KSP Launches vs IRL Launches

[Geschosskopf]

So if I was to use FAR and aerodynamic rockets I should follow a more real life launch profile?

I use FAR and find the above not only advantageous but essential to getting off the ground at all. But FAR is something of a mixed bag when it comes to realism IMHO. While it does require you to make realistic-looking rockets (at least much closer to real life than the pancake asparagus nonsense you can get away with in vanilla), it goes too far the other way IMHO. Still, all in all I much prefer FAR to vanilla so recommend you try it.

Some observations from my experience with FAR and ROCKETS (so don't talk to me about spaceplanes ). Far is beneficial in some ways for rockets although it also imposes numerous penalties. But to get the most out of it, you really need tweak your ascent profile to something more realistic. This is mentioned in the docs. You want make your ascent in a long, gradual curve starting at pretty low altitude. After much testing with MechJeb's ascent profile tweaker, what works best for me is setting the turn to start about 6km, use the 70% curve, and finish the turn at 80km. MechJeb will either fly this itself on autopilot or will project aiming markers on the navball is you want to fly it yourself. The latter is pretty much essential if you really want to get the most out of the whole exercise.

Anyway, doing all this means I routinely get to 100-150km LKOs with only about 3200 dV; my record is 2800 dV. Thus, for payloads of up to about 60 tons, my ascent stage is usually nothing but a stack of 2 of the biggest fuel tanks with 1 of the biggest engine underneath, with MAYBE a couple of small-medium SRBs. No asparagus, no crossfeed. HOWEVER, you pay for this in other ways, which I shall now list:

First off, a requirement of FAR is that the CoL be as close to the bottom of the rocket as possible. You need rockets that look like arrows, long and skinny with lots of tail feathers. Sticking little winglets on won't cut it. I need 4 stock delta wings with standard control surfaces to do this, and any SRBs need at least winglets, sometimes full delta wings, too, depending on payload shape.

A stack of 2 of the biggest fuel tanks leaves little room on top for payload before you hit the VAB ceiling. This significantly limits the size of your payload, as does the need to put it in a fairing if it's lumpy. This all means that you'll have to dock stuff together in orbit to get anything big in space. You'll also need a mod with fairings, or figure out how to get Fairing Factory to work with 0.20 AND the rest of your rocket.

While all the tail feathers are usually enough for adequate control in the lower atmosphere, you REQUIRE copious RCS to turn these long rockets above about 40-50km. Like to face prograde to circularize the orbit. A good rule of thumb I've come up with is that for each big fuel tank in the ascent stage stack, the rocket as a whole needs 4 RCS clusters on each end to get facing the right way before it passes apoapsis, and of course the fuel to run them through at least 1/2 the ascent. Unless you're just going to Mun, your payload will probably be docking, so will of course have 8 RCS thrusters of its own and a mono tank. That takes care of the top end so you need 8 more RCS clusters on the bottom (I usually mount them on the tips of the delta wings) and a BIG RCS tank in the ascent stage. On the way up, the payload RCS tank drains first. Thus, it's good practice to build your ascent stages so you can circularize the orbit with a tad of dV to spare. Then you can pump whatever RCS fuel is left out of the ascent stage into the payload before ditching the ascent stage, so the payload will be able to dock. Forgetting to do this is very annoying .

Because the ascent is a long, constant curve, you're never holding still. This complicates dropping SRBs. To avoid them smashing into you, you need to use a lot of Seprotrons, all carefully placed to throw the boosters far away without tumbling, which can be tricky if the SRBs have wings on them.

You'll probably have to get (I THINK it is) TT's latest Modular Multiwheels mod, the one with the "heavy strut" part. This strut can stretch a VERY long way. You often need this to go from the top of your payload through the bottom cup of the fairing onto the big fuel tank below, to keep the payload from wobbling through the fairing. This is only really a problem if you're launching long, lumpy payloads that have docking ports on both ends.

