Defeated by KSP 1.0 aerodynamics -- need counseling...

[Hotel26]

(The subject of vehicle design ought probably to have its own section in the forum; I'm posting here... sorry if I've misposted)

Background

I discovered KSP in Dec 2014 as a result of a Xmas gift. From Dec 25 until the release of KSP 1.0, I spent more time

flying in KSP than I spent at my full-time engineering job that pays the bills. With the release of 1.0 and discovering that

my growing stable of single-architecture vehicles would no longer fly, I stopped playing KSP -- cold turkey -- thinking that

I would "solve the problems" later. Well, I'm back, but struggling and still failing.

I would really appreciate general and specific advice from anyone who has had some success making the transition to

1.0.

Deeper Background

(I just read a post about "what kind of Kerbonaut are you?") I suppose there are two kinds of players in general. Those

who like to think that astro-engineering in KSP means they could cut it at NASA as real astrojocks. Maybe they want

more and more reality until they can't cut it any more and, hey, it was fun while it lasted. Then maybe there are those

like me who want a simplified environment that teaches you real things by offering graduated challenges but leaves

plenty of room for imagination, creativity, innovation and fun.

As engineering is my day job I can tell you it is detailed, hard, tedious, repetitious, boring, frustrating and exhausting.

Whichever kind of Kerbonaut you are (see above), KSP has taught you much more about e.g. solar systems or orbital

mechanics than you ever could imagine have learned anywhere other than NASA and given you that education broadly

as well (which you would not get at NASA which is why engineers there relax at home (or at work ) flying missions

in KSP!).

I don't want to avoid worthy challenges! (These are what make recreational pastimes fun.) Orbital mechanics and

particularly rendez-vous and docking are enormously hard (the first Russian cosmonaut to try to dock spent a very

frustrating hour and then ran out of RCS fuel after having been as close as 1 meter. And he had no idea what he was

doing: a space cowboy with the Right Stuff making it up as he went... The Russian-language version of KSP had not

yet been released...) Orbital mechanics is challenging but it is my friend because if I understand it, it helps me get

to where I want to go. Conversely, atmospheric re-entry is a foe because it does absolutely nothing for me except

oppose my progress.

While the first kind of Kerbonaut might very well tend to build spacecraft by consciously or unconsciously copying

NASA ("mmm, yeah, this looks real"), the second kind of Kerbonaut may produce outlandish and even butt-ugly,

preposterously impossible craft that are entertaining only if you don't puke on first sight. (Scott Manley has a

couple of early videos of space-suited Kerbonauts sitting atop rockets in the open that demonstrate the concept

pretty well!) The first kind of Kerbonaut probably should have tuned out of this post long ago -- and my apologies to

any who have stayed with me this far!!

Example

Before I fell off KSP when 1.0 came out, I had been developing a common architecture for a whole family of multi-

purpose vehicles to facilitate exploration of the whole Kerbol system. Yes, they were butt-ugly. But underneath,

in my opinion anyway, there were concepts and insights and innovations that I had seen nowhere else. I was

having fun. In addition, I came across someone else's craft labeled simply 'Duna Explorer' that I thought was

ingenious and inspiring.

None of this worked any more with KSP 1.0 and the typical problem was stable ascent for a while, followed by

instability and tumbling.

It's time to reveal a picture of exactly what I mean by the aerospace term: "butt-ugly". Please take a look at

http://forum.kerbalspaceprogram.com/threads/113874-Neptune-Space-Labs and follow it to the Youtube site,

Neptune Space Labs. You'll only get to see Neptune Gemini (space station launcher) which is a craft launched

in pairs to make a six-spoke space station. It's a member of the Neptune family and the architecture is based

upon three tiers:

1. upper deck is payload and varies between family members depending upon mission(s)
   
2. middle deck is the structural base for the whole vehicle and is also the deorbit stage
   
3. lower deck contains the staged boosters for ascent and orbital insertion.
   

Neptune tends to be tall which leads to the problem in KSP 1.0.

Failed Remedies:

• I've put fairings onthe high-drag upper deck components
  
• I've put extra RCS especially at the top extremities
  
• I've put totally big-ass fins at the bottom of the craft (and everywhere else)
  
• I've shortened the length of the craft by distributing the payload (spokes) around the girth
  
• I've employed dynamic (and static) shifting of fuel mass to control the CoM so that the upper deck does not become comparatively fuel-heavy
  
• I've eschewed use of the MechJeb Ascent Guidance to opt for simple, stable, vertical ascent in the initial phase
  

Nothing works. KSP 1.0 aerodynamics has defeated me. Unfairly, I feel because the totally fat-ass tail fins should have totally prevented end-

swapping (hmm, well maybe not). To boot, I have lost my mind. (Obviously, when you rerun the tape at Neptune Space Labs...)

Some of you will say, "yeah, that's ugly, Squad should revoke your KSP license and refund your purchase price and ban you for life."

