Planning Hohmann transfers and getting an encounter

[Lunar Sea]

I've been having a lot of trouble with this lately. Take Eve for example. I wait for the transfer window, have my craft in orbit over Kerbin and try to set a maneuver node. I can never get an encounter going by the book. I always have to either spend extra DV or set another maneuver further down and spend more there. I know inclination of course has to be corrected for at some point but I still feel I'm spending too much. Last time I waited til exactly the alarm clock showed the transfer window and it seemed I wound up just behind the planet I was after. 

Advice on planning encounters when heading out from Kerbin please\?

[Plusck]

I think you just have to accept that you'll very rarely get a perfect encounter. The launch window planners give you the best possible figures, but since you cannot have an infinite TWR you can never spend exactly that amount, even if you managed by miracle to set up a maneuvre node with the exact given values.

The main trick is just to drag the node along the orbit, while watching "through" the sun. Once you get an enounter, reduce the node value until the encounter disappears, then drag the node around a bit more to see if you can get an encounter again. If you can, great, if not, add some more prograde and drag the node around some more.

For Eve (and Jool, and Dres, and Eeloo...) you may need to set an inclination-correcting node halfway there to get an encounter to start with.

Still, the absolute minimum is not necessarily the best bet, since this might require a correction later to get a low-altitude flyby for your capture burn. So once you have the "best" encounter, you really need to switch focus to the target planet to see how close you are, and probably add some more to your burn to get closer to the actual planet.

Ths all tends to get tiresome, so mostly I just plot an approximate minimum burn, drag the node around until I get a hit, then zoom out a little and add to the burn until my post-flyby ejection trajectory appears to be longest and appears to give the most significant change to my orbit. It'll all be slightly different at the end of the burn anyway... so I prefer to spend a little more fuel and avoid spending 10 minutes tearing my hair out.

 

As for the timing issue, that should sort itself out when dragging the ejection burn node along the orbit: if you're late, you'll end up ejecting with a slight angle towards the other planet's orbit, and you'll cross paths with the target earlier. If you're sure about the timing but still seem to be missing the planet, you're probably burning too much in the wrong direction. Setting up a mid-course correction burn then dialling back your initial burn (and dragging along the orbit again) should sort that.

[Superfluous J]

I long abandoned looking at the direction and speed figures that all transfer window planners gave me, and for years now just use them to tell me the TIME of a good transfer. When that time comes (or, using precisenode, long before that time comes) I just drop a node and fiddle until I get an encounter. After a few (dozen) times, you get a feel for it.

Basically, what @Plusck said.

[Geschosskopf]

6 hours ago, Lunar Sea said:

Advice on planning encounters when heading out from Kerbin please\?

An actual Hohmann transfer always requires a mid-course correction burn if the target has an inclined orbit (and is sufficiently far enough away for that to matter).  You do a transfer burn in the plane of your original orbit, then do a plane-change burn when you reach the ascending or descending node of your transfer orbit and the target's orbit.  Thus, when you leave Kerbin, you will never have an encounter with a standard Hohmann tansfer when you start out, unless it just so happens that  the target is at one of its nodes with Kerbin's equator when you get there.

To get an encounter right off the bat from Kerbin orbit, you have to do a ballistic transfer, which combines the orbit-changing and plane-changing burns into one.  This is usually a bit more efficient than a true Hohmann because you're doing the plane-change further from the target, and gives you more confidence from the get-go that you'll actually hit the target.  However, even this will almost certainly require a slight mid-course tweak to fine-tune  your encounter, due to inaccuracies in creating the node, executing the burn, changes in coordinate systems when you leave Kerbin's SOI, and inherent floating point errors in KSP.  However, if you do a reasonably good job on the transfer burn, this tweak burn won't be enough to worry about.

There's an excellent tutorial on doing ballistic transfers here:  

 

[Lunar Sea]

Thanks folks that is really helpful info. Think I will print out this page for a reference sheet So far (interplanetary-wise) I have just done one Duna trip, which was comparatively easy.  Now it's getting trickier. 

Edited May 18, 2016 by Lunar Sea

[Geschosskopf]

14 hours ago, Lunar Sea said:

Thanks folks that is really helpful info. Think I will print out this page for a reference sheet So far (interplanetary-wise) I have just done one Duna trip, which was comparatively easy.  Now it's getting trickier. 

