Your opinion please

[Ron Devu]

I am starting to think that it is a general principle in KSP that bigger rockets do not go farther, they just lift more, and in order to go farther you don't need larger rockets so much as better designed, more efficient ones.

 

What's your opinion on this ?

[Snark]

13 minutes ago, Ron Devu said:

I am starting to think that it is a general principle in KSP that bigger rockets do not go farther, they just lift more, and in order to go farther you don't need larger rockets so much as better designed, more efficient ones.

There's a large measure of truth in that. "How far can you go" is roughly equivalent to "how much dV do you have". And the logarithmic nature of the rocket equation means that simply piling on more mass (additional fuel) gives very rapidly diminishing returns in terms of additional dV.

So if you're designing to try to pack more dV into your rocket... yeah, you gotta build smart. Building smart will yield far more reward than just throwing more mass at it. This is true IRL too, not just in KSP, because of the way the math works.

So yeah, you've pretty much hit the nail on the head. Welcome to KSP, and the "tyranny of the rocket equation".

[Vanamonde]

Yep, this is how it works.

[T-Bouw]

Your pretty much right.
In the years of playing KSP, I've come to believe that rockets don't become much bigger once you leave Kerbin. A manned munshot rocket only needs a little extra fuel tank to make it interplanetary. The low dV requirements of outer space make it so.

[Pecan]

A few 1.25m parts are all you need for everything but multi-crewed missions.  For those you need 2.5m because that's the size of the capsule (and hitchhiker).  Less than 65 tonnes on the launchpad will take a set of communication satellites anywhere in the system, for instance.

It has often been said that in rocket science less is more and, yes, you often get better results by taking things off than putting more on.

[Superfluous J]

That's a great rule of thumb, yes.

[Zhetaan]

8 hours ago, Ron Devu said:

I am starting to think that it is a general principle in KSP that bigger rockets do not go farther, they just lift more, and in order to go farther you don't need larger rockets so much as better designed, more efficient ones.

Others have generally agreed, and @Snark mentioned the rocket equation specifically, so I will go into detail and explain the limits of this point--but in a spoiler, because I don't want to derail anything.

Spoiler

 

The rocket equation depends on two things:  the engine efficiency and the mass ratio (which is not exactly 'size' but when you say 'bigger rockets', you're still referring to this more than engine efficiency).  You can pick more efficient engines, but only to a point because KSP doesn't offer ways to design better engines:  better rockets, yes, but better engines, no.   The most efficient engine available in the game has a certain predetermined efficiency and it will only ever have that efficiency, so that sets a limit.

Mass ratio is defined as the wet mass divided by the dry mass and refers to the amount of fuel available in the stage to push the rocket from place to place.  The dry mass of the rocket is the mass of everything but the fuel for that stage (this dry mass can include fuel for upper stages that is not available to the lower stages), and the wet mass is the mass of everything, period.

It may not be obvious, but it is possible for a larger (that is, more massive) rocket stage to have the same capabilities as a smaller one, provided that the mass ratios and engine efficiencies are the same for each rocket.  Identical efficiency is easy to achieve:  use the same engine on both rockets.  Identical mass ratio is more difficult, but the idea is that if, for example, you have a 3-tonne rocket made up of one tonne of hardware and two tonnes of fuel, then you will get identical delta-V from a three-hundred tonne rocket that is one hundred tonnes of hardware and two hundred tonnes of fuel, because the mass ratio of both rockets is 3.  There are differences between the rockets, but raw capability in terms of available delta-V isn't one of them.

The reason that 'better designed, more efficient' rockets get better performance is because every gram of payload that you remove improves the mass ratio for a given load of fuel.  Imagine the performance of that three-hundred-tonne rocket if you could reduce the payload to that of the one-tonne rocket, but keep the two hundred tonnes of fuel.

However, just as the fact that KSP doesn't let you improve engine performance sets a limit on engine efficiency, the fact that fuel is only available in predetermined fuel tank sizes sets a limit on the mass ratio, too, because once the fuel is spent, you are left with the dry mass of the empty tank.

We cannot add fuel tanks without limit (computer memory capacity prevents it if nothing else does), but let's assume that we could.  We cannot add fuel without also adding fuel tanks, so we have to account for the fuel tank mass ratio.  However, if we add enough fuel tanks, then their mass dwarfs the mass of the rest of the rocket, so we can say that the theoretical limit of performance for the best possible rocket stage is one where the mass of the rocket is effectively nothing but fuel tanks.  In that case, we can simply use the mass ratio of the tank itself (one tank has a certain amount of fuel mass and tank mass; two tanks have double the fuel but also double the tank and so the same mass ratio, and so on) and get the theoretical limit of performance for the rocket.

