How do you do your intercepts/transfers to other bodies?

[EdFred]

Assuming you are not using MechJeb and are creating your maneuver nodes manually:

Do you...

...try and establish your plane of orbit and periapsis distance with your ejection/transfer burn and make no other corrections?

...get a reasonable encounter with your initial burn, and then make your final adjustment en route?

...get a reasonable encounter with your initial burn, and make your final adjustments while in the SOI?

or something else?

I started using MechJeb early on, but I've since gone back to doing it "old school."

Also assuming no mods to place exactly where the node should be, how do you do it? I have Kerbal Engineer installed, so it tells me my burn is in, say 67 degrees, but it's still hard to place the node for an optimum burn.

What techniques do you incorporate?

[SlyReaper]

This is a copy-paste of an earlier post I made on how to do the maths for interplanetary transfers. Apparently, some folks found it useful, so here it is:

You start with the equation that orbital period is proportional to the semi-major axis to the power 1.5. For a circular orbit, that's radius to the power 1.5. So T = k*(a^1.5) where T is time to complete one orbit, a is semi-major axis (or radius, which is usually close enough), and k is some proportionality constant that doesn't matter because it cancels out. In the case of Earth, we can set all the constants to be equal to 1. The Earth's semi-major axis is 1 AU, the orbital period is 1 year, etc. So if we say Earth is 1 Astronomical Unit from the sun and know that Mars averages about 1.52 AU from the sun, then we know that Mars completes one orbit in time T where T = 1.52^1.5 = 1.87 years which is about 684 days. Actually it's 687, but close enough since we're using approximations, and mid-course corrections don't tend to cost a lot of fuel.

Now imagine you want to do a hohmann transfer orbit between Earth and Mars. That means your orbit will have a perihelion of 1 AU and an apohelion of 1.52. That makes your semi-major axis equal to 1.26 AU. That means your orbital period is 1.26^1.5 years = about 1.4 years. You want to rendezvous with Mars, so you're interested in the outbound leg of this journey, so total transfer time is half that; about 0.7 years. So you want to launch 0.7 years before Mars reaches the point where it'll be 180 degrees from your present location. That's 0.7 / 1.87 = 0.37 of a Martian orbit. Times that by 360 to get 136 degrees. So Mars should be 136 degrees behind the point that's directly opposite to the sun from where you are now. Subtract that from 180 degrees, and you get 44 degrees. So you want to launch when Mars is 44 degrees ahead of you in its orbit.

So, to apply that maths to the Kerbal universe, you need to measure Kerbin's orbital distance from the sun, and call that "one Kerbal astronomical unit", or 1 KAU. It's always worth setting known values to 1 because it makes the maths so much simpler. Then, all other planets' orbital distances can be measured in KAUs. Plug in those numbers, and you'll get your optimal launch timings. The same maths works for orbital rendezvous, you just have to remember that semi-major axis is measured from the centre of gravity, not from the surface. So if you're orbiting 100km above Kerbin (and Kerbin has a radius of 600km), then your semi-major axis is 700km.

[regex]

...get a reasonable encounter with your initial burn, and then make your final adjustment en route?

For me this is how it ends up in practice because of inaccuracies introduced during the initial burn. I use a launch window planner for all my interplanetary trips and wrote a mod to handle placing the maneuver node at the right spot to get the ejection angle correct. If using Engineer, wait until you have just passed the spot on your orbit where you need to burn and place the node right behind your ship; burn on the next orbit. In lieu of a mod you can generally just eyeball it and adjust the node after setting the delta-V.

[r4pt0r]

i burn to escape kerbin, either prograde or retrograde depending on inner or outer planet as destination. then i place a node and slide it around/push/pull it, adjust it till i get a good intercept. minor course corrections halfway and one quarter there. then aerobrake or retrograde burn to orbit. its simple, crude, and kerbal. i dont have time to calculate launch windows or dv or weight. build and go!

[ArmchairGravy]

I used to use the Interplanetary Calculator and eyeball the phase angles, but have recently installed Kerbal Alarm Clock and am now relying on it. For the ejection angle I just slap a maneuver node down and tweak its position until I get an encounter or have an orbit overlap in the case of inclined targets. I'll then make the final adjustments at the ascending/descending nodes. Not the most efficient in delta-v, but I find it a reliable system.

[Tw1]

I use protractor for phase angles, and a set of periapse kicks, timed so I'll reach the periapsis/burn point just a little later than wanted. Leaving slightly later lets me adjust better with another burn just outside Kerbin's SOI.

[SpaceSphereOfDeath]

I memorise the transfer angles and just launch straight to the target, doing an inclination adjustment if necessary.

