When to start Gravity Turns

[sephirotic]

At Kerbin you should start your gravity turn at 10kms, always. The curve itself through the atmospheric should always be a perfect regular parable till the 70km mark, doesn´t matter if your final desired orbit is 150km, the difference will only be noticeable by few degrees at the end of the parable as you will just burn for longer almost at horizontal.

However, I seen many people here saying that it "dependeds of the rocket, if it is too slow or heavy, start later."

There is some fundament in saying that heavier rockets should have a different behavior when piloting for the gravity turn, however, the actual ascent PATH should be ALWAYS THE SAME. What happens tough, is that a rocket with a LOW TWR (or better, maximum acceleration, e.g. 15m/s²) if turning the direction at the same rate/angle per altitude as a rocket with a higher TWR, it'll generate a differente parabole as the prograde marker will fall faster. What you SHOULD consider when doing a gravity turn, is keeping the PROGRADE MARKER, at the desired ideal angle, NOT THE CENTER/NOSE marker!

Just turning 45 degrees after 10km's is far from ideal, what i've found to be a good "simplified" ascent path is:

10-20kms: Keep the prograde marker above 60 degress

20 -30kms: Keep the prograde marker between 60 and 45 degrees

30-45kms: Keep the prograde marker close to 25 degrees

45+kms: Get as close as possible to 90 degrees while keeping an eye for not overshooting the 70km apoapsis mark.

Doesn´t matter if the rocket has a 60m/s² maximum TWR or just 15m/s², the inclination will be different, but you just need to look at the prograde marker.

By the way, i've successfully made some return vehicles from Eve, took me several tries to discover the ideal turn profile for it's surface, at Eve you should start your gravity turn at 23km , until 35 around 60degrees, until 50km around 30 degrees and etc.

In this video i don´t make an ideal turn profile in Eve but you can check it out anyway:

Edited September 16, 2013 by sephirotic

[allmhuran]

Like giggleplex, I tend to use velocity as the indicator for when to start a turn. Giggleplex said he used 200 m/s as a base, I tend to use about 250 - 275 m/s. On conventional designs where you able to maintain an optimum ascent speed (equal to terminal velocity) this translates to a turn at about 10km, which is I assume the reason why Mr Manley indicates this as a good rule of thumb altitude.

However, a consideration that I have not yet seen mentioned in this thread is what happens when you are building really stupidly large rockets. These probably won't be hitting 275 by 10000, but that's not the real issue.

A much more important concern is the safety of staging. If you attempt to stage radial columns while pitched over but not travelling at a sufficiently high speed you run a very high probability of having the staged columns crash into the rocket. Therefore I would suggest that for very large rockets with a lot of radial stages it is much wiser to ascend slightly less efficiently but much more reliably by maintaining a vertical ascent until you have either shed your most risky stages, or exceeded 300m/s. For my monster rockets this tends to occur by about 15000m, at which point I obviously perform a more aggressive turn directly to 45 degrees.

[Vector]

There is some fundament in saying that heavier rockets should have a different behavior when piloting for the gravity turn, however, the actual ascent PATH should be ALWAYS THE SAME. What happens tough, is that a rocket with a LOW TWR (or better, maximum acceleration, e.g. 15m/s²) if turning the direction at the same rate/angle per altitude as a rocket with a higher TWR, it'll generate a differente parabole as the prograde marker will fall faster. What you SHOULD consider when doing a gravity turn, is keeping the PROGRADE MARKER, at the desired ideal angle, NOT THE CENTER/NOSE marker!

I find this idea interesting, that the path should be the same regardless of the craft. It sounds reasonable that something should be invariant to rocket design and it is definitely not the center/nose marker as a function of altitude. An invariant prograde angle as a function of altitude sounds plausible but I can't prove it.

Keeping the prograde marker along a fixed path would mean that for underpowered rockets, the nose would be in between prograde and zenith, with some extra thrust spent on gravity. For a rocket with plenty of thrust, the nose would either be between prograde and horizontal, or it would be precisely at prograde with less than 100% throttle, truly a gravity turn.

[allmhuran]

An invariant prograde angle as a function of altitude sounds plausible but I can't prove it.

It is indeed an interesting idea. I'm not sure how useful anything outside of actually doing the math would be to answer this question, but in any case, here's one consideration....

The ISP of a rocket increases as atmospheric pressure decreases. Therefore there is something beneficial about getting into thin atmosphere as rapidly as possible (ie, going straight up), with the opposing consideration that the "ideal" launch trajectory is actually horizontal, because that way all of your thrust is being directed into your orbital speed instead of being wasted fighting gravity directly.

A similar argument to the ISP argument can of course be made regarding drag.