Anyway, I like FAR (at least for rockets--it infuriates me with spaceplanes). Just say good-bye to launching huge rovers and T-shaped station components with pancake rockets.

[DerekL1963]

Atlas and Titan, as used in Mercury and Gemini, didn't use a roll program--I'm not certain they even had roll control--and thus the launch sites had to be "aimed" in the direction they were to be launched--there was a small crossrange capability, but not much, maybe a couple of degrees. This was because they were done to quickly get reliable ballistic missiles into service; later ICBMs had the same roll control as purpose-designed space launchers, because it meant that the orientation of the launch site wouldn't restrict targeting or tell the enemy what the missile's target was.

Yes, the Atlas birds of that era had roll control - that's part of the function of the little bitty vernier motors located on the side of the sustainer. It's also part of why Titan had two engines on it's first stage.

Pretty much all liquid birds have/had roll control (even going back to the A4/V2) because roll a) makes your guidance problem much harder (because your pitch and yaw commands now have to be split into two components and you risk tumbling the platform), and roll makes the propellant move to the walls of the vehicle which both does funny things to your center of mass and inertial moment, and tends to starve the engines of fuel because the outlets are generally located at the bottom center of the tank. Reason a) applies to solid fuel rockets as well... Polaris and Poseidon both had positive roll control throughout all phases of flight up to RV deployment. (The Trident -I and -II guidance electronics OTOH have sufficiently accurate platform stabilization electronics and sufficiently fast gimbal actuators that they don't need roll control until post boost flight.)

[asmi]

So if I was to use FAR and aerodynamic rockets I should follow a more real life launch profile?

You will have to. Turning more than 5-10 degrees away from velocity vector is a recipe for guaranteed fireworks with FAR.

[asmi]

<skipped>

I use MJ's Smart ASS Surface mode to steer rockets during ascent and even before lift off I put 88 pitch angle, so I'm starting gravity turn right at the liftoff. Then SLOWLY decrease pitch as your velocity vector goes lower to avoid it getting farther than 5-7 degrees away from your orientation.

Yes, FAR clearly prefers tall rockets with vertical staging and so having EditorTools installed is practically a must (it expands VAB/SPH so you won't have to worry about hitting ceiling). The real problem is thrust - you have to do few test launches to depermine optimal thrust profile. In ideal case once initial pitch over is completel, you shall just direct MJ to follow prograde vector and use only thrust to control your profile (more thrust = rocket turns slower, less thrust - faster). This way you won't need any RCS at all for the ascent.

And after playing with FAR I simply can't play stock without banging my head and shouting "This is impossible!!! This can not be!!!" But I agree, it will require modest time to re-learn how to build and fly vessels.

[AndreyATGB]

You will have to. Turning more than 5-10 degrees away from velocity vector is a recipe for guaranteed fireworks with FAR.

I would like to use FAR but it hates ASAS as it is right now. Smart ASS from MechJeb is useless with FAR, it'll just wobble until it inevitably stalls or breaks apart. I might use it in .21, it should cope a lot better with FAR.

[Geschosskopf]

@ Asmi:

Thanks for the tip on Editor Tools. Never heard of that before, sounds like it will solve my main problem with FAR--hitting the ceiling .

How much delta-V does your ascent stage / profile need to reach LKO? And at what altitude are you going pure prograde? I find that if I'm already prograde while still in the atmosphere, I never make orbit, which is why I wait to about 80km. Seems to take more than the 3200 dV I usually pack if I go prograde lower. But turning prograde between 70-80km means all my tail feathers are useless so I have to have RCS to finish the turn, even if it only needs another 10-20^ of pitch to do it, before passing the apoapsis.

And if the payload is really long and wobbly, I find RCS all the way up sometimes unavoidable. If the only control forces are coming from the tail feathers, this increases the wobble at the payload end, sometimes to fatal levels, but if the payload is burning RCS, the rocket wobbles a lot less.

