True or False, 10 Kilometers is the Best Place to Start Your Gravity Turn

[Titan Space Agency]

This will be a new serries posted every sunday (it will not always be true or false but will just be siencey discusions)

I did an experement to see if Scott Manley (if you are reading this Scott hi ) was true that you that the best time to start you gravity turn is at 10 kilometers. Here is my data, the left is the g turn point, the left is the fuel left. ( i used the kerbal X with mechjeb launching it)

6km-616 fuel

7km-617 fuel

8km-617 fuel

9km-616 fuel

10km-614 fuel

So it turns out 7/8km is the best place to start your g turns. Thats all I have to say there is not much logic here I have to explane

[illectro]

But My manual launch had 624 Fuel Left!

The thing that's missing here is the parameters for the curve setting, you need to setup a 3 dimensional grid of paremeters -

Start Turn Altitude

Final Turn Altitude

Curve %

[Francesco]

yeah it's not enough to just say, "start at 10 km".

also, while it is true that 10km is a good rule of thumb, every rocket tends to have its own optimal ascent profile:

7/8 km is good if your TWR is higher than average.

[rdfox]

AND, if you're someone who likes to get into the habit of following Best Practices before they're actually necessary and thus take Range Safety concerns into account how, so that when it becomes troublesome to drop spent stages on the launch site, you'll be used to it already, you should probably start your turn at about 1km or even lower, to make sure that you're well out over the ocean before you start jettisoning boosters or stages.

[drakesdoom]

There is no "best," many rockets will have varying optimal flight profiles. T/w including as fuel burns and staging are the two biggest factors. 10k is still a good starting point to test from for most rockets just try to hold 2g acceleration until 90% of the atmosphere is behind you.

[legoheli]

Usually, I go for a real-world style immediate 5 degree pitchover. This means my lateral velocity is increasing with my vertical velocity instead of essentially being at zero when you begin the pitch-over maneuver. I am not sure how much of a difference this makes in fuel consumption, but I do not think that it's a disadvantage at all.

[Moon Goddess]

As soon as I clear the launch clamps I tilt to 85, then 80, then 75, etc etc etc. by the time I'm around 50 km I'm tilted to 0 degrees

Seems to work

[Motokid600]

I've always wondered if I'm waisting fuel by doing it wrong. I start my g turn at 10k no matter what craft. But the pitch I turn over on is what varies for me. Very large craft pitch over slowly to 0 degrees. About 5 degrees every few hundred m/s. Then once approaching 1k m/s I level the craft off. Smaller craft turn right to 45 then pitch over.

[Diche Bach]

As soon as I clear the launch clamps I tilt to 85, then 80, then 75, etc etc etc. by the time I'm around 50 km I'm tilted to 0 degrees

Seems to work

Since I'm learning to use the RemoteTech Flight Computer, I've started doing something like this and anecdotally it does seem to spare more fuel.

What are the actual ship parameters that determine the "optimum" ascent path? Would it be possible to use an 'eyeball' and intuition method with any given spaceship, and adjust your flight path accordingly?

[Joshington]

I wait until I'm at 15-20km, turn 45 degrees, then when my apoapsis is at 70km I turn 45 degrees again. I'm sure it's not the most efficient gravity turn ever, but it works.

[tavert]

What are the actual ship parameters that determine the "optimum" ascent path? Would it be possible to use an 'eyeball' and intuition method with any given spaceship, and adjust your flight path accordingly?

Thrust-to-weight ratio is the big one, since in KSP's current drag model, drag force is proportional to craft mass, and most parts have the same 0.2 drag coefficient. Also how your TWR changes over time, which is a function of engine Isp, and any discontinuities due to staging.

Low TWR craft will need a bit more pitch to maintain the same trajectory, and possibly a more gradual gravity turn to provide more hang-time to get up to orbital speed.

[Diche Bach]

Thrust-to-weight ratio is the big one, since in KSP's current drag model, drag force is proportional to craft mass, and most parts have the same 0.2 drag coefficient. Also how your TWR changes over time, which is a function of engine Isp, and any discontinuities due to staging.

Low TWR craft will need a bit more pitch to maintain the same trajectory, and possibly a more gradual gravity turn to provide more hang-time to get up to orbital speed.

