Gravity Turns

[Paul G]

I'm still very new at this, I have a couple of lunar orbits and returns under my belt, and have looked through a fair few YouTube videos, but haven't got a huge amount of flight time under my belt.

Now I get the point of what a gravity turn us, why you would use them, and that without some mods, the manoeuvre I'm doing after takeoff is not a true gravity turn, but all the faq's and YouTube clips consistently say that the manoeuvre that we know as the gravity turn shouldn't be started until after 10km.

My question is - why so late? Photos of real life launches show that things start moving away from vertical much sooner. The photo below (borrowed from Cmdr Chris Hadfields twitter feed) shows what i mean.

http://twitter.com/Cmdr_Hadfield/status/435808451973758976/photo/1

I don't think the rocket nose is following this arc (at least not at the start), as this shot really just shows the arc that the engine flame travels, but there is horizontal travel occurring much earlier than on Kerbal.

Is this a limitation of the physics engine, or am I missing something?

Thanks

Paul

[Blackstar]

The reason is that in stock KSP, the lower atmosphere is very thick and "soupy". The atmospheric density there limits your terminal velocity to the point where doing the gravity turn early is inefficient and wastes a lot of delta V fighting the air resistance (or more accurately, having to fight gravity more than you should because your airspeed is limited and you can't get enough upward momentum while also flying at an angle).

You simply can't go faster than a couple hundred m/s in the lower atmosphere, so there's no point in trying to build up lateral velocity so soon.

Once you get above 10K, the air gets really thin really fast, so by turning there you can start building up to orbital velocity.

Having FAR installed changes the atmosphere of Kerbin to more closely resemble Earth's atmosphere but at a smaller scale. With that mod, the atmosphere is thinner earlier, but with less of a sudden falloff as you get higher. Starting your gravity turn right off the pad does make a lot of sense in that case and that's how many people launch rockets with FAR. Note that FAR's change to the atmosphere makes achieving orbit require less effective deltaV for rockets, essentially making the game easier. If that's not what you want, there is an ISP adjuster mod you can install to compensate, or you can do what some of us do and add more challenge mods (Deadly Reentry, RemoteTech, etc) to add the difficulty in other places instead.

EDIT: Or of course, you can stick with stock, armed with the knowledge of Kerbin's unique atmosphere and what the most efficient ascent path is for it. Whatever is fun to play.

Table of terminal velocity on Kerbin at various altitudes is here:

http://wiki.kerbalspaceprogram.com/wiki/Kerbin#Atmosphere

Edited February 19, 2014 by Blackstar

[Taki117]

The pic is actually pretty accurate for nose travel as well, but as Blackstar said it's because the aerodynamics in KSP is very wonky, making anything below 10km more like pea soup and less like atmosphere. FAR changes both Atmospheric conditions and how drag is calculated, meaning that if you design "Real-world" rockets (Rockets that are plausible in the real world) they will perform much better in FAR than the traditional pancake rockets.

[Hodo]

As Taki said, FAR fixes most of that nonsense that is the stock KSP aerodynamics.

[Greenfire32]

Yeah, Kerbin's atmo is very thick until about around 10km. The reason real-life launches execute their gravity turns earlier is mostly because Earth's atmo allows for it and party because dropping boosters on mission control is generally frowned upon.

[Superfluous J]

I start my gravity turn at 5k altitude, and after that I hit certain milestones on the navball at certain apoapsis milestones.

10k ap = 80-70 degrees

15k ap = 60-50 degrees

20k ap = 40 degrees

30k ap = 30 degrees

40k ap = 25 degrees

50k ap = 20 degrees

60+k ap = horizon/prograde, which is essentially the same thing.

It's pretty darn efficient, though I'm sure it could be more so. I'm usually almost in orbit when my apoapsis is at 80k which is what I like. This is about when I ditch my final lifter and my nukes complete the orbit, so obviously it's 99% complete by then.

[Kasuha]

Turn right 45 degrees at 10 km is decent rule of thumb for newbies as any rocket with sufficient fuel reserve can get to orbit that way without losing too much to atmospheric drag. Later on, you'll find out that doing the turn gradually is better and that different rocket designs require different approach. Unless you'll start using automated tools for it such as MechJeb in which case you likely won't care anymore.

[Claw]

Yeah, Kerbin's atmo is very thick until about around 10km. The reason real-life launches execute their gravity turns earlier is mostly because Earth's atmo allows for it and party because dropping boosters on mission control is generally frowned upon.

Also, the thing most people gloss over, is that in the real world if you wait till you're traveling at high speed and yank your rocket over to 45 deg of AoA, the aerodynamic loads would cause it to structurally fail.

So rockets in the real world tend to tip over earlier so that your effective AoA is zero as a balancing act against the parasitic drag.

Also, I see lots of complaints about how "suddenly" the KSP atmosphere changes at 10km. There certainly is a marked change in performance, but the Earth experiences a similar

Earth:

Kerbin:

The scales of the planet are different, but the basic idea of the curves are the same. Kerbin's atmospheric pressure drops off faster than Earth, but it's also scaled differently. I don't know anything about FAR, so I didn't look into comparing with that. It's not ideal, but it isn't a horrific mess either.

