Should you orbit Kerbin first on an interplanetary journey?

[Ninety-Three]

Just recently, I realized that I'd been doing my interplanetary flights the same way I do my Munar ones: Fly up, establish a 100 km orbit, burn at PE until your AP is where you want it (out of the SOI, or intersecting the Mun, depending on where I'm going). It occurs to me that if I'm going to say, Duna, it seems more efficient to just burn straight up until my trajectory is leaving Kerbin's SOI, without bothering to orbit first. However, I don't actually understand any of the principles behind efficient rocketry (I've sort of just been cribbing from youtube videos and memorizing the practices), so I thought I'd ask the forums: Is that the right way to go about it?

[Kosmo-not]

It depends on how much thrust to weight your rocket can put out. Since we're probably talking realistically here, it is generally more efficient to insert into low Kerbin orbit first. An exception would be if you time your launch in such a way that you will be accomplishing your interplanetary burn before where you would normally be doing your first engine cutoff for inserting into low orbit. The reason for this is that we want to do all of our burn as low as possible to minimize gravity losses.

If you travel from 40km in altitude to 70km, a specific amount of energy is taken away from your kinetic energy. While that is constant and doesn't change, what we can change is how much velocity is lost through that altitude transition. If you double your velocity, you increase your kinetic energy by a factor of 4. If you think about it a little, you will realize that less velocity is lost when you are traveling faster and subtracting the set amount of energy loss from altitude change than if you were traveling slower.

Going straight up with a rocket, you pass through all that altitude while not travelling as fast as you would be by getting some horizontal speed going. Depending on how tight of a fuel budget you have, this fact might be the difference between coming home or being stranded in space.

[Daze]

It's cheaper to do the Hoffman in One burn only if you have calculated the phase angle of the planets...

[paulthebob]

I only have a very roundabout view and I'm sure someone in here will give a more indepth answer, but think of it this way;

When you want to get distance on a thrown ball, you throw it at 45 degrees because if you throw it 90 degrees (straight up) to the gravity, then it has to counteract all of the gravity, whereas if thrown at an angled trajectory, it can gather more momentum and thus travel farther (gravity has less of an effect).

For this reason, you ideally want to exit the planet at an angle to it. Ideally, you want to begin 'gravity burning' at around 10km height, so that the crosshair on the navball is just at the tip of the prograde vector, and constantly, with altitude, keep it like that until the crosshair reaches the 90o mark on the navball.

Err, yea, I think that's about it. Either way, straight up trajectory out of the planet is very inefficient.

[Kosmo-not]

It's cheaper to do the Hoffman in One burn only if you have calculated the phase angle of the planets...

Hohmann. Who started that "Hoffman" thing anyway?

[Alistone]

It's a tradeoff.

By burning straight up into your ejection from Kerbin you can benefit from the hohmann effect of doing your accelerations low in the gravity well; BUT you will be burning straight up which means your effective thrust will be reduced by the force of gravity.

The higher your T/W ratio the less the effective of "gravity drag" will be. Thus it can be better.

But if you have a lower T/W ratio then you are better off establishing a low kerbin orbit and burning out from there. Ejection burns from orbit tend to be horizontal and do not suffer "gravity drag". Also, the path tends to keep you in low kerbin orbit for longer (curved arc of a circle) which gives low T/W ratio rockets a greater hohmann effect than a direct vertical burn (short straight line).

Considering the relatively low power (and corresponding low T/W ratio) for the fuel efficient rockets generally used for interplanetary travel; and considering the better control you have of using a maneuver node to plan your burn from low kerbin orbit, the vast majority of people find it easier and "better" to establish orbit first.

Edited April 29, 2013 by Alistone
Spelling errors

[Ninety-Three]

Thank you, Alistone, putting it in terms of gravity drag vs realistic TWRs helped me understand it. I did a test with this... slightly overpowered launcher:

I found that the "Straight up" method left me with slightly more fuel, but that was moving a 25 ton ship with four non-nuclear rockets. I can see how attempting that burn with a single interplanetary nuke engine would be a lot worse.

[foamyesque]

Depends on the rocket. In principle, with high TWRs and good aim, not bothering to stop in a parking orbit will be more efficient. However, a parking orbit will generally make your transfers a lot easier to aim-- and lets you use much lower TWR rocketry for your ejection burn.

