Help understanding fuel efficient launches in relation to atmosphere density / terminal velocity

[Atheistic_Logic]

So way back when I started playing ksp I remember a wiki article on what your max velocity should be at various kerbin altitudes. It went something like 110m/s for 1000m, 120m/s for 2000m, 160m/s for 5000m, 260m/s for 10000.

I remember testing full burns vs controlled burns, and noticed that it is more efficient to follow this chart.

Thing is, I now restarted the game, and am experimenting with the basic solid boosters. I am finding that it is less efficient to lower the thrust limiter on the bacc thumper booster.

At 100% thrust, it gets me a 42 second burn, to an altitude of about 296 000m. If I do a 75% thrust limit burn, I get ~56 second burn, altitude of 267 000m.

Can anyone explain why the 75% doesn't get me higher? I'm sure it's something obvious I'm overlooking. The pattern above is also true the more you lower the thrust ( I tried it at 50% thrust as well)

[moogoob]

Different atmospheric model compared to the souposphere before 1.0. We can now break the sound barrier at low altitudes!

[Gaarst]

With very high TWR, the air resistance is stronger in the lower parts of the atmosphere, but you spend less time there. You get high very quickly and do most of your burn in higher atmosphere where air resistance is weaker.

A TWR a bit lower might be enough to make you spend more time in low atmosphere, and therefore lose more speed there, and in the end not go as far as you did before.

Accounting for air resistance makes predicting optimal TWR/speed during ascent extremely difficult: to get an accurate result, you'd have to solve multiple-variable differential equations to several degrees. And the result you'll get will depend on basically everything: rocket, ascent profile, speed, heading, position of launch, altitude, mass...

The best thing to do is follow the result generally accepted by the community as one of the best which is to launch with a TWR of 1.5, and break the sound barrier a bit below 10km (approx 300m/s at this altitude). But then again, this will depend on the design of your rocket: if your first stage is huge, you might end up having a TWR of 10 at 10km, and on the other side, if it's small compared to the rocket, you might never get over 3 with full throttle.

Edited October 10, 2015 by Gaarst

[StevenRS11]

Getting a totally optimized ascent profile is very hard, but getting a good one isn't so bad. You really have two things to balance- air resistance and gravity losses. Obviously, you loose energy to the air whenever you are in it depending on how fast you go. You loose energy to gravity whenever your acceleration vector is not perpendicular to gravity, aka non sideways.

In a airless environment, you want to accelerate at the maximum possible rate at the maximum possible horizontal angle that keeps you from impacting terrain. So from say minmus, you pretty much thrust a tiny bit up and then just go sideways. You want to minimize the time in which you accelerating against gravity. If you want, its fun to hyperedit launches from Tylo and practice a bit there. With only a single variable, its really easy to optimize your 'gravity turn'. Put a single mainsail under a single orange tank and get the maximum velocity you can before you reach your apoapsis. That's pretty close to the optimal turn for that TWR ratio on Tylo. It gives you a good intuitive feel for what's 'right'. If you want to be even more precise about it, build a rocket with 3,100 m/s dv and escape tylo. That allows for only ~30 m/s of gravity losses, or just 4 short seconds of thrusting directly against gravity.

Accounting for air resistance isn't that hard if you guess and check. At various heights during your launch, quicksave, throttle all the way down, and look at your accelerometer. Any acceleration you see is caused by air resistance, and if its greater than the fraction of the gravity vector that's opposing your thrusting vector, its probably much. Atmospheric drag should ideally be 1/5 to 1/10th of the gravity drag, with smaller rockets having more atmospheric drag relative to gravity drag.

Also don't forget that any signifigant angle of attack drastically increases your atmospheric drag. I generally try to use fixed winglets on my rockets and use very limited thrust vectoring to start my gravity turn. My best launches are when I don't have to do anything but stage after that. The less steering the better, especially in denser atmosphere.

Edited October 10, 2015 by StevenRS11

[cantab]

Yeah, now we have a realistic atmosphere drag losses aren't a concern. Ascent speed tends to instead be limited by aerodynamic instability. Many rockets are mildly unstable aerodynamically and if you fly too fast too low you'll flip out of control, slowing down reduces the aerodynamic forces. Many real rockets including the Space Shuttle throttle down to reduce "Max Q", as the maximum aerodynamic stress is known. Practical launch TWR is also limited, in real life and KSP alike, because engines are heavy and cost money.

A wide range of liftoff TWRs can work. Looking at real rockets the Saturn V had 1.15 TWR on the pad, while the Space Shuttle had rather higher. New Horizons was blasted off with 2.15 liftoff TWR, though that was an extreme case.

[StevenRS11]

Yea, it's really easy to get massive TWRs in KSP. Unless it's something like a sounding rocket, when you have a TWR over 3 things start to get a bit silly. That said, I'll sometimes add a few short burning SRBs for my rockets that have a TWR under 1.25 so they get a bit of aerodynamic stability going sooner.

Anyway, I usually play with FAR, and lately RSS. The difference is unbelievable.

[Vanamonde]

This question about how to play the game has been moved to Gameplay Questions.

[MarvinKitFox]

Air drag losses are of much less import than gravity loss, under the current physics model.

The sole exception would be that you should avoid riding on/very near the speed of sound, as this causes significant additional drag. (and other funky control issues).

If your rocket is capable of smashing past mach 1, then do so.

If it gets stuck at mach 0.99 for an extended time, you might as well drop the throttle to 50% and idle at mach.90 for 20 seconds, then throttle up and go supersonic.