Help with Eeloo Transfer

[Pahimarus]

**Update added to thread**

 

Hi all,

Playing the deepest into career mode that I've ever gotten and I'm trying to pull off an Eeloo intercept and failing, miserably. I've read a bunch of guides and posts but I haven't been able to sort out my issues with them.

I basically have three questions:

1.) How to get to Eeloo generally but ideally under my conditions.

2.) How to match the maneuver node trajectory more precisely.

3.) Is there a better way (mod?) to more precisely plot interplanetary transfers?

What I have attempted:

It is day 257 02:00 I am within 1 hour of the transfer window from Kerbal alarm clock. One day late for the optimal transfer window in AlexMoon's launch planner. 

I plot a prograde burn with the listed DV (2054m/s). I then eyeball the ejection angle listed (109 degrees to prograde). 

This generates and change of SOI with jool past apoapsis. I fiddle the DV slightly and manage generate an intercept with Eeloo at 25 million km closest approach. This seems far so I fiddle for an hour and never come up with anything better. My idea based on other guides is to call this good enough and then try to adjust it closer by matching inclinations at the AN/DN.

I then attempt the burn 5 times. Each time I can generate the change of SOI with Jool but I can't match the trajectory precisely enough to generate an Eeloo intercept. 

Reasons I am having trouble matching the maneuver node's trajectory. My rocket has 971m/s remaining in a chemical stage and then 1778 m/s in a NERV stage. The chemical stage burns out in 55s and the nuclear stage burns for almost 8 minutes. I have extensive experience flying around the kerbin system and I usually point toward the maneuver node and try to divide my burn evenly around either side of the maneuver node. I've struggled to do that here for several reasons. 

1.) Do I want to divide my DV or my burn time around the maneuver node? I thought logically DV but when I do that the trajectory has been more off than when I try to divide time.

2.) Do I want to burn prograde or toward the maneuver node? I've tried both and I can't tell. I'm limited in my ability to burn toward the maneuver node too far in advance of the maneuver because I end up burning down into the atmo since I am so far around the orbit from the maneuver.

So, any advice on how to sort this out would we welcome. OR Did I make some fatal mistake in my rocket DV or timing and I should convert this to a Jool bound mission instead? I have 3600m/s of DV remaining on the craft and I am in LKO 80km x 80km currently.

I ask is there a better way to do this because well honestly I love playing KSP, even somewhat tedious, repetitive stuff like orbital rendezvous but this is my second attempt at interplanetary intercept (have 3 craft on their way to successful Duna encounters) and holy cow is planning a maneuver using the standard maneuver nodes just agonizing and all around not fun.

Thanks in advance!

 

 

Edited July 22, 2020 by Pahimarus
Flag that updated answer is below

[Grogs]

I would say that 8 minutes is really pushing it in an 80x80 orbit. A rule of thumb I've seen about long burns is to not exceed 1/6 of an orbit, i.e., 60 degrees. At 80 km your period is around 35 minutes so you're well over that. It's probably doable, but it's going to be a less efficient burn. If I were planning the mission from scratch I would either aim for a higher starting orbit or else try to divide up the burn over multiple orbits (periapsis kicking). From an 80x80 orbit you could preburn around 900 m/s, leaving you a bit over 1100 to finish. Unfortunately the orbit would take you well over a month to return to periapsis so you'd have to plan ahead.

One thing you can do if your first stage has a higher TWR is to lower the thrust of the engines so that they match. This gives you a continuous thrust over the whole burn, but unfortunately in your case it would make it a good bit longer. For a more reasonable length of burn I'll often do that to keep the burn balanced.

 

As for your specific questions:

1. The maneuver nodes assume an instantaneous burn right at the node. The best way to approximate that is to start the burn when the time to the node is half the burn time, i.e., half the burn before the node and half after.

2. What I usually do is switch back and forth. Start burning towards the maneuver node and watch your periapsis. When it starts getting too low (say 71 km), switch to prograde. Once you pass the periapsis you can safely switch back to the maneuver node.

