What can you tell about transfer windows?

[Spricigo 1,146]

Quote

The key to getting anywhere is performing the transfer at the correct time.

Figuring out the correct time to depart to another planet is not a trivial task. Yes, some veteran make it looks easy, but most us struggled with this at some time. ..some still have trouble.

So, what you can tell about transfers windows?

What is your technique to find it out? Or you rely on some tool? Or are you still struggling with it and have your own questions about it? In any case, you are welcome to share your thoughts. 

This is not a typical question, of someone looking for help with an specific issue it rather meant to be an opportunities, for some people learn about an important thing in KSP, for other to share what they know about it. I hope to be right in my assumption that both are things this community enjoy.

[Snark 14,343]

20 hours ago, Spricigo said:

Figuring out the correct time to depart to another planet is not a trivial task. Yes, some veteran make it looks easy, but most us struggled with this at some time. ..some still have trouble.

So, what you can tell about transfers windows?

Let's say you're at planet 1, and you want to go to planet 2.  The thing you're trying to figure out is the phase angle at launch, i.e. you should launch when planet 2 is how many degrees ahead or behind planet 1.

Well, this can be done in a mathy way, with equations and such.  I can go into that, if you like. 

However, here's a relatively straightforward way to do it visually, just by eyeballing things with maneuver nodes in map view:

1. Get into low circular orbit around your origin body.
2. Plop down a maneuver node in any ol' place, and give it some dV.  Give it enough that it would put you on an escape velocity with a reasonable amount of margin.  (For example, if you're in low Kerbin orbit, you could put around 1000 m/s for starters).
3. Move the maneuver node to the right place.
   • If you're going to a "superior" planet (i.e. one that's farther from the sun than your origin), then you want to position the maneuver node so that at the point you exit the SoI, you're traveling in the same direction as the planet's orbit around the sun (i.e. solar-prograde).
   • If you're going to an "inferior" planet (i.e. one that's closer to the sun than your origin), then you'd position the node the opposite way (i.e. so that you exit the SoI in the solar-retrograde direction).
4. Now zoom out so you can see enough of the solar system to see the sun, your origin planet, and your destination planet.  Rotate it so that you're looking straight down at the sun's north pole.
5. Since your maneuver node shows you as being on an escape trajectory, then when you get to this view, you'll see your projected solar orbit after escape.
   • If you're headed for a superior planet, then this means your solar orbit will have its Pe at your origin, and its Ap a little bit higher than that (but probably lower than your target).
   • If you're headed for an inferior planet, then this means your solar orbit will have its Ap at your origin, and its Pe a little bit lower than that (but probably higher than your target).
6. Give your maneuver node more dV (i.e. drag its handle) so that you raise your Ap (for a superior target) or lower your Pe (for an inferior target) until your projected orbit just touches the target's orbit.
   • Note that as you do this, you may need to go back to step 3 to re-adjust the position of the maneuver node, to keep your ejection direction from the origin pointed perfectly solar-prograde or solar-retrograde.  This is because as you add dV, your post-ejection trajectory within the SoI won't curve as much, so you'll need to readjust your aim a bit.
7. At this point-- when your projected orbit touches the target's orbit-- the game should be showing you the light blue closest-approach markers.
8. Suppose that when you do this... you see that the closest-approach markers are really close together already!  Well, that would mean you just got really lucky and you actually are in the transfer window right now.  Launch!  Launch! Launch! 
9. However... probably you're not super lucky.  Probably, the closest-approach markers are separated by quite a bit. In which case, see below.

Probably at this point you're looking at something that's more or less like this example (here I'm going from Kerbin to Duna):

See, the "lucky case" where I'm already at the launch window would be if the two closest-approach markers were both right there at Ap, where I've got the green arrow pointing in the above picture.

As you can see, I'm not so lucky here.    Well... look at where the two markers are (red arrows).  See that at closest approach, Duna is ahead of where my ship will be, by quite a bit-- looks like just under 90 degrees.  So... that means I need to catch up to Duna some more (by close to 90 degrees), before I launch.  That tells me roughly how far away from the launch window I am, at least in terms of the relative angle of Earth and Duna.

So... at this point, what I'd do is warp ahead by quite a bit, then repeat the above process.  I probably wouldn't go for the gusto and advance Kerbin almost-90-degrees in one whack, just 'coz it would be kind of annoying if I overshot the window.    So I'd warp, say, two-thirds of what I think I need to do, then re-do the above process to verify that it's looking good, and kinda work my way in from there.

