Start with lower orbit to transfer to higher one or straight to desired one?

[W2C]

So still pretty new and learning a lot of these things. Just started putting my first relay network in place.

But I'm wondering what the most ideal way to get into an orbit of around 320k would be. The obvious response to this might be how easy it is for me to just go test it for myself, but seeing as how I'm new, my launches are far from uniform and it's difficult for me to tell if a dv savings was from the maneuver or from me launching differently at this point thus why I'm asking here instead to see if I can find out what I should be aiming for. Should I be trying to circularize a lower orbit from 80-100k before transferring up to my desired orbit or should I continue aiming straight for it and circularizing at 320 without bothering with the lower one?

[KawaiiLucy]

For low orbits it probably doesn't change much in dv used, however it is usually recommended to do Hohmann-transfers. Basically circularize at 70-75km in LKO and then raise apoapsis to 320km and finish circularizing . That would theoretically be the most efficient way to reach your orbit. Of course it depends whether you use a satellite without a comnet (yet) and only relay on sparse groundstations, I'd recommend burning enough so that your apoapsis is over the next groundstation and you won't be out of control of your probe.

 

[Streetwind]

The thing about ideal ascents is not where your apoapsis ends up, but rather where your periapsis ends up. You want it to be as close to positive as possible before you have to turn off your engine and coast to apoapsis. The reason being, if your periapsis is already positive, or close to it, then you have achieved close to all the horizontal speed you need, and the orbit insertion burn will be fairly negligible. It means that not only did you spend the majority of your dV further down in the gravity well, where it is more efficient to do so (-> Oberth effect), but also that you burned sideways a lot instead of upwards, which minimizes gravity losses, and that you always burned reasonably near your apoapsis, which is the most efficient way of raising your periapsis.

And you'll find that if you follow that guideline, the two solutions you propose... kind of start to look the same. All you do is let the engine run just a little bit longer before turning it off and coasting to apoapsis. Which, if done while you are close to apoapsis and having a high periapsis, will involve flipping that apoapsis around to the other side of the planet. Which is functionally similar to inserting into orbit and then performing a Hohmann transfer.

So as long as you get the fundamentals right, you don't really have to worry about this question.

[vv3k70r]

1 hour ago, W2C said:

Should I be trying to circularize a lower orbit from 80-100k before transferring up to my desired orbit or should I continue aiming straight for it and circularizing at 320 without bothering with the lower one?

Cargo to the tech You have is an issue.

If You are aiming for 320km and You cant get a balanced rocket that get to LKO in a propper manner first just go above atmosphere any way You can (I guess it would be high but still stratospheric) and circulirize it (more or less) by flying to hotizon during ascend, then transfer.

After You build some rockets and fly them You will easilly get to orbit even on 72km just by giving them small angle on given speed between 60-120m/s, setting prograde and walking after snacks when it fly to next staging, but it will come with practice.

Edited December 1, 2020 by vv3k70r

[W2C]

Thanks for the responses guys. I'm going to add a bit more detail so I can get a bit better clarity from the answers. 

So firstly, getting to orbit itself isn't an issue for me. I have a ship that I've used to get me into 320k orbit for 5 or 6 satellites now. I've spent time learning the maneuvers to adjust my inclination and eccentricities as one of my mods (GPP) begins me on a planet off the equator and my RemoteTech contract required a relatively precise placement over the equator. I'm using a manned rocket to place my satellites to ensure I don't have to deal with signal blackouts and therefor don't have to rush to circularize before that happens. I begin doing a gravity turn almost right away, getting between 50-60k for my AP from my first stage boosters though my actual height is usually only 20-30k by the time I jettison them and begin my second stage. I then use my second stage which I think was a valiant or reliant engine, I can't remember which, but considering I'm still in atmosphere I need it to raise my AP up about 280k and then my final stage with my tarrier engine to get me the rest of the way, then to circularize my burn, followed by another to adjust my inclination and eccentricities before deploying the satellite and then plenty of DV left to deorbit me and a burn prior to entering atmosphere to lose the speed for easy and safe reentry. 

