KSP Rocket Design Process failings

[MaverickSawyer]

I'm making a flowchart to help new folks to troubleshoot their designs, but I need more than just my own experiences to fill in the metaphorical blanks in terms of common reasons for a rocket to fail. So, I'm asking you folks, the community at large, for some of your most headdesk inducing, facepalm causing, or otherwise derptastic moments to help with this endeavour...

[Temstar]

I think the main thing is delta-V and TWR. You need more than 1 TWR to lift off and you need somewhere about 3500m/s of vac delta V to reach orbit with safe margins for a new player, assuming of course you're not using vacuum only engine on first stage.

Barring extreme cases like a flying barn door or clearly structurally unsound design, if you have these two numbers you automatically know if a rocket can reach orbit or not without even have to do test flight. Unfortunately Squad is steadfastly against giving these two numbers away so one would have to rely on mods like KER.

[MaverickSawyer]

Well, I'm talking more about things like "I didn't know that clipping the tail fins into the fuel tank would summon the Kraken on staging" kinds of things. Trying to keep it simple, not getting into things like TWR and dV. Just "Did it fly? Y/N If No, add boosters" kinda stuff.

 

EDIT: Here's the flowchart so far...

Edited December 22, 2015 by MaverickSawyer

[monophonic]

My number one reason for rockets failing to leave pad is failing to set throttle to full, followed by the sticky launchpad bug. Latter has thankfully become quite rare with 1.0.5.

[Bandock]

One critical failure of any rocket is if it is too top heavy (especially once the center of mass is altered by fuel usage during flight) and starts flipping over like crazy (even with winglet support!).  Good remedy for that in some cases is to make it wider instead of making it too tall.  Larger parts do have fewer problems with top heavy issues, but still need to watch out for it.

[swjr-swis]

• Did you check the staging order? (KSP default staging of added parts is wacky to say the least) -> fix staging order
• Do all parts of a stage remain attached to their stage? (wouldn't be the first to attach winglets to boosters that are the first to be decoupled) -> reattach parts to the intended stage
• Are radial boosters hitting the center/other parts when decoupled? -> place sepratrons to counter rotation
• Does the rocket flip over halfway through a stage? -> Empty that stage's tanks in VAB and check CoM; redesign to keep CoM towards the top even with empty tanks
• Does the rocket keep rotating or veering off course? -> check symmetry, you may have tried moving/replacing a part and forgot to set the correct symmetry
• Is the rocket still not keeping course? -> check that CoM and CoT are where you expect them to be (generally in one line through the rocket center)
• Still not keeping course? -> test your winglets' reactions to steering before launch, maybe one or more need to be inverted
• Rocket reached orbit but it's now immobile? -> Add RCS for orbital maneuvering
• Rocket went dead somewhere in orbit? -> add a power source (solar panels or RTG)

Edited December 22, 2015 by swjr-swis

[Plusck]

39 minutes ago, Bandock said:

One critical failure of any rocket is if it is too top heavy (especially once the center of mass is altered by fuel usage during flight) and starts flipping over like crazy (even with winglet support!).  Good remedy for that in some cases is to make it wider instead of making it too tall.  Larger parts do have fewer problems with top heavy issues, but still need to watch out for it.

I used to think that until someone told me it was flat-out wrong.

Top-heavy is good and doesn't flip: it makes the nose droop towards the horizon maybe but it makes it fly like a dart.

Bottom-heavy makes it flip because the momentum of the rear (engine + full fuel tanks) joins forces with the drag of the front, so it flips. Fuel usage in flight is from the top down, so the worst flipping time starts about half-way through a stage, especially if you have a draggy top.

Edited December 22, 2015 by Plusck

[KickAssAndGiggle]

Are you planning on landing? Do you have landing legs?

Are you planning on landing on the Mun or similar? Are you relying on a parachute to do so? Don't.

[GoSlash27]

^ What Plusck said; top heavy is good. Top draggy is bad.