Note that the above description is simplifying things a bit-- for example, it's assuming that you're pretty close to coplanar with the target-- but it works pretty well for most cases.

(But honestly, what I'd really do is just go to http://ksp.olex.biz and get the answer from there in a couple of clicks.) 

 

3 hours ago, king of nowhere said:

- the cochlea manuever: I don't know it it already has an official name.

I believe the term you're looking for is bi-elliptic transfer.

[OhioBob 3,473]

The first thing we must be able to do to travel interplanetary is to recognize when the time is right to launch our mission.  The time when a mission can be launched is called a launch window.  Launch windows occur at periodic intervals when the planets are correctly positioned relative to each other.  For instance, launch windows to Duna occur about every two years.

For a typical interplanetary mission, we attempt to intercept the target planet when it is approximately 180 degrees from that launch point.  That is, let’s say we have a clock face with the Sun located at the center, and Kerbin located at the 6:00 position.  Our spacecraft will travel around the Sun counterclockwise and intercept the target planet near the 12:00 position.  (In practice the intercept point will likely be somewhere between 1:00 and 11:00, but as close to 12:00 as possible.)  For this to work, it is vitally important that the target planet be located at the 12:00 position when our spacecraft arrives there.  To make this happen, we must launch when the target planet is in a specific location relative to Kerbin.

For an inner planet, such as Eve, we must launch when the target planet trails Kerbin in its orbit.  This is because the inner planet is in a smaller orbit and is traveling faster than the spacecraft.  So in the time that the spacecraft travels 180°, the inner planet will travel more than 180°.  Let’s say that in the time it takes the spacecraft to travel 180° the planet travels 225°.  In that case we must launch when the planet is 45° behind Kerbin in its orbit.  In other words, when the planet is at the 7:30 position on the clock face.

For an outer planet, such as Duna, we must launch when the target planet leads Kerbin in its orbit.  This is because the planet is orbiting more slowly than the spacecraft and will sweep through a smaller angle than the spacecraft during its cruise phase.  Let’s say that in the time it takes the spacecraft to travel 180° the planet travels 135°.  In that case we must launch when the planet is 45° ahead of Kerbin in its orbit.  That is, when the planet is at the 4:30 position.

The angle between Kerbin and the target planet at the time of launch is called the phase angle.  For Eve and Duna, the correct phase angles are roughly -45° (behind Kerbin) and +45° (ahead of Kerbin) respectively.  Unfortunately those numbers vary because the planets’ orbits aren’t perfectly circular.  The orbital speeds of the planets vary, as does the distance the spacecraft must travel to reach the intercept point.  Therefore the phase angle may be more or less for a specific launch window.

Using the phase angles above should be close enough to get an intercept with the target planet, but it may not be the most efficient transfer.  Calculating the exact phase angle for a specific launch window is more complicated than I care to explain here, but if you want to know more about it, additional reading is available here (warning - lots of math):

http://www.braeunig.us/space/interpl.htm

The easiest way to determine launch windows and phase angles in KSP is to use the mod Transfer Window Planner, or its online version here:

KSP Launch Window Planner by alexmoon.
 

Edited February 16 by OhioBob

[Brikoleur 4,026]

Alexmoon or KAC FTW. It's frankly a bit silly that the feature isn't available in stock because it really is so essential for any interplanetary mission.

I did my first ones by eyeballing, and that was just silly, made those things completely unnecessarily more difficult.

There are a few destinations however that are different, notably Moho, Dres, and Eeloo. The orbits are so inclined that with them I just eyeball the An or Dn relative to the body I'm leaving, and then burn for the opposite one; then I burn there to get an intercept.

[Hotel26 2,202]

45 minutes ago, Brikoleur said:

The orbits are so inclined that with them I just eyeball the An or Dn relative to the body I'm leaving, and then burn for the opposite one; then I burn there to get an intercept.

An interesting approach (and I am by no means an expert on Moho).  If I understand you correctly, noting that Moho and Kerbin share a mutual AN and DN  (of course), you depart Kerbin at one of those, lowering your PE (at the opposite node) to the orbital altitude there of Moho.  Burn there to reduce AP to get an intercept with Moho one orbit later.  At this point you've made no inclination change (except possibly trying to incorporate one in the Kerbin departure burn (very difficult to judge)).  The idea then is to enter the Moho SOI and handle the inclination there (at much lower relative speed)?  Have I understood that correctly?