The Prize

I think I've set the stage now with a pretty impossible challenge.. and packed the grandstand with observers. Any one in the very helpful and

very knowledgeable KSP community who can identify in particular what is my major malfunction in KSP 1.0 all of a sudden is gonna deserve

to comport themselves henceforth on two legs with a John Wayne, or perhaps I should say, Chuck Yeager, style swagger.

More seriously... guys. what are the biggest lessons you learned when you hit the KSP Mach 1.0 aerodynamics barrier? How did you break

through?

[John FX]

Turn early, turn slow.

Tall, thin rockets with fins at the base.

Have your main engine be one that gimballs and don't do any dramatic changes in atitude once you are past around 250 metres per second.

Try not to have too much payload on top of a very empty rocket if the rocket does not have enough control authority this is especially important around 30 kilometres high which is a point when you are likely to be staging an empty stage with lots of control authority and moving to a smaller stage which has less control authority.

EDIT : I see that you made your craft shorter thinking that would make it more stable when in fact the opposite is true. Make your craft longer and thinner then the fins will have a larger moment of force and therefore will control the attitude of your craft more efficiently as they are further from your centre of mass.

Edited November 10, 2015 by John FX

[Starhawk]

The new aerodynamics model changed everything. Drag matters now. Streamlining and nosecones matter now. We use a much more real-world ascent profile now.

With the changes in 1.0, many people had difficulty learning how to manage getting stuff through the atmosphere. Reams were written.

If I may direct you to one of the threads in question, perhaps reading some of Kuzzter's How Not To Flip could be of some assistance.

You can still build butt-ugly (I consider myself a master of butt-ugly) stuff that works great, but you have to shepherd it carefully through the atmosphere now.

Happy landings!

[MarvinKitFox]

....

Failed Remedies:

I've shortened the length of the craft by distributing the payload (spokes) around the girth

I've employed dynamic (and static) shifting of fuel mass to control the CoM so that the upper deck does not become comparatively fuel-heavy

I've eschewed use of the MechJeb Ascent Guidance to opt for simple, stable, vertical ascent in the initial phase

And right there you show that you do not have the faintest clue what forces are involved.

You need to sit back, throw away old preconceptions, and think about what forces are involved in aero.

Hint:

Go skinny, not squat. Yes, bigger turning forces on a long vehicle, but also much greater turning moment. Allows you to react in much slower time.

Go Nose-heavy, not tail heavy. You want drag to pull you straight, not push you off-prograde.

Aim very very close to your *surface* prograde vector. I.e. start turning very early, and turn very slowly. Not so early that gravity-induced sideslip becomes a problem though.

Steerable engines are awesome for control, rather than extra RCS, for when you get into the high atmosphere.

[Warzouz]

(I can't see you vids from here)

Your rocket should mostly turn by itself up to 35km after you do a forced turn at around 50 to 100m/s.

- If it flips : tails fins are to small

- If it's hard to turn : tail fins are too big.

Don't hesitate to use wing parts

SAS is usually counter-productive at low altitude (it fights the natural gravity turn of you rocket and increases wobble). Use it only after 30/40km when aerodynamic has nearly no effect.

Staged launchers are harder to use because TWR and aerodynamic change a lot at each stage. I prefer using SSTO rockets which I find easier to fly. Further more SSTO are easier to recover.

Here is my generic SSTO launcher family from 15T to 600T payload (not 1.0.5 certified)

http://forum.kerbalspaceprogram.com/threads/123195

[paul23]

While of course gimballed-engines are good while in space etc. I found that for the same increase in price during the first 10km using active control surfaces is better than adding a liquid core + swivel. Advantage of active control surfaces is that they become more effective the higher the speed, so it counteracts the increase in instability.

The biggest problem I see however with these things is that they tend to overcorrect. Especially when the whole system starts having a memory (wobbling) the overcorrection might lead to resonance if using "sas". (I've had a rocket destroyed by this already).

[Jovus]

If you still have the craft file, and want to post a link to it, I'll be happy to make the necessary changes in everything but the payload, get that thing to orbit, and then explain how I did it, if you think that would help.

[Pecan]

And my summary of the above ...

Darts are made heavy at the front with fins at the back for the same reason as rockets - so they point the right way.

Tall and thin is good, but short and fat still works if you follow the important rules.

1. Mass at the front (top)
   
2. Drag at the back (bottom)
   
3. Start gravity turn early (1km or lower)
   
4. Turn slowly (heading within 5-degrees of prograde) all the way up
   
5. Which means a TWR no higher than about 1.3
   

If you don't have the mass and drag correct the ship will probably flip.

If you turn too sharply it'll probably flip.

If you turn too late but keep it gentle you're left with a huge circularisation burn.

If you turn too far too early nothing else matters much because you'll just run out of fuel from all that draggy low-atmosphere stuff.

[fourfa]

follow the important rules.