8 hours ago, Temstar said:

Unfortunately Squad is steadfastly against giving these two numbers away so one would have to rely on mods like KER.

Well... you don't have to rely on mods like KER to get those numbers, but they are the easiest way.

Best,

-Slashy

[MaxwellsDemon]

My facepalm moments are typically of two types:

A )  Rocket starts flexing/twirling, usually around the point of Max Q...  sometimes throttling back helps, but too much and I'm putting my path to orbit at risk.  Three-step approach:  1) throttle back to 70% or so and see if it dampens out sufficiently to get me to the next stage (when the problem usually disappears), then gradually throttle up again.   2) If that doesn't work, and the thing is about to break up or does break up, declare it a "simulation" and revert to VAB for moar struts!  (This is easier using KCT when I'm actually doing a simulation.)  3) Back to the drawing board to make the thing shorter and fatter.

B )  This doesn't happen so much anymore since I've learned to keep a close eye on TWR and deltaV during design phase, but I used to have a big problem getting larger payloads to orbit... I'd get them into about the upper ionosphere and then I'd hit my apogee and nothing I could do could prevent me from falling back to the ground.  Triangulating the right mixture of TWR, deltaV, and burn time is still a challenge, but one I'm getting steadily better at.

ETA:  OH... almost forgot C)... solid booster explodes too close to the rocket and blows its tail off.  I used to very carefully position the boosters and time the roll during launch so they would drop clear.   Then I discovered Sepratrons, and have rarely had that problem since.

 

I rarely have issues past the launch phase, although when I do, they're usually because I 1) don't have enough solar panels to keep the lights on, or 2) I neglected to put on enough RCS and/or reaction wheels to maneuver.  As long as I've kept an eye on those, getting to LKO is still the main obstacle.

Edited December 22, 2015 by MaxwellsDemon

[MaverickSawyer]

6 hours ago, swjr-swis said:

• Did you check the staging order? (KSP default staging of added parts is wacky to say the least) -> fix staging order
• Do all parts of a stage remain attached to their stage? (wouldn't be the first to attach winglets to boosters that are the first to be decoupled) -> reattach parts to the intended stage
• Are radial boosters hitting the center/other parts when decoupled? -> place sepratrons to counter rotation
• Does the rocket flip over halfway through a stage? -> Empty that stage's tanks in VAB and check CoM; redesign to keep CoM towards the top even with empty tanks
• Does the rocket keep rotating or veering off course? -> check symmetry, you may have tried moving/replacing a part and forgot to set the correct symmetry
• Is the rocket still not keeping course? -> check that CoM and CoT are where you expect them to be (generally in one line through the rocket center)
• Still not keeping course? -> test your winglets' reactions to steering before launch, maybe one or more need to be inverted
• Rocket reached orbit but it's now immobile? -> Add RCS for orbital maneuvering
• Rocket went dead somewhere in orbit? -> add a power source (solar panels or RTG)

Thank you! That's a fairly comprehensive set of checks.

[swjr-swis]

2 minutes ago, MaverickSawyer said:

That's a fairly comprehensive set of checks.

What can I say, I'm the Edison of Space. That is to say: I never fail at going to space - I just keep discovering 10000 ways that prevent me from getting there.

[The Optimist]

If your rocket looks dirty, it'll make your kerbal's pants brown.

 

heehehehehehhehehhehh

[322997am]

My top failiure is when I have radially attached boosters and I for some reason get afraid of crashing and I spam the space bar and then the boosters fly right into the cockpit and boom.

[Bandock]

22 hours ago, Plusck said:

I used to think that until someone told me it was flat-out wrong.

Top-heavy is good and doesn't flip: it makes the nose droop towards the horizon maybe but it makes it fly like a dart.

Bottom-heavy makes it flip because the momentum of the rear (engine + full fuel tanks) joins forces with the drag of the front, so it flips. Fuel usage in flight is from the top down, so the worst flipping time starts about half-way through a stage, especially if you have a draggy top.

