To Duna and Back - Out of Phase and as quick as possible

[WanderingKid]

Hey all.  I've poked around a bit and I'm having a hard time figuring out something, and I'm quite sure someone's pulled this off before.  It's particularly about Duna, but it's a more generic question for roughly all the planets for a career mode.

In career mode, I've got a Duna First World's Contract pop up.  They show up eventually, of course, but as usual these usually include a 'and return to Kerbin' clause, so you can't pick up any more until you've swung past X planet, then came home.  Now, my plan is to stuff as much dV into a probe as I can get off my level 2 Launchpad and fling it at Duna out of phase for as quick a possible flyby.  The idea being when I'm ready to send a serious mission (IE: Land on Duna) I'll be IN phase (or closer) and the mothership I end up sending for that won't require over-engineering due to phase positions.  That's not the hard part, it just takes extreme dV and really odd departure angles.

Now, where I'm running into trouble is getting back in a timely manner.  My best guesstimate forces me to establish an orbit (which I'm trying NOT to do so I can keep that world's first banked for a contract later) and then reverse fire myself at Kerbin, again at extreme dV cost and odd angles.  As far as I can tell, no gravity assists or anything of that nature will significantly help, as I'm on the right Solar Periapsis for Kerbin, I'm just WAY off phase.  EDIT: This was misleading.  My periapsis may be on Kerbin's plane, but I'll be traveling out to Eloo before I come back, which defeats the point.

I'm really hoping one of you will tell me 'Oh, that's easy, just do this!', and I'll have to noodle on it for a week before I understand it but I can get it done.

I have this random idea that if I can bank myself around Duna, without actually orbiting but coming in for an Oberth, and then burning like mad at the right angle to push the phase around I might be able to pull it off, but I just can't figure out how that should be done.

Edited February 16, 2017 by WanderingKid

[Streetwind]

31 minutes ago, WanderingKid said:

I'm really hoping one of you will tell me 'Oh, that's easy, just do this!', and I'll have to noodle on it for a week before I understand it but I can get it done.

I'm afraid to tell you, good sir, that while what you are trying to do is certainly possible (dV budget permitting), it's no easy feat. You will need to do maths and guesstimate some values and use a mod and use an external tool to pull it off, and you will have to do it on the fly - you cannot plan the return transfer before you have executed the transfer burn to Duna. Which is why I said "dV budget permitting". There's a chance that you'll find yourself without enough dV left for an expedited return after you make your Duna transfer burn, because the cost of return depends 90% on the initial transfer. Therefore, you need to make a best guess at how much to throw at the initial transfer to accelerate it, and how much to keep in hopes of being able to also accelerate the return. There exists no tool for KSP which can precalculate this for you - I recommend a named quicksave for trial and error purposes.

On to the how-to, then:

1.) Go to https://alexmoon.github.io/ksp/

This tool is available as an ingame mod, but you'll be using the website, because you're going to need its advanced features later, which the ingame tool doesn't have. For starters, use it to plan a high energy transfer to Duna suitable for your desired departure date and dV budget. Make sure your probe is in a circular, equatorial Kerbin orbit, and set the Initial Orbit altitude accordingly. Check the "no insertion burn" box since you're planning a flyby, set Earliest Departure to your current ingame date, select "Optimal" from the transfer type dropdown menu, click on Show Advanced Settings, and enter a Latest Departure to your taste. Setting this latest departure date close by allows you to constrain the result space and keeps the planner from suggesting the next ideal window instead of a suitable near-term solution.

2.) Let it calculate a solution. In the colorful plot at the bottom, travel time is on the vertical axis, and departure date is on the horizontal one. Mouse over the plot; you'll see a dV cost next to your mouse cursor that represents the particular date/traveltime combination you're hovering over. Find a solution that's suitable for your needs. If you want, you can adjust your latest departure time and the travel time window in the planner settings and recalculate, to make the plot more focused on specific areas. When you have found your solution, click on it, then press "Refine Transfer" at the bottom.

3.) Now, make a maneuver node using the information given and execute it.