So you're either spending extra fuel fighting gravity, or you're spending extra fuel because of more time spent fighting drag and more time spent at lower rocket efficiency.

It wouldn't completely blow my mind if someone came along and showed that the math worked out such that the common factors in these two considerations cancel out and therefore that the optimum trajectory is independent of TWR, velocity, whatever, meaning that there is indeed one ideal ascent path for any vehicle regardless of its velocity at any particular altitude.

But for what it's worth - which is very little - my gut says that's not the case, and that rockets with lower TWR should ascend vertically to a higher altitude than rockets with higher TWR before turning... and that this is particularly true of rockets with very different ISP and/or TWR across multiple launch stages.

Edit: My gut also says that it might be possible to demonstrate this one way or the other without having to do the precise mathematics, but rather by a logical reductio (start with something like "Imagine a rocket whose second stage has an ISP of 10 million in vacuum and an ISP of 1 at sea level").

Edited September 16, 2013 by allmhuran

[Kasuha]

10-20kms: Keep the prograde marker above 60 degress

20 -30kms: Keep the prograde marker between 60 and 45 degrees

30-45kms: Keep the prograde marker close to 25 degrees

45+kms: Get as close as possible to 90 degrees while keeping an eye for not overstating the 70km apoapsis mark.

Sounds like an interesting idea. I have a loaded lifter which I struggled very very hard to get to orbit yesterday, I'll try your approach with it and test if it gets any easier.

[sephirotic]

A similar argument to the ISP argument can of course be made regarding drag.

So you're either spending extra fuel fighting gravity, or you're spending extra fuel because of more time spent fighting drag and more time spent at lower rocket efficiency.

Indeed the point of going straight up is too get away from the thick atmosphere. The scale of atmosphere (and drag) is exponential, you can look the table of "terminal velocity" from the wiki:

Altitude (m) Velocity (m/s)

75 100.9

1000 110.5

2000 121.9

3000 134.5

4000 148.4

5000 163.7

6000 180.6

7000 199.3

8000 219.9

9000 242.6

10000 267.7

12500 342.4

15000 437.8

20000 716

32000 2332

You can see that Between 9~12km is the sweet spot to start the gravity turn by this table. Anything beyond that and the resistance will be already to low to justify the waste of fuel of going straight up.

my gut says that's not the case, and that rockets with lower TWR should ascend vertically to a higher altitude than rockets with higher TWR before turning... and that this is particularly true of rockets with very different ISP and/or TWR across multiple launch stages.

Sorry to contradict your gut, but TWR makes absolute no difference to influent the ascent profile whatsoever. The only factor that could justify starting the curve a little later (but the difference would be very small, mind that), is the specific drag coefficient of the vessel in question. If the vessel has a very HIGH drag (overral close to 0.3, most vessels stay around 0.2, 0.23) , then it may be worth consider starting the gravity turn a little late, like around 12~15km, but hardly any higher than that and wouldn´t be surprise if this .07 of extra drag reflect even less on the ascent profile (like starting at 10.500 instead of 10km) at all.

About the ISP argument, again, since by 10km you only have 10% of atmosphere left, the isp of most rockets is already very close to it's maximum, (LV-N is 725s vs 800s while the Skipper is 342 vs 350s) so the wasted fuel (d/v) of transversing an extra 2 or 3km up MUCH higher (more than 200m/s) than the small benefit of saved fuel regarding ISP.

Nevertheless, In my favourite ascent profile that i mentioned early, even tough i start to curve at 10km height, i don´t make much of a steep curve, only getting at 60 degress by 20km of height, with means much of the transverse through this region of the atmosphere were more vertical than horizontal anyway.

Edited September 16, 2013 by sephirotic

[M5000]

It really depends on the rocket you're flying.

Look, if your rocket has a higher TWR, you can start earlier, if it's a lower TWR, I'd suggest starting a bit higher, since the atmosphere thins out logarithmically.

If your rocket like, takes off like really fast, you can probably afford a 7Km beginning on a gravity turn. If you've got a lower TWR, you may need to start later at like, 13Km.

The only exception to this type of general guideline is if you have lots of side boosters and you plan on losing some TWR in staging them away, so you may want to factor that in.

[Kasuha]

10-20kms: Keep the prograde marker above 60 degress

20 -30kms: Keep the prograde marker between 60 and 45 degrees

30-45kms: Keep the prograde marker close to 25 degrees

45+kms: Get as close as possible to 90 degrees while keeping an eye for not overstating the 70km apoapsis mark.

Sounds like an interesting idea. I have a loaded lifter which I struggled very very hard to get to orbit yesterday, I'll try your approach with it and test if it gets any easier.

