How reliable are your rockets?

[Prince of Rockets]

When I first started playing, I was lucky to get rockets off the launch pad... I think this is because I wanted my rockets to "look" cool. But as I started to get a feel for the VAB and piloting rockets, I started putting things through various quality control checks and tests. For example, on landing missions I'll make separate files for the lander and payload sections. With each section, I'll do tests until I'm reasonably confident they will do the job. Once I'm happy with their respective designs, I'll make a sub-assembly of one and add it to the other. I've adopted this mentality with all my projects now and rarely do I encounter failure during the actual missions.

So, how do you guys do it? Do you just build something and hope for the best when you hit space bar? When you have something that almost works, do you simply launch it over and over until it does? Or do you critically analyse things, read the mission logs when there are failures, and employ reliability engineering techniques when something is flaky until you have a reliable design?

[Whackjob]

This reliable.

[1812overture]

Javascript is disabled. View full album

[/1812overture]

[Akhar]

Really darn reliable. Like, they actually work a whopping 20% of the time. That's at least more reliable than when I started.

[astropapi1]

This reliable.

~snip~

The tags just made it perfect.

[TheDarkStar]

I don't usually have trouble once I finish bug-testing them. Getting to that point is a different story; one that involves excess explosions and midair disassembly.

[Decho]

I tend to use an informal engineering pipeline for my main designs, so they're pretty reliable. Bespoke launchers for odd payloads tend to be a bit flakier, but most have better than 1 in 5 failure rates.

[Prince of Rockets]

ROTFL Whackjob, the one with 100+ mainsails and X200-8 fuel tanks that looks like it's folding into itself is pure awesomeness! Something similar happened to me recently when I was trying to make a 270 ton lifter.

[Whackjob]

ROTFL Whackjob, the one with 100+ mainsails and X200-8 fuel tanks that looks like it's folding into itself is pure awesomeness! Something similar happened to me recently when I was trying to make a 270 ton lifter.

I did eventually get that one to launch properly.

[lajoswinkler]

Actually pretty damn good. I'm mostly dumping lander probes around the system.

[Prince of Rockets]

I did eventually get that one to launch properly.

Wow... O.O From the looks of it, you well exceeded the area of the launch pad. Did you have tons of stabilizers all around the sides holding it up? Or did all the struts just flex when it hit the ground and it stayed together? Bah haha, that has to be over 1,200 parts.

EDIT: And I just noticed the smoke trails, very cool.

Edited February 4, 2014 by Prince of Rockets

[Jouni]

My medium lifter (40 tonnes to LKO) has had tens of succesful launches and no failures with reasonable payloads. The track record of my heavy lifter (60 tonnes to LKO) is a bit worse: sometimes it collapses on launchpad, and sometimes it snaps during a particularly hard gravity turn. Those failure modes are quite rare, however, and I've had tens of succesful launches with it as well. Both rockets follow the same general design: a single-stage lifter core that gets the payload into orbit and then deorbits itself, and four Mainsail-powered boosters in asparagus configuration.

Once you learn to use struts effectively, there are no longer unreliable rockets, only difficult payloads. With smaller rockets, even something as simple as the cupola module can qualify as a difficult payload, as I learned today.

Landers are an entirely different matter. If you avoid too ambitious missions, there are some basic patterns that get the job done reliably without too much testing. With more ambitious goals, complex designs, or parts with weak connections (such as Rapiers and nuclear engines), all bets are off.

[ZedNova]

98% reliability with Approved loads under 150 tons.

[Prince of Rockets]

98% reliability with Approved loads under 150 tons.

98% is pretty reliable. For a while I only tested things 5 times, which allowed me to measure reliability to degrees of 20%. When I started getting into 300+ ton rockets, I found that sometimes things would test successfully 4 times in a row, and fail the 5th time for no apparent reason. Then, I started testing things 10 times, giving reliability in degrees of 10%. For something to have a 98% reliability, that means you would of had to test it 50 times.

