Getting Into Orbit: What Happens Above 10K?

[The Flying Kerbal]

Without blowing my own trumpet, over the months since I first starter playing KSP, I have to admit I've definitely got better.  I suppose when you have as many rockets blow up, disintegrate, or simply fail to leave the launchpad, as I've had, you really MUST learn something.   However I still think I could do better flight profiles from launch to orbit, so maybe you'll be able to give me a few pointers?

OK... So I throttle up to maximum on most occasions and launch the rocket.  At somewhere between 50 and 100 m/s I nudge the vessel over to about 5 degrees.  Then I gradually increase the angle until it's sitting on 45 degrees at 10 kilometres.  This is now where things starts to get a bit more hazy, what do I do next?  What I actually do is now sit there at this angle while the Ap rises up towards the altitude I'm aiming for, and when it's about 10-20K short of this, I start to push the rocket on over until it finally points horizontally.

I keep reading about keeping the time to Ap around 45 seconds, some saying between 45 secs. and a minute is OK, but for me to do that usually means I have to actually point below the horizon which is surely a bad idea?

I do appreciate no two rockets are the same and what works for one could see another tumbling back down to Kerbin, but a few general pointers on how to fly above 10K would be really useful and very much appreciated.

[Kryxal]

What sort of TWR are you working with?  Also, do you have a mod that shows you that sort of info?

Oh, and the "what next" is you follow prograde as best you're able (if you do it right, except for the initial turn you can follow prograde the whole way).

Edited June 12, 2017 by Kryxal

[Weywot8]

I'm assuming that if you have to point below the horizon, you have a pretty good TWR to begin with (at max throttle).

TL;DR answer:

• set your SAS to prograde
• continuously manage your throttle up/down so that you are always maintaining a comfortable amount of time* before hitting Ap - Ap will continue to climb
• continue until you hit the Ap you want (as visible in map mode)
• kill throttle, coast
• set up circularisation node
• burn as appropriate.

*comfortable = enough time to set up the node and burn (presumably 40-45 sec for the average kerbalnaut).

Longer answer with explanation:

Burning prograde during ascent makes the most of the dV used, converting it into both vertical and horizontal velocity while gravity changes your direction (ignoring drag). Pointing in other directions leads to "steering losses" although others might disagree, depending on how they define the term. Anyhows, smarter minds have proven somewhere in the forums that for any given orbit, (one of?) the best ascent profiles dV-wise would be to provide just enough thrust all the way throughout ascent such that at Ap, you can shut down the engines in a circularised orbit, the wonders of a well executed pure gravity turn.

The "burn prograde, 40-45 sec to Ap" rule of thumb tries to capture most of that while maintaining room for kebal/human limitations. 

I suspect that the current launch profile with 45 degrees at 10km could be better - sounds like there is alot of active pointing away from prograde - too steep or too fast at initial launch? Depending on ship size/drag, I can get to 75-80km orbit with 2900-3100dV. With my smallest ships, directly launching at a 15-30 degree angle, I've used ~2750dV before.

Of course, all this only matters if you want an efficient launch - otherwise, MOAR boosters can always do the trick! Seriously though, SRBs are cheap and you can dump them at 3-20km after they've gotten you off the ground and past the draggy atmosphere. (which means you can use a lighter engine, have more fuel)

See the discussion below.

 

 

Edited June 12, 2017 by Weywot8
spellling

[ForScience6686]

The gravity turn should be a gradual turn.  Most times I add no input after the initial nudge.  Usually just set to follow prograde or turn off sas altogether.  I wouldn't hold the angle after 10k, continue to let it turn towards the horizon.  Facing prograde contributes the least drag, so you want to be pointing dead on while in the atmosphere.  If you stay to increase ap to fast reduce throttle, and finally complete the orbit at ap.

[CSE]

Keep pitching over after you pass 10km altitude, with the intention of reaching horizontal somewhere between 50-60km. This is regardless of the orbital altitude that you're aiming for. You can think of it as gaining altitude without lifting your rocket at all - instead, you're using your horizontal speed to make the planet curve away underneath you faster than you're falling towards it.

Use as much throttle as your rocket's strength and heat tolerance can withstand. Saving weight by using a less powerful engine is good rocket design when you're in the VAB, but when flying the rocket: if you've brought it then use it. Get the horizontal speed that gives you free altitude.