[Kosmo-not]

It's a tradeoff.

By burning straight up into your ejection from Kerbin you can benefit from the hohmann effect of doing your accelerations low in the gravity well; BUT you will be burning straight up which means your effective thrust will be reduced by the force of gravity.

The higher your T/W ratio the less the effective of "gravity drag" will be. Thus it can be better.

But if you have a lower T/W ratio then you are better off establishing a low kerbin orbit and burning out from there. Ejection burns from orbit tend to be horizontal and do not suffer "gravity drag". Also, the path tends to keep you in low kerbin orbit for longer (curved arc of a circle) which gives low T/W ratio rockets a greater hohmann effect than a direct vertical burn (short straight line).

Considering the relatively low power (and corresponding low T/W ratio) for the fuel efficient rockets generally used for interplanetary travel; and considering the better control you have of using a maneuver node to plan your burn from low kerbin orbit, the vast majority of people find it easier and "better" to establish orbit first.

You're getting confused between a Hohmann transfer and the Oberth effect.

A Hohmann transfer is when a spaceship transfers from one orbit to another using only 2 burns.

The Oberth effect is a statement about rockets producing more power the faster they are travelling.

[Alistone]

You're getting confused between a Hohmann transfer and the Oberth effect.

A Hohmann transfer is when a spaceship transfers from one orbit to another using only 2 burns.

The Oberth effect is a statement about rockets producing more power the faster they are travelling.

You are correct... But I will defend my mistake by pointing out the efficiency of a Hohmann transfer comes from its use of the Oberth effect. In a Hohmann transfer the majority of the energy is exerted at the higher speed lower orbit... so I hope it's a forgiveable mistake.

... at least I didn't call if hoffman...

[Francesco]

Oberth, Hohmann, Hoffman (?)... we're messing with German surnames here

but yes, everything has already been explained.

to make it really simple: straight to Kerbin escape = more efficient, LKO insertion first = easier for you.

[Francesco]

You are correct... But I will defend my mistake by pointing out the efficiency of a Hohmann transfer comes from its use of the Oberth effect.

actually, there are some situations where a Hohmann transfer is less efficient than a bi-elliptical transfer.

so, no.

[Alistone]

actually, there are some situations where a Hohmann transfer is less efficient than a bi-elliptical transfer.

so, no.

And the bi-eliptical transfer essentially just using two hohmann transfers to get a greater oberth effect... so yes?

[Francesco]

And the bi-eliptical transfer essentially just using two hohmann transfers to get a greater oberth effect... so yes?

nope, Hohmann is 2 burns, bi-elliptic is 3

[Alistone]

nope, Hohmann is 2 burns, bi-elliptic is 3

... not sure if trolling...

So two hohmann transfers would have 4 burns...Burn from orbit A to orbit B, Burn at orbit B, Burn at orbit B to orbit C, Burn at orbit C.

And if you realize that the "burn at orbit B" and the "Burn at orbit B to orbit C" could occur back to back as essentially the same burn you would notice 3 burns.

The point is to be thrusting horizontally and preferably as much as possible as low as possible in the gravity well. Those primary attributes are shared whether you do it as a single hohmann transfer or a bi-eliptic transfer.

Edited April 30, 2013 by Alistone

[Endeavour]

... not sure if trolling...

So two hohmann transfers would have 4 burns...Burn from orbit A to orbit B, Burn at orbit B, Burn at orbit B to orbit C, Burn at orbit C.

And if you realize that the "burn at orbit B" and the "Burn at orbit B to orbit C" could occur back to back as essentially the same burn you would notice 3 burns.

The point is to be thrusting horizontally and preferably as much as possible as low as possible in the gravity well. Those primary attributes are shared whether you do it as a single hohmann transfer or a bi-eliptic transfer.

That's not how a bielliptic transfer works though unless I misunderstood your explanation. The goal is to go from one circular orbit to another, so you burn your apoapsis higher than the final orbit, then burn your periapsis up to the desired altitude while at apoapsis. Then you circularize for a total of 3 burns. This is only better when ratio the semi-major axis of final to starting circular orbit is greater than 12 I believe.

[Francesco]

... not sure if trolling...

So two hohmann transfers would have 4 burns...Burn from orbit A to orbit B, Burn at orbit B, Burn at orbit B to orbit C, Burn at orbit C.