LFO fuel tanks in KSP have a mass ratio of 9; that's eight parts fuel plus one part tank for a full fuel tank.  The most efficient LFO engine in KSP is the Poodle with 350 seconds of I_sp.  This gives, through the rocket equation, a theoretical absolute best performance of 7,541.6 m/s of delta-V.

I could go into detail about the Nerv and Dawn but it's not necessary to make the point.

 

The point is that for any rocket you might build in KSP, its performance in terms of delta-V is going to fall within a range of possible values.  For LFO rockets, the range is, at most, between 0 m/s and 7,541.6 m/s.  Practically, you can expect a well-designed, efficient rocket (with a Poodle or similar-efficiency engine) to have about half of that--it is extremely difficult to get into the upper reaches of that range and impossible to get to the maximum.  Thus, although you may want to try for the upper end of that range, the rocket nevertheless has to stay in the range.  The only way to improve the rocket's performance past that limit is to somehow convert payload mass into fuel--which is exactly what staging does--or to shed dry mass entirely.  Staging does that, as well.  By splitting the total delta-V into smaller pieces and shedding unneeded dry mass, a staged rocket can operate at close to its maximum mass ratio for longer, so even though stages add mass, the overall effect is more of an efficiency boost than a size boost.

To summarise, there is a maximum amount that making a rocket bigger can help you.  There is also a maximum that making a rocket more efficient can help you, but that is a higher limit and thus it is easier to see useful improvements.

[Snark]

41 minutes ago, Zhetaan said:

For LFO rockets, the range is, at most, between 0 m/s and 7,541.6 m/s.

Well, not quite-- be careful not to over-generalize.  The analysis you give in the spoiler's a nice one ... but that's the effective upper limit of an Isp 350 engine in a single stage.  Multi-stage rockets can and do get better than that total amount of dV; it's why multi-stage rockets are a thing.  You can always add dV to your rocket by adding another, bigger stage under it.  It's just that because of the rocket equation, adding dV this way is going to require each successive stage to be exponentially bigger, which makes things less and less practical the more dV you add.

It's technically possible to build an Isp 350 rocket that gets 25 km/s of dV, for example.  It would just be stupendously, inconveniently gigantic for a tiny payload, is all.

So there's not any one hard-and-fast cutoff for the maximum dV that a particular set of parts can get, just exponentially diminishing returns the farther you go.  So the observation stated in the OP is correct in essence... but be careful about trying to hand out one specific number, since it's a bit messier than that.

[Zhetaan]

28 minutes ago, Snark said:

Well, not quite-- be careful not to over-generalize.  The analysis you give in the spoiler's a nice one  ... but that's the effective upper limit of an Isp 350 engine in a single stage.  Multi-stage rockets can and do get better than that total amount of dV; it's why multi-stage rockets are a thing.  You can always add dV to your rocket by adding another, bigger stage under it.

Yes, I thought that was made clear both when I defined it as the 'theoretical limit of performance for the best possible rocket stage' and again at the end when I mentioned that staging overcomes the limit.  Admittedly, I do not use bold and italics for emphasis the way you do (you really do have a talent for that), so perhaps I could make that a bit more obvious.

It is an interesting thought problem to consider how taking a rocket and dividing it into stages can make it more efficient, but I wanted to focus on single-stage rockets for three reasons:  first, because the maths are much easier; second, because an example that depends on a constant mass ratio better makes the point; and third, because people usually don't think of staging as dividing an existing rocket to get better performance out of it:  instead, they think of it as adding boosters that detach when expended and leave the original rocket.  There is nothing specifically wrong with that, but since that makes the rocket bigger as well as better, it doesn't really answer the original question.  That's not even getting to the point that adding stages without the knowledge of what makes a good rocket stage rapidly devolves into cries for 'more boosters!'--and that falls into the same trap of thinking that bigger is necessarily better.

I sought to provide an example that highlighted the original point by showing how bigger rockets can't, of themselves, compete with better-designed and more efficient ones, and I cannot think of a better way to do that than to show that two rockets with different masses, but with the same engines and the same mass ratio, have the same delta-V regardless of what the masses involved in that ratio happen to be.