Basically I'm getting into orbit successfully, although my gravity turns aren't perfect and my DV is inconsistent between burns but I'm accomplishing the mission just fine. I've just been wondering if the way I've been doing it is really the best way to be doing it or not. Every video I've watched shows them aiming for a lower initial orbit but I'm not sure if that's just because it's more convenient for their missions or if there's an actual DV savings to be had from a lower initial orbit before Hohmann's transferring up to the higher one. Am I wasting DV by burning straight for the desired height (using gravity turn to do it) or am I saving DV by only having to circularize once at an altitude where it takes relatively little DV to do it?

[Spricigo]

That's a tricky question you shoot at us. And the only answer we can give without any risk of being wrong or, at least, imprecise is:

Depends on a lot of variables.

Let's elaborate a bit on that.

 The Gravity Turn is regarded as the most efficient way to get into a orbit just outside the atmosphere. In a nutshell: give a small nudge towards west shortly after taking off and let the rocket steer itself into orbit and control the throttle to go neither too fast (because aerodynamic drag will be an issue), neither too slow (because some of the power/deltaV will be used to counter gravity until the craft get up to orbital speed).

There is also the most efficient way to go from a low orbit to a higher orbit, the Hohmann Transfer. In a nutshell: a single burn prograde to raise your orbit so it touches the higher orbit in a single point, then another burn prograde at that point to match the orbit.

In a ideal world, we could just combine both the Gravity Turn and the first leg of the Hohmann Transfer in a single maneuver to have the most efficient way to go to a higher orbit. But Kerbin is not ideal, it have an atmosphere. And the atmosphere will take a toll if you combine the maneuver, both in increased losses due aerodynamic drag and reduced Isp of engines. We also don't have ideal engines (weighting nothing and capable of instantly execute the maneuvers) thus resulting in further losses either way. And all that assuming heating will not be an issue. And it will be different for each different craft you launch.

So, the most efficient way is probably something along a Hohmann Transfer right after a quite aggresive Gravity Turn. But I guess most players will not consider trying to find that precise trajectory as the most efficient way to expend his/her game time. Considering the chance of being too aggresive, maybe doing a regular gravity turn and a Hohmann Transfer once in space is a better idea.

 

[AHHans]

The short answer is that intentionally focusing on getting into a circular LKO is not efficient.

The slightly longer answer is like @Streetwind wrote: if you do it right to begin with then at the point where you reach orbital velocity - i.e. raised your PE to your current altitude - you are effectively in an LKO. So it doesn't matter if you just keep burning burning or coast for a while and then continue your burn. For maximum efficiency you want this "intermediate LKO" to be as low as reasonably possible, maybe even still in the atmosphere. (Well not too deep in the atmosphere, and you wouldn't want to coast in the atmosphere.)

The IMHO practical answer is: concentrate on getting your launch ascent right! You are much more likely to loose significant amounts of dV with a bad launch than you loose by first going to LKO and then to your target orbit.
Also: if you are aiming for a specific target orbit (e.g. when you want to rendezvous with a station there) then it might be better to first go into LKO to make a better adjusted transfer to your target, e.g. by combining the transfer burn with a plane change.

[paul_c]

The issue isn't really overall delta V - I dare say the 2 approaches are going to be very close there. 

BUT....put simply, once you are in orbit, you can use an engine of TWR 0.00001 to change orbit to...anything, really. Yes in practical terms you would want something a bit more spritely, but you don't need the TWR that you needed to get to orbit. (And yes, I know in the later stages of that phase, you can use a TWR<1 while completing the gravity turn to 70km, but you have a limited time to put the energy in, so it still needs to be a reasonably sized engine then).

You can maximise efficiency by thinking about the different needs of each phase of the flight, and building a vehicle which is best/cheapest/lightest at that. For example:

* first 1000 delta V - you don't need to steer, you're going straight up - use SRBs. Needs to be TWR>1
* to orbit (say 2500-3000 delta V) - initially needs to be TWR>1 (later bits if staged could be below though). Needs to be controllable and responsive
* orbital maneuvering - can be TWR<1; doesn't particularly need to be fast/responsive at changing direction either

So if you had a contract with a lot of orbital maneuvering needed, eg a big Ap and Pe, then it makes a lot of sense to "build" a satellite/probe as small as possible but with a decent fuel tank and a tiny engine (eg the Ant engine...only 20kg). In your case, a circular orbit of 320km won't take up that much delta V from LKO so its a bit moot though. (A really basic probe and a bag of fuel (the doughnut) gives you about 2400m/s delta V, which is more than enough for pretty much any changes and future contracts too).