And if you realize that the "burn at orbit B" and the "Burn at orbit B to orbit C" could occur back to back as essentially the same burn you would notice 3 burns.

The point is to be thrusting horizontally and preferably as much as possible as low as possible in the gravity well. Those primary attributes are shared whether you do it as a single hohmann transfer or a bi-eliptic transfer.

I don't always troll on the KSP forums...

...you seem to have some confusing ideas there: what's this "burn at orbit B" anyway? you burn at B to go to C, it's not like first you circularize and then you do another burn.

so it's only 3.

and you don't burn as low as possible: in fact, when you are at "B", you are as high as possible in the gravity well, since "B" is your Apoapsis.

[Alistone]

I don't always troll on the KSP forums...

...you seem to have some confusing ideas there: what's this "burn at orbit B" anyway? you burn at B to go to C, it's not like first you circularize and then you do another burn.

so it's only 3.

and you don't burn as low as possible: in fact, when you are at "B", you are as high as possible in the gravity well, since "B" is your Apoapsis.

If you choose to use a bi-eliptic to go from a 80km circular orbit to a 200km orbit and you just happen to chose your middle burn as 100km

Then your first burn at 80km periapsis raises your apoapsis to 100km.

Your 2nd burn at 100km apoapsis raises the periapsis up to 200km.

During that burn, as the periapsis crosses 100km it becomes the apoapsis AND you have just circularized your orbit as you continue to burn.

Finally at your 200km apoapsis you burn to circularize in a 200km orbit.

I fully understand that to be more efficient it will usually go the other direction; you burn from 80km to lower your periapsis to 70km, then you burn at periapsis lower in the gravity well to push your apoapsis out from 80km to 200km. In this latter example the orbit is never circularized in the middle. So it's not exactly 2 hohmann transfers. But it's really not that different since you can choose to do a bi-eliptic transfer at whatever orbit you like (but efficiency varies.) I know the middle step can vary, but I'm not saying they are identical. A bi-eliptic transfer versus 2 hohmann transfers aren't really that different either.

Edited April 30, 2013 by Alistone

[drakesdoom]

I bi-elliptic transfer by definition uses a mid burn point well above your target orbit. You start from a low circular, burn for a high apoapsis, burn to bring periapsis up to target, and finally burn retrograde to circularize.

https://en.wikipedia.org/wiki/Bi-elliptic_transfer

You don't choose to have your mid burn in the middle, that would be a stupid transfer not a bi-elliptic.

[Nao]

Returning a little to the topic from the bi-hoffman dillema...

I only have a very roundabout view and I'm sure someone in here will give a more indepth answer, but think of it this way;

When you want to get distance on a thrown ball, you throw it at 45 degrees because if you throw it 90 degrees (straight up) to the gravity, then it has to counteract all of the gravity, whereas if thrown at an angled trajectory, it can gather more momentum and thus travel farther (gravity has less of an effect).

For this reason, you ideally want to exit the planet at an angle to it. Ideally, you want to begin 'gravity burning' at around 10km height, so that the crosshair on the navball is just at the tip of the prograde vector, and constantly, with altitude, keep it like that until the crosshair reaches the 90o mark on the navball.

Err, yea, I think that's about it. Either way, straight up trajectory out of the planet is very inefficient.

The idea of not going to standard orbit is interesting.

For example AFAIK to land on the Mun with least Dv burned you use suicide burn (going straight down when entering SOI and burning just before ground), and ascent is landing in reverse so in theory vertical ascent into escape velocity could be using smallest amount of Dv.

The problem with this (especially if we start wondering why NASA isn't using this kind of launch) is that the Dv required for ascent is directly tied to thrust (TWR), with best results at instant speed change and engines weight a lot.

Thus even by spending more Dv going into proper orbit first we do it in a smaller (lighter) ship with less TWR thus reducing costs.

In KSP if we use Nerva's for interplanetary flight it's a no brainer that going to orbit and burning to escape velocity from there is the way to go.

But due to how KSP drag model is different from RL and gravity is dropping faster with distance than it does on Earth, vertical ascents, while kind of unintuitive might, be quite good.

