What is the most efficient amount of stages for a rocket? (Answered)

[Spaced Out 8]

How many and what size for each? I know that one large 1st stage and a small but efficient second stage is great but is there any better configuration for maximum delta-v?

Edited February 5, 2018 by Spaced Out
Mispelled a word

[Snark 13,158]

21 minutes ago, Spaced Out said:

How many and what size for each? I know that one large 1st stage and a small but efficient second stage is great but is there any better configuration for maximun delta-v?

It basically comes down to this:

The design of each stage is going to result in a certain mass ratio, i.e. what fraction of the stage's total mass is fuel.  You don't want that to be too low (because then you're wasting lots of dead mass, and doubling the fuel would nearly double the dV)... but you also don't want it to be too high either (because then you're making your stage too big, and too much of your mass is dead weight of big empty fuel tanks).  So usually you end up with a mass ratio somewhere in the 1.5-to-2.5 region.  Given the typical Isp of KSP engines, that usually means you end up with something on the order of 1400 to 2000 m/s of dV in each stage, for "conventional" LFO engines (I'm not including the LV-N or ion engines).

So, that's a handy rule of thumb for efficiency:  each stage gives approximately 1400 to 2000 m/s of dV.  (Higher for high-Isp engines, lower for low-Isp engines.).  So... the way to answer the question "what's the best number of stages?" is to take the total dV that you need based on your mission profile, and just keep adding stages until you reach that amount.

The reason that "3 stages" is such a common design is that "1400-to-2000 m/s per stage" usually gives a well-designed three-stage rocket somewhere north of 4500 m/s of dV, which is enough to get to LKO with sufficient spare dV to go interesting places and do interesting things.  That's enough dV to get to Mun or Minmus orbit, or take a one-way trip to Duna or Eve, for example.

If you're only going to LKO, then two stages are generally enough, if designed well.  If you're doing something that's more dV-consumptive than "orbit Mun and return" or "one-way trip to Duna", then adding a 4th stage will get you nearly anywhere in the solar system.

[GoSlash27 4,258]

20 minutes ago, Spaced Out said:

By maximum delta-v, I meant for an all-purpose vehicle, but thanks for the tip.

Spaced Out,

 There is no efficient "all- purpose" vehicle. An efficient vehicle is engineered for a very specific purpose. For a rocket stage, there are 3 questions that need to be answered:

1) How much payload does it need to accelerate
2) What is the minimum acceptable acceleration rate
3) What is the minimum total change in velocity it must provide

Only a stage that is designed to the mission requirement for the least mass and/ or cost can be considered truly "efficient".

Best,
-Slashy

[Cpt Kerbalkrunch 1,061]

I don't wanna sound flippant, but whatever you need to do the job is the best configuration. I usually use three stages, but it largely depends on where you're going, what you're doing, and what you're bringing with you. If you're going far and bringing a lot, you're gonna wanna dump dead weight as often as possible to maximize your efficiency. Someone  else may be able to give you a more technical answer. Not sure how else to phrase it. I just see what I need, and try to travel as light as I can without sacrificing anything I want or need to bring with me.

Edited February 4, 2018 by Cpt Kerbalkrunch

[Starman4308 1,699]

I'm going to quote myself a bit here:

19 hours ago, Starman4308 said:

As with many questions of optimal booster design, there is simply way too much "it depends" and non-analytical math going on to make absolute answers impossible.

In the end, there is a lot of dependence on mission profile, payload size/mass, available engines, and what you're trying to optimize for.

For example, if optimizing on mass, you'd tend to use all liquid-fueled engines... but if optimizing for cost, first-stage solid rocket boosters are often an excellent choice. They have mediocre specific impulse and full/empty ratios, but they're cheap.

In the end, a lot of it boils down to trial and error: design your payload and work backwards through the mission, trying out several different choices until you're satisfied with it.

[5thHorseman 12,749]

That's like asking what's the best car.

Are you racing, or hauling logs? Or are you a soccer mom?

Same with a rocket, though maybe not so extreme (no soccer games in orbit... yet).

In general, though, I like to have 2 stages to orbit, each with about 1700 m/s dV. Or, each with 1500 and boosters for the extra 400 or so off the pad.

[Snark 13,158]

1 minute ago, 5thHorseman said:

no soccer games in orbit... yet

I beg to differ. 

http://www.pcgamer.com/kerbal-space-program-official-mod-brings-the-world-cup-to-space/

[foamyesque 409]

It depends entirely on what you want to do and what you have available to do it. A launch to LKO generally only requires two stages (.6TWR/1400m/s dV minimum upper, 1.2TWR/2400m/s dV minimum lower), but what your mission may require beyond that can be basically anything. Do you need an interplanetary ejection stage(s)? Landing stage(s)? Return stage(s)? You can easily wind up with machines running around with eight, ten, or more stages once you start accounting for drop tanks, boosters, crossfed staging, etc.