@ AndreyATGB:

The FAR docs say not to use SAS/ASAS during launch and I agree. Really, you don't need it if you built your rocket properly so the Center of Lift is way down at the bottom end (again, per the FAR docs). With the CoL way down there, which is because you put some big tail feathers on the rocket, it will fly straight as the arrow it looks like.

[Sof]

How much delta-V does your ascent stage / profile need to reach LKO? And at what altitude are you going pure prograde? I find that if I'm already prograde while still in the atmosphere, I never make orbit, which is why I wait to about 80km. Seems to take more than the 3200 dV I usually pack if I go prograde lower. But turning prograde between 70-80km means all my tail feathers are useless so I have to have RCS to finish the turn, even if it only needs another 10-20^ of pitch to do it, before passing the apoapsis.

And if the payload is really long and wobbly, I find RCS all the way up sometimes unavoidable. If the only control forces are coming from the tail feathers, this increases the wobble at the payload end, sometimes to fatal levels, but if the payload is burning RCS, the rocket wobbles a lot less.

Prograde is the direction you are currently travelling. I assume you mean the direction of prograde orbit (the 90 degree line border between blue and orange), which you shouldn't be doing in atmosphere, because that means you aren't going up at all. If your prograde marker isn't pointing horizontal, and your apoapsis is above the atmosphere then you can burn horizontal to make an orbit faster.

As for your RCS woes, try using a rocket with high gimbal for the upper stages. If your payload is too wobbly, then you haven't used enough struts.

[Whackjob]

And if the payload is really long and wobbly, I find RCS all the way up sometimes unavoidable.

Take inspiration from insects. Specifically, add an external series of girders and struts that can be blown free when no longer required. Just remember to use girders end-to-end and to avoid direct connections with their radial side points.

[Geschosskopf]

Prograde is the direction you are currently travelling. I assume you mean the direction of prograde orbit (the 90 degree line border between blue and orange), which you shouldn't be doing in atmosphere, because that means you aren't going up at all. If your prograde marker isn't pointing horizontal, and your apoapsis is above the atmosphere then you can burn horizontal to make an orbit faster.

Yeah, my bad. I was nearing the end of a long shift at work when I wrote that, so had almost sobered up. But now I'm home with a couple drinks in me and realize the error of my ways, terminology-wise . But anyway, I typically come out of the top of the Kerbin atmosphere with about 10-20^ of pitch between me and being tangent to the high point of my suborbital ballistic parabola. At this point, the tail feathers that kept me flying true in the air are just an ankle weight and without RCS I can't line up for the burn to achieve orbit.

As for your RCS woes, try using a rocket with high gimbal for the upper stages. If your payload is too wobbly, then you haven't used enough struts.

Well, those FusTek station masts with the 16 large-size radial docking ports are ornery critters and put up a real fight, especially if they're sitting on a docking point that's sitting on the KW Rocketry 3.75m fairing cup stack separator. They still wobble worryingly even with 24 of TT's Heavy Struts tying them down. Can't fit any more than that--they keep running into other bits of the mast module instead of reaching the fuel tank below the fairing separator. But with 24 I can do it, although it needs RCS stabilization all the way from about 8km to lining up for the apoapsis burn.

[Whackjob]

KSP launches versus IRL Launches?

Why can't we have both?

[asmi]

@ Asmi:

Thanks for the tip on Editor Tools. Never heard of that before, sounds like it will solve my main problem with FAR--hitting the ceiling .

You're welcome! In addition to increasing VAB/SPH (which alone makes it worth trying IMHO ), there are few other features which I'm sure you'll like I'm not going to spoil them - check it out yourself - make sure you read readme!

How much delta-V does your ascent stage / profile need to reach LKO? And at what altitude are you going pure prograde? I find that if I'm already prograde while still in the atmosphere, I never make orbit, which is why I wait to about 80km. Seems to take more than the 3200 dV I usually pack if I go prograde lower. But turning prograde between 70-80km means all my tail feathers are useless so I have to have RCS to finish the turn, even if it only needs another 10-20^ of pitch to do it, before passing the apoapsis.