Okay! More-or-less what I expected. I guess in the real world, drag is far more complicated. I could infer from your post that, at this point, aerodynamic parts such as conical nose cones do not actually reduce drag and increase efficiency?

Also when you say "more pitch" you mean that, they need to pitch away from vertical earlier in their ascent (typically toward zero eh?).

If I am correct then, outside of a model that reflected the real life aerodynamic stresses of turning while under thrust, is there really any downside to engaging in a more stepwise pitching to east process like Moon Goddess described for a ship of _any_ size?

I would imagine that for a small turn, there is a bit of lost velocity but if you are doing say, 5-degrees to east every 1000, or 1500m of ascent, I would think that, in the relatively simple model you seem to be describing for KSP, that ascent pattern would be optimum for effectively all ships, though perhaps more critical for attaining orbit for a low TWR ship?

My past few orbits I've been following exactly this type of ascent path (ascend 1000 or 1500 m, turn 5 degrees east; repeat). This is so far with ships that have adequate surface TWR (2.09 with 7.8 Max TWR and 4535 Atmo Dv) and 28.6 ton mass. I am pretty confident I'm getting into a low orbit (75 or 80km) with more fuel left over using the stepwise ascent path.

Edited August 7, 2013 by Diche Bach

[tavert]

Okay! More-or-less what I expected. I guess in the real world, drag is far more complicated. I could infer from your post that, at this point, aerodynamic parts such as conical nose cones do not actually reduce drag and increase efficiency?

Correct. They're purely for looks in the stock game right now. There's a mod called Ferram Aerospace Research that completely changes the drag system, but I've actually never played with it since it leads to some different performance with rockets. From what I've heard it makes planes a lot more fun though, and I believe it fixes infiniglide and the overpowered-ness of jets and intakes.

Also when you say "more pitch" you mean that, they need to pitch away from vertical earlier in their ascent (typically toward zero eh?).

Sorry, should've said "higher pitch." With a lower TWR, you'll need to point higher up to fight gravity, and won't be able to put as much of your thrust into horizontal speed if you want to gain the same amount of altitude. You may want to start the gravity turn slightly earlier since you'll have lower drag losses at low TWR, and you'll need the gravity turn to be slower and end higher in order to maintain an ascent at low TWR.

If I am correct then, outside of a model that reflected the real life aerodynamic stresses of turning while under thrust, is there really any downside to engaging in a more stepwise pitching to east process like Moon Goddess described for a ship of _any_ size?

I would imagine that for a small turn, there is a bit of lost velocity but if you are doing say, 5-degrees to east every 1000, or 1500m of ascent, I would think that, in the relatively simple model you seem to be describing for KSP, that ascent pattern would be optimum for effectively all ships, though perhaps more critical for attaining orbit for a low TWR ship?

My past few orbits I've been following exactly this type of ascent path (ascend 1000 or 1500 m, turn 5 degrees east; repeat). This is so far with ships that have adequate surface TWR (2.09 with 7.8 Max TWR and 4535 Atmo Dv) and 28.6 ton mass. I am pretty confident I'm getting into a low orbit (75 or 80km) with more fuel left over using the stepwise ascent path.

It is advantageous to change pitch smoothly and continuously, in order to reduce steering losses by keeping your attitude fairly close to the prograde marker. 5 degrees likely won't make too much difference early on, as far as the trajectory goes what tends to be more important is the rate of downward pitch at medium altitudes from 10-30 km, and your altitude/speed when you reach horizontal.

If you use MechJeb, a set of parameters that has been shown to work very well for typical-TWR craft is starting the gravity turn at 8 km, ending at 37 km but with final flight path angle of 5 degrees above horizontal, turn shape 66%, limit to terminal velocity and limit acceleration to 22 m/s^2, corrective steering disabled. Low TWR craft usually need a higher turn end altitude.

Edited August 7, 2013 by tavert

[Diche Bach]

Ah thanks for clarifying all of that Tavert! Great fun stuff

I've actually enjoyed playing and replaying the "Alpha" mission and the "Sputnik I" mission in the NT Space Program and the Stock mission pack for Mission Controller. Basically just, send up a probe to 70 or 75 km and land it safely. Because MC uses a budget there is an incentive to use as cheap of rocket as you can, and also to land it on land-best of all flat land where it won't get damaged-because you can recycle it for more.