Once you're horizontal, don't worry about the time to Ap being too long. Cutoff and coast when you need to. Time to Ap is a good instrument to watch as you're pitching over - the faster it's rising, the more you pitch - but when you're horizontal with the time rising then you've basically made it to orbit at that point.

[Spricigo]

Ideally you launch at full throttle, tilt* the rocket just a bit and do nothing until the orbit is circular.

45° @ 10km is just a rule of thumbs for mid coarse evaluation. This works 'ok'  for a 'typical'  rocket and makes for a 'safe'  quick advice to someone struggling with gravity turn.  If you already reach orbit with reasonable efficiency is time to go past that rule. Rather than aiming for being at X° @ Ykm, you should consider how good are you going so far and how your adjustments will affect your trajectory. 

Suppose the trajectory is too steep. You can orient the craft more horizontal this actively lowe your path but increase drag. Or reduce your thrust,  giving time to gravity pull the trajectory down but also slowing down (forfeiting some Oberth effect) 

It's a matter of balancing all these factors.  Which one is more of a issue varies from situation to situation. 

*or launch it already tilted. 

[The Flying Kerbal]

Loads of really terrific answers there guys.  I didn't get a chance to do any KSP related activities last night, but hopefully I'll try some of these suggestions and techniques tonight.

If my plans work out for this evening, I'll build an incredibly simple and straight forward rocket and just play around with it over a number of launches, and see what happens.

Thanks to everyone for posting so many excellent replies and I'll let you know what happens.

[The Flying Kerbal]

Just as a little afterthought, would all this hold up when launching from other planets with atmospheres, or would the figures, timings, etc. change depending on gravity, atmospheric density etc?

[Spricigo]

36 minutes ago, The Flying Kerbal said:

Just as a little afterthought, would all this hold up when launching from other planets with atmospheres, or would the figures, timings, etc. change depending on gravity, atmospheric density etc?

It will change depending on circumstances,  not only celestial body but also,  more commonly, rocket design. 

After doing experimenting with a basic rocket,  try high thrust,  low thrust,  high drag variations. 

 

[severedsolo]

I know most people disagree with me on this, but I don't use the throttle at all on ascent. I just set it to full and go.

My lifters are usually two stage designs, designed to drop the main lifter into the atmosphere as either (depending on the design) my PE reaches 0 (if I've over-engineered) or when the fuel runs out.

So what I do, is I make a note of the TWR of the main launch stage. I then aim for a "median" surface TWR of 1.3. So if the main stack has a TWR of 1.2, I set the boosters to 1.4 (using the thrust limiters), if it were 1.4 I'd set the boosters to 1.2 (it's usually under it has to be said).

The end result is usually a rocket that basically follows the "standard ascent profile" - I personally aim for 45 degrees at 10-12km. but if the main stack has a particularly low TWR it may be higher than that, but when the main stack kicks in, it pitches over quickly (using "aim at prograde" in SAS).

If I'm pitching over too quickly, I turn the boosters up a little, or if they are already at 100% I boost the main a little. If I'm not pitching over fast enough, I turn them down. Then it's just "T", "Z" "Space" - wait to hit 100m/s and then pitch over 5-10 degrees depending on the design. Hold Prograde all the way to a 100k AP, don't have to touch the controls again until it's time to kill the throttle and set the maneuvre for circularisation.

Now, I will say, this may not be the most efficient launch profile. I use an average of 3.6k - 3.8k of delta-v on the way up, the Delta-V map claims it's possible in 3.4k.

Edited June 13, 2017 by severedsolo

[The Flying Kerbal]

I really like the sound of your launches Severedsolo, although I suspect to achieve it would take a bit more skill in rocket design than I've got.  Any chance you've posted videos of such launches, I wouldn't mind seeing this in action.

[severedsolo]

48 minutes ago, The Flying Kerbal said:

I really like the sound of your launches Severedsolo, although I suspect to achieve it would take a bit more skill in rocket design than I've got.  Any chance you've posted videos of such launches, I wouldn't mind seeing this in action.

Here you go, I just filmed one for you. This is a standard early Career (on my way to Mun) rocket. As I said, payloads are expected to circularise themselves in orbit.

You'll notice I had to correct the inclination a bit early on, that's down to the bad aerodynamics on the boosters (they are on for so little time, it's easier to sort it out later, than add weight making them aerodynamic).