If so, I would add the following: one of those nodes has a lower solar altitude than the other and the relative speeds at intercept will be quite a bit lower than at the node at the higher altitude.  (In fact, intercept at Moho's perihelion is the optimal point.)  And since Moho's orbit is 102d, it'd only be a max of another 51d to wait for the better node.  However, this does not detract from you general point, esp. as Dres and Eeloo have much longer orbits.

The interesting take-away is to stage the intercept without bothering to remove the inclination beforehand??

Edited February 16 by Hotel26

[Brikoleur 4,026]

Yep, correct. It's the same way I intercept Minmus -- or RV with a spacecraft, essentially. And yes, I don't do an orbital plane change, that's mega expensive for Moho. This isn't the cheapest possible way to do it but it's not horrendously expensive and it's pretty easy. 

Lately I stop by Eve -- capture in a highly elliptical orbit, then burn for Moho at the Eve/Moho An or Dn. That saves a fair bit of dV. I'm not good enough with gravity assists just to slingshot straight off Eve at Moho, although I am assured that that can be done too if you're a real daredevil at orbital mechanics.

[Spricigo 1,146]

4 hours ago, Brikoleur said:

I did my first ones by eyeballing, and that was just silly, made those things completely unnecessarily more difficult.

Similar experience there, I did the hard way a few times to prove to myself that I can, then I started to use mods and other resources to make it easy.

3 hours ago, Brikoleur said:

Lately I stop by Eve -- capture in a highly elliptical orbit, then burn for Moho at the Eve/Moho An or Dn

 Kerbin<>powered assist at Eve<>Moho is my preferred method. But I screw up sometimes.

[king of nowhere 653]

to find out the launch window, i generally use alexmoon planner. fairly straightforward.

some other tricks i learned when this does not work so well, though...

- orbital plane management: matching orbital planes is straightforward  but expensive. it is often cheaper to make a smaller manuever to move the ascending and descending nodes so that you will meet the planet on a node.

- the moon-to-planet slingshot: when you are orbiting a moon, you can fall down on the planet and make a burn at planet periapsis for an interplanetary destination. you will get the full oberth effect from the planet, while still keeping most of the energy you spent to reach the moon in the first place. for this same reason, if you're using a mothership-and-shuttles approach, it is convenient to park the mothership around moons

- the cochlea manuever: I don't know it it already has an official name. It's a way to reach a planet when you are too late for a transfer window; it's more expensive than launching in the window, but it is cheaper than a normal transfer out of window. to make a transfer after the window, if you go for an inner planet, you have to lower your periapsis a lot, so your orbit will be faster and you'll catch the planet on your way back. this results in high cost for lowering periapsis, and high intercept speed. the cochlea manuever makes an additional burn at periapsis to lower apoapsis. this allows for a higher periapsis, and a lower intercept speed, making up for the extra burn. when going to an external planet, you start with apoapsis higher than the planet and then you raise periapsis

- the two stage transfer: I am sure this one has an official name; when you are way off a transfer window, you make a transfer manuever to touch the target orbit, and then you make a burn there to change your orbital time and meet the target planet on the next orbit. it is basically a rendez-vous manuever applied to a planet. this manuever is very useful when using gravity assist, because the planets will rarely be aligned right to reach your target immediately. it is also useful when you have a radiation mod and you are in a hurry to leavethe jool radiation belt, but it's not a good time to transfer.

- the duna aerobrake assist: when going from an outer planet to an inner planet or viceversa, it is sometimes convenient to stop at duna first, on an ike orbit (see moon-to-planet slingshot): using ike for the moon-to-planet slingshot ensures that you don't lose too much energy in the stop. the benefit is that you can insert into duna orbit for free with aerobraking (if your ship can do it, at least), and this will let you cancel a lot of high apoapsis or plane change for free. for example, when coming back from dres, i had 1 km/s of speed along the plane that i would have had to pay with rocket braking at kerbin; instead, at duna i was able to aerobrake that speed away, and then i had a much smaller intercept speed when going to kerbin.

I'm  sure i know other tricks, but those are the ones i can think of right now

[Hotel26 2,202]

3 hours ago, king of nowhere said:

the cochlea manuever: I don't know it it already has an official name.

This one sounds like a high-energy transfer...?  (I can't find a better reference for the moment.)  It spends energy (more than the Hohmann) to get there sooner than waiting for a transfer window.

 

[Snark 14,343]

7 minutes ago, Hotel26 said:

(I can't find a better reference for the moment.)

(see immediately above at the end of my post)

[Hotel26 2,202]

20 minutes ago, Snark said:

(see immediately above at the end of my post)

Great tip in your post, by the way, Snark!