I usually budget 3.5 km/s for ascent, but this includes healthy margin in case I screw something up, which I do sometimes As far as optimization goes, my personal record is around 2900 m/s, but achieving this was hard and required several retries. Since we aren't exactly short on fuel (I play with Mission Controller mod which makes me pay for everything I use, and still fuel is pretty cheap), I pack more just in case.

And if the payload is really long and wobbly, I find RCS all the way up sometimes unavoidable. If the only control forces are coming from the tail feathers, this increases the wobble at the payload end, sometimes to fatal levels, but if the payload is burning RCS, the rocket wobbles a lot less.

The trick is to control throttle in way that ensures you get to apoapsis with orbital velocity. In this case gravity does turning for you, so you just have to let it do so (ASAS/MJ off).

@ AndreyATGB:

The FAR docs say not to use SAS/ASAS during launch and I agree. Really, you don't need it if you built your rocket properly so the Center of Lift is way down at the bottom end (again, per the FAR docs). With the CoL way down there, which is because you put some big tail feathers on the rocket, it will fly straight as the arrow it looks like.

Yea ASAS it generally useless (if not worse) during ascent as usually the way it works (assuming rocket is aerodynamic) is you either on a correct trajectory and so aerodynamic forces will keep it there without any input from you, or you're in a wrong attitude and so aerodynamic forces WILL overpower whatever RCS/SAS/ASAS you've got anyways and the mission is a failure regardless of what you do.

[Lexif]

And at what altitude are you going pure prograde? I find that if I'm already prograde while still in the atmosphere, I never make orbit, which is why I wait to about 80km. Seems to take more than the 3200 dV I usually pack if I go prograde lower. But turning prograde between 70-80km means all my tail feathers are useless so I have to have RCS to finish the turn, even if it only needs another 10-20^ of pitch to do it, before passing the apoapsis.

I only have about a dozen launches with FAR under my belt, so take this with a grain of salt... But I find a good way to get to your orbit in one burn is to get a tool like Kerbal Engineer or Mechjeb that can show your time to AP. This works also without FAR.

When the apoapsis rises above 40 km, and when I'm out of the thickest of the atmosphere -- maybe above 25 to 35 km -- I throttle to keep my time to apoapsis from rising. Starting from less than one minute, I try to throttle so that the time slowly decreases. I generally aim to keep it at maybe 20--30 s at about 50 km, going down from there. (Assuming you're aiming for a low orbit like 75 km.) That way, I don't increase the apoapsis to much by burning to far away from apoapsis. You got to get a feel for it, it's kind of hard to describe.

I'm aiming at orbit prograde throughout the whole procedure. In my opinion, burning off the prograge vector is just wasted energy once atmospheric drag doesn't matter anymore. That way, I generally don't have to do any special orbital injection burns anymore as I never shut the engines down. Also, you can profit from thrust vectoring for the whole ascent, reducing the RCS fuel needed.

I found one problem in my early FAR rocket launch attempts was that I didn't limit the travelling range of the control surfaces. That made them stall at about 10 km height, when the larger control deflections needed because of the thinner air made them stall in the transsonic range (around mach 1).

Edit: For the first part of the launch with FAR, I generally point slightly below prograde, at most at the edge of the surface prograde marker. That means that my rocket points about 20 degrees below vertical at maybe 15 km, if I remember right. At about that altitude, I switch to the orbital velocity marker, as I found that the orbital and surface velocity get pretty close to each other around there. That's what I center on above that point.

Edited July 6, 2013 by Lexif

[Van Disaster]

One thing about FAR is you need much less TWR - and in fact too much thrust with FAR can tumble your rockets - and if they don't tumble then the shearing force can make things very wobbly - if you start turning in the thicker part of the atmosphere. You can launch ok with say, 1.4 TWR.