If your TWR is outside the 1.1-1.5 range you have probably over/under engineered.

It's also worth noting that the "booster and stack" method works best on 1.25m parts. On larger engines (Skipper/Mainsail in particular) the TWR can get away from you quite quickly. On those engines it's worth taking the hit on a lower launch TWR and using the speed when you get higher, or you won't pitch over enough (which is probably why people say to use the throttle... but it's just down to designing around it)

 

Edited June 13, 2017 by severedsolo

[The Flying Kerbal]

That's really awfully good of you, I can't view it at work but as soon as I get home it'll be the first thing I do.  Really looking forward to seeing this... thanks so very much!

[severedsolo]

7 minutes ago, The Flying Kerbal said:

That's really awfully good of you, I can't view it at work but as soon as I get home it'll be the first thing I do.  Really looking forward to seeing this... thanks so very much!

It's no bother. I had KSP open anyway, and I've just bought a graphics card that supports Shadowplay - I've been waiting for an excuse to use it

[Kryxal]

Another metric of how good a launch you managed is the height of your PE once your AP is as high as you want ... the more velocity you have horizontally, the higher it's going to be and the smaller the final burn to establish orbit.  Of course, if your design has a lower TWR you're going to use more delta-v, but that may not be less efficient if you're trading expensive engines for cheap fuel.

[Superfluous J]

4 hours ago, severedsolo said:

I know most people disagree with me on this, but I don't use the throttle at all on ascent. I just set it to full and go.

I don't disagree with you (much) on this. I tinker with my throttle in the VAB, not in ascent.

However, once you're mostly horizontal, throttle really doesn't matter because you're not fighting gravity anymore.

For me, I don't use my altitude at all but instead the altitude of my Apoapsis, as that accounts better for my speed, altitude, and AOA than a simple altitude number. So long as I follow the rules from the start (and have decent TWR, > 1.2 but < 1.5 on the pad and not going crazy high in ascent) this almost never fails to get me a nice efficient orbit.

• Full throttle, straight up at launch.
• Tilt 5 degrees (to 85) when Ap is 1km
• Tilt 5 more degrees (to 80) when Ap is at 4km
• Keep tilting 5 degrees every 2km of Ap, or 10 degrees per 4km. You'll not hit it perfect, but it's not that bad.
• That's 75° at 6km, 70° at 8km, 60° at 12km, 50° at 16km, and 40° at 20km.
• If you did this all correctly, you'll be well over 10km by the time your Ap is over 20km, and you'll have followed prograde most of the way, probably chasing it a little or pulling it down a little.
• After you hit 20km, switch to 5 degrees every 5 km.
• 35° at 25km, 30° at 30km, 20° at 40km, 10° at 50km, and horizontal at 60km and above.
• At this point (and actually once you're over 40-50km really) throttle doesn't matter much anymore as you're not really fighting gravity anymore. If you're low enough, though, air drag may encourage you to throttle down. Heat related explosions will assuredly encourage it, if you have any

This sounds really complicated, but it's not. It breaks down to "5° every 2km to 20km, then 5° every 5km to 60km. At 20km, you want to be at 40°."

[Weywot8]

9 hours ago, The Flying Kerbal said:

Just as a little afterthought, would all this hold up when launching from other planets with atmospheres, or would the figures, timings, etc. change depending on gravity, atmospheric density etc?

• In general, both the TL;DR answer and the suggested very horizontal burn with long coast to circularizing works well once you get out of the 'draggy' part of any atmosphere.  To figure that out, check the atmospheric pressure. For example: 0.01 atm on Kerbin is ~25km. I'm usually at Mach 3.5 speeding up to Mach 5 at this point. Drag for any given speed is ~100x less than at launch and dropping rapidly. On Eve, 0.01atm is at ~45km (KSPWiki)
• Working backwards, the real problem fun is figuring out how to get to 45km on Eve. The initial ascent profile and rocket design is where all the KABOOMS tend to occur . Gravity losses vs drag vs overheating.

A longer walkthrough on initial launch speed and angles if you have time. KSP aerodynamics kinda models real life drag effects at trans-to-supersonic speeds, so: 

• optimal speed is terminal velocity(V_t). That changes with air pressure. The higher you go, the faster your optimal speed. So getting up high, fast is a good idea. And really burning from 20-25km onwards if your angle of ascent permits looks good to.