The Earth atmosphere experiences a sudden change at 14km-ish, illustrated below. Effectively making flight above 42,000-ish feet a different experience.

Maybe someone can point me to the error of my ways?

Edited February 20, 2014 by Claw

[Sternface]

With FAR installed, (and RSS), when I start my gravity turn is totally dependent on the design of my launch vehicle.

With larger and heavier rockets, I tend to start my turn later, especially if it has a high TWR that cannot be throttled back (like when using the Real Fuels mod). Without throttle control, you need to get out of the lower atmosphere as soon as possible before dynamic pressure builds too much. Larger rockets, especially longer ones that don't use asparagus staging, are simply more susceptible to transverse aerodynamic stress. Having too large an AoA (Angle of Attack, which is sometimes confused with attitude) when dynamic pressure is high can easily force my rocket into an uncontrollable downward pitch, or even worse shear it apart. Keeping an almost zero angle of attack while turning would solve this by starting the gravity turn very soon after launch before velocity builds, but I usually want to get out of the lower atmosphere ASAP.

With large Saturn V-like launch vehicles, I basically start my pitchover maneuver when dynamic pressure (represented by Q in the FAR flight info) drops below a certain threshhold, rather than by a specific altitude. Depending on the rocket, this can vary from 4000m to 16,000m. This allows me to risk a non-zero AoA before my velocity vector lines up with my pitch without tearing apart my rocket. Usually I wait for Q to be under 8000. In real life, large rockets usually wait to break through Q-max before pitching over and beginning their downrange acceleration, which is one reason they don't begin the gravity turn immediately. (this is what happens if you turn too quickly

)

Another reason why I don't just start my turn right away to negate the need to manage Q and AoA is because I like to stage my boosters while the vehicle is facing up. Don't like them slamming into me.

When you do your gravity turn, keep the NAV ball on surface mode and watch your vector. You want to keep your AoA (deviation from your vector through the atmosphere) as close to zero as possible. It will follow your turn if you do it slowly. Deviate too much and transverse stresses can ruin your day.

P.S. When I say I start later, I mean when I pitch over to less than 75-80 degrees attitude. It is always good to start right off the launch pad, but very slowly.

Edited February 20, 2014 by Sternface

[technion]

I know for my early experiments with FAR, turning as early as people suggest often ends up pushing the prograde marker to ninety degrees and trying to navigate the globe at 15km, even when the rocket is still pointing at 45 degrees.

[Taki117]

Sternface makes a good point when using FAR, but my question is how do you determine Max-Q?

[Sternface]

I know for my early experiments with FAR, turning as early as people suggest often ends up pushing the prograde marker to ninety degrees and trying to navigate the globe at 15km, even when the rocket is still pointing at 45 degrees.

You want to start early, but very slowly. The first 10-20km of my turn rarely is at an attitude less than 60 degrees. Once dynamic pressure drops you can really pitch over. But yeah, if you pitch over too quickly, transverse stress will force your rocket down. It also helps to keep your eye on your vertical speed, and keep it from dropping too low. I try to keep my Ap about a minute away at all times, and vary my pitch to keep it there. Pitch up if it starts dropping, and down if it starts rising. No need to coast to circularization that way, which helps if using the engine ignitor mod. You could also vary thrust instead of pitch, or both.

[Sternface]

Sternface makes a good point when using FAR, but my question is how do you determine Max-Q?

If you are using FAR, it is listed in the flight into popup. I don't think you can stock tho.

P.S. Maybe if you click on the nose and watch the drag? I believe it should be proportional to Q, but I could be wrong, and it could be modeled differently in-game.

P.P.S. Does stock even tell you drag? Can't recall.

Edited February 20, 2014 by Sternface

[Blackstar]

You want to start early, but very slowly. The first 10-20km of my turn rarely is at an attitude less than 60 degrees.

^^^ This. The initial pitch bump shouldn't be more than a degree or two. If your rocket is balanced, then you can take your hands off the keyboard and it will just about fly itself. The only thing you might have to tweak is the throttle, depending on how heavy the rocket is.

One question for technion is does your rocket have tail fins? They are pretty much a must to do a proper gravity turn in FAR. The fins catch the air and help keep you on your prograde vector. Even static ones without control surfaces will do. It's possible to do it without fins with lots of manual control, but much more difficult.

[technion]

One question for technion is does your rocket have tail fins? They are pretty much a must to do a proper gravity turn in FAR. The fins catch the air and help keep you on your prograde vector. Even static ones without control surfaces will do. It's possible to do it without fins with lots of manual control, but much more difficult.

They do now. I've learnt that lesson the hard way. .Except where I couldn't afford the weight of a fin.

[Sternface]

One question for technion is does your rocket have tail fins? They are pretty much a must to do a proper gravity turn in FAR. The fins catch the air and help keep you on your prograde vector. Even static ones without control surfaces will do. It's possible to do it without fins with lots of manual control, but much more difficult.

