craft planning

[gslarmour]

It's occurred to me that a lot of the time I spend playing KSP is spent aimlessly fitting parts together. So, I'm asking the community, how do you ask plan your vehicles? I've watched Spooty throw things together, but I've also seen the crazy plans that ej_sa has written up. What works for you?

[Superfluous J]

I start with my return craft (assuming I am returning). I then build my payload around/under it and make sure it has the oomph to land where I'm sending it. Then I build a stage to get it to that location from LKO. Finally, I build something that can take that into LKO from the launch pad.

I keep telling myself I'll save a bunch of subassemblies of lifters so I can just grab one that's the right size for the job, but they keep getting out of date because I keep unlocking better rockets and fuel tanks.

[Spartwo]

Throwing something together then spending ages refining it.

[Giggleplex777]

Occasionally, I would draw diagrams before I built a ship, otherwise I plan my designs in my head (somehow, I always get great ideas just before I fall asleep).

[EndOfTheEarth]

My flight infrastructure is built around experience with a single nuclear tug design that I developed almost half a year ago for the sake of sending a probe to Eve. The tug was just overbuilt enough that if I replaced the lander probe with a docking port, then I could haul something the size of one of my normal Mun landers almost anywhere in the Kerbol system. So the plan is usually like this:

1) Rebuild the nuclear tug. Stare at it for a while and shift minor things like docking ports, monoprop, and RTGs around until I'm satisfied that everything is balanced, and that docking an emergency fuel mission to it would be easy.

2) Develop the lander. This usually involves comparing the gravity at my destination to the gravity of the Mun, and then making modifications to my Mun lander accordingly. Again, balance things and wonder how much weight can be omitted.

3) Launch 'em. Dock 'em. Refuel 'em as needed.

I've only recently begun to deviate from this design, but that's only because I've been encountering more complex destinations, or unique opportunities. My recenty Gilly mission, for instance, I knew that hauling a massive lander would be unnecessary, so I integrated a crew cabin and landing legs into the tug:

[Dispatcher]

Similar to gslarmour. I start with my ultimate payload (usually that's intended as a two way trip and ends by parachuting back). Then, if its to land at a target (which usually is an airless body), I put a lander under it. Since I'm still playing Career, and I've yet to unlock sturdy docking ports, that lander must also be able to return and deorbit. Then I put a single stage under that which goes one way from LKO to LKO. Finally I put the multistage launch assembly under that. I do have some sub assemblies for that, but as 5thHorseman says, by the time I'm ready to use them, I have new parts which I incorporate into a new design. When I do use a sub assembly, I end up tweaking it by adding or removing some boosters, until it will do the job without too many extra parts.

I do use Kerbal Engineer so that I don't need to do extreme amounts of experimenting or over building. That's a choice for each person to make (or not).

[_Aramchek_]

1. Rough idea.

2. .......

3. Build something that fits the parameters I wanted.

4. Profit.

[CalculusWarrior]

I generally design backwards, which means for an Apollo-style mission I would design my return stage first (command pod with parachutes), then service module, then lander, then transfer stage, and so on. This also has the added bonus of most things being kept in the right order on the staging list.

For more static structures, like base modules or station parts, I know which features it will require and thus know which parts I need to add. I then consider the basic form of the module, and also its delivery mechanism. Having a basic idea, I can build the basic structure of the part, adding docking ports/features as necessary.

[I_Killed_Jeb]

Almost invariably build top to bottom, like most everyone who's posted so far. The lander/orbiter is almost always the most interesting part of the design process and as such for me it takes 90% of the work. Then it's just a matter of strapping on enough dV to take that into orbit.

[regex]

Sometimes I build the transfer stage first to fit whatever booster I'm using or simply to size, like if I'm building a reusable tug for a variety of payloads or if I have a specific delta-V I want. I'll then throw on one of those flat plates and stack fuel tanks on top to find my desired payload/delta-V/TMR, adjusting until everything looks good. Then I'll build a payload within the mass limit I specified.

Most of the time, though, I build the payload first.

[I_Killed_Jeb]

Sometimes I build the transfer stage first to fit whatever booster I'm using or simply to size, like if I'm building a reusable tug for a variety of payloads. I'll then throw on one of those flat plates and stack fuel tanks on top to find my desired payload/delta-V/TMR, adjusting until everything looks good. Then I'll build a payload within the mass limit I specified.

Most of the time, though, I build the payload first.

Do you always launch payloads with the thrust plate configuration?

[regex]

Do you always launch payloads with the thrust plate configuration?

Never. I just use the plate to prevent fuel flow so that Engineer uses the correct delta-V with the fuel tanks I'm using for dead weight. It's only used in the VAB.

[I_Killed_Jeb]

Never. I just use the plate to prevent fuel flow so that Engineer uses the correct delta-V with the fuel tanks I'm using for dead weight.

oo that's an excellent trick

[Aragosnat]

Throw things together. Only have 2 things saved in sub assembly. One of which is a rover and the other is a bisic skipper asparagis. But, mainly work from scratch with not a care in the world. At least till budgest gets implemented.