4.) Your probe is now on a flyby trajectory past Duna, putting it into a solar orbit afterwards. It is this solar orbit from which you must return to Kerbin. Therefore, you need to configure Transfer Window Planner to recognise this orbit. This is where it gets complicated.

Press "Add Body". Switch body type to Vessel, give it a name. Note down what the solar apopasis and periapsis altitudes will be after your Duna flyby. Select the Sun in map mode, pull up its info window, and look for the radius. Add the Sun's radius to both apoapsis and periapsis, and use only the modified numbers for all further colaculations. Calculate your semimajor axis by adding together your apses and dividing by 2. Also calculate your eccentricity = 2 / ((apoapsis height / periapsis height) +1). Put these numbers into the planner. Mouse over the periapsis, note how much time it takes to arrive there, and add that time to the current ingame date. Try to be precise despite the numbers constantly changing. Enter the date you calculated under "Time of Periapsis Passage" in the format "year/day hh:mm:ss" - for example "1/235 04:53:18".

Inclination is a problem. You can easily determine your current inclination after your departure burn by targeting the Mun, which is perfectly equatorial; but, your Duna flyby will change your inclination. Especially if it's a close flyby, and especially especially if Ike is getting in the way (which you know it will). I think you'll have no other choice than to check your current inclination against the Mun, then eyeball how your post-flyby solar orbit differs from your transfer orbit. I know it's easier said than done, but try to be precise.

That leaves two fields: the longitude of ascending node, and the argument of periapsis. Duna has an inclination relative to Kerbin, so your orbit does too, so they must be specified. Unfortunately, there is no way for you to find them for your final solar orbit. The best you can do is to use a mod of your choice (KER, MechJeb, etc) to display them for your current transfer orbit. They will be wrong for the final orbit, especially for close Duna/Ike encounters, but there's no way around that which I know of. And ideally the error is small enough to not influence your final return trajectory cost by more than a few hundred m/s. It's because of these two errors that I told you to be as precise as possible with the other values... you want to avoid cumulative errors growing out of control.

5.) Save your custom vessel orbit, and select it as origin. Set destination to Kerbin, and check No Insertion Burn (you don't have dV to spare). Look at your probe's trajectory, find the time to encounter Duna's SoI, add that to your current ingame date, and set that as Earliest Departure. Set Latest Departure a hundred days later. Transfer Type optimal, and calculate.

Now is the big moment: the planner will have preselected the cheapest possible route in this time window. Is the dV cost within your probe's remaining dV budget? If yes, you can attempt to find the shortest possible transfer your budget allows for, tightening the Latest Departure and Travel Time settings for more precision. If not... well, you can attempt to relax the settings to see if you can return at all. But most likely, you'll have to reload and do the whole thing all over again with less dV spent on the initial transfer.

6.) If you have a result: try to make a node at the appropriate time. Do you get a Kerbin encounter? I'm willing to bet you don't, because your origin orbit wasn't set up correctly. The question is, what (and how much) do you have to change to get your encounter? This is left as an exercise to you; I cannot give advice here. You must rely on your skills in maneuver node editing. Once you have your encounter, check again: can you still afford it? Remember, just an encounter alone isn't good enough, it must also intersect Kerbin's atmosphere (or leave enough fuel for a later course correction to that effect). If not, try moving the node to earlier or later in the orbit, and see if that makes it cheaper. When all else fails, reload and repeat.

[bewing]

The way I figure it goes like this:

A Hohmann xfer from Kerbin to Duna is an ellipse with the Pe at Kerbin's orbit, and the Ap at Duna's orbit. Therefore, if you pick up no gravity slingshot energy at all at Duna, then you come back to a Pe at Kerbin's orbit again.

You are going for a less-efficient version. So your Pe is going to be somewhat inside Kerbin's orbit -- with the Ap still at (or a little beyond) Duna. If you want your return leg to get you back to Kerbin, then you only want to pick up a little energy at Duna. A deep Oberth dive will get you a lot of energy, and put you in totally the wrong orbit. Ike, of course, causes issues if you get too close -- but in this case, you might be able to keep your Pe above Ike's orbit.