The dirty little secret of interplanetary flight is that if you can get to Minmus, you can get everywhere else (except Moho, which usually requires a different method).  The maneuvers you do are all the same no matter where you go (except Moho).  The only differences are, the burns are bigger, the trips take longer, and you have to wait longer on transfer windows, but how you do the burns is exactly the same.  So have some faith in your abilities (or your Kerbals, if you prefer).  You've got this

What makes interplanetary trips challenging is not the journey (which is, after all, merely "warp-over country") but what you do at the destination.  And that's a challenge you set for yourself depending on what you want to do there.  Do you want just flags and footprints or do you want to conquer the alien planet and populate it with hundreds of Kerbals?  The latter (or similar grandiose schemes) is what folks end up doing once they figure out just getting there is really no harder than going to Minmus

[OhioBob]

It might help to supplement Geschosskopf's explanation with some pictures.

On ‎5‎/‎18‎/‎2016 at 0:27 AM, Geschosskopf said:

An actual Hohmann transfer always requires a mid-course correction burn if the target has an inclined orbit (and is sufficiently far enough away for that to matter).  You do a transfer burn in the plane of your original orbit, then do a plane-change burn when you reach the ascending or descending node of your transfer orbit and the target's orbit.  Thus, when you leave Kerbin, you will never have an encounter with a standard Hohmann tansfer when you start out, unless it just so happens that  the target is at one of its nodes with Kerbin's equator when you get there.

On ‎5‎/‎18‎/‎2016 at 0:27 AM, Geschosskopf said:

To get an encounter right off the bat from Kerbin orbit, you have to do a ballistic transfer, which combines the orbit-changing and plane-changing burns into one.  This is usually a bit more efficient than a true Hohmann because you're doing the plane-change further from the target, and gives you more confidence from the get-go that you'll actually hit the target.  However, even this will almost certainly require a slight mid-course tweak to fine-tune  your encounter, due to inaccuracies in creating the node, executing the burn, changes in coordinate systems when you leave Kerbin's SOI, and inherent floating point errors in KSP.  However, if you do a reasonably good job on the transfer burn, this tweak burn won't be enough to worry about.

[Lunar Sea]

11 hours ago, Geschosskopf said:

The dirty little secret of interplanetary flight is that if you can get to Minmus, you can get everywhere else (except Moho, which usually requires a different method).  The maneuvers you do are all the same no matter where you go (except Moho).

Tell me about this... Moho. 

[moogoob]

I just have to add two things:

1. Minmus is indeed great training for approaching inclined orbits, is you leave Kerbin from an equatorial orbit.

2. I'm spoiled by Mechjeb's Porkchop-plot planning for interplanetary transfers. It's the one maneuver I never plan manually.

[FullMetalMachinist]

3 hours ago, Lunar Sea said:

Tell me about this... Moho. 

Basically, Moho has three things going against it when it comes to trying to get there with minimal dV. 1) It's close to the sun, so the burn to get your solar PE that low is pretty expensive. 2) It's orbit is fairly elliptical, so it matters a lot where it is in it's orbit when you try to intercept. And 3) It has a pretty inclined orbit, so same problem as #2.

What all this means is that traditional transfer windows don't work so well. 

[Geschosskopf]

10 hours ago, Lunar Sea said:

Tell me about this... Moho. 

@FullMetalMachinist defined the problem.  Those factors make it so that while you CAN of course do a Hohmann or ballistic transfer to Moho, doing it right requires a higher degree of precision both in node placement/creation and then execution than the inherent inaccuracies of the game itself, must less the player, are usually capable of.  A very slight error, which would amount to nothing going anywhere else, will totally hose you at Moho.  Not that you won't get there, but you'll get there crossing Moho's orbit at a steep angle instead of being almost parallel with it, and that means your capture burn will be huge.  5000-7000m/s is not uncommonly reported.  That's why it's usually better to use a different method when going to Moho.

The consensus of the community is that you do it in 2 steps.  First, you totally ignore all traditional transfer windows.  Instead, you leave Kerbin only when Kerbin is at either the ascending or descending node with Moho's orbit.  The ultimate goal is to encounter Moho at or close to its other node on the opposite side of the sun from where you left Kerbin.  If you can hit Moho there, you don't need to worry about the huge plane change needed to match Moho's inclination.