My reading of the bi-elliptic transfer is that it is a low-energy transfer in a particular case, in fact, in which it takes longer to arrive but pays off by using *less*  energy than the Hohmann.  It would have its own transfer window, so wouldn't be the way to "go anytime

4 hours ago, king of nowhere said:

but it is cheaper than a normal transfer out of window

Therefore, I think you may be right.  The above  statement from the poster is the litmus test.  Everything else in the maneuver sounds more like a high-energy transfer.  We'd have to ask the poster whether the intention in his case was to "go anytime" or arrive cheaper.

But he does also say this: "it's more expensive than launching in the window".

Perhaps the resolution is that "bi-elliptic transfer" is, like, "high-energy transfer", a very generic term, implying nothing detailed about application.

Your reference does say: may, in certain situations, require less delta-v than a Hohmann transfer maneuver"; keyword being 'may'.  However, it does also specify that the entry to the second eliipse is performed at the apoapsis of both.  In a HET, you have the option to ride around the inner trajectory (lowering AP below the target orbital altitude) for a while to catch up, which is its value.

We got to discuss both, so that is good.  (More on HET: Banish transfer windows.)

Edited February 16 by Hotel26

[Snark 14,343]

2 minutes ago, Hotel26 said:

My reading of the bi-elliptic transfer is that it is a low-energy transfer in a particular case, in fact, in which it takes longer to arrive but pays off by using *less*  energy than the Hohmann.  It would have its own transfer window, so wouldn't be the way to "go anytime

Not quite go any time, no.

It does, however, widen the options for when you can do a reasonable launch.  And if you're well outside the window for a Hohmann transfer, then a bi-elliptic can be substantially cheaper than a direct-from-origin-to-target-in-one-burn trajectory.

 

[king of nowhere 653]

1 hour ago, Snark said:

I believe the term you're looking for is bi-elliptic transfer.

 

1 hour ago, Hotel26 said:

This one sounds like a high-energy transfer...?  (I can't find a better reference for the moment.)  It spends energy (more than the Hohmann) to get there sooner than waiting for a transfer window.

 

it's definitely not a bi-elliptic transfer (i forgot to mention those). it is a form of high energy transfer, because it reduces time at the expence of fuel. it is more expensive than a regular hohmann, but still within a reasonable budget.

unfortunately, i have the game open with a very big ship that takes several minutes to load, so i don't want to exit it, or i'd snap some pictures. let's try to describe an example.

say i want to go to from Kerbin to Eve, but i missed the transfer window. in a transfer window, i lower solar periapsis to 9.8 M km, and when i reach periapsis, there i find eve.

but not now. since i missed the window, now if i do that i reach periapsis and eve is still ahead. to reach eve, i would have to lower my periapsis to, maybe, 6 M km, this way my orbit will be faster, and after reaching periapsis i will go on towards apoapsis and will intercept eve while crossing its orbit. but this kind of setup results in an insanely expensive intercept.

the cochlea manuever instead lowers periapsis to about 7 M km, and then at that periapsis it makes an apoapsis lowering manuever to 9.8 M km. and then, if it the periapsis was set correctly, i will meet with eve at apoapsis. i call it cochlea manuever because my trajectory looks like a spiral.

this manuever has a fairly expensive burn in solar orbit, but the intercept is very cheap, and it does work after the regular transfer window has passed.

Edited February 17 by king of nowhere

[Hotel26 2,202]

15 minutes ago, king of nowhere said:

bi-elliptic transfer

(I was just starting to update my post above to abbreviate/make it more helpful/productive -- and clearer about Snark being correct -- but I'll let it stand.)

In the general case, a bi-elliptic is simply a series of two Hohmann transfers, but linked at the common apoapsis/periapsis.

I think a bi-elliptic to a lower orbit is always high-energy (meaning, more than a Hohmann).  And a bi-elliptic to a higher orbit is mostly always high-energy  (and a waste of time because you could have waited at home for the window and got there sooner), except in the special case described in Snark's reference, in which it works out better (making it "low energy", as an exception).

I think Snark's term may fit the maneuvers you describe more closely (especially given your description, "cochlea").  But maybe "high-energy transfer" describes the intentions, trade-off and costs usefully.

Edited February 17 by Hotel26

[paul_c 424]

A very general/vague rule of thumb I use, is that if the ratio of the radii of the orbits are >12, its worth considering a bi-elliptic (to change the size of the orbit on roughly the same inclination). 

(Also if you want to do a BIG inclination change, like 60deg or more...)

Neither of these cases occurs with Kerbin-->another planet, so, its not considered for interplanetary.