[EdFred]

Whoa, whoa, whoa. We can slide the node along our orbital path?

(incomprehensible swearing and cussing)

[NecroBones]

I use this to get my window:

http://alexmoon.github.io/ksp/

Then, I do my initial burn to get a relatively rough encounter, but I'll try to tune it to get as low of a final periapsis as I can reasonably do without spending too much time on it (mostly sliding the maneuver back and forth on the orbit, adjusting the prograde burn, and doing small tweaks to the normal/radial vector as well). Usually the burn isn't perfect, so I do a second burn a few minutes later to get it back to where I had intended.

I'll usually end up doing a correction again at about the midpoint of the transfer trajectory, and then fine-tune after entering the SOI. I may also do an inclination correction at an ascending/descending node, if there is one (a lot of the ideal windows put the interception time within the same plane as Kerbin's, so either I'll skip this step, or try to alter inclination a little with the initial escape burn).

This is a copy-paste of an earlier post I made on how to do the maths for interplanetary transfers. Apparently, some folks found it useful, so here it is:

You start with the equation that orbital period is proportional to the semi-major axis to the power 1.5. For a circular orbit, that's radius to the power 1.5. So T = k*(a^1.5) where T is time to complete one orbit, a is semi-major axis (or radius, which is usually close enough), and k is some proportionality constant that doesn't matter because it cancels out. In the case of Earth, we can set all the constants to be equal to 1. The Earth's semi-major axis is 1 AU, the orbital period is 1 year, etc. So if we say Earth is 1 Astronomical Unit from the sun and know that Mars averages about 1.52 AU from the sun, then we know that Mars completes one orbit in time T where T = 1.52^1.5 = 1.87 years which is about 684 days. Actually it's 687, but close enough since we're using approximations, and mid-course corrections don't tend to cost a lot of fuel.

Now imagine you want to do a hohmann transfer orbit between Earth and Mars. That means your orbit will have a perihelion of 1 AU and an apohelion of 1.52. That makes your semi-major axis equal to 1.26 AU. That means your orbital period is 1.26^1.5 years = about 1.4 years. You want to rendezvous with Mars, so you're interested in the outbound leg of this journey, so total transfer time is half that; about 0.7 years. So you want to launch 0.7 years before Mars reaches the point where it'll be 180 degrees from your present location. That's 0.7 / 1.87 = 0.37 of a Martian orbit. Times that by 360 to get 136 degrees. So Mars should be 136 degrees behind the point that's directly opposite to the sun from where you are now. Subtract that from 180 degrees, and you get 44 degrees. So you want to launch when Mars is 44 degrees ahead of you in its orbit.

So, to apply that maths to the Kerbal universe, you need to measure Kerbin's orbital distance from the sun, and call that "one Kerbal astronomical unit", or 1 KAU. It's always worth setting known values to 1 because it makes the maths so much simpler. Then, all other planets' orbital distances can be measured in KAUs. Plug in those numbers, and you'll get your optimal launch timings. The same maths works for orbital rendezvous, you just have to remember that semi-major axis is measured from the centre of gravity, not from the surface. So if you're orbiting 100km above Kerbin (and Kerbin has a radius of 600km), then your semi-major axis is 700km.

This is excellent, and very nicely written up. Thanks!

Edited April 21, 2014 by NecroBones

[lazarus1024]

This is a copy-paste of an earlier post I made on how to do the maths for interplanetary transfers. Apparently, some folks found it useful, so here it is:

You start with the equation that orbital period is proportional to the semi-major axis to the power 1.5. For a circular orbit, that's radius to the power 1.5. So T = k*(a^1.5) where T is time to complete one orbit, a is semi-major axis (or radius, which is usually close enough), and k is some proportionality constant that doesn't matter because it cancels out. In the case of Earth, we can set all the constants to be equal to 1. The Earth's semi-major axis is 1 AU, the orbital period is 1 year, etc. So if we say Earth is 1 Astronomical Unit from the sun and know that Mars averages about 1.52 AU from the sun, then we know that Mars completes one orbit in time T where T = 1.52^1.5 = 1.87 years which is about 684 days. Actually it's 687, but close enough since we're using approximations, and mid-course corrections don't tend to cost a lot of fuel.

Now imagine you want to do a hohmann transfer orbit between Earth and Mars. That means your orbit will have a perihelion of 1 AU and an apohelion of 1.52. That makes your semi-major axis equal to 1.26 AU. That means your orbital period is 1.26^1.5 years = about 1.4 years. You want to rendezvous with Mars, so you're interested in the outbound leg of this journey, so total transfer time is half that; about 0.7 years. So you want to launch 0.7 years before Mars reaches the point where it'll be 180 degrees from your present location. That's 0.7 / 1.87 = 0.37 of a Martian orbit. Times that by 360 to get 136 degrees. So Mars should be 136 degrees behind the point that's directly opposite to the sun from where you are now. Subtract that from 180 degrees, and you get 44 degrees. So you want to launch when Mars is 44 degrees ahead of you in its orbit.