Since staging works by changing the mass ratio and possibly the engines as well, I can't include that in an analysis that depends on the same engines and a constant mass ratio.

That being said, of course adding a (well-designed) stage to a rocket will make that rocket better.  Adding an efficient stage to an efficient rocket will be better yet.  But that's not because bigger is better; it's because more efficient is better.  I don't need multiple stages to show that; I can instead make the point by showing that even an utterly impossible infinite rocket stage (i.e., the most booster--) still has a finite amount of delta-V.

[Snark]

58 minutes ago, Zhetaan said:

Yes, I thought that was made clear both when I defined it as the 'theoretical limit of performance for the best possible rocket stage'

Yes, and your explanation was a really good one-- my bad for not noticing the staging comments later on.

A few thoughts on the nature of explanations in spoiler, since I'm starting to wander afield from the topic of the thread, which is the OP's question.

Spoiler

Remember that the target audience here is "folks new to KSP", and I was triggering on the comment

3 hours ago, Zhetaan said:

The point is that for any rocket you might build in KSP, its performance in terms of delta-V is going to fall within a range of possible values.  For LFO rockets, the range is, at most, between 0 m/s and 7,541.6 m/s.

...which doesn't include the comment "per stage" and could potentially cause confusion if someone's skimming.

1 hour ago, Zhetaan said:

That being said, of course adding a (well-designed) stage to a rocket will make that rocket better.

Yes, it's "of course" to you and to me and to folks who have been playing KSP for a while and are old hands at this. 

There are a lot of things like that, which aren't necessarily "of course" to a newbie, though.

Overall, I liked your explanation, I thought it was great.  And I see now that you did include the comment about staging later on-- my bad for not spotting that.  As a person who tends to write very lengthy and detailed explanations myself ... I have had some experience with the sorts of explanations that tend to go over well, versus ones that might get past people.  And one thing that I've noticed is that a very common mode that people get into, when reading Gameplay Questions threads, is "I just want one specific answer, I'm just gonna read until I find the answer to my question and then stop."

So explanations that make a statement that may be misleading when taken out of context, and then later provide qualifying caveats that are designed to prevent that misunderstanding... can often go awry, because people get to the sentence that appears to be the explanatory "payload" of what they wanted to know, and then stop reading and don't get to the later bit.

Apologies for the goof on my part, and certainly I don't mean to criticize your explanation, which was a good one-- hope I didn't give that impression.

 

[vossiewulf]

[Empiro]

Also, efficient planning is a big part of getting where you want to go with a reasonable sized rocket too, especially if you want to make multiple stops.

Edited August 9, 2019 by Empiro

[Geschosskopf]

On 8/7/2019 at 10:36 PM, Ron Devu said:

I am starting to think that it is a general principle in KSP that bigger rockets do not go farther, they just lift more, and in order to go farther you don't need larger rockets so much as better designed, more efficient ones.

What's your opinion on this ?

The payload at the top of the rocket determines all else.  You design the payload to do whatever job you've assigned it for the given mission.  This includes 1) the equipment the mission calls for (Science! and/or resource exploitation), whatever crew space it needs (and associated life support for the mission duration if you're into that), TWR sufficient to land (and maybe take off, too) from where you send it (if that's part of the plan), and sufficient fuel to go the distance you want it to go (which might be all the way home from wherever the previous stage left it).

Once you have the payload capable of doing the job you need it to do, then you just built the rocket necessary to get the payload where it needs to go and it can't get to itself with its own fuel and engine combination.  In general, you can divide this rocket into 2 parts:  a transfer stage (from wherever the lifter leaves it to wherever the payload takes over), and a lifter (to get the payload and transfer stage off the ground to where they start their journey).  Both of these main parts can have multiple stages as needed and to take advantage of getting out of Kerbin's atmosphere.

Ultimately, the lifter needs to get the whole mass of the payload and transfer stage to at least orbital velocity.  This requires it to have a sea-level TWR of about 1.5 at a minimum (and this also works well as an initial maximum).  This determines size and configuration of the lifter stages.

Now, you can add some optional constraints on all this yourself.  Such as, instead of each transfer stage and lifter being custom-built for the specific payload, you can have standard lower and upper stages that can only handle so much payload mass.  But that's just the same thing in reverse.  The constraint then switches from the payload to the lifter, so you have to design the payload to a mass your rocket can fly.