So, you leave Kerbin with your new solar Pe at one of Moho's nodes.  Because you totally ignored Moho's own position in its orbit when you left Kerbin, odds are it won't be anywhere close to your Pe when you get there.  But no problem.  Once at your solar Pe, retro burn to pull your Ap down somewhere between the orbits of Eve and Moho.  Then you go around the sun 1 or more times until Moho's shorter orbital period eventually makes it meet up with you at or near your Pe (which is at or near one of its nodes with your orbit.  You'll probably have to do a relatively small burn near your Ap at some point to turn a distant closest approach into an encounter, which won't usually be quite on the node, but close enough for government work.  Once you have this encounter, the capture burn will be quite reasonable.

[Warzouz]

On 17/5/2016 at 0:48 AM, 5thHorseman said:

I long abandoned looking at the direction and speed figures that all transfer window planners gave me, and for years now just use them to tell me the TIME of a good transfer. When that time comes (or, using precisenode, long before that time comes) I just drop a node and fiddle until I get an encounter. After a few (dozen) times, you get a feel for it.

Basically, what @Plusck said.

Agreed. The default manoeuvring node is very hard to deal with interplanetary encounters.

Here is how I do

• Get the approximative window using KAC
• Launch ship a week before
• Set a node anywhere (don't bother about ejection angle)
• Delay the node to be in the window (in MJ you can add a number of days)
• Set a prograde burn (any value)
• Delay the node 10s in 10s to get the orbit as far at possible. When you orbit is very far, you ARE at the correct ejection angle, even if not knowing it)
• Tweak prograde to touch target orbit.
• Finally tweak with radial or normal to get an encounter. It's usually better to do a second node at AN/DN

I don't use MJ calculator either but I use "Node Editor" MJ's feature (equivalent to precise node)

[Chaos_Klaus]

4 hours ago, Warzouz said:

Delay the node 10s in 10s to get the orbit as far at possible. When you orbit is very far, you ARE at the correct ejection angle, even if not knowing it)

This is really the main point many beginners do wrong. It does matter in which direction you leave kerbin. For a hohmann transfer you need to make sure that you leave the Kerbin SoI in parallel with Kerbin's orbit! That is equivalent with what you wrote. 

If you love this up and leave in another direction, you will not get an encounter. If you try to correct that by adding more delta v to the burn, you'll love up the timing of the intercept. It's bad. You really need to care about the ejection angle.

[Warzouz]

2 hours ago, Chaos_Klaus said:

This is really the main point many beginners do wrong. It does matter in which direction you leave kerbin. For a hohmann transfer you need to make sure that you leave the Kerbin SoI in parallel with Kerbin's orbit! That is equivalent with what you wrote. 

If you love this up and leave in another direction, you will not get an encounter. If you try to correct that by adding more delta v to the burn, you'll love up the timing of the intercept. It's bad. You really need to care about the ejection angle.

Yes but my technique doesn't require the player to know about ejection angle.

[Superfluous J]

3 hours ago, Warzouz said:

Yes but my technique doesn't require the player to know about ejection angle.

Much like how catching a baseball doesn't require integrating cosines of velocity vectors. It's why the method works so well: You only need to do it a couple times until you "get it" and then it's second nature.

Edited May 20, 2016 by 5thHorseman

[Warzouz]

9 hours ago, 5thHorseman said:

Much like how catching a baseball doesn't require integrating cosines of velocity vectors. It's why the method works so well: You only need to do it a couple times until you "get it" and then it's second nature.

Also this technique requires to have a proper Node editor too which is able to change the time of the node easily, not like stock...

[Chaos_Klaus]

On 20.5.2016 at 8:27 PM, Warzouz said:

Yes but my technique doesn't require the player to know about ejection angle.

Well ... you just need some visual cue. I'd say getting your SoI departure parallel with the planets orbit is a better que then the maximal apoapsis hight, because apoapsis changes very little when you are close to the optimal ejection angle. However, it's quite easy to see whether you are leaving the SoI in the right direction.

None of this requires anything that is more complaicated then the stock nodes. People are just fat fingering them all the time. They can actually be tuned pretty well with the mousewheel one of the two directions is making smaller adjustments then the other, so you set it to roughly the right point and then dail it back with the more precise control.

The only place where this is not precise enough is whith scaled up solar systems.