Once you get back out of Duna's SOI, you can use a small Kerbol "surface" radial-in or -out burn to change the timing so that you intercept Kerbin again. It's going to be a fairly high-speed intercept, but you can use a deep Munar Oberth dive to slow you down a bunch.

 

[Streetwind]

@bewing The way I read his request, he wants to go really fast. His AP is going to be beyond Dres' orbit, possibly beyond Jool's. And while it's perfectly possible to use a gravity assist with a bit of correction burning to put his PE in a place where he will encounter Kerbin again eventually, he doesn't want to wait. He wants to propulsively reverse his direction inside Duna's SOI, so he goes right back to Kerbin directly - without getting carried into the outer solar system for a year or two. He wants to chain two high energy transfers back to back without orbit capture in between, basically. And for that, he needs more than just a PE adjustment.

Generally sound advice though to only skim the outskirts of Duna's SOI. It minimizes errors in the approach I selected, too. It means no low orbit science, though, so if getting that is part of the plan, he'll have to go in close and personal...

Edited February 16, 2017 by Streetwind

[Reusables]

What about high speed rocket powered airplane skimming through Duna's atmosphere? I think the escape velocity is small enough to allow the airplane flight, and just reverse lift is needed. Also considering L/D ratio(>1), it will certainly help the direction change manu ever.

[Streetwind]

40 minutes ago, Abastro said:

What about high speed rocket powered airplane skimming through Duna's atmosphere? I think the escape velocity is small enough to allow the airplane flight, and just reverse lift is needed. Also considering L/D ratio(>1), it will certainly help the direction change manu ever.

Hmmm... I wonder. Doing such a thing would make the return transfer essentially unplannable, since you have an unpredictable direction change at Duna. But technically, aerobraking at Duna would help. No need for a plane, just aerobraking with a heatshield. It'll drop solar apoapsis somewhat, which makes the return burn cheaper. However, you then have to wait until after you've exited Duna's SOI until you can use Transfer Window Planner to plot a course, which loses you all Oberth help. I can't say if aerobraking would save you more than the Oberth effect. Of course, there's always the possibility to manually eyeball the return transfer... but since it's a highly unusual trajectory, this is hard to pull off. There's no way to know the correct burn angle, for instance, apart from tediously trial-and-erroring your way towards it.

All in all I'm not strictly sure it's a good idea... but I could well be wrong.

EDIT: Actually, now that I think about it. You're changing frames of reference on an SOI boundary, but does that actually affect the readout of orbital parameters like argument of periapsis by info mods, if you're already on an escape trajectory? I've made that assumption but I've never checked if it's true. If it was not true, then you could plan the return transfer with TWP right after coming out of Duna aerobraking, which would be ideal.

Edited February 16, 2017 by Streetwind

[Reusables]

43 minutes ago, Streetwind said:

Hmmm... I wonder. Doing such a thing would make the return transfer essentially unplannable, since you have an unpredictable direction change at Duna.

Using control surfaces, I think it's possible to control the ejection direction. Aside from that, I got that getting return transfer will be tricky.

Edited February 16, 2017 by Reusables
Edited a bit

[Streetwind]

28 minutes ago, Abastro said:

Using control surfaces, I think it's possible to control the ejection direction. Aside from that, I got that getting return transfer will be tricky.

Yes, you can control the direction, but you don't know which direction you should aim at. Because you cannot predict the exact amount of braking force you experience while going through the atmosphere, and therefore, you cannot predict the shape of your post-flyby orbit. Whch is the thing you need to determine the direction you need to go.

This is also why aerobraking alone would do the trick. There's little need to steer when you don't know where you want to go until after you cannot steer anymore. Instead, just go for maximum drag and then compute the way home afterwards. But that only works if you can see the argument of periapsis and longitude of AN of your solar orbit while still in Duna's SOI. And due to the patched conics approach, I'm not sure you can (but will accept empiric evidence to this end).

Edited February 16, 2017 by Streetwind

[Kryxal]

I'd give serious thought to an Ike gravity assist, to set up for the return ... but that would get really complicated.