[Scarecrow71 1]

After spending countless hours trying to eyeball transfer windows and calculate dV required and determining correct ejection angle...I decided to use MechJeb to do this for me.  In a nutshell, using the Maneuver Node creator in MJ does the following:

1. Determines the appropriate time to do the burn to transfer to another planet;
2. Determines the correct point on your current orbit to place the created node;
3. Determines the correct amount of dV to burn;
4. Determines the correct amount of time to do the burn.

I can literally save my game, launch a vessel and use MJ to calculate the transfer, and then write down all of the information I need.  Then I reload the saved game and not only fast-forward to the appropriate time, but then make sure the vessel I'm crafting has enough "oomph" to get to the target.  Not necessarily there and back, but at least there; I don't care if I strand a Kerb or three on some remote rock for a while because hiring new astronauts is cheap enough AND you can always go pick them up.  If you remember you stranded them in the first place.  Which I seldom do until I run out of pilots and then go look in the AC and see that I've got 5 of them out there in space somewhere.

I realize that, for some people, it's not in the spirit of the game to get the calculations and then load a save to get it right.  But, this is what works for me, and it's how I like to play.  Your mileage may vary.  Greatly, in some cases.

[jimmymcgoochie 1,684]

I just use MechJeb’s advanced transfer maneuver planner mode to do all the calculations for me, make a nice pork chop plot to show the combinations of time until burn, time to destination and delta-V required and then pick what transfer I want to go for; sometimes there are high-energy transfers that are much faster to get to the target at a cost of a bit more fuel, which work well if you want a probe flyby, or you can pick the ideal point to launch that will save every drop of fuel by using the most efficient timing for your transfer, both with and without considering the capture burn. A few clicks and it’s all done for you, node created exactly where and when it’s needed plus MJ can execute the node for you with reasonable accuracy (longer burns are inherently less accurate no matter who does them).

[paul_c 424]

5 hours ago, Scarecrow71 said:

; I don't care if I strand a Kerb or three on some remote rock for a while because hiring new astronauts is cheap enough AND you can always go pick them up.  If you remember you stranded them in the first place.  Which I seldom do until I run out of pilots and then go look in the AC and see that I've got 5 of them out there in space somewhere.

I realize that, for some people, it's not in the spirit of the game to get the calculations and then load a save to get it right.  But, this is what works for me, and it's how I like to play.  Your mileage may vary.  Greatly, in some cases.

I got so adept at "rescue" missions, I ended up with 28 Kerbals in the 12 capacity Astronaut Complex. If I lose one of these 'rescues' on a distant planet, its even?

[MechBFP 1,132]

On 2/15/2021 at 7:24 PM, Spricigo said:

Figuring out the correct time to depart to another planet is not a trivial task. Yes, some veteran make it looks easy, but most us struggled with this at some time. ..some still have trouble.

So, what you can tell about transfers windows?

What is your technique to find it out? Or you rely on some tool? Or are you still struggling with it and have your own questions about it? In any case, you are welcome to share your thoughts. 

This is not a typical question, of someone looking for help with an specific issue it rather meant to be an opportunities, for some people learn about an important thing in KSP, for other to share what they know about it. I hope to be right in my assumption that both are things this community enjoy.

My preference is to use the "Transfer Window Planner" mod to find launch windows. Lately I have also been using MechJeb to auto-create the maneuver node as well as I have simply got tired of fiddling with maneuver nodes to get those interstellar transfers. Due to the inaccuracy of manual burns it also requires planning another correction node on the way there and its just a "create a work project" at this point for me so I try to avoid doing it. For me the challenge is designing/building a rocket, launching it, landing it, and completing the goal of the mission. The time spent in space to get there has just become tedious for me so I avoid it as much as I can.

Edited February 18 by MechBFP

[chd 28]

Admittedly I've done this by hand (using the visual approach with maneuver nodes that Snark described) a few times, but now I just use MechJeb to calculate a porkchop plot for me. If you go that route, a couple of things to be aware of...

If your burn has a staging change it in, or if it's very long because your thrust is low, the result may not be all that accurate and you may need to do it twice. (I usually wait until I'm in the sun's SoI and then do an additional Hohmann transfer maneuver from there.)

If you have some reason to want to choose a less optimal departure window than the ideal (you need to leave sooner for some reason, for example), you'll need to take more fuel than might be immediately obvious. Not only will you need more dV as shown in MJ's calculation, but you'll need even more than that because you'll enter the SOI of your target moving quite a bit faster relative to it than you would have in the optimal window. So you'll need extra dV to slow down and capture as well.