the first stage. I think I'll need to impose a minimum percentage of vectored thrust to keep it controllable. Also, mhoram, I tweaked my original design a little for another few hundred m/s about an hour after initially posting, hoping no one would notice.

When you add a category for 100 ton mass, here's my entry for it at 21580 m/s (and 195 parts, was optimizing for mass not part count though):

Can you clarify exactly what you mean by "powerful enough to leave Kerbin (Kerbin orbit or Kerbin-escape)"? Does initial TWR have to be >= 1?

I did it in 2.022 tons (not counting the Kerbal) a few versions ago, and with MechJeb flying: http://forum.kerbalspaceprogram.com/threads/39106-Exploiting-quirky-ladder-physics-to-get-to-the-Mun-and-back-in-2-022-tons You can do it in just under 2 tons now using a 48-7S, FL-T100, probe, decoupler, a few intakes, and a jet. If we get tweakables in the next version, using a few round-8 tanks without oxidizer might allow you to go a little bit lighter. @Johnno orbit is doable in 0.8 tons (or even lighter by exploiting infiniglide), go a few pages later in that thread you linked (using parts that were from KSPX at the time I posted, but are now stock - and have even more thrust than they did in KSPX).

I recommend you folks read through this thread from a few months back: http://forum.kerbalspaceprogram.com/threads/46194-I-need-someone-help-me-do-some-math-for-launch-optimization?highlight=launch+optimization 3 main approaches for this optimization problem: pseudospectral methods, local discretization, or using the Pontryagin principle to express the solution in terms of a boundary value problem. Alterbaron got a pseudospectral method using PSOPT (and possibly local discretization via an option setting with the same software?) working for the single-stage problem, fixed design. Multiple stages are trickier, you have to split up the time horizon for each stage, scale the time variable by a different interval in each stage, and stitch together the dynamic conditions between the end of one stage and the start of the next.

That video is an extremely low TWR, yes. But it demonstrates the type of trajectory that I'm calculating for. The lower the TWR, the more difference the type of trajectory makes. The ratio of final mass to initial mass can easily be calculated from delta-V and Isp using the rocket equation. The payload fraction data for Tylo I'm showing here is quite simple to derive from the initial delta-V numbers, you just need to plug in the engine thrust-to-weight data and the fuel tank mass fraction (9 units wet per 1 unit dry, for all but the smallest stock fuel tanks). You need to consider not just fuel mass (and dry tank mass), but also engine mass. Higher TWR means you need less delta-V to land, but you need to carry more mass in engines.

Yes they are. From that same page, And as I said, test it. The numbers in the cfg files and VAB/SPH tooltips are not always interpreted the same way by the game's physics engine.

Yes. If you're landing on Tylo with a single stage, here's what you can expect the payload fraction to be, meaning ((final mass) - (engine mass) - (propellant mass used in landing) * 9/8) / (initial mass): If you want to land and take back off with the same stage, your payload fraction will be quite a bit lower (this is hard to do), and the best engine to use is LV-N at an under-1 starting TWR (it will go higher than 1 as you burn fuel): Things change a bit as you stage, your TWR changes when you stage, and it can change either up or down depending how many engines and fuel tanks you drop.

There are very, very few situations in which an LV-909 is a better choice than a 48-7S, if part count isn't an issue. http://forum.kerbalspaceprogram.com/threads/45155-Mass-optimal-engine-type-vs-delta-V-payload-and-min-TWR

Test it. Make a small rocket with TWR just barely over 1. Note its mass in the info panel on the map screen on the launch pad. Go back to the VAB, add a whole bunch of landing gear, and try again. Same mass, same TWR, everything flies more or less the same.

Lots of math on constant-altitude landing (and takeoff) here: http://forum.kerbalspaceprogram.com/threads/39812-Landing-and-Takeoff-Delta-V-vs-TWR-and-specific-impulse The lower your initial altitude and vertical speed during the landing burn the better, gravity losses are your enemy at low TWR.

In the stock game, the landing gear are currently massless. Ignore the indicators in the SPH after you put the landing gear on.

Here's the first thing I threw together, design's not totally optimized but took plenty of tweaking to get the ascent trajectory right. 18.65 ton payload, 95.62 tons on the pad, 9 stages, 19.50%.

It's not a big deal if it's just a tiny fraction of the payload, but it would make a substantial difference if the whole payload is made up of low-drag parts. You have to draw the line somewhere.