You know I always thought the same thing. And I always used tail fins with FAR. However this last week I tried using zero tail fins on my larger rockets and I found them to be so much more stable, especially under high Q. The only thing is that you must keep AoA very small throughout the entire turn, and mistakes are much less forgiving. Without tail fins, I was actually able to use SAS on a 1000 ton+ rocket for the first time without catastrophic results.

Never going back to them!

[hijack]

P.S. Blackstar, if you haven't tried out the engine igniter mod or Real Fuels, I highly recommend them if you want to change up how you do your launches. The Real Fuels mod changes engines to have limited throttling, with lower stage engines often only able to burn at 100% power. And the igniter mod limits the amount of ignitions you can do. So you are forced to use pitch as the only real means of speed control.

[/hijack]

Edited February 20, 2014 by Sternface

[Paul G]

It was the booster dropping whilst still going straight up that made me start to question this. I read (or watched) somewhere that even going straight up results in a slightly parabola shaped flight path relative to someone on the ground due to the spin of the planet, but even so, it would seem bad form to drop a whole stage back down on the launch pad

[EuSouONumero345]

This *might* help:

[Alchemist]

There's always a tradeoff between drag and weight. You have gravity losses until the point you get your orbit and only getting lateral speed reduces the gravity losses, so you want to start the gravity turn as early as possible. But when you fly in dense air and have high airspeed you get huge drag, so you want to climb as fast as possible while going really slow. To balance that the optimal solution is keeping just below your terminal velocity, but it is increasing exponentially with you gaining altitude. So the moment to start the gravity turn is when with you flying straight up at current terminal velocity, the terminal velocity starts growing faster than you can accelerate. At this point you don't need to gain more vertical speed (just ensure it doesn't drop too fast) and you can afford to start gaining the lateral velocity without losing too much energy to drag.

As for going couple degrees off right from the start, that's done in real life to avoid what happened during the second launch of N1

[Runboy]

Ill try to keep this as simple as possible.

Ok,so in KSP the lower atmosphere is much thicker than Earths,and thus if you start your gravity turn earlier than 10km you will waste Delta V by spending more time trying to push through the thicker lower atmosphere.On the other hand,if you start your gravity turn early on Earth,you will not waste as much DeltaV because the lower atmosphere on Earth is thinner than on Kerbin.

[rkman]

The goal of optimal ascent is to spend the least amount of fuel.

A rule-of-thumb for Kerbin is that an efficient ascent to orbit takes about 4500m/s delta-v. Afaik the record stands at a little over 4300. (without FAR)

My advice is to experiment with the starting altitude for gravity turn, and with acceleration along the middle- and upper part of the ascent trajectory.

I regularly achieve some 4400 to 4500m/s by starting the gravity turn a little below 10k (7 ~ 8km), and reducing acceleration to some 14m/s² during the middle- and upper part of the ascent.

Sternface makes a good point when using FAR, but my question is how do you determine Max-Q?

Mechjeb. See terminal velocity.

[DrD]

I've aslo read that rocket engine thrust increases with decreasing air pressure, for example the Saturn I-C went from 7.5 million at takeoff to 9 million pounds of thrust at stage seperation which occured at about 65 kilometers altitude. Is this modeled in KSP?

[Wayfare]

My question is - why so late?

The answer is - it's bogus

Giving your rocket a very slight (2-3 degrees) nudge soon after liftoff, then heading to about 10 degrees off vertical once you're about 4000 meters up will do wonders for your gravity turns. The specifics do depend on your TWR curve (asparagus designs keep a fairly even TWR while serial designs tend to build up TWR as their fuel drains, then drop off sharply as they stage) but I find going by surface velocity to be a good rule of thumb:

- Halfway to 45 degrees at 250m/s

- 45 degrees at 450-500m/s

- halfway to horizontal at 750-800m/s

- Horizontal at 1200-1400m/s

This will give you a very efficient profile and result in a very flat trajectory. In fact on my most efficient launches, by the time I push my Ap to 80km my Pe is almost peeking above the surface of Kerbin, making circularization as easy as pie. The Munshine Lifters were built around this profile and it yields delta-v savings of over 10% compared to the standard 10km turn, 4500m/s profile.

Edited February 20, 2014 by Wayfare

[Wayfare]

A rule-of-thumb for Kerbin is that an efficient ascent to orbit takes about 4500m/s delta-v. Afaik the record stands at a little over 4300. (without FAR)

Hogwash Here's an Imgur album of a Munshine V lifting over 55 tons to an 80km orbit at under 3900m/s detla-v. All stock except for Kerbal Engineer and Kerbal Alarm Clock.

Javascript is disabled. View full album

If that sets a record I demand to know

[sjwt]

I've aslo read that rocket engine thrust increases with decreasing air pressure, for example the Saturn I-C went from 7.5 million at takeoff to 9 million pounds of thrust at stage seperation which occured at about 65 kilometers altitude. Is this modeled in KSP?

Yes, but for background reasons its modeled as a change in ISP (use the right click on engine parts to see the details)