It is actually somewhat challenging to get a ballistic missile to come down more or less where it took off from! Very interesting to fiddle around and see how little course and velocity changes can have dramatic changes on the shape of the parabola, but I should stop as I risk sending this thread off topic.

[Nuke]

false. different launchers require a turn profile that best fits the capabilities of the vehicle.

[Agent86]

I usually start my gravity turn when my forward velocity exceeds 500m/s, regardless of altitude. I figure there's not alot of point doing a gravity turn if your behemoth is only moving at 200m/s at 10K.

[InfinityArch]

I play with FAR, so I start turning almost immediately. You can generally get going a fair bit faster lower in the atmosphere with FAR assuming you've got an aerodynamic design, which makes delta-v to orbit a great deal more variable depending on your crafts aerodynamic properties, ranging from about 3.5 km/s to about 5 km/s.

[NeoMorph]

false. different launchers require a turn profile that best fits the capabilities of the vehicle.

This for sure... bigger launch vehicles need to get to where the air is thinner before turning or else you will get a hell of a lot of air drag eating into your fuel supply. So yeah... false.

Actually I deliberately waste fuel on some launches? Why? So that when the main booster stage cuts out we are at 85% of orbital velocity and it drops into the ocean instead of leaving it in orbit. Without the extra drag the booster could reach orbit and leave a huge assed navigation threat for later missions (and yeah, I actually had a VERY close incident of hitting a booster that I left in low Kerbin orbit. Was a "WHOA... WHAT WAS THAT?" situation as it went flying past (well, the launch vehicle was going fast for orbit when we passed the old booster).

[magnemoe]

Usually, I go for a real-world style immediate 5 degree pitchover. This means my lateral velocity is increasing with my vertical velocity instead of essentially being at zero when you begin the pitch-over maneuver. I am not sure how much of a difference this makes in fuel consumption, but I do not think that it's a disadvantage at all.

As I understand its two reasons for the 5 degree pitchover in real world, one is that you want to rocket to point directly in the direction of flight to reduce drag, doing an turn 10 km up will generate a lot of drag, second is safety if the rocket fails its less likely to crash into the launch platform.

[Tyren]

My 760 tons Monster rocket is just not able to turn before 5k...due to the density of the atmosphere (and FARs drag model) it will instantly flip over and you have a 225 pieces Meteor shower incoming

with a TWR about ~2 in the first stage (12800 kN Thrust) i dont bother burning up to 10-12 km before starting the turn. I love this rocket, pulls up 70-80 tons of payload without breaking to sweat.

Dont think you can say that right over a thumb, as mentioned before.

[Nuke]

if your upper stage uses nerva or ion engines, then a higher that usual apokerb is usually required so that those engines have time to complete the orbit. the gravity turn in that case is hardly a turn, you might turn to 45 degrees or even stop at 60 degrees tops, and then hold that angle with the upper stage until an orbit is certain. but if your top stage uses something like a mainsail, might as well start turning at 4000-8000 meters and then level out completely when your apokerb is 75% of your desired apokerb. sometimes you need to keep a first stage vertical till its used up, just for stability (true for very large launchers or when you are launching with a lot of drag up front). there is never a rule of thumb.

[HeadHunter67]

I usually start my gravity turn when my forward velocity exceeds 500m/s, regardless of altitude. I figure there's not alot of point doing a gravity turn if your behemoth is only moving at 200m/s at 10K.

If you're already at 500 m/s before you start your turn, you are inevitably facing one of two problems:

Either you are going too fast at a low altitude and losing a significant portion of that thrust to increased atmospheric drag that you are creating, or

you are high enough in the atmosphere that drag is not an issue, but you'll use a significant amount of that thrust to go from a vertical ascent to an orbital vector.

If you are doing 200 m/s at 10k, you're actually doing it right. I highly recommend you look into the concept of "terminal velocity" and see why you've been wasting fuel and losing thrust with your current technique.