Yeah, you're right.  It's usually a problem when it drains fuel from the top fuel tank.  I forgot when it gets too bottom heavy (for a super tall rocket).

One helpful way to remedy that is build a rocket that is radially constructed, reducing the risk of flipping.  I could be wrong though.

[sdj64]

On 12/22/2015 at 9:13 AM, MaxwellsDemon said:

A )  Rocket starts flexing/twirling, usually around the point of Max Q...  sometimes throttling back helps, but too much and I'm putting my path to orbit at risk.  Three-step approach:  1) throttle back to 70% or so and see if it dampens out sufficiently to get me to the next stage (when the problem usually disappears), then gradually throttle up again.   2) If that doesn't work, and the thing is about to break up or does break up, declare it a "simulation" and revert to VAB for moar struts!  (This is easier using KCT when I'm actually doing a simulation.)  3) Back to the drawing board to make the thing shorter and fatter.

Often it means you have too much control authority.  Throttling back helps partly because it reduces the effect that your engine gimbals have on your rocket.  One of my favorite mid-career satellite lifters uses a single Kickback first stage and ends up being really tall if the satellite is going far away, and it gets to orbit just fine with no struts if flown carefully.  Solution: reduce gimbal limits, remove or lock control surfaces, stiffen with struts as you said, or put your "control from here" core closer to the middle rather than the very top.

Another problem - first stage doesn't turn / doesn't turn well:  add (sparingly) control surfaces, remove some fixed fins, or reduce speed.  And the opposite problem, first stage turns too quickly or is jerky:  remove control surfaces, add fixed fins, use soft controls (caps lock, I forget the real name of that feature).  Adding reaction wheels to the rocket stages is usually not the way to go for lift rockets, though reaction wheels in the payload can help control or hurt by contributing to wobble.

Flight path contributes to all of these problems, if you fly far from prograde or turn in jerky movements it can make it much worse.

 

Remember that there is a bug with the center of lift of fairings being too far forward in 1.0.5, causing some designs that should work to be overly flip-happy.  Stock Bug Fixes corrects it but of course many people are against mods in general.

Edited December 23, 2015 by sdj64

[Joshua A.C. Newman]

On 12/22/2015 at 3:56 AM, Bandock said:

One critical failure of any rocket is if it is too top heavy (especially once the center of mass is altered by fuel usage during flight) and starts flipping over like crazy (even with winglet support!).  Good remedy for that in some cases is to make it wider instead of making it too tall.  Larger parts do have fewer problems with top heavy issues, but still need to watch out for it.

You've got this backward. Winglets on the front are adding drag to the front — exacerbated by the CM moving aft as fuel is burned up because it's a longer and longer lever. Of course, once atmospheric drag is gone, they have no effect, so it's OK to put them only on the bottommost stages.

[White Owl]

If it looks sort of like a real world rocket, there's a good chance it'll fly sort of like a real world rocket.

[Evanitis]

I totally love the 'Kerbal' way of engineering - as in eyeballing designs and field testing hundred tonnes of rockets a few times to see if they blow up when ignited. Your flowchart got me to smile - I'd totally hang it on my wall if it was a colorful, illustrated supplement. I also prefer to 'add moar boosters' and 'moar struts' as a go-to problemsolving method. But I also know it's the way I find the most fun, and not the ideal one.

I know if I had difficulties, the best way to overcome them would be efficiency and math. The process is described very well in this neat, fresh tutorial. If someone told me he can't get to space / orbit / Mun / etc, I'd probably directed them there - even though I think the 'figure out by trial-and-error' way is the most enjoyable solution.

So my point? Dunno, forgot it while writing this post. But I'd turn your finished flowchat into a big, colorful, illustrated poster once it's done - if you didn't originally plan to do so.

Edited January 2, 2016 by Evanitis

[selfish_meme]

Strutting payload inside fairings, howto

Dont add more SAS, use fins or vernors if necessary