 

Anyway, Eeloo is a tough target. If you can pull off an encounter with the initial burn you're doing really well. I think I had to do two correction burns before I finally got an intercept. With this long burn it's going to be even harder. You need to allow a decent margin of extra fuel (10-20%) to make up for the inefficiencies.

I use MechJeb to do most of my interplanetary transfers. It seems to do a pretty good job most of the time. I also make frequent use of the info windows which display things like degrees to prograde and relative inclination with the target.

[bewing]

Yes, the long burn time is probably what's messing you up. My suggestion will be to break it into two burns, since you have an hour remaining. Use the first burn to make sure you come back to your maneuver node at the correct time. That is, burn to set the period of your orbit so that you come around your Pe at the right time, and force your maneuver node to be at the Pe. The fact that your orbit will take twice as long will mean that your second burn will be a much smaller fraction of your total orbit, and therefore will not mess up your angles nearly so much. And yes, burn to the maneuver node in this case and not just prograde. Getting to Eeloo requires precision, and the maneuver node gives you precision. Burning prograde gives you efficiency without precision.

 

[Srpadget]

What bewing said about  splitting your burn over 2 orbits.  That'll knock time off the longer, trickier burn.

For the 2nd burn, I'm going to espouse heresy:  Ignore what the transfer-window planner tells you for all except the day (ish) of departure.  Your high-efficiency, low-thrust transfer stage will not be able to make the desired burn with sufficient precision.

I've had success getting to Eeloo and Dres with the following (and my apologies in advance for the Giant Wall O' Text):

1) Set up a "dummy node" from LKO with a prograde burn that gives you an Ap at about the right distance from the Sun and an ejection from Kerbin SOI that's parallel/tangent to Kerbin's orbit.  That tells you where you need your FIRST burn in order to get its Pe in the right place.  (You won't get an encounter; you likely won't even get 'closest approach' markers.  Don't sweat it.  Just rotate map view so you're looking "down" on the whole solar system and eyeball it.)

2) Reduce the magnitude of that burn to give you the first burn that bewing mentioned.  (I'd recommend setting this up for a day or two before the optimum launch day, and a dV of around 800-850 m/s; this will put you a bit late for the optimum launch window, but that shouldn't be critical--Eeloo wants precision in the trajectory burn vs trajectory planning, but it doesn't need to-the-hour precision timing for departure in my experience).

Once you've made that burn, set up your transfer:

3)  Repeat Step 1; this node should be ~800-850 m/s smaller (and correspondingly shorter duration) than what you had in Step 1 in order to achieve a similar trajectory.  Again, you won't get an encounter; don't sweat it.

4)  Drop a node onto the transfer orbit (the simpler and more reliable method is to put is on the ascending/descending node, but it's often more fuel-efficient to put it earlier.  That's REALLY fussy, though, and I don't recommend it if you've got enough dV--especially for your first time).  Set up a plane change to bring your intercept point up/down as needed to get closer to an encounter.  Your goal here is to get a "closest approach" indicator.

5)  Once you've got that, fuss with BOTH of the nodes from steps 3 and 4 (the "plane change" node will likely need some prograde/retrograde and radial components), driving the closest-approach distance ever smaller until you get an encounter,and then focus view on Eeloo and use the same technique to drive Pe smaller.  (This last part should be familiar to you from fine-tuning arrivals at easier planets.)  Ideally, try to achieve a collision course so as to have the most room for error later, but don't get TOO fussy about it.

6)  Once those 2 burns are arranged to your satisfaction, prepare to execute the Kerbin-escape burn.  In map view, get the plane-change node from Step 4 as near the "camera" as you can reasonably manage--you'll be watching it like a hawk for the last 100 m/s or so.

7) Execute the Kerbin escape.  For a long-duration, low-acceleration burn like you're going to be doing, I kind of 'split the difference" between burning prograde and burning to the node:  I'll burn basically prograde, offset about 10-15 degrees in the direction of the node.  You'll need to adjust your pointing over time, and you'll eventually shift from burning somewhat radial-in to somewhat radial-out.