Another thing...for efficiency, NEVER make a plane change when launching the satellite (except for a small correction which might be required). ALWAYS launch precisely in the launch window (so the path of the orbit passes directly over KSC) and 'drive' the rocket in the right direction to be on the correct inclination from the outset. The 'cheapest' time to do that, is when you're going slowest horizontally, which is earlier rather than later in the launch. ie once you've just started the gravity turn, also look at the inclination and tweak the direction you're flying.

[Laie]

5 hours ago, W2C said:

But I'm wondering what the most ideal way to get into an orbit of around 320k would be.

All things being equal, you can save a little dV by going for directly for the target altitude and circularize there.

Little experiment: you can go from 80km circular to 320km circular in 10km increments, circularizing at 90, 100, 110...km; or you can go from 80km to 320km in one leap. The latter will turn out to be more efficient.

In a similar fashion, and for the same reasons, launching to target altitude rather than going through an intermediary orbit is more energy efficient.

The difference will be relatively small, though, I'd guess on the order of 50m/s or so. As you say that your launches are far from uniform, it's quite possible that you won't really notice the difference.

[Superfluous J]

Basically, the entire reason to get into LKO first is because LKO is the cheapest you can get into *an* orbit, and once you're in any old orbit you can plan your next steps at leisure. It's not all that much more expensive to launch to LKO and then later use a carefully planned maneuver node to transfer to Mun, than it is to launch directly from the KSC to a transfer orbit directly to Mun. And the uncertainty involved in the direct route means that the direct route may end up being *more* costly due to the need for mid-course corrections.

You know the orbit you want to be in, so launching directly into it is better. If you were instead trying to rendezvous with a satellite in that orbit, it may be better to launch to LKO first so you could plan out the rest.

[W2C]

Awesome. Again thank you guys. It's nice to know that both routes are basically equally viable but the point a few of you have made about the lower step making it easier to insert yourself into an encounter is a point I was considering as well. I know that orbits of different heights go at different speeds so if I launch into a bad window I can always wait until the lower orbit is in a better position to insert into the higher one where I want. Might be a good idea to get used to lower initial orbits just to give more flexibility as was mentioned. I was a bit hesitant with having to launch into too low an orbit because I kept falling back into the atmosphere during my burns but that was before I upgraded buildings to get maneuver mode's combined with using a relatively weak engine at the time. You guys have definitely been a big help though.

[paul_c]

14 minutes ago, W2C said:

 I know that orbits of different heights go at different speeds so if I launch into a bad window I can always wait until the lower orbit is in a better position to insert into the higher one where I want. 

In the above - if its a transfer from one circular orbit to another, it doesn't matter. But if its a transfer to an elliptical orbit (very different Ap and Pe) then its important to burn at the right time. Basically, you'd do this as a general rule:

1. Be in a lower, circular orbit
2. Burn prograde at the Pe of the desired orbit; or more accurately, imagine a line joining the desired orbit's Ap and Pe, where this line intersects your current orbit, burn then. That will raise your Ap to the same point in space as the desired orbit's Ap.
3. At the new Ap, burn to raise(/lower) the Pe as needed.

(The above assumes the inclination is the same).

Personally I'd always aim for at least 70250m as a 'parking' orbit, the reason being in KSP it assumes there is a boundary between space (where there is no atmosphere) and within the Kerbin atmosphere. Of course, in the real world, this is a much less well defined thing. Once above 70250m KSP doesn't include drag, and nothing else which could peturb an orbit is taken into consideration either, so you can effectively leave it there and know it will stay there.

Maneouver nodes help - for a while, I got muddled with normal/anti-normal and quickly sketching out a planned maneouvre helps clarify.

