Why to orbit?

[Wanderfound]

VOID and Kerbal Engineer Redux also have DV readouts. VOID also has the nice HUD displays. I'd recommend using a non-autopilot information mod before getting into MechJeb. Learn first, MJ second.

OTOH, ain't nothing says that you have to use the autopilot features of Mechjeb. I use it almost purely for the customisable flight data displays; the only time the autopilot comes into play is if I'm launching identical rockets a dozen times in a row for some reason. Manoeuvre Planner also comes in handy; I know that I can get the same result manually, but why would I spend five minutes faffing about with manoeuvre nodes when I don't have to?

But, yes, learn first. If you start relying on autopilots too early you'll never learn how to fly manually. Do it yourself until you know that you can do it, then automate the bits that you find tedious.

[Renegrade]

OTOH, ain't nothing says that you have to use the autopilot features of Mechjeb.

This is true, but my main point was the second bit here:

But, yes, learn first. If you start relying on autopilots too early you'll never learn how to fly manually.

I actually used to calculate delta-v and TWRs and such by hand (well calculator and notepad), and it actually gave me an appreciation for certain things (ex, dry mass matters quite a bit, would have taken me a while to really learn that otherwise). I let KER take care of that in the VAB now (have you checked out the 1.0 version? It's quite nice), and VOID in space (a recent bug was fixed in VOID, it's usable again with Engineering calculations enabled! yay!)

I still tend to fly entirely by hand (although I did putter around with MJ's enhanced RCS features, since stock's RCS control system needs an overhaul), although there are times when doing my usual 7-rocket launch for BTSM planetary exploration that I kinda wish I wasn't so anti-autopilot hehe.

[Pds314]

You're not wasting fuel by achieving orbit first. Once you're in orbit you're no longer fighting gravity.

You can test it yourself very easily. Build the biggest rocket you can...... point it directly at the mun, and see how far your rocket gets before it runs out of fuel and loses the battle with gravity.

Actually, that's not quite correct, once the atmosphere is relatively thin, mathematics says you are best to accelerate so as to put yourself on a transfer to the Mun or Minmus or other targets immedietely while the Oberth effect is working in your favor. Don't circularize at 100 km, or even 70 km first.

But there is a disadvantage here. You have little control over what where you go, so you must plan your launch time ever so precisely that it is rarely worth it to burn directly for whatever target.

Remember though, don't do this until your terminal velocity is higher than your current velocity.

[Empiro]

Actually, drag tends to be proportional to velocity squared--or was it cubed? Well, in either case, going faster in the air is detrimental. A straight up launch does have an advantage though that it goes through less atmosphere.. in FAR, however, this benefit is kinda minimal as the overall losses from air are small (just like in real life; Nathan told me that Apollo is something like ahh.. I think it was roughly 2000-ish gravity losses and 100-ish air losses)

I'm very very certain that it's because I spent more time in the gravity field. On a direct ascent, I can just absolutely burn the crap out of my ascent, minimizing the time I spend having my velocity stolen. Every second you spend in the gravity well is 9.82 m/sec that's lost (well, that number declines as you climb, but the thought holds). A thought experiment to illustrate this is to ask oneself: How much fuel does it take for a rocket with a maximum TWR of 1.0 or less to reach orbit?

One other thing is that I waste NO energy circularizing (the final course looked like two parallel lines extending out from KSC just before it changed to an escape course) -- raising the PE doesn't help in any way for this sort of course, and DOES cost energy. Flying horizontally in the atmosphere during a climb is basically a cleverly disguised circularization.

I just want to point out that this isn't correct -- gravity loss comes when you thrust against gravity. If you thrust perpendicular to gravity like when you are in a circular orbit, there is no incurred gravity loss.

You said earlier that your design has an absurdly high TWR of 4 to 6+. This does reduce gravity losses by quite a bit, but then I believe that in this case you're optimizing for the wrong thing. If you made a smaller rocket with a more modest TWR, you'd still be able to reach Duna. Even though it would cost more Delta-V, it would cost less resources overall (in terms of quantity of fuel and engines) because your rocket is so much smaller.