8)  As you get down to the last minute or so of the burn, shift to a basic prograde burn and go to map view (if you aren't already there) where you preset your view for the plane-change burn in Step 6.  Your goal is to place your actual trajectory right smack on top of the planned trajectory at that node.  IGNORE WHAT THE NAVBALL SAYS ABOUT HOW MUCH BURN YOU HAVE LEFT--you want to match the TRAJECTORY, and that low-accel means you won't exactly match the dV of KSP's "instantaneous-velocity change" node planning model.  Cut engines when you're as close to nailing the trajectory as you can manage--rotate ship and make a short retrograde burn if you overshoot. (I am also shameless about knocking the thrust limiter down to 0.5% for tiny final adjustments; it's cheaty, but OTOH the stock game doesn't provide any REALLY low-thrust maneuver/attitude jets, so I justify it that way.) Once you've managed that, delete the Kerbin-escape node. 

KSP will now lay your planned midcourse node onto your ACTUAL transfer trajectory rather than the initially-planned transfer--on a trip as long as Eeloo, there WILL be errors.  That's okay.

9)  In map view, focus view on your destination (Eeloo).  See how close you got.  If you're very lucky, you may still have an encounter.  Even if you don't, you've probably at least got closest-approach markers, and should be able to adjust your midcourse burn to regain the encounter and place your Pe where you want it.

10)  When it's time for your midcourse plane change, focus map view on Eeloo and watch THAT for the last 100 m/sec (-ish) of the burn, with the goal of eyeballing the last bit of the burn so you match the planned trajectory as best you can.  Again, that's more important than precisely matching the planned burn time/dV. 

That should get you there.  It's not a technique for the impatient, as there's a LOT of fussing about re-fine-tuning nodes that you've already fine-tuned once before.  But Eeloo isn't MEANT to be easy.

Let us know how it comes out!

[Streetwind]

6 hours ago, Pahimarus said:

Reasons I am having trouble matching the maneuver node's trajectory. My rocket has 971m/s remaining in a chemical stage and then 1778 m/s in a NERV stage. The chemical stage burns out in 55s and the nuclear stage burns for almost 8 minutes.

The others have already mentioned periapsis kicking and mid-course corrections as a viable way of dealing with low-thrust propulsion, so I'm going to focus on this part here. Because yes, stepping from one stage to another with wildly different TWRs is a problem. The game will factor this TWR change into your burn time, but you still get a noticable error in your maneuver node because you expend too much dV too fast at the beginning.

What you do instead is this:

When flying the vessel, click your total vessel dV readout. This will pop out the advanced view for all of the stages. (You can also click each individual stage to expand or contract it.) Ideally, one of the numbers in this advanced view shows you the TWR for the stages which have engines. If this number is missing, you unfortunately need to go back to the space center and enter the editor in order to configure the stock dV tool. It's a bit odd in that regard. Anyways -

Rightclick on your chemical engine. Take the thrust limiter slider and push it down. Observe as your TWR in the stage info drops. Adjust the TWR to be slightly less than that of your nuclear stage. Not much less, but slightly so. After all, your chemical stage will shed mass relatively briskly as it burns, increasing TWR again, and you wish to minimize the TWR difference when staging.

And yes: you will now have an even longer node burn time. You definitely need to employ periapsis kicking here, which means missing the transfer window by possibly a day or two as you loop around repeatedly. But the thing is: because your destination is Eeloo, and Eeloo is really far (in physical distance terms) out there and moving really slowly, your transfer window has a lot of leeway. Moho is the planet that gives you headaches because missing the window by an hour already impacts your transfer. Eeloo is super forgiving. A tiny mid-course correction at the proper place in solar orbit should nudge you back on the right track.

Edited July 21, 2020 by Streetwind

[king of nowhere]

i also suggest that you not do the correction manuevers at ascending or descending nodes. that's for matching orbital plane, which is not what you need to do here. what you want is simply to intersect eeloo's orbital plane. the reason for delaying the correction manuever is that you get more precision if you do it later, and you get more efficiency if you do it about halfway. i'm not sure on the exact reason, though i assume it's a compromise between making a manuever earlier meaning a tiny change in speed has more time to produce a large deviation, and making the manuever when your ship is slower being more efficient. Anyway, i haven't gotten to eeloo yet, but i guesstimate that the best place for such a course correction would be bewteen duna and dres orbits. from there, make a normal/antinormal acceleration to intercept eeloo's orbit, and then fiddle around with prograde/retrograde and radial/antiradial until you get an intercept.