[fastbikkel]

With my humble career, only been as far as minmus. However i have launched a decent amount of rockets, labs, stations, just to name something. I have very good results. Only at the very beginning i crashed a few. And if kerbals are on board, my safetyrecord is pretty good i might say. Cannot remember losing one lately. (yeah i do take pride in keeping the little buggers alive)

[Captain Sierra]

Too reliable.

I can't make anything fail spectacularly without purposely screwing something up, and even then, not very impressive.

I need to start builing 100+ ton serially staged lifters.

[RSwordsman]

My lifters are quite reliable. Within acceptable payload masses, there's no reason why they shouldn't work every single launch, save some uncontrollable physics glitch. But even that is infrequent. How easy (difficult) they are to fly is another matter, but the fact of the matter is that they work.

Whackjob, that album is some of the most magnificent destruction I've laid eyes on.

Edited February 4, 2014 by RSwordsman

[RocketPilot573]

Reliable? What's that?

[Finwolf]

Over 70% reliability on all my ships built this year. 95% after I got my new computer (no lag with 800+ parts on a station and docking a 100 part ship). My heavy lifter (120t to 150x150 orbit) has not failed a single time.

[Duxwing]

I custom-build each craft for each mission and test due to a quicksave-ban test my lifters scores of times. I also fly my entire mission in sandbox mode first, ensuring that all systems within safety margins function. No craft in my present career save has ever killed a Kerbal, and my lifters are stable and forgiving.

-Duxwing

[Prince of Rockets]

Too reliable.

I can't make anything fail spectacularly without purposely screwing something up, and even then, not very impressive.

I need to start builing 100+ ton serially staged lifters.

You should try a 1,500+ ton Eve mission, those train wreaks can sometimes take minutes to play out. You've also got Matrix like slow motion going on if your computer isn't a beast, increasing the effect.

[SkyHook]

I have staging errors 1 out of every 20 launches or so, but that's because I'm too lazy to add seperatrons to my stages. Otherwise, things are pretty eventless on liftoff. That's how I like it too.

[maccollo]

I prefer brute force serial staging, so usually it's impossible for my stages to collide with each other.

The thing that usually do my rockets in is when I build them so large that physics just begin to arbitrarily fail (even with unbreakable joints).

[ZedNova]

98% is pretty reliable. For a while I only tested things 5 times, which allowed me to measure reliability to degrees of 20%. When I started getting into 300+ ton rockets, I found that sometimes things would test successfully 4 times in a row, and fail the 5th time for no apparent reason. Then, I started testing things 10 times, giving reliability in degrees of 10%. For something to have a 98% reliability, that means you would of had to test it 50 times.

I got a million backup systems and i have waaaay too much extra fuel.

[CaptainArchmage]

Recently I have had some 100% reliable rockets, which means I am either playing the game very right or very wrong.

Yes, I have excess fuel on some of these, and recently changes have been "adding more boosters to increase delta-v from Kerbin orbit" or "I'm going to put in better control systems and optimise the number of gimballing engines vs. lightweight engines", so the ships seem overengineered.... until you have to perform powered flight over the Mun to find a flat landing ground, that is. The Mun rocket design I have could probably make it to Eve or Duna orbit and back with an optimal trajectory.

Edited February 4, 2014 by CaptainArchmage

[OrbitusII]

Have a gander at Maiden 1:

100% reliable since 0.20! I've improved it over the months and never a major issue. Newer versions feature extra fuel on the side boosters and an extra poodle/Rockomax 16 stage after the Mainsail runs out. The only problems I have with the Maiden line are my reconstructions of it (I typically forget struts) each version. It's not the greatest at rolling but is super-stable and can take about 20 tons to orbit.

Maiden-series rockets have built things like this

and this...

The newest version eliminates the side boosters and is designed explicitly for launching this crew transport vehicle:

I think Maidens were used to launch these parts, but I'm not sure. I'll have to check that next time I can play...

Overall, the Maiden Launch System is my most reliable rocket ever. It looks good, flies well, and has a long history of success.