Rocket based "reusable" SSTO's could be easy to do with this way of launch

[Temstar]

Burning straight up against gravity is a huge waste of delta-V regardless of your TWR. Any time you're burning straight up you lose 9.82m/s (slightly less at higher altitude) of delta-V to gravity drag. Where as if you burn horizontally relative to gravity all of your delta-V spent actually goes into changing your ship's velocity. That's the whole idea behind gravity turn - by beginning your horizontally thrust earlier (compared to the "shoot straight up, wait till AP then fire horizontal) you lose less to gravity drag and more of the work your rocket did actually goes into increasing it's final orbital velocity. Remember by the time you're in stable orbit all of that work your rocket did in the form of vertical thrust have been lost due to gravity pulling you down - it's only the horizontal velocity that's keeping you in orbit.

Ejection burn is the same. The goal of an interplanetary transfer is to apply some kind of velocity change to your ship in either Kerbin kerbolar orbit prograde (for superior planets) or retrograde (for inferior planets) direction. You want all of your delta-V spent firing your engines going into changing in velocity in one of these two directions. Thrusting straight up relative to Kerbin surface will most likely not be adding the full delta-V in the correct direction and so delta-V in other directions only modify your Kerbolar orbit inclination and orbital phase rather than changing AP/PE, never mind the gravity loss.

Unless you're referring to the idea of going to the Mun or interplanetary in one continuous burn from Kerbin surface without using a Kerbin parking orbit? If that's the case the correct terminology for this is either "Direct Injection" for going to Mun or Minmum or "Direct Ejection" if you're going interplanetary. This is slightly more efficient compared to using a parking orbit (you save delta-V that would otherwise be used for circularising) but at the cost of very narrow launch windows and high difficulty of pulling it off accurately.

Edited April 30, 2013 by Temstar

[SZDarkhack]

Ok fellows, here's how it goes. Burning straight up all the way is not the most efficient trajectory. You are fighting gravity drag throughout the burn. What needs to be done is to accelerate perpendicular to gravity, and also low enough into the gravity well to take advantage of the Oberth effect.

The optimal trajectory to do so is to launch, do a gravity turn, turn almost horizontal (when appropriate) and burn directly into a hyperbolic escape trajectory. This burn doesn't suffer from gravity drag, and since your altitude remains low throughout (since you're burning horizontally), you take advantage of the Oberth effect for the maximum amount of time possible.

Now, this is only theory of course, in that it doesn't account for everything. For instance, the ease of a straight up burn might make the dV loss acceptable, in fact i've launched most of my interplanetary satellites like that. Similarly, the actual way that you fly will change the numbers dramatically. If your gravity turn introduces high steering or atmospheric drag losses, you've already lost any benefit.

The best option for most cases is, thus, the middle ground, i.e. a parking orbit. You lose a small amount of dV for shaping the orbit correctly and for starting the escape burn higher up, but on the other hand orbits are easy to make, you don't need to worry about launch timing, you can set up your maneuver with plenty of time to spare etc.

All in all, i would advise the following. If you have an overpowered rocket and you care about convenience over efficiency (which is totally legitimate on occasion), just launch straight up. In any other situation just get in a low orbit first, you're going to be saving some fuel this way. And finally, this is KSP so money and lives are not at stake. Just try both methods and measure which works best for you. Although, if straight up turns up better, you might need to work on your orbit skills a bit

Edited April 30, 2013 by SZDarkhack

[Francesco]

If you choose to use a bi-eliptic to go from a 80km circular orbit to a 200km orbit and you just happen to chose your middle burn as 100km

Then your first burn at 80km periapsis raises your apoapsis to 100km.

Your 2nd burn at 100km apoapsis raises the periapsis up to 200km.

During that burn, as the periapsis crosses 100km it becomes the apoapsis AND you have just circularized your orbit as you continue to burn.