[Pahimarus]

Y'all I have to be missing something. Here is an update I mostly tried to follow Sir Padget's detailed advice.

So I actually have two vessels in orbit of Kerbin with intent to launch to Eeloo. The first the one contains commsats and is the one that I discussed in my initial post. I was unimpressed with the DV that I got from the NERV and so for the second one I switched back to a pure chemical rocket carrying an orbiter and a lander.

Side note, I didn't realize that the NERV burned only fuel and not oxidizer (makes sense based on how a NERV is supposed to work but I didn't figure the designers would make it that opaque). Good news is that I was able to crossfeed the oxidizer into my partially emptied 3rd stage tank that will get discarded after I start running the NERV transfer stage so I "found" about 600 DV with that move.

I decided to try to setup the intercept with the pure chemical rocket as it avoids some of the problems I was having with the low-thrust awkwardly staged one previously described. I figure if I can learn a process to get an intercept with this one, then I can apply that back to the other one.

1.) I setup the initial escape burn using the numbers from the Alexmoon transfer planner. With this I was able to generate a closest approach of about 25,000,000 km. 

2.) I scaled the burn a back to about 1/3 of the total DV.

3.) I executed a periapsis kick burn around the scaled down maneuver node leaving me with about 1400 m/s to go for the full burn.

4.) I setup the escape node again.

5.) I setup a plane change node on the AN of the orbit, out near Ap. I fiddled until I got an Eeloo encounter with a periapsis of about 50,000 km which was the lowest I could get. I honestly can't remember which handles I had to pull to manage that or how much DV it was. Update: Here is before the escape burn.

6.) I executed the escape burn per Sir Padget's description above and achieved this orbit. (Note: I also used the trick shared of changing my max throttle to 0.5% to tweak the burn to get the absolute closest encounter that I could while burning prograde/retrograde only).

7.) I didn't quite have an Eeloo encounter any more and so I started fiddling with different midcourse correction nodes. Somewhere I lost the plane change node that got me closest to Eeloo. 

8.) I fiddled and fiddled and fiddled and fiddled and fiddled.................................................................................................................................................................................................................

9.) Fiddled some more.

10.) Managed to get a midcourse correction node near Ap that got me an Eeloo encounter but with an unacceptably high DV (like almost 2k more DV and wouldn't have enough left to capture then). 

11.) Fiddled and fiddled. More fiddling.

12.) Managed to replicate the node from #10. Yep, still unacceptably high DV.

13.) Fiddled and fiddled and fiddled. 

I am just utterly lost on what I am supposed to be doing when constructing a "mid course correction node." I would think it would be similar to trying to rendezvous with a spacecraft that changing inclination at the AN or DN to match would help me get closer but nope anytime I change that I lose my closest approach altogether. The only successful node had me pulling mostly radially at apopsis.

Anyway, feeling lost and super frustrated. All the advice that got me to this point is appreciated. Any further advice would be much appreciated. 

 

Edited July 22, 2020 by Pahimarus
Added screenshot before the escape burn

[bewing]

Generally, maneuver nodes are worthless for plotting mid-course corrections. The corrections typically are at most a few m/s of deltaV, and a maneuver node stinks for trying to plot out such delicate adjustments. Mid-course corrections are not about matching planes. They are about creating a nice intercept even when your planes don't match.

First, adjust your engine down to minimal thrust.

When you get halfway to your destination, look at your orbit and closest approach marker, from the POV of your destination (move your camera so you are looking from your Ap back toward Kerbol). Does your orbit pass above or below your target's orbit?

Step 1: If above, burn due South. Directly toward the pole on your navball. If below, burn North. With gentle burns, make your orbit line intersect the orbit line of your target.

Now you need to adjust the timing to achieve an intercept.

Step 2: If your closest approach marker shows that you are arriving too early, then you need to slow down.  So burn some combination of retrograde or West or radial in. If you are arriving too late, then do the opposite.