I tried it and no, I couldn't get that ship to the orbit this way. It all went smooth till 30 km but after that there was no way to keep the prograde marker high enough even if I accelerated directly upwards.

I did get it to the orbit before using different approach though. So I guess it still has to be tailored for the ship.

[Tripzter]

For most crafts 10k, for very slow, top heavy, heavy crafts at 15-20k.

[AletzN1]

I usually start it at 10k and aim at 45 degrees. Then at +30k I aim it below the horizon so I start getting a periapsis. And as far as gravity turns go thats my usual one.

[Dooz]

i Tend to Gravity turn at 7km

[EuSouONumero345]

Scott Manley made a video about this

[ArmchairGravy]

If someone could check me on this, I'd appreciate it. I do 80 degreees off the pad, turn to 60 at 10k, and go ballistic (45 degrees) at 20k until apoapsis hits 70k. I then turn off engines until 20 seconds or so before apoapsis then burn at horizon until I have orbital velocity. If my orbital velocity is below 1k/s I burn earlier, above 1.5k/s I burn later. It works, but is it efficient?

[allmhuran]

It works, but is it efficient?

Due to the extreme density of Kerbin's atmosphere at lower altitudes it's probably better to go straight up for a while until it thins out... hence the recommendation to stay pretty much vertical until about 10000m. The rest of your technique is good. It is actually possible, if you do everything perfectly, to burn all the way up to apoapsis and end up with a nice circular orbit, but the overall delta-V saved is not worth worrying about. It's more a "yay I did it perfectly" thing.

[sgt_flyer]

my g-turns for most of my rockets is usually as such : i start a smooth curve at 8km, aiming to get to 45° around 15km, then i continue my smooth turn to reach 30° at 20km, 20° at 30, and i usually finish between 5° and 0° around 45000m, then i continue burning in this direction until i get my apoapsis at the altitude i want, before coasting to it for circularization.

[sephirotic]

I decided to have some empirical data after so many people just throwing they random opinions over the table. I got specially concerned about some guys recommeding starting the curve over 15km of height, in my experience that is clearly a bad choice and a waste of fuel. Doesn´t matter if your rocket is too large heavy and weak.

I did over 60 launches with 2 vehicles and 3 sub settings of engines for each with mechjeb and differenc ascenting profiles.

I expected a very consistent data, it was, but with some small surprises.

I'll organize everything and share the full report and conclusions here by the end of the day. What i can say tough, is that the sweet spot is indeed between 8~10km of height for starting the gravity both for large and small rockets, with either high TWR low ISP, and the opposite: LOW THRUST high ISP.

Another interesting fact: The Sweet spot of acceleration, is 25m/s², anything above or bellow that is a waste of fuel.

[Kasuha]

It might be useful to do it with asparagus-staged rocket. With it you are losing thrust as you ascend and if you are carrying heavy enough load the last stage may be unable to lift it on its own so it may rely on impulse given by previous stages.

[sephirotic]

@Kasuha

One of the two ships i tested on, is indeed a large asparagused rocket with 6 lateral orange tank boosters and a final nerva stage.

pic related.

I have to go to work now, by the end of the day i organize the results and post it here.

[LeadMagnet]

Everyone has some pretty interesting idea's. With "evidence" and "proof" of what you should and should not be doing. If the point of playing KSP for you is trying to save fuel, all these suggestions will get you in the ball park.

Just remember, it's a game. Try playing "in-effeciently"! Waste that fuel!! If going straight up to 70Km then hanging a right works, great! If you want to "fly the nav ball", super! Whatever works, works, and what don't, don't.

I'm a "what do I feel like today" kinda player. 6K, 10K, 15K and sometimes....25K!!! Just to see what happens! For the moment, fuel is free, so I'm burning all I can pack in!!

[sephirotic]

@LeadMagnet

Re-read the Op's question. He objectively wants to know the most efficient possible gravity turn.

I'm a "what do I feel like today" kinda player. 6K, 10K, 15K and sometimes....25K!!!

Yes, of course we all know this just a game and ultimately as long as we have fun, anyone can play as they like. But you can´t CLAIM, that a 15km gravity turn is an efficient turn because that is just simply wrong. And we are not condemning you by wanting to waste fuel either, you can of course play as you want.

Just like when i'm bored and just feel like wasting fuel spinning around on a rocket or over engineering things, i do it shamelessly, that's just a game, but knowing the ideal ascent curve, could be a lifesaver if we want to role-play and do difficult missions with large launches that have limited fuel and d/v, so this is a relevant matter to discuss.

There is no point in being relativist on objective matters such as this.

[LeadMagnet]

@LeadMagnet Re-read the Op's question. He objectively wants to know the most efficient possible gravity turn.

*snip*

There is no point in being relativist on objective matters such as this.

