Intermediate Staging

[paul_c]

I've put "intermediate" in the title because its more involved than simply "this is how staging works...." (hopefully)

Principle 1

We do staging because it doesn't make sense to be hauling 10t of sheet metal (empty fuel tanks) and using the hulking great 6t engine that lifted us off the ground. (I know.....SSTOs....) in the late stages of a flight or mission.

Principle 2

We aim to reduce the weight as much as possible. It might have the side benefit of increasing simplicity, reducing part count (to under 30, in early gameplay...) etc So for smaller rockets, carrying the interstage equipment and extra engines seems counterintuitive. (This may conflict with Principle 1).

Principle 3

Once we have a spaceship with separate stages, we optimise the fuel/engines in each to maximise the deltaV and produce the required TWR (and probably not much more.....because a bigger engine is a heavier engine). Being mindful that a vacuum-optimised engine isn't great at low altitudes too, etc.

Principle 4

We 'set' the stages to be defined amounts each, so that each stage corresponds with a phase of the mission profile. Eg liftoff from Kerbin is typically 1000m/s dV straight up, so we use SRBs for that. Then to orbit needs a reasonable motor; then the next (which might be, eg Minmus insertion) need not be so powerful etc etc. We don't want to be half way through injecting for Minmus and needing to mess about with staging, leaving a debris fuel tank going half way there too etc. This might conflict with Principle 3.

 

Does anyone violently disagree with the above principles? How do you resolve when they conflict? Personally I've tweaked and tweaked things on several iterations, generally in the order 1) it needs to do the job 2) 'efficient' whatever that means, ie an 18t rocket can carry a good payload; or no obvious weirdness like taking a million m/s dV to reach orbit; or having lots of dV left for later/extending the mission(s). 3) fly nice. As in, consistent, repeatable launches not relying on daredevil piloting or luck 4) low cost. If there's something that can do it for £11k when the previous did it for £12k....

But the NCD challenge has taught me - due to its limited funding and grindy nature - you need to "come to the table" with a decent design in your head, build it and fly it with no real dramas or issues straight out the blocks. So any tips, ideas or discussion is much appreciated.

 

[king of nowhere]

i would disagree that liftoff is 1000 m/s straight up, as i generally start turning around 50 m/s with the objective of hitting 500 m/s with a 45 degrees inclination... but then you mentioned caveman. so you flew a lot of small rockets with inadequate parts and no fairings, so you have much greater problems of aerodinamics. in which case it makes sense to launch straight up for a while.

[Spricigo]

First, as my not answer, I'd say it is  about the efficiency, reaching the objective with small expenditure of resources. 

Now, there is the catch: you are the one that decide what the objective is and what resources are valuable for you. Someone playing in science or sandbox would not care for Funds and even in career many are more worried about their game time. Find what works for you and go for it.

now "an actual answer":

Thinking about mission phases and making the stages somewhat correspondent with those phases makes the design process a lot easier. It allow to focus in the different aspects the are important for each phase/stage separately, favoring the creation "specialized modules" that can be reused for similar tasked in the future.

It makes sense to design the stage in the reverse order of use, since the first stage will need  to carry the later which will then be used "in standalone". For the same reason , small adjustment in a later stage will probably have a cascading effect all the way to the launch vehicle.

Finally, the hallmark of a good design is simplicity. If any part is not necessary remove it. (But notice that in KSP "needs" is largely a consequence of "wants". having some extra fuel for a safety margin, a stronger reaction wheel for faster steering or some 'cool parts' is not a problem, unless it is not allowing to do something else that you want more.)

 

oh, if a old but still good tutorial interest you, this one will probably give insights that goes along with the principies that you elected.

 

[DeadJohn]

6 hours ago, paul_c said:

Does anyone violently disagree with the above principles?

I don't disagree with those principles. I'll add that you should almost always build a rocket from the top then work downwards, and a "DV map" or spreadsheet is essential.

For example, let's say I want a one-way robotic probe, capable of landing on the Mun and doing 1 or 2 biome hops. My thought process is such:

1. Create the uppermost payload stage. I start with a probe core, solar panels, science experiments, and an antenna to transmit data. From there, I then find a fuel and engine combo that has enough DV to land from Mun orbit and do the biome hops. Minimize mass, since every kg in the upper stage burns more fuel in all the stages below.
2. Create the transfer stage below that. Assume that this stage will begin just short of low Kerbin orbit (LKO), so it needs to circularize in LKO parking orbit, then intercept the Mun and establish orbit there. I try to leave spare DV to handle the initial part of the Mun landing, letting this transfer stage crash into the Mun surface to minimize space junk.
3. Create the lifter stage(s). 2 stages are typical. I try to have just enough DV to raise apoapsis to orbital altitude yet leaving periapsis below 25km. That lets my 2nd lifter stage burn up after release. The transfer stage described above handles the final part of getting to orbit. Another option is to let the lifter itself achieve full circular orbit but include some sepratrons mounted backwards that activate with staging, so the lifter deorbits itself. LKO is the only place I get this fussy about space junk.

I do a rough draft of that complete rocket. Then I recheck the TWR and DV of each stage and the total rocket. Remember to check the lowest stage's DV and TWR at sea level, and the rest in vacuum.

After proving that the design works, I'll note the mass of the payload + transfer stages, then save a copy of just the lifter stages with a description of the lifting capacity. That lets me reuse old lifter designs for new payloads.

Edited February 26, 2021 by DeadJohn