[Spricigo 843]

1 hour ago, Spaced Out said:

 for maximun delta-v?

That's something you shouldn't go for when designing a vessel.  You are essentially guaranteeing to always be inefficient. 

To increase deltaV you need more fuel, and because of this fuel mass you need stronger engines and more fuel to move it. And if your mission don't requires that fuel the only effect it had is to make your burns slower.

 

So, stop to think about "maximum" deltaV. Take how much deltaV your mission require and design your vessel to have that amount of deltaV with just a safety margin, no more. 

Notice that a similar reasoning apply to others performance parameters. Make your vessels with enough performance,  don't worry about maximum (is too costly).

[Spaced Out 8]

15 minutes ago, Spricigo said:

That's something you shouldn't go for when designing a vessel.  You are essentially guaranteeing to always be inefficient. 

To increase deltaV you need more fuel, and because of this fuel mass you need stronger engines and more fuel to move it. And if your mission don't requires that fuel the only effect it had is to make your burns slower.

 

So, stop to think about "maximum" deltaV. Take how much deltaV your mission require and design your vessel to have that amount of deltaV with just a safety margin, no more. 

Notice that a similar reasoning apply to others performance parameters. Make your vessels with enough performance,  don't worry about maximum (is too costly).

By maximum delta-v, I meant for an all-purpose vehicle, but thanks for the tip. 

[Starman4308 1,699]

1 hour ago, Spaced Out said:

By maximum delta-v, I meant for an all-purpose vehicle, but thanks for the tip. 

 

59 minutes ago, GoSlash27 said:

Only a stage that is designed to the mission requirement for the least mass and/ or cost can be considered truly "efficient".

Best,
-Slashy

I think there's a difference in philosophy and gameplay style here. I'm pretty sure GoSlash is one of those who custom-designs a launcher for each new mission. This can maximize cost efficiency, since that means you have the absolute least amount of booster necessary to get payload to the target, but can be a bit time-consuming.

I tend towards designing a set of boosters for common payload ranges, saved as subassemblies, and grab the smallest one that I think will be able to execute the mission. This is significantly reinforced by running heavily modded careers (including RP-0 on my latest career), where there are a million things to remember when designing boosters. Ullage, restarts, maneuvering, communication, the works: much simpler to not have to worry about that for each craft, by sticking a pre-designed booster under each payload.

The downside to this approach, of course, is that you can't absolutely optimize the booster for the payload and mission. If the payload is 5.01 tons, and you have a 5-ton-payload and a 10-ton-payload booster, you're either going to run the margins thin on the 5-ton booster or shell out for the 10-ton booster.

[putnamto 187]

8 hours ago, 5thHorseman said:

 

Same with a rocket, though maybe not so extreme (no soccer games in orbit... yet).

 

Lol, challenge  accepted, well  sorta.

 

I  want  to  build  a  square  space  station, then  launch  a  ship  with  a  srb  for  its  nose  and "fire" the  srb  from  a  distance  from  the  square  space  station  and  try  to  make  a  goal.

[Spricigo 843]

15 hours ago, Spaced Out said:

By maximum delta-v, I meant for an all-purpose vehicle, but thanks for the tip. 

Vessels are tools. You don't pick a screwdriver to drive a nail  or a hammer to drive a screw.

You may build an "all-purpose" vehicle but it will only be good for limited purposes and even in this cases probably not as good as an purpose-build vehicle. Most of the time you will be driving a screw with a hammer, and big, expensive and unyielding hammer.

 

Notice, is ok to use the same vessel for similar missions (Minmus flyby and Mun flyby) but most mission are not like that (Minmus flyby and Tylo landing)

 

 

[GoSlash27 4,258]

16 hours ago, Starman4308 said:

I think there's a difference in philosophy and gameplay style here. I'm pretty sure GoSlash is one of those who custom-designs a launcher for each new mission. This can maximize cost efficiency, since that means you have the absolute least amount of booster necessary to get payload to the target, but can be a bit time-consuming.

Starman4308,

Not for me
I have a spreadsheet that does that. I just plug in the requirements for the stage (payload, DV, minimum t/w) and it tells me exactly what I need to build.

Since my payloads are reusable and modular, my launchers do wind up being sort of standardized anyway.

Best,
-Slashy

[kerbalfreak 39]

Like many people said, there ins't a straight answer, and there's a lot of factors to consider. You can easily go mad trying to build the best rockets possible, for every situation. In that case you will probably end up with SSTO Spaceplanes.

I like to make with 2 stages, mostly thinking in reusability, simplicity and low part count. The first one is relatively short, and I simply drop into the ocean, with airbrakes to fall near the beach and to safely open the parachutes 3-5 seconds later (using Smart Parts). The TWR gets very high when finishing the orbit, which is not necessary but make things much smoother with very heavy payloads, and allows for an quick, powerful and precise reentry burn. I recover everything, expect for the fairing.