 

[king of nowhere]

7 hours ago, W2C said:

 I was a bit hesitant with having to launch into too low an orbit because I kept falling back into the atmosphere during my burns but that was before I upgraded buildings to get maneuver mode's combined with using a relatively weak engine at the time.

a low orbit is best if you have the TWR, but with a low twr you need a higher orbit to gain the time to circularize before falling back. i had the same issue with a nerv-powered vehicle, it turned out i needed an apoapsis over 100 km to have the time to enter orbit. but that's an edge case. most of the time, low orbit is better.

[vv3k70r]

9 hours ago, paul_c said:

BUT....put simply, once you are in orbit, you can use an engine of TWR 0.00001 to change orbit to...anything, really.

Because no drag. Otherwise low orbit would do just above mountains (they have high drag).

8 hours ago, Superfluous J said:

Basically, the entire reason to get into LKO first is because LKO is the cheapest you can get into *an* orbit

& not carry overpowered, heavy engines any longer.

7 hours ago, W2C said:

It's nice to know that both routes are basically equally viable

Diference in the game is miniscule. Do not worry much about it - fuel is cheap. It is not a nasty planet that take 9500dV to leo.

8 hours ago, W2C said:

I kept falling back into the atmosphere during my burns but that was before I upgraded buildings to get maneuver mode's combined with using a relatively weak engine at the time.

After You play a while You not gonna need manouver nodes for anything more then estimates for given vessel. Once You get into randevouz You find that there are intuitive ways to get where You want by navball.

 

 

[paul_c]

3 minutes ago, vv3k70r said:

Because no drag. Otherwise low orbit would do just above mountains (they have high drag).

 

 

 

No, not necessarily "no drag" - its because an orbit, by definition, has a horizontal velocity component that means it does not descend and hit the ground again.

Think of leaving Mun/Minmus - you NEED a TWR > 1 otherwise you're just making a barbeque initially. If the planet/moon were glass smooth, no terrain whatsoever, you'd still need TWR>1 to lift off the ground, but the gravity turn element would be very short and then you could get by with TWR<1. You need the lift to overcome the gravity and you need to do that long enough to get the horizontal speed up  to orbital.

[Superfluous J]

9 minutes ago, vv3k70r said:

& not carry overpowered, heavy engines any longer.

You can drop those before LKO, or even before leaving the atmosphere.

[Spricigo]

2 hours ago, paul_c said:

No, not necessarily "no drag" - its because an orbit, by definition, has a horizontal velocity component that means it does not descend and hit the ground again.

Incidentally, the deltaV lost fighting gravity is sometimes referred as gravity drag.  Also notice that low TWR will result in long burn and, which is in itself inefficient (because of steering losses).

In fact, the whole issue is that one can only go so far in reducing some cause of inefficiency before something else start to increase.

Edited December 1, 2020 by Spricigo

[vv3k70r]

11 hours ago, paul_c said:

No, not necessarily "no drag" - its because an orbit, by definition, has a horizontal velocity component that means it does not descend and hit the ground again.

You can have this component just above the surface.

11 hours ago, paul_c said:

Think of leaving Mun/Minmus - you NEED a TWR > 1 otherwise you're just making a barbeque initially. If the planet/moon were glass smooth, no terrain whatsoever, you'd still need TWR>1 to lift off the ground

On the glass smooth surface (mean ignoring friction?) You can get to the speed that will have a proper component and You are in orbit on the surface (means You are falling on this glass ball as fast as it curve). And You still getting speed making this vector longer and thus away. TWR>1 is a practical solution because such surrealistic circumstance dosent occur.

We consider gasous atmosphere as issue only because we can solve it, but it could be fuild or even solid. It is way mountains have a high drag, but it dosent mean that with propper TWR part of the vessel will not go throu.

11 hours ago, paul_c said:

You need the lift to overcome the gravity and you need to do that long enough to get the horizontal speed up  to orbital.

Sphere has no end, You have all the time You will as long You accelerate. With speed You will get to the point where Your component of speed is higher then loss on gravity until touchdown, and because of lack of friction bouncing will not stop You from loosing horizontal on subsequent, longe bounce (even if considered completly rigid body that dosent deliver any component from spring) untill You miss the surface.

