How do you come up with your rocket design?

[Stealth2668]

Since I'm a noob, I just put together a bunch of fuel tanks and rockets and put random stage sections in there without knowing why I'm not just doing a SSTO rocket. I know it's not the most efficient method, so I'm wondering what goes through the mind of KSP vets when building a rocket?

How do you decide when your current stage is enough and when to start another. How do you decide if you want to go with a dual, tri, quad tank design? How do you know when you have enough DV to get to where you want to go and what changes when you want to go somewhere close vs. going to the furthest planet? What changes in your design if you need to lift something heavy? What are some general tips a noob should know in order to make a "good" rocket?

Thanks!

[Whackjob]

70% of all Kerbish design research involves throwing random things into the wall and see what sticks. The remainder 30% is watching things fail over and over again and puzzling out why.

[Goldham]

I come across a Federation starship and repeat "Can you help us make us go?" over and over again. Then we become strong!

[Pineapple Frenzy]

Get Kerbal Flight engineer. Launch TWR between 2.0 - 2.5. Find a DV chart to figure out the rest.

[The Destroyer]

Practice.

Lot's.

Of.

Practice.

[DisarmingBaton5]

It takes time. You will learn.

You will learn that designing a rocket takes a lot of "why is this exploding"

[Aragosnat]

Lots of practice. Lots of trying to make it look good to my eyes and do over kills.

[Redrobin]

Kerbal Engineer or Mech Jeb can help you with seeing how much TWR and delta-v you have.

[Rocket Farmer]

Personally I fire up excel usually design the rockets in there first.

First I look up where I am going and what the various stage DV requirements will be. I add a bit of a margin for safety as I play without mods so no mech Jeb for me.

Then I determine what science/capsules/rovers/satellites/return capsule etc are going. This becomes the 1st payload. I then calculate using ln(full weight / rocket without fuel) x 9.81 x ISP of the rocket. I also calculate acceleration as thrust / mass to determine if its enough. With excel in a couple of minutes I can quickly test all rockets ad fuel amounts to determine my optimum setup.

For the transfer stage start with the lander as your payload and repeat the above calculations.

With the lifter I figure out the combined lander and transfer star mass and divide it by 2.5. That forms the number I skippers I will need (1 skipper with 54 tonnes of fuel can comfortably put 2.5 tonnes of payload in orbit so a pancake of power with 49 skippers can lift 123.5 tonnes. For safety sake at this size strap on some boosters....

[Vanamonde]

Build a payload. Build a rocket under it. If it doesn't go fast enough, add an engine. If it doesn't go far enough, add a stage under it. Repeat as needed.

[Mobjack]

The top stages should be small and the bottom ones should be big. As a general rule, each stage should be at least twice as big as the previous.

[Anton P. Nym]

Trial and error. Lots of trial and error... sometimes very big trials with very big errors. But when the dust settles and the payload's in orbit, it's worth it.

-- Steve

[Omicron314]

I find it useful to examine historical designs for inspiration. After that it is just a lot of tweaking around with your design until something works. You can do it!

[Stealth2668]

I've got mechjeb already and found this DV map: http://i.imgur.com/NKZhU57.png

Now I guess it's time to put random stuff together and fail my way to success.

[Jouni]

I try to keep lifter rockets independent from payloads. As long as the payload is reasonably shaped and between 50% and 100% of the nominal capacity of the lifter, the rocket should be oblivious of whatever it's carrying, get the payload into orbit, and then deorbit itself. Sometimes that's not possible, but it's a good thing to aim for.

When designing a new lifter rocket, I start by building a dummy payload with mass as close to the nominal capacity as possible. Then I add something like this under the payload:

The "brain" of the rocket includes a decoupler for separating the payload, a probe core for controlling the rocket after separation, reaction wheels, batteries, and solar panels. I usually add MechJeb as well, as it greatly speeds up the design process.