Finally at your 200km apoapsis you burn to circularize in a 200km orbit.

no again. that's not a bi-elliptic transfer.

there is no point in terms of efficiency in what you are saying: to go from a 80km orbit to a 200km one, you just do a simple Hohmann transfer: to take advantage of the Oberth effect, you burn lower, at 80km, not at 100km.

raising your AP first to 100km and then, from there, to 200km, will cost you more in terms of delta-v.

but we're digressing here: what the OP was asking is what Temstar said, Direct injection/ejection.

as already stated, that maneuver is more efficient, but also more challenging to accomplish.

a real word example of this is: http://en.wikipedia.org/wiki/Chang%27e_2#Lunar_mission

[Nao]

Burning straight up against gravity is a huge waste of delta-V regardless of your TWR. Any time you're burning straight up you lose 9.82m/s (slightly less at higher altitude) of delta-V to gravity drag. Where as if you burn horizontally relative to gravity all of your delta-V spent actually goes into changing your ship's velocity. That's the whole idea behind gravity turn - by beginning your horizontally thrust earlier (compared to the "shoot straight up, wait till AP then fire horizontal) you lose less to gravity drag and more of the work your rocket did actually goes into increasing it's final orbital velocity. Remember by the time you're in stable orbit all of that work your rocket did in the form of vertical thrust have been lost due to gravity pulling you down - it's only the horizontal velocity that's keeping you in orbit.

The idea of vertical launch directly to escape velocity sounds bonkers but if we look at it from energy perspective it's not that bad.

Firstly even if we circularize first and then accelerate to escape velocity, we will still encounter quite a lot of "gravity drag" as after the escape burn our velocity vector will rise and we will be slowing down.

That's a standard orbit mechanic of exchanging kinetic energy for potential one, but what we count in the end is velocity at the edge of SOI (there is a term for this i forgot how it was called), so all that horizontal energy gained from orbiting first will still be eaten up by that "gravity drag".

The vertical burn will have much more gravity drag from the start but also will not "waste" energy by trying to get into orbit first.

So i think in both cases it comes down to getting enough speed in direction opposite to gravity vector to reach escape trajectory. And in the end as long as our burn was in the direction of velocity vector only, we will reach it and the actual performance will only be influenced by how much we can get from Oberth effect, how closely we follow the velocity vector, and how much we lost to aero drag (and of course many other things that have smaller effect).

In the end i made a rocket: Mk1 pod, FLT-800 + LVT-30 second stage and 2x FLT-800 + LV-909 first one (with fuel lines) total Dv of 5851m/s.

Launching Vertically into 400m/s speed at SOI edge left me with 280m/s Dv

Launching into 90x90 orbit first and then into 400m/s speed at SOI left me with 330m/s Dv

Launching into 27 deg angle gave 456 m/s (one burn with end attitude of ~27deg)

Of course i don't want to argue that vertical is better or anything silly like that but i think it's great food for thought stuff .

Ejection burn is the same. The goal of an interplanetary transfer is to apply some kind of velocity change to your ship in either Kerbin kerbolar orbit prograde (for superior planets) or retrograde (for inferior planets) direction. You want all of your delta-V spent firing your engines going into changing in velocity in one of these two directions. Thrusting straight up relative to Kerbin surface will most likely not be adding the full delta-V in the correct direction and so delta-V in other directions only modify your Kerbolar orbit inclination and orbital phase rather than changing AP/PE, never mind the gravity loss.

I think it's all vectors in the end so it would be possible to get the exact same angles at the edge of SOI just by starting burns at the proper times. The positions (relative to Kerbin in Kerbol frame of reference) would be different but that's very very small difference)

So both types of launches could achieve the same results (kerbol orbits) by burning only in the velocity vector direction.

I don't see how burning straight up always in the direction of velocity vector would not add *every* drop of Dv into escape velocity, the only thing to watch out for would be timing of launch to get proper escape angles.

Unless you're referring to the idea of going to the Mun or interplanetary in one continuous burn from Kerbin surface without using a Kerbin parking orbit? If that's the case the correct terminology for this is either "Direct Injection" for going to Mun or Minmum or "Direct Ejection" if you're going interplanetary. This is slightly more efficient compared to using a parking orbit (you save delta-V that would otherwise be used for circularising) but at the cost of very narrow launch windows and high difficulty of pulling it off accurately.

Yep, direct injections are the way to go if the ship has sufficient thrust to complete the burn in time

Cheers!

[lump]

I think of it this way, if rocket has enough thrust to do the burn at such speed they lose less than 100m/s delta-V (a 10 second burn) it is probable not worth a parking orbit, if though you are making a ship that uses LV-N engines for example, it is way more efficient to get into orbit as some of these burns can be ten minutes long, if you where going straight up that would be 6000m/s lost (if my maths holds up) now pretty much anything is more efficient than that