Kerbin Terminal Velocity Table

Altitude (m)	Velocity (m/s)
75	100.9 (vs ~340m/s, speed of sound)
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 (vs ~295m/s, speed of sound)
20000	716
32000	2332

 

• but terminal velocity also depends of the drag coefficient (C_d) and C_d changes dynamically with speed. IRL example :

KSP's debug menu used to let you see their version of the curve but it's not an option anymore.

 

Since I'm not that patient with launches, on goes the cheap, high thrust SRBs for a high TWR that naturally increases further as they burn. This gets over the trans-sonic hump ASAP, before 15km (and dumping the SRBs). The speed of sound drops by ~15% at 15 km so for the same speed, I'm travelling at a higher Mach number & experiencing relatively less drag . Which means for the same TWR, I'm accelerating MOAR. I'm still experiencing higher drag but over much smaller time. The only reason why it works out OKish dV-wise is that while drag increases exponentially with velocity, KSP and real life atmo has density decreasing exponentially with altitude. And for a fixed acceleration, distance is exponential with time.

Within the 25-30km range, I'm already at optimum velocity and burning here gets the most out of the Oberth effect - I can use a lighter rocket engine with lower TWR and pack more fuel/payload. Only using up to ~3100dV but not as hands free as other methods.

And that works out well with a high initial TWR setup because the rocket can tilt at a much steeper angle instantly after launch and keep tilting under gravity as we climb and drag drops. How low can we go? Because the rocket is already moving rather fast at 20-25km up rather than starting cold, try getting to that altitude with the rocket slowly tilting (gradual active tilting helps, doesn't incur much dV loss at this point) from launch to as much as the angles suggested for various TWR's in the post below. Quite possibly much more if you are up to it... Have fun!  

P.S. As long as it doesn't blow up, overheating is fine. Push through the flaming atmo effects .

P.S.S.  very cool severedsolo. very cool indeed. My over-engineered launches are for >>20 ton payloads using BACCs and LV-T30s, sending 10 tourists to orbit in one ship early in career. Your way is much more impressive.

 

Edited June 13, 2017 by Weywot8
severedsolo gets applause

[Spricigo]

@Weywot8 I think the analises Norcalplanner did on gravity turn on kerbin is more relevant to the topic at hand than the one you linked. Granted that both are very interesting and usefull.

_______________________________

Anayways, take all those advices and discussion with a grain of salt. not because it may be wrong but because there are some cases where it is not so relevant anywaw. Case example

Spoiler

A small sattellite (probecore, antenna, solar panels, battery, thermometer, fuel tank and engine) on top of a SRB with a fairing and some basic fins for aerodynamical purposes.

The usual advice simple don't work for this one. A initial TWR of 1,5 result in not enough control autorithy to prevent it to drop the nose to the ground and crash.How it is set the fins mealt and the temperature peaks at 92% in the fairings. The second stage need to be activated at an altitude that is neither too low not to high, otherwise it'll not reach orbit.

A small difference and instead of an orbit we get a catastrophic failure. It's not what first comes to mind when people talk about efficient gravity turn, but it reaches anywhere within kerbin's SoI for just 2600\F It can be furhter improved? Probably, but given the Pareto Principle, I refrain to do so.

[Weywot8]

5 minutes ago, Spricigo said:

@Weywot8 I think the analises Norcalplanner did on gravity turn on kerbin is more relevant to the topic at hand than the one you linked. Granted that both are very interesting and useful.

5 minutes ago, Spricigo said:

 

Very true - had never seen this particular thread, having worked out a theoretical profile from first principles and being pleasantly surprised when it actually worked in real KSP.  Great to know from well done Kerbal Science that the practical minimum dV isn't too far off - I couldn't be bothered to find further efficiency because I reckoned the fast launch criteria would limit any gains. Not a fan of long ascents.

I only went down this particular rabbit hole because I also dislike docking and wobbly docked-together-structures/having to EVA strut & end up sending very, VERY heavy payloads to orbit later in career mode. At this extreme, a few hundred dV savings to ascent simplifies rocket design: weight, cost and most importantly, part number! Not so much an issue in 64-bit KSP. Not yet anyway .

[The Flying Kerbal]

Some super posts here guys, though I'll have to wait until tonight to study them in depth.  Unfortunately my boss doesn't do KSP...