Fully agree with you. I did read the OP, which is why I found the argument interesting. Being relativist is never a bad thing. If I can save 50% more fuel than you, doing it my way, because I can't get your way to work, then why not?

My point is, there are so many variables, ya just gotta go with your gut sometimes. I re watched Scotts clip on gravity turns and thought calling the "un-named posters" morons was a bit harsh. That was their way, and it worked. Scotts way is (admittedly) spot on, but only for that one rocket design. But did he save 100's of units? Nope. By all means, listen to the forum, but test it yourself. I just re-watched a video of a guy doing a 150m/90 degree gravity turn from Minmus!! (I am so trying that!) Can you say "thread the needle!".

And your right again, in Career Mode, you will most certainly want the most efficient burn, for a given design, tested in Sandbox, implemented in Career. It only makes sense.

On a side note, I didn't see anyone ask if his accent turns towards 90. So if it's towards 180, how does this change peoples thinking? Kinda defeats the term "gravity turn", but who says you have to go East anyway!?

Any-hoozzle...M2CW

Edited September 17, 2013 by LeadMagnet

[wasmic]

I just re-watched a video of a guy doing a 150m/90 degree gravity turn from Minmus!!

That is actually the most efficient way - actually, the most efficient ascent from an atmosphereless body is a 0m/180 degrees turn. That, however, won't work as you'll just crash into the ground trying. When I ascend from mun, I'll usually go 500 meters up, then turn over to 180 over a fewe seconds. on Minmus, it depends on whether I'm right next to a mountain or some distance away. If there are no mountains right next to me, I'll just tip over to 180 at about 100 meters.

Oh, and you don't have to restrict your throttle to a TWR of 2.0 on gravity-less bodies, as terminal velocity doesn't exist there. Any less power than full power is just gravity losses.

Also, you don't have to gravity turn east. It'll just require a little more delta-V if you do it to the west instead.

[rkman]

There is more to ascent efficiency than at what altitude to start the gravity turn.

Other factors are:

- at what altitude should craft attitude be horizontal

- (max) acceleration during ascent

Going by the ascent of the winner of the launch Efficiency Exercise in the Challenge forum: http://forum.kerbalspaceprogram.com/showthread.php/39196-Launch-Efficiency-Exercise-Updated-for-0-21-1/page13

(Nao: http://forum.kerbalspaceprogram.com/showthread.php/39196-Launch-Efficiency-Exercise-Updated-for-0-21-1?p=526321&viewfull=1#post526321 )

(also see "Best ascent path for mechjeb" http://forum.kerbalspaceprogram.com/showthread.php/33570-Best-ascent-path-for-mechjeb?p=484014&viewfull=1#post484014 )

The most fuel efficient ascent to LKO appears to be:

- start gravity turn between 5 and 10km alt

- gradually turn so that the craft is horizontal at 35 to 40km alt

- limit acceleration to ~22m/s

It can make as much as ~200m/s difference in expended delta-V.

Edited September 17, 2013 by rkman

[Alistone]

That is actually the most efficient way - actually, the most efficient ascent from an atmosphereless body is a 0m/180 degrees turn. That, however, won't work as you'll just crash into the ground trying. When I ascend from mun, I'll usually go 500 meters up, then turn over to 180 over a fewe seconds. on Minmus, it depends on whether I'm right next to a mountain or some distance away. If there are no mountains right next to me, I'll just tip over to 180 at about 100 meters.

.

Those 180 degree turns always seem to end with "you'll just crash into the ground trying." Very Kerbal.

(90 would be turning to the horizon; 180 would be turning completely around and pointing back into the surface)

Gravity turn between 10-15km depending on the T/W ratio. Lower T/W means I usually wait longer to start my turn; but this is often because I want to empty and eject radial tanks from a stage while I'm still going vertical. If I lean over first they may hit my main rocket.

[tavert]

Look carefully at those links. They actually don't end horizontal, they end at 5 degrees above horizontal at 37 km. This results in lower steering losses during the final part of the ascent.

There is an absolute minimum-fuel trajectory for a hypothetical craft with infinite maximum thrust, and imposing a constraint of lower maximum thrust will both change the optimal trajectory and increase the minimum amount of fuel required to ascend to the same orbit.

Have a look through this thread if you want to see a lot of math and progress towards solving for the optimal trajectory for a given craft numerically: http://forum.kerbalspaceprogram.com/showthread.php/46194-I-need-someone-help-me-do-some-math-for-launch-optimization

The thread starts out talking about the pure-vertical Goddard problem, but later on gets into the full ascent-to-orbit gravity turn problem as well. It's a challenging nonlinear constrained optimal control problem, and the results will absolutely depend on the parameters of a given rocket.