[Renegrade]

I just want to point out that this isn't correct -- gravity loss comes when you thrust against gravity. If you thrust perpendicular to gravity like when you are in a circular orbit, there is no incurred gravity loss.

No, it's still affecting you, otherwise you wouldn't need to thrust at all. The 'perpendicular to gravity' thing only works when you're assuming you're going directly to orbit. I'm pretty much skipping that here. My final orbit just before it goes hyperbolic is a narrow arch.

Also when you're pushing up the PE in orbit, that is fighting gravity, regardless of where the PE is. Plus if you're heading to orbit, you're plowing through a lot of extra atmosphere. Really, really thick atmosphere in stock (less so in FAR but still present).

Note that I used the same rocket to do an actual orbital approach to a Duna transfer, and it significantly underperformed. It even beat out a paper rocket based on a near-perfect orbital approach.

You said earlier that your design has an absurdly high TWR of 4 to 6+. This does reduce gravity losses by quite a bit, but then I believe that in this case you're optimizing for the wrong thing. If you made a smaller rocket with a more modest TWR, you'd still be able to reach Duna. Even though it would cost more Delta-V, it would cost less resources overall (in terms of quantity of fuel and engines) because your rocket is so much smaller.

Yes, but the discussion here was about saving delta-v, not cost or anything. A smaller rocket (although it was pretty small already--also only 3.6 twr on the launchpad) on a less efficient course will indeed save a bit, but that doesn't change the fact that the course is less efficient.

Now, just to be clear, I'm not saying we should all abandon orbiting and classical injection burns (especially since all the tools are oriented towards the classical orbital approach, and it's so much safer and easier) and all start doing missile launches instead, but rather understand the potential tradeoffs (y'never know, the tradeoffs could flip around at some point and a fusion/plasma rocket may very well be capable of putting out a highly efficient, 30 TWR..).

[Empiro]

No, it's still affecting you, otherwise you wouldn't need to thrust at all. The 'perpendicular to gravity' thing only works when you're assuming you're going directly to orbit. I'm pretty much skipping that here. My final orbit just before it goes hyperbolic is a narrow arch.

Also when you're pushing up the PE in orbit, that is fighting gravity, regardless of where the PE is. Plus if you're heading to orbit, you're plowing through a lot of extra atmosphere. Really, really thick atmosphere in stock (less so in FAR but still present).

Note that I used the same rocket to do an actual orbital approach to a Duna transfer, and it significantly underperformed. It even beat out a paper rocket based on a near-perfect orbital approach.

You're always under the influence of gravity, but that's not gravity drag is.

Gravity drag is the lower effective DV you get from burning against gravity. For example, if your TWR is 4, if you burn against gravity for 10 seconds, your effective acceleration is 3G, and your resulting speed is about 300 m/s. If you burn perpendicular to gravity, your resulting speed is about 400 m/s.

The extra cost of raising the PE is mostly due to the reduced Oberth effect. When you circularize your orbit, there's a time between cutting your engines when your AP reaches 70+km and when you circularize. During that time, you slow down a lot. When you perform a direct ascent, you burn all your fuel at once without cutting your engines.

Your approach definitely has lower loss due to atmospheric drag. Your rocket is pretty small, so atmospheric drag makes a big difference, but for bigger rockets, drag has relatively less of an effect.

Basically, direct ascent has the following advantages:

-Lower atmospheric drag loss

-Better use of Oberth effect because you burn earlier and when travelling faster

The one big disadvantage is gravity drag loss.

For very small rockets with high TWR, the advantages probably outweigh the loss from gravity drag.

Yes, but the discussion here was about saving delta-v, not cost or anything. A smaller rocket (although it was pretty small already--also only 3.6 twr on the launchpad) on a less efficient course will indeed save a bit, but that doesn't change the fact that the course is less efficient.

Now, just to be clear, I'm not saying we should all abandon orbiting and classical injection burns (especially since all the tools are oriented towards the classical orbital approach, and it's so much safer and easier) and all start doing missile launches instead, but rather understand the potential tradeoffs (y'never know, the tradeoffs could flip around at some point and a fusion/plasma rocket may very well be capable of putting out a highly efficient, 30 TWR..).