Then I start fooling around with fuel tanks, engines, and boosters, until the numbers look right. The total delta-v of the rocket should be at least 4300 m/s in atmosphere and 4750 m/s in vacuum. TWR should be between 1.5 and 2.0 at liftoff, stay above 1.4 all the time, and not get above 3.0 during the first couple of minutes.

Because engines are tall and heavy, I usually don't do vertical staging, but aim for a single-stage lifter with 2-4 boosters instead. If the boosters use liquid fuel, I usually add fuel lines from them to the core lifter for some extra delta-v. As I prefer using a small number of powerful engines over a larger number of smaller engines, the core lifter usually has a Skipper or a Mainsail, while the boosters can use Skippers, Mainsails, or LV-T30s. Larger lifters may have some radial engines near the bottom of the core for some additional thrust, or at the top to improve stability. If top engines are used, they must be gimbal locked, as SAS can't use thrust vectoring properly above the center of mass.

As a final touch, I add struts and launch stability enhancers, until the rocket feels solid enough. For larger lifters, I may build some kind of exoskeleton for connecting boosters and the payload to the core, as well as connecting individual fuel tanks within the core. Then it's just a matter of testing and fine tuning the design, until the lifter starts feeling reliable enough.

Edited February 5, 2014 by Jouni

[badpandabear]

First I was just copying real world designs like the Titan, Atlas and Saturn V. Copying the Gemini and Apollo missions was a good way to learn how to dock.

But now I've moved onto using a spreadsheet to calculate delta-v for my ships. First I design the payload, then either use an existing lifter or design a new one based on the delta-v needed to get to orbit. I tend to build interplanetary ships by docking the pieces together in orbit.

[SkyRender]

1. Figure out what I want to do.

2. Figure out the minimum size and form factor to do it with.

3. Build the payload accordingly.

4. Build the drive and/or launch stage to support it.

Simple in concept, complicated in practice.

[Starwaster]

I usually have some idea in my head what I want to put up. Then I just throw a bunch of random rocket parts together then slowly take away anything that doesn't look like what's in my head.

[boxman]

Since I'm a noob, I just put together a bunch of fuel tanks and rockets and put random stage sections in there without knowing why I'm not just doing a SSTO rocket. I know it's not the most efficient method, so I'm wondering what goes through the mind of KSP vets when building a rocket?

How do you decide when your current stage is enough and when to start another. How do you decide if you want to go with a dual, tri, quad tank design? How do you know when you have enough DV to get to where you want to go and what changes when you want to go somewhere close vs. going to the furthest planet? What changes in your design if you need to lift something heavy? What are some general tips a noob should know in order to make a "good" rocket?

Thanks!

I kinda started out like you but eventually just learned to judge what is enough depending on what I am lifting and where i am going.

This is why I would personally not use any delta v calculators since you will just become dependent on it.

[Seret]

Build payload > Check ÃŽâ€v for destination > add control gubbins > keep adding tanks and engines until total ÃŽâ€v is enough to hit the destination.

Generally the choice of stages will be decided by your mission's profile. For example, if you're going interplanetary with a reasonable payload you're probably going to want a transfer stage with a good vacuum engine (eg: LV-N) for that part of the mission, then you'll need at least one more stage to get that into Kerbin orbit.

Generally staging is good, as you don't have to drag the empty tanks along with you, the problem is that you have to drag the extra engines along until you can light them. Get something like MechJeb or KER and have a play around, see how adding stages affects ÃŽâ€v. Another big consideration is that different engines are good for different things. An engine that's good for powerhousing your way off the pad is unlikely to be good for orbital maneuvers, and vice versa.

[AlamoVampire]

I come across a Federation starship and repeat "Can you help us make us go?" over and over again. Then we become strong!

the Pakleds approach is not the nicest approach, tends to get Geordi La Forge hurt lol

as to how I do it? same as you pretty much, at first, then, well, I just go with what looks good, works and blows up the least other than that, I just play it by ear and LEGO it

[fireflower]

I come across a Federation starship and repeat "Can you help us make us go?" over and over again. Then we become strong!

TNG was so weird sometimes.

How do you decide when your current stage is enough and when to start another.