1. Mass at the front (top)
   
2. Drag at the back (bottom)
   
3. Start gravity turn early (1km or lower)
   
4. Turn slowly (heading within 5-degrees of prograde) all the way up
   
5. Which means a TWR no higher than about 1.3
   

Interesting - any comments on applying these ideas to a mechjeb ascent profile? Limiting acceleration to 13m/s (ie ~1.3TWR) and 5-deg AoA yields a suborbital splashdown near the KSC. 5-deg AoA with no acceleration limit yields a big circularization burn, as it can't come close to finishing the turn. I usually let it go to 15 degrees on a 40% path with no acceleration limit, and get to orbit on 3500-3600m/s like everyone else...

I'm guessing that 1.3TWR means only at the sea level pad, and let it go up during flight?

[GoSlash27]

Hotel26,

In addition to what's been posted up- stream, I'd like to recommend a fundamental change in design philosophy.

If it doesn't look like it'd work in real life, it probably won't work in KSP anymore.

I'm a big proponent of letting form follow function. Your payload mass, DV requirement, and minimum t/w are the basic criteria that the launch vehicle must meet. So work that out and then engineer the launcher to meet the requirement.

Good luck!

-Slashy

[Pecan]

I'm guessing that 1.3TWR means only at the sea level pad, and let it go up during flight?

Correct, should have read Launch TWR. For a SSTO that'll increase quite a lot, however, if you have a two-stage design the second should generally have an initial TWR of less than 1.1. From launch you should, I suggest, be starting your gravity turn by the 1km altitude and/or 100m/s velocity - which should be around the same sort of time with that launch TWR. 5 degrees is not a hard limit either, you can turn further from prograde while you are relatively slow because i) there simply isn't the atmospheric drag to make you flip yet, ii) prograde will move towards your current heading quickly because there isn't much existing 'prograde' to adjust. The important thing is not to be turning too sharply once you've built-up serious velocity. Hence a slow (5-degree limit) turn through the main part of the ascent. Since you're turning slowly you have to start early just to get pitched (or yawed) over far enough before your apoapsis has fully developed.

For what it's worth the old '10km altitude then pitch to 45-degrees, then pitch to 20-degrees when Ap reaches 50km' advice still provides good milestones, it's just that you want to be doing the whole thing as one smooth turn instead of in two sharp jerks. So, maybe kick your gravity-turn off with a 10 or 15 degree turn while you're low and slow if need be, then continue to turn slowly within the 5-degree limit so you're passing 45-degrees around 10km altitude. A continuation of the same thing should, more or less, have you at a very shallow pitch by the time your Ap reaches orbital altitude, leaving you with only a short circularisation burn.

If you still have a big circularisation burn with the 5-degree limit then you've simply got too much thrust, getting you too high before the turn gets far enough. I've never used MJs acceleration limiter, just adjusted engines as required. Incidentally, the TWR limit is only there so the 5-degree limit has time to turn you and that's only there so the vehicle doesn't flip. If it's happier (aerodynamically stable) with a sharper turn then by all means use one - what works is good. As has been mentioned in other threads though, optimising for launch-to-orbit dV is generally not useful compared to cost or payload ratio.

[Hotel26]

Thanks! I am very happy with all the helpful response and I will consider everything carefully.

Most helpful to my sanity is to know that I am not alone in going through this and that it is a big subject: thanks!!

The suggestion to post the craft file so anyone interested can lay hands on it is great. Here it is:

https://mega.nz/#!ZsBwFIqB

[Hotel26]

Well, guys... thank you. I would never have guessed you could make this so easy for me.

I have just performed two successful tests.

With the Duna Explorer, I made no changes but started an early gravity turn and then let SAS handle it by tracking Surface Prograde.

Easy to orbit and lots of fuel left, too.

With Neptune Gemini, I added steerable winglets, unlocked all the engine gimbals, initiated the turn before reaching 100 m/s and tracked Sfc Pro.

Nice and steady, longer in the fiery atmosphere than I liked but it too orbited to 100K with lots of fuel left. I didn't even put fairings on the payload.

I want to thank you all: John FX, Starhawk, Warzouz, paul23, Jovus, Pecan, fourfa, GoSlash27 for your insightful xepertise and professional

courtesy.

Brilliant, guys, thank you so much!

[Razgriz1]

...longer in the fiery atmosphere than I liked...

For what it's worth, that "fiery atmosphere" that you experience on ascent is not quite the same as what you're experiencing upon reentry. This fire on the way up is KSP's attempt at mach effects that are much lower than those that would actually create plasma like it looks like. This means that it really isn't doing much to heat your craft. It isn't nearly as destructive as reentry effects.

Furthermore, a proper gravity turn, like the one you've started taking, will almost always produce these fiery effects and it's nothing to worry about.

[Pecan]

I have just performed two successful tests.

Applause.

Once you hear the news and it clicks into place everything suddenly looks easy again doesn't it? :-)

(Right up until whatever the next personal challenge is, that's the fun of KSP - "we choose to do these things ... not because they are easy but because they are hard")