[Buzzed Aldrin]

1. Establish mission parameters / goal.

2. Calculate necessary dV for goal +~20%

3. Design craft in head and on excel spreadsheet.

4. Build in Vab. Tweak. Test. Tweak.

6. Launch mission.

[Space Peanut]

1. Establish mission parameters / goal.

2. Build lander, launcher

3. Craft crashes on takeoff, redesign

4. Craft gets to orbit after using all of its fuel, redesign

5. Craft does not have enough fuel to complete transfer burn, redesign

6. Craft makes it to destination, crash lands

7. Go back to the start picking a new destination

Thats about how I operate! Obviously calculating the quantity of fuel necessary is all important but i've always been too lazy to figure out how to do it.

[bsalis]

Not sure it's planning, more of a style or approach for me.

For Rockets

I will build the payload I want first. Nowadays, it's typically single stage with ports to dock with other crafts in orbit. Main consideration for the lifting stages is having adequate TWR for various phases of ascent, and not undershooting the dV. I will typically build for a target of LKO or Jool (in the case of LKO, I tow it to Jool after docking).

Never got around to building standard lifter subassemblies. Since i can put a rocket together for modest payloads quickly, and often the payload needs more of a bespoke rocket built around it. I have some very odd payloads that are basically encased in the lifting stages.

SSTO Spaceplanes

Oh gee - could write an article as long as this thread so far. Especially on ones with VTOL capability. In summary though, main considerations when planning an SSTO Spaceplane is around balance, forces, and proportions. Anything you change will impact something else. It's a design puzzle that keeps me coming back and pushing the envelope even further.

[Kasuha]

I usually spend some time just thinking about it, putting parts together in my mind rather than in the game. It's much faster and I can go through way more designs in short span of time that way. I never use calculators and guessing whether it will work or not does not depend on whether I can see it on my screen or not.

Once I have a potentially working design in my mind, I go implement it in VAB/SPH.

Sometimes I find I thought of something that can't be done to can be done better, then I either switch to the better way or return to thinking how to do it better.

[Ravenchant]

Depends on what's being designed. If part of the craft is supposed to come back, I'll start with that, and build the rest around it. That works well for ascent stages and orbiters, at least. If it's a probe, I'll think about the scientific returns first, and if it's a mothership, I'll draw some diagrams. This is the only time the "top to bottom" rule is broken, as I usually design a transfer stage first, check how much payload it can get to a chosen destination, then plan the rest of the craft.

Engineer is a neccessity, though. You can only plug the numbers into Tsiolkovsky so many times before going mad.

[Silverchain]

Sometimes I use a program (it seems to me that relying on other people's tools to do that sort of thing does people no good as a whole) that runs through every possible combination of fuel tanks and engines for given payload, velocity change and acceleration requirements and returns me a set of lowest mass options, then carefully build something to that specification.

Sometimes if I see something I like and try to blatantly copy it.

The rest of the time I come up with a silly idea and throw parts together and see what happens, then run it through a few launches. Usually it gets abandoned, but if it works really well I start making evolutionary changes. With my present save I've got rover parks and plane parks (er, and debris fields) outside physics range of the Launchpad for all the early models, and LKO is full of successive iterations of spaceplane that I haven't bothered to land...

[el_coyoto]

I usually start with the landing/return module and wonder how I should plan the mission.

Is it a short hop to the Mun, and does it involve a long trip in a big modular space ship with plenty of habitable space and fuel?

Once that I've outlined the mission profile, I can start thinking about my ship specifications. This process usually involves a piece of paper, a pen, a coffee and cigarettes. I make the ship can deal with all the aspects of the mission, essentially playing the mission in my head : refuel, docking, landing.

Once that I have my specs (fuel, living space, docking ports, etc) I head to the VAB build the ship, test it on the ground if it's a lander and then build the lifter.

[Richy teh space man]

My usual development cycle goes something like:

1 Think carefully about what I want the craft to be able to do.

2 Put something together and think this is the answer.

3 Launch it.

4 Revert to hangar/VAB when it fails.

5 Repeat steps 2 to 4 until it works

6 ???

7 PROFIT

In general, I don't design ships to be stand alone. I usually design a small fleet that works together to accomplish a goal, and usually leave something aound the planet/moon so that in the future it can be reused as and when necessary.

Edited January 16, 2014 by Richy teh space man
Typos

[Comrade Jenkens]

I decide on the parameters I want and a basic shape I want my craft to look like. Then I just experiment until I'm happy. After that its just a case of added all the systems and aesthetic pieces in.

[Vanamonde]

I design the mission payload carefully, but as we all know, KSP's physics is quirky. To paraphrase Clausewitz, "No ship design survives contact with the launchpad." So I'll build a transit stage for that payload, launch/test/revise, then build an upper stage for that, launch/test/revise, then a launch stage, launch/test/revise. I have no idea what the final ship will look like until I see it, as it is whatever was needed to get that payload into space.

[MajorThomas]

My usual development cycle goes something like:

1 Think carefully about what I want the craft to be able to do.

2 Put something together and think this is the answer.

3 Launch it.

4 Revert to hangar/VAB when it fails.

5 Repeat steps 2 to 4 until it works

6 ???

7 PROFIT

In general, I don't design ships to be stand alone. I usually design a small fleet that works together to accomplish a goal, and usually leave something aound the planet/moon so that in the future it can be reused as and when necessary.

This is what I do, almost exactly.

It also helps if you develop a decent idea of how much fuel is needed at different stages of the mission, so you don't do things like waste upper stages, or run out of fuel during the return trip, etc. You can get a sense of this by flying missions to other planets.