Surface is the point where gravity component dosent stack any making vertical contribution 0. But it is surrealistic. There are mountains and they have "drag".

Curvature of the sphere "deliver" vertical component. I made simulation of this years ago also on flat, never ending surface.

11 hours ago, Superfluous J said:

You can drop those before LKO, or even before leaving the atmosphere.

Often do (last I had a robotic vesel that was safer to unfold in vacum so I detached last booster from between spider legs after deployment). But the only reason is that vessel is overpowered untill it leave atmosphere, if it be perfect it will keep exact acceleration counteracting issues just to the edge. Cannonball leave its engine after ignition on a proper TWR.

8 hours ago, Spricigo said:

In fact, the whole issue is that one can only go so far in reducing some cause of inefficiency before something else start to increase.

Yes, it is about practial solution in given circumstances. The rest is theoretical.

 

Edited December 2, 2020 by vv3k70r

[AHHans]

42 minutes ago, vv3k70r said:

It is way mountains have a high drag, but it dosent mean that with propper TWR part of the vessel will not go throu.

Well, in this case it's not just TWR that determines how far or fast you can go through mountains. The material properties of the mountain and your craft play a major role. (Aaand we're back at the topic of fuel tanks made out of armor plate.)

[vv3k70r]

25 minutes ago, AHHans said:

The material properties of the mountain and your craft play a major role.

I was about to state that mountain made of iridium or something that class could be nice, but it would be another offtopic.

However celestial colision go throu mountains, planets core and stars so it is possible (with some mass lose on the object).

If we go back to this miniscule TWR on curvature that will stack to orbital speed there is one hidden problem that differ reality, simulation and formula. Formula does not slice time in instances, simulation do (timing to next iteration on variable stacks and because of this clock gap result in long term differ from pure math equation) and reality... who knows - human mindset tend to refer to persistent hypertime instead of multiple, interacting instances because we exist in events that are enough local. I read about idea of implementing n-body solution to KSP2 and because I made such simulation I know issues of making it playable, especialy with people that gonna run it forward and backward to see what previous states lead to such a configuration and how would it colapse (results of forward and backward would differ because of ceil/floor cuting data type). Any blend to rail formula (Keplers) would be invokation for Kraken exploits.

Most of these blend cases are Kraken exploits curently and lot of disapointment when aplying parts to eforts not ment by design. And there are many workarounds visible in cheats (easy in phy) or camera orientation, VAB/SPH construction (specialy if You look in files). One of them is topic here - atmosphere is cuted on 70km alowing to depend on miniscule thrust to get any orbit making gravity turn changed from curvature to flat in that 70km instance. This behaviour is not expected in reality, and it is not included in gravity turn formulas we have and apply. This are quirks of simulation, specialy simplicited to real time game.

Edited December 2, 2020 by vv3k70r

[paul_c]

We are not considering sliding along the surface, but flying - ie clear of the terrain. In KSP you'll blow up if you go too fast on a surface.

My original point remains - you NEED TWR>1 to "lift off" then a sufficient TWR to transit from sub-orbital to orbital before returning/hitting the surface again. (Once you ARE orbital, the criteria which sets a minimum TWR is different, so it can be much lower).

The "glass smooth" analogy was to eliminate terrain, nothing else......

Edited December 2, 2020 by paul_c

[vv3k70r]

3 hours ago, paul_c said:

then a sufficient TWR to transit from sub-orbital to orbita

If transit - whatever, but higher the better - burn is shorter and thus less eliptical.

If randevouz - there is a third paramater - control of burn time execution. Because we can asume to play manualy without any autopilot mod more control mean more eliptical low thrust burn to get to corect place. Means less efective - more control.

If specific orbit with argument of periapsis - more short burns (inclination, PA, AP, argument corection in between PA & AP.

More control - less thrust, less efectivity. Monoprops have low thrust.

Efectivity - high thrust.

Edited December 2, 2020 by vv3k70r

[paul_c]

Well......yes and no. Higher is nicer and simpler to 'drive', but it will inevitably weigh more. And if its the final/upper stage, a 1kg saving in this means many kgs saved in earlier stages.