TWR for each stage should be greater than 2 in the gravity well it is to operate in. Total ÃŽâ€v for the launcher should be equal to ÃŽâ€v budget for the mission plus a fudge factor. I like the diver's rule of thirds, when means I try to have 30% more ÃŽâ€v than I need.

How do you decide if you want to go with a dual, tri, quad tank design?

Do you need two, three, or four LV-T30's to make the stage in question have a TWR greater than 2?

How do you know when you have enough DV to get to where you want to go

I programmed rocket equations into excel so now I just plug and grind. There are also cheat sheets available on the internet by googling "ksp delta v map."

and what changes when you want to go somewhere close vs. going to the furthest planet?

ÃŽâ€v requirements. Really the biggest difference in a mission to Eeloo and a mission to Mun is the former uses a LV-N. I even powered my Eeloo mission using solar!

What changes in your design if you need to lift something heavy?

I use dozens of Mainsail engines staged asparagus. Or I dock the objects in Kerbin orbit. Or I dock them in orbit at their target destination. Or a combination of all three.

What are some general tips a noob should know in order to make a "good" rocket?

Something I've noticed from your questions is that you're designing the rocket from the bottom-up. You need to design it from the top down. First, decide where you want to go. This will tell you what you want the rocket to do, and you can look up the ÃŽâ€v figures for the mission. Then create your payload. If you're going to have it land on Minmus for instance you'll notice you don't need a lot of thrust to have a TWR >2 nor will you need a lot of ÃŽâ€v in the lander itself, so you can make do with LT-5 Micro Struts. It'll also need a probe core, solar panels, and a battery. Maybe even lights. Then you design the stage that will perform the transfer burn. Then you design the stage that will get that out of the upper atmosphere. Then you design the biggest stage that will get it all off the surface. Once you've done it enough, it's entirely possible for complete mission architecture to appear in your mind fully-formed in a flash of inspiration, or to come to you in dreams. Be especially sure to check your dream-math while awake. Speaking of, some sort of flight computer is indispensable for designing your rocket in the VAB. You can't do rocket science by intuition. I don't calculate TWR and ÃŽâ€v manually every time I place a new part, MechJeb does that. I monitor its data to ensure the construction is proceeding to plan. I can also notice if there is a problem in staging that will end in tears if not corrected.

[SkyHook]

Looking to the past for inspiration has helped me a great deal. If it's a particularly ambitious project that I'm stuck on, then google images is my best friend. Otherwise it's Payload, destination, launcher. Once all three are checked, it's time to launch.

[Stealth2668]

Thanks for the replies everyone. Seems like it'll take some time to get the hang of it so I might only get serious in summer when my Uni studies are done. I'll have to stick to simple games like BF4 until I have the time to invest in KSP.

[InterCity]

Step 1: Create payload (that thing on top of the rocket that is supposed to get away from Kerbin's atmosphere.)

Step 2: Add struts

Step 4: Add stack decoupler.

Step 5: Pick one to four most efficient engines possible (Usually nuke or KW rocketry's SPS), and attach it under the payload.

Step 6: Add fuel tanks above the engine as long as your TWR is higher than 1,00. The lower TWR, the better.

Step 7: Add struts.

Step 8: Repeat steps 5-7 with more power, using 2,5 or 3,75 m stacks.

Step 9: Add boosters (Typically 30-50)

Step 10: Add tweaks, such as RCS, struts, winglets, rocket holders, struts, Jeb into cockpit, struts, MechJeb, struts, chutes and struts.

Step 11: Fly the thing.

Step 12: Crash the thing.

Step 13: Add struts.

Step 14: Fly the thing again. Let's start with LKO, hope you have enough Delta-V to get there

Step 15: Check for solar panels. If you still have them, get ready. You will control that thing a lot longer than expected. Kouston has decided to send it to Eve.

This is the fifteen steps KESA engineers' guide to rockets. Similar with spaceplanes, although i only use them for atmospheric travel.

Always works.

P.S. Step 12 is NOT optional