So, to apply that maths to the Kerbal universe, you need to measure Kerbin's orbital distance from the sun, and call that "one Kerbal astronomical unit", or 1 KAU. It's always worth setting known values to 1 because it makes the maths so much simpler. Then, all other planets' orbital distances can be measured in KAUs. Plug in those numbers, and you'll get your optimal launch timings. The same maths works for orbital rendezvous, you just have to remember that semi-major axis is measured from the centre of gravity, not from the surface. So if you're orbiting 100km above Kerbin (and Kerbin has a radius of 600km), then your semi-major axis is 700km.

Which is extremely useful and helpful...but to be honest, I never bother. Only because its annoying to me to time warp (not as annoying now that I can do it in the tracking center, which I haven't really played with yet TBH).

I generally just launch, then punch in manual maneuver nodes and if it takes me an orbit or three to line up, so be it. Or at least that is what I do with destinations that are not cis minmus. Other than the rare screw up, I pretty much manage within about 2 orbits, occasionally 3.

[KvickFlygarn87]

Number 2. Not quite as efficient, WAAAAAY easier.

[cantab]

I run Alexmoon's planner and input what it tells me into Precise Node. Then I play with it a little to adjust things: periapsis nice and low for an insertion, or appropriate change to the orbit for a flyby.

For asteroids my approach is a bit different. I've found that aiming to leave Kerbin's SOI in the same direction as the asteroid is predicted to enter it works well. Usually I'll do an initial ejection burn, for which the magnitude isn't too critical, a bigger ejection burn simply gets me an earlier encounter. Then I'll make a normal burn somewhere between the Mun and Minmus's orbit. Of course if I'd thought ahead I'd just have launched into the right inclination to start with. To be honest, I'm not sure how efficient this is, it seems somewhat delta-V hungry, but my thinking is that if I just want to deflect the asteroid then the earlier I get to it the less I need to push it by.

Edit: Make that seems really delta-V hungry. I'm finding I've got a nearly 2000 m/s burn to match speeds with the asteroid on one mission now

Depending on the ship I may or may not need to make a later correction. Little probes are easy to get right first time. Big asteroid tugs, not so much.

Edited April 21, 2014 by cantab

[Superfluous J]

I do 99% of my burn at #1 to get in the planet's SOI, then at the AN/DN do the final 1% with a small burn to get my "real" encounter with the planet in question. If I'm doing aerobraking, I do what is usually a 1-10m/s burn to get my periapsis right, and then after aerobraking a 10-20m/s burn to actually capture my ship because I suck at aerobraking.

So my answer is all 3

[brdavis]

Whoa, whoa, whoa. We can slide the node along our orbital path?

(incomprehensible swearing and cussing)

Grin. Yes. New feature in 0.23 (IMS). Hover the mouse over the "ring" of an expanded node, and when just the ring highlights, click and drag it along the orbit. Yes, it makes life much easier

[Uncle Mort]

I have a list of optimum transfer days for the planets. I eyeball the angle for the manoeuvre node and get a close approach on the first burn. I normally do another manoeuvre once I leave Kerbin's SoI to get an encounter. In the target SoI I circularise, then correct the inclination and then drop the periapsis for aerobrake orbit(s) and finally circularise again. Not the most efficient but it works.

[EdFred]

Grin. Yes. New feature in 0.23 (IMS). Hover the mouse over the "ring" of an expanded node, and when just the ring highlights, click and drag it along the orbit. Yes, it makes life much easier

Oh, how I wish I knew this. I would create a node, see that it was off 5-10 degrees or so, delete, try again. This will make a world of difference!

I already knew when the phase angle was, where the ejection burn was to occur, it was just a matter of placing the node correctly without having to burn additional radial dV to shift the orbit. Well, I know what I'll be experimenting with tonight!

[Pecan]

Lots of burns.

1) Plane to zero degrees

2) Ejection/rough intercept

3) Plane-match at AN/DN

4) Mid-course correction

5) Tuning on SOI change - mostly because the orbit path/markers tend to bounce around so much they can't be relied on earlier (big wobbly rockets!)

6) Orbital injection/final corrections

7) Adjust Ap/Pe

8) Plane change

9) Adjust Ap/Pe

10) Plane change

11) Adjust Ap/Pe

12) Plane change - I'm a bit obsessive about my space-stations' orbits ^^

What? I don't know, there's something about an idle engine that just cries out for a poke.