Sure, but optimizing purely for DV is the wrong approach to take when rocket designs and flight plans go hand in hand. Maybe with future technology things will be different, but in KSP today, getting into orbit first and then going to your destination is usually a good idea.

[Renegrade]

Gravity drag is the lower effective DV you get from burning against gravity. For example, if your TWR is 4, if you burn against gravity for 10 seconds, your effective acceleration is 3G, and your resulting speed is about 300 m/s. If you burn perpendicular to gravity, your resulting speed is about 400 m/s.

...

Basically, direct ascent has the following advantages:

-Lower atmospheric drag loss

-Better use of Oberth effect because you burn earlier and when travelling faster

Hmm, I see what you're saying here. And I'm starting to think it would be fairly simple to calculate (err ignoring atmospheric effects, those are never simple outside of stock) where a missile launch crosses over in (course) efficiency over a traditional approach (5thHorseman's test rocket at 1.6 TWR had the reverse results, so that's obviously on the 'classic' side of the crossover).

Calculating for the atmospheric effects with non-stock air would be problematic but still doable-ish.

Sure, but optimizing purely for DV is the wrong approach to take when rocket designs and flight plans go hand in hand. Maybe with future technology things will be different, but in KSP today, getting into orbit first and then going to your destination is usually a good idea.

I'll agree with that. Often practicality trumps theoretical things. Being able to set up a stable orbit and calmly assess maneuver nodes, ship readiness, perform last-minute resupply etc, is definitely worth a few hundred dv, especially when none of the tools support the missile launch approach and some Funds savings are to be had.

[Pi_]

off topic...

It's the same idea as the "burn at munrise" trick that was used before maneuver nodes existed

The Munrise burn is still the only method I use for Mun missions. It just seems, idk... more elegant.

[toadicus]

VOID and Kerbal Engineer Redux also have DV readouts. VOID also has the nice HUD displays. I'd recommend using a non-autopilot information mod before getting into MechJeb. Learn first, MJ second.

There's a map at the Kerbal wiki, in the Cheet Sheet page.

IT's a nice map for starting with, as it has a safety margin built into the numbers.

Here's an example trip, using the Cheet Sheet's map's numbers:

1. Launch from Kerbin, get to "Low Orbit" (80km altitude). That's the blue line at the bottom, 4550 dv.

2. Burn 860 m/sec prograde at the right time (maneuver nodes FTW!) and you'll be going to a Mun Intercept (grey line, going up and left from Low Orbit)

3. Brake into Mun orbit (it says 210 on that line)

So far that's 4550+860+210 using a plain, unmodded KSP. That totals 5,620 dv. So if your rocket has that, it SHOULD be able to reach Mun orbit.

If you want to return, you would have to burn back across the "210", costing 210 again. However, the little white arrow tells you that the next step can use aerobraking (ie parachuting into the atmosphere), so you can ignore the 860 and 4550.

If you instead wanted to land on the mun and return to Kerbin, you would have to go across the 640 line to land, and back again to get into Mun orbit, and then across the 210 to return .. that's ummm.. 7,100 total on the launchpad.

Technically speaking, that "210 m/s" is incorrect for an optimum return from Mun. It's only correct that you need 210 m/s to escape Mun's influence; you need a bit of additional delta-V to push your periapsis back down to Kerbin's atmosphere. The key to minimizing a Munar return is in making sure you burn in the correct place. You want to place your escape burn from the Mun such that your orbital velocity is going to naturally put you into a Kerbin orbit with a lower periapsis. For an equatorial Munar orbit, that location is typically about 10-30° counterclockwise from the Mun's "prograde" orbital vector, IIRC. The cheat sheet doesn't really have a good way to estimate that, but the very cheapest I've ever made that burn is something like 232 m/s.

Edit: The more I think about it, the less sure I am about that number; it might be more like 320 m/s? It's been too long.

Edited August 18, 2014 by toadicus