You probably need another rule that the payload must consist of >= 0.2 drag coefficient parts. With 48-7S's and a payload of RCS blocks, or if those are forbidden even when unfueled then aircraft cockpits, 20% should be quite possible. Standard-drag I'm not so sure what it'll max out at, will give it a shot shortly.

I think I tried flying it in several configurations, with MJ trying to hold a direction, with just SAS, and with nothing, and it would always go snaking around like crazy. The gimbal direction bug certainly doesn't help though. And the fact that the last stage was just drop tanks (that's what the optimizer told me) meant it fell off immediately...

This morning I was playing around with Eve ascent configurations taking advantage of the inexplicably uprated 48-7S, and with this linear staging trick it should be doable in 3 stages, each with 1 ton of fuel and 1 48-7S. You can add an ant-and-oscar final stage, but it's only good for another 400 m/s or so total. Single digits for the round trip is absolutely doable now, just pick how few ion engines you have the patience to get away with.

Here's a theoretical solution that came out of a MINLP solver after a few days: The now-even-more-overpowered 48-7S results in the part count constraint being active, potentially giving another 1500 m/s or so beyond what I came up with in 0.21. It looks like I could have used up to 10 seconds longer burn time since MechJeb actually rounds down the numbers it displays, which I wasn't taking into account in my optimization. This takes Leonon's "linear staging" concept to a ridiculous extreme. I've actually made use of this glitch unintentionally before to mount landing gear below engines, but never tried mounting whole stages below running engines. If there's a physics-less strut right below the engine, it works fine with no blocked thrust or part damage. Definitely the lowest-mass, lowest-part-count staging method I've seen, though it results in crazy tall rockets (hence this one was put on its side and pointing backwards in the SPH). Sadly, it tears itself apart in flight. I might have to go back to a more conventional horizontal asparagus setup with prescribed symmetry in each stage and leave more room in the part count for struts...

I'll second that. How tough would it be to extend your map-extracting plugin to obtain the Biome info? Your overlay GUI would be perfect for showing biome maps.

MechJeb's fuel flow simulation is a bit more sophisticated when it comes to SRB's and mixing SRB's with liquid engines. The TWR on other planets/moons is pretty simple to scale, atmospheric dV on Duna/Laythe slightly less so. Otherwise, it's mostly a matter of personal style preference. Or inability to resist clicking the automation buttons in MJ...

My dislike of imperial units is certainly coloring my argument that the convention of including the acceleration of gravity on the surface of Earth in the conversion between effective exhaust velocity (which has a perfectly consistent physical definition and interpretation, with no arbitrary magic-number conversion constants thrown in whatsoever) and specific impulse as a time is a bad convention. We can leave imperial units out of it for a second, if we're ignoring where the convention came from in the first place. If, as you had said earlier, specific impulse were defined as the amount of time 1 kilogram of propellant would last if burned at a thrust of 1 Newton, then specific impulse as a time might make sense - at least the conversion factor in there would be 1 m/s^2, which from F = m a is the acceleration that 1 kilogram would undergo if 1 Newton were applied to it. But this is not the case. There's a factor of g0 arbitrarily inserted into the definition. That factor has no business being there if we're trying to talk about universal physical behavior. Or going even further down the rabbit hole, even kilograms, meters, and seconds have anthropocentrically chosen values, so 1 m/s^2 would also be a non-universal choice. If you want to call specific impulse a time, you have to at some point choose a ratio of thrust to mass (so an acceleration) to define as your reference value. There's no way to make that choice in a purely physical way that isn't biased by the values we've chosen to use, from our history and home planet. Unless we all used Planck units instead of SI, but those have crazy values and we don't know Newton's constant with enough precision for Planck units to be scientifically useful... So yeah, effective exhaust velocity is the only purely-from-physics unbiased convention that could rightfully be used here.

Well there's going to be a certain design-dependent minimum amount of gravity loss that is required to get from stationary on the surface to the lowest-altitude stable circular orbit. If you consider the minimum-altitude stable circular orbit your reference point, you have to climb out of the gravity well (or the lack-of-horizontal-speed well...) to reach that point. From then on, around that reference, you can go into elliptical orbits and things will fluctuate up and down.

Think of it as (ending speed) - (starting speed) = (delta-v applied) - (drag losses) - (steering losses) - (gravity losses) The gravity losses aren't really permanent, in an elliptical orbit the kinetic/potential energy shifts back and forth so speed is higher at some points, lower at others. Gravity losses there depend on flight path angle gamma, which is angle to velocity vector, not thrust angle alpha. Thrust perpendicular to position vector would be alpha = 0...