[Pulstar]

Actually you could drag maneuver nodes along trajectories since they have been added in 0.18.

Also personally I just use Kerbal Alarm Clock to know when the launch window appears, then I fiddle around with manuever nodes to get a proper planned manuever for an encounter. Usually the burn is inaccurate or I don't have the right inclination so I do a correction burn at an ascending or descending node. A final correction burn is done before I enter the target's SoI to adjust any aerobraking/aerocapture I might be attempting to do and to get me into an equatorial orbit matching the planet's rotation around its own axis.

[_Aramchek_]

I do not create maneuver nodes as this actually hinders me, I plot ballistic trajectories by eyeball and high i.q. shenanigans. (And very high TWR's)

[Brotoro]

Nowadays I let Kerbal Alarm Clock remind me when a transfer window occurs (previously I'd used the various online calculators).

I do my escape burn targeted to get as close an encounter as possible, but I generally do NOT include the plane-shift in this burn (unless Kerbin happens to be right at the node at this time). If the orbital inclinations are such that no encounter shows up, I'll drop in a second maneuver at te ascending/descending node and verify that a plane-shift maneuver later will let me encounter the planet.

If the escape burn is long, I'll do the escape burn in two parts for better efficiency.

I then carefully take the ship across Kerbin's sphere of influence boundary at 1X timewarp to be sure my trajectory doesn't go wonky by crossing the SOI at high warp.

I then do a plane-shift burn at the ascending or descending node. I set a reminder up in Kerbal Alarm Clock as to when this will occur (if I'm going to be handling other things in the meantime).

Immediately after the plane-shift burn, I do my fine targeting burn. I use conic draw mode 0 to be sure this trajectory is targeted precisely at the destination planet, and going in the correct direction.

Then I set an alarm in Kerbal Alarm Clock to warn me before the ship crosses the SOI of the target planet so I can sneak it across that boundary at 1X as well.

I know that it is more efficient to combine burns (such as the plane-shift burn and the fine-targeting burn), but I find it much easier to do the fine targeting once the plane has been sorted out already, so it's worth the little inefficiency for me there.

[regex]

I then carefully take the ship across Kerbin's sphere of influence boundary at 1X timewarp to be sure my trajectory doesn't go wonky by crossing the SOI at high warp.

I used to do this too until I found out that 50x (and, I think, 100x) is safe as well. Anything over that, though, will introduce errors.

[FenrirWolf]

I establish parking orbit then use this method to eyeball the phase and ejection angles

Javascript is disabled. View full album

(EDIT: Not my album, found it on some website who's name I don't recall)

Once I have my intercept I adjust my approach with RCS or lots of turn-and-burn

Super secret .23.5 protip: once you have an intercept with the target planet/moon, click on the body in the map view and click focus view. The camera centers on the body and sets the patched conics relative to its SoI. Makes it a ton easier to adjust your path and to see if you're coming in on a prograde or retrograde flyby.

Edited April 24, 2014 by FenrirWolf

[ayana]

I try to be precise. I have Kerbal Alarm Clock stop warp 24 hours before the transfer window and start setting up my maneuver. I just get my encounter to either directly above or below the target and worry about inclination later on during the trip. It's easy to get every encounter right that way.

I've gotten good enough at it that I can get direct captures by Laythe even before I hit my Jool periapsis.

[Rabada]

I establish parking orbit then use this method to eyeball the phase and ejection angles

http://imgur.com/a/S8Ija

Once I have my intercept I adjust my approach with RCS or lots of turn-and-burn

Super secret .23.5 protip: once you have an intercept with the target planet/moon, click on the body in the map view and click focus view. The camera centers on the body and sets the patched conics relative to its SoI. Makes it a ton easier to adjust your path and to see if you're coming in on a prograde or retrograde flyby.

This is a very interesting method! I'm going to have to try it! This will make my life a whole lot easier, Thankyou!!

[NecroBones]

I establish parking orbit then use this method to eyeball the phase and ejection angles

http://imgur.com/a/S8Ija

Once I have my intercept I adjust my approach with RCS or lots of turn-and-burn

Super secret .23.5 protip: once you have an intercept with the target planet/moon, click on the body in the map view and click focus view. The camera centers on the body and sets the patched conics relative to its SoI. Makes it a ton easier to adjust your path and to see if you're coming in on a prograde or retrograde flyby.

Very good. I've been using one of the launch-window calculators, but this is a nice straightforward method for those who want something straightforward, instead of trying to save every last possible kilogram of fuel.