I avoid doing very sophisticated landings with the first stages (using FMRS). I think that role, in KSP, is more for SSTO Spaceplanes. Sophisticated, but expensive if you fail, much harder to design and fit payloads. But no stages and 100% recovery except for the fuel.

2 stages rockets ideas are also popular IRL. Falcon 9, BFR, Sea Dragon. Even the Space Shuttle, which only dropped the E.T. to avoid debris.

My designs:

1. A central liquid fuel booster with two SRBs merged with two tanks in an simple asparagus configuration. The SRBs and the tanks finish almost at the same time, and then drop to be recovered, while the central booster is full.

2. First stage only a big SRB (from SpaceY), and second stage a Rhino or similar (from mods). The SRB takes the Rhino at an altitude where it's ISP is very high, and it fires for a quite long time. Since it's an inline rocket, without side boosters, the aerodynamics seems to be better. 

3. And last, something similar to the NASA's SLS. An central core with engines that work well in any altitude (Vector, Mammoth, Kite and Buzzard (Near Future), and the SRB's. It launches with the central core at maximum, to minimize gravity loses. I set mechjeb to limite MaxQ at 20,000, so after ~20 seconds the central core shutdown, and then it goes with only the SRBs. When the SRBs are almost finishing, the central core starts again. When the SRBs drop, the central core goes to maximum. That first 20 seconds with high TWR helps A LOT, and uses only 10-15% of the fuel from the central booster, so I don't need to bother to make an asparagus (and is more realistic to avoid the fuel lines).

The second stages, after deploying the payload, have and inflatable heat shield at the top (that I use "inverted", like a mushroom, it helps a lot with engines overheating, stability and slowing the rocket), and around 800-1500 dV left. So I do an 600-1000 m/s retroburn for an very easy reentry, and a "Falcon 9 style" landing at the KSC, sometimes I use parachutes.

I have rockets with this setups for 15, 30, 70, and 110 tons. And some Sea Dragon inspired crazy ones for 150 and 200 tons  So I just make the payloads fitted for one of these, attach, and go boom (in a good way).

Hope it helps!

[Spricigo 843]

31 minutes ago, kerbalfreak said:

You can easily go mad trying to build the best rockets possible, for every situation. In that case you will probably end up with SSTO Spaceplanes.

I'd like to know what is the base of your assumption that SSTO spaceplanes are superior to anything else (Staged rockets, SSTO rockets, Staged Spaceplanes). 

Or maybe not.

 

[kerbalfreak 39]

1 hour ago, Spricigo said:

I'd like to know what is the base of your assumption that SSTO spaceplanes are superior to anything else (Staged rockets, SSTO rockets, Staged Spaceplanes). 

Or maybe not.

 

It's kinda funny how you reply that in a post where I tell about the rockets I use, and where I mentioned 3 disavantages of SSTOs Spaceplanes in KSP 

[Spricigo 843]

3 minutes ago, kerbalfreak said:

It's kinda funny how you reply that in a post where I tell about the rockets I use, and where I mentioned 3 disavantages of SSTOs Spaceplanes in KSP 

Nice that I managed to be funny. 

And yet, the part I quoted can be interpreted in a way that gives the impression that you consider SSTO spaceplanes superior.  If it's not what you mean, maybe is in your interest to clarify that. 

OTOH if that is really your opinion, may be interesting to see the strong points you think SSTO spaceplanes have.

 

 

 

 

[Gaarst 3,113]

22 hours ago, Snark said:

Given the typical Isp of KSP engines, that usually means you end up with something on the order of 1400 to 2000 m/s of dV in each stage, for "conventional" LFO engines (I'm not including the LV-N or ion engines).

I'm curious, how do you obtain this value?

I've tried to determine the optimal dV for a rocket stage (ie: most dV per unit mass) and obtained between 2200 and 2600 m/s for LFO engines (obviously higher for NERV or Ion engines).

[Snark 13,158]

7 minutes ago, Gaarst said:

I'm curious, how do you obtain this value?

I've tried to determine the optimal dV for a rocket stage (ie: most dV per unit mass) and obtained between 2200 and 2600 m/s for LFO engines (obviously higher for NERV or Ion engines).

Not a question of "obtaining", more a matter of observation that that's what it typically seems to work out as, in practice, based on a few years of playing KSP.

"Optimal" is a highly slippery concept, since it depends so much on exactly what one is "optimizing" for, and also on large amounts of context about the intended mission of the ship.

Sometimes I'll end up with an LFO stage that gets somewhat higher, even occasionally up in the 3000 m/s range, for an upper stage with minimal payload and low TWR, but I find that 2000 m/s and below tends to be more common, particularly for lower stages that tend to need higher TWR and tend to have engines with lower Isp.