[WanderingKid]

5 hours ago, Streetwind said:

@bewing The way I read his request, he wants to go really fast. His AP is going to be beyond Dres' orbit, possibly beyond Jool's. And while it's perfectly possible to use a gravity assist with a bit of correction burning to put his PE in a place where he will encounter Kerbin again eventually, he doesn't want to wait. He wants to propulsively reverse his direction inside Duna's SOI, so he goes right back to Kerbin directly - without getting carried into the outer solar system for a year or two. He wants to chain two high energy transfers back to back without orbit capture in between, basically. And for that, he needs more than just a PE adjustment.

Generally sound advice though to only skim the outskirts of Duna's SOI. It minimizes errors in the approach I selected, too. It means no low orbit science, though, so if getting that is part of the plan, he'll have to go in close and personal...

This is absolutely correct.  The angle of approach combined with the speed of the transfer is enough to send me interstellar depending on exactly when/how I do it.  Due to the nature of the problem (flyby of planet only) the low vs. high science is of little consequence to me.  This is essentially a funds run, not a science adventure.

I want to thank you for your highly detailed walkthrough of using the AlexMoon's transfer calculator.  I've used it before, but never to that detail level.  It's going to take me a few read-throughs and some practice to fully understand the moving parts, yet I figured that was coming. XD

7 hours ago, bewing said:

You are going for a less-efficient version. So your Pe is going to be somewhat inside Kerbin's orbit -- with the Ap still at (or a little beyond) Duna. If you want your return leg to get you back to Kerbin, then you only want to pick up a little energy at Duna. A deep Oberth dive will get you a lot of energy, and put you in totally the wrong orbit. Ike, of course, causes issues if you get too close -- but in this case, you might be able to keep your Pe above Ike's orbit.

I'll come back with pictures.  This is the ideal scenario, but not the one that's occurring.  It's all because I'm being inefficient with dV and energy to trade for time.  If you want a simpler scenario, think of what the departing orbit for a 4 day Minmus transfer looks like after you leave its SOI.  The mechanics are equivalent, just a lot deeper and lot less forgiving.

[WanderingKid]

So I finally built the Duna Whip, and as promised... pictures!  Here is stage 3 of the Whip after achieving circular orbit ~74.5k.

On day 102 we left Kerbin orbit for Duna, 121 days prior to good window.

After digging around with maneuver nodes, I finally came up with a return to Kerbin at a significant, but acceptable, timetable, for a reasonable dV cost all things considered.  Running at 4,500 dV (before braking at Kerbin), I purposely avoided doing an aerobrake for this attempt so I can see how well I can aim with maneuver nodes and not even worry about what the aerobrake will do to me.  I really want to avoid an accidental orbit.

The first minor node is simply a cleanup of the approach to get close to Duna and reasonably equatorial:

Then, upon entering Duna's SOI, I brake, HARD.

Close to the descending node, I clean up the inclination with a small burn and then get my grav assist to stop aiming for Dres:

Finally, after leaving Duna SOI, I do a major anti-radial burn with some more retrograde tossed in to get myself a new intercept with Kerbin

Finally, 1 year and 61 days later, if all goes well, I'll have a new intercept with Kerbin's SOI.  I'll clean up from there and land hot.

Thanks for the help everyone!

I'm going to keep mucking with the return leg for a while, but figured I'd throw this together as my starting point in case anyone's got some great ideas. 

Edited April 9, 2017 by WanderingKid

[FancyMouse]

My early Duna crew always arrives at ~1y370d or something, and from the TWP plot, I have to depart from Duna by 1y400d, or the fuel penalty will be really high.

A few tricks:

1. When using any TWP tool, make sure that you have equal interval for departure date and trip length - this way, you could draw a 45 degree line to evaluate the best trajectory on the same arrival date.

2. You need to design the ship for both Kerbin return heat (with my trip plan, this one will be ~4km/s at periapsis, need to slow down to at least 3km/s) and Duna aerocapture (this one will be almost 2.5km/s which is also dangerous, and you need to slow to 1.2km/s to be captured).

Edited April 9, 2017 by FancyMouse