This will slightly mess up the orbital intersection that you got from step 1.

So now you refine your intercept by going back to Step 1 again.

[Pahimarus]

Ok that is super helpful, I guess I really didn't understand what people were describing with mid-course corrections.

So I went back and followed Sir Padget's directions again and managed to generate an intercept this time without losing my node. 

First of all. <pauses> Huzzah!

I've got about 2700 m/s of DV left in the tanks. It looks like that won't be enough to execute this maneuver and still have enough DV for a capture burn (estimates I've seen 1000-1500 m/s for Eeloo capture). 

Should I keep that node and just go for the flyby or could I expect that an earlier mid-course correction as described by bewing might be able to generate an encounter with less DV?

**Further updates**

Using same process as described by Sir Padget managed to generate an intercept with my nuclear rocket as well as well. I setup my plane change node a little closer to Kerbin (about halfway as suggested by some posters). This allowed me to generate an intercept with an acceptable amount of DV ~500 m/s which leaves me enough to make a capture. With this experience and orbit for reference, I feel confident I could move the node on my other rocket and replicate a similar lower DV maneuver. 

So to recap:

1.) Periapsis kicking worked beautifully to make my escape burns more manageable. 

2.) Adjusting the throttle limiter was also quite helpful for #1 as well as for fine-tuning orbital maneuvering.

3.) Sir Padget's method appears to be (as he mentioned in his post) somewhat wasteful of DV but it was critical to helping me visualize what I needed to do to the orbit in order to make the intercept with Eeloo happen. 

I'm still trying to get some clarity in my mind around "mid-course corrections."

Bewing's explanation was helpful. 

Would it be accurate to state that Sir Padget's "plane change node" (described in #4 of his step-by-step instructions) is, in essence, doing the same thing as a "mid-course correction"?

It seems that because the DV required is higher when this is positioned closer to the intercept point that the maneuver nodes become usable for visualizing it. Whereas when the mid-course correction is far from the maneuver node the amount of DV is so small that it is too fiddly to visualize using the maneuver node?

Thanks again everyone for your helpful input. Your advice was super helpful, any failures were purely end user error.

Edited July 22, 2020 by Pahimarus
Added additional updates

[FalconHeavy2019]

Mechjeb

It will help you with your transfer burns

 

[Streetwind]

6 hours ago, Pahimarus said:

I'm still trying to get some clarity in my mind around "mid-course corrections."

A course correction is theoretically any burn that is a.) noticeably smaller in magnitude than the original transfer burn, and b.) helps you reach your destination. As you already discovered, there are multiple reasons why you would make a course correction. And they all require different approaches.

• Plane change. Sometimes you have a transfer window that is located far away from the intersections of your orbit and the target's orbit (Ascending Node/Descending Node). If the difference in inclination between your location and your destination is significant enough, this makes it impossible to force a single-burn solution unless you mean to go full-on torch ship. You'll have to make a second burn at AN or DN in order to reach your target. Sometimes these burns can be quite costly in terms of dV.
• Fixing your transfer orbit. When you use a low-thrust propulsion solution, your final transfer orbit may be off compared to the simulations, because all the simulations assume an instantaneous change in velocity. Rather than waggle about with your engine on after your node is already completed, you instead cut thrust and make a new node. Maybe fifteen to twenty minutes ahead of your ship's current position. Use this node to get the intercept back that you wanted. It should typically only require single-digit dV, rarely going into the low double digits.
• Refining your intercept. Most missions that want to insert into orbit at the destination generally want to do so with the lowest periapsis possible, but it may be hard to get a nice low periapsis with a maneuver node from really far away, and then also hit that course exactly with your manual throttle control. It's usually better to do this when you're closer. Personally, I like to do it right after I've entered the destination's sphere of influence, as you are slowest relative to your destination at that point. Then you can use radial and normal directions to push the periapsis closer to the surface, and adjust whether you're passing over the poles or by the flanks of the destination (or whererver you want to be passing). You typically don't use prograde/retrograde here except for precision adjustments; slowing down is what your periapsis is for. The cost of such a refinement node will vary greatly depending on your destination, how you came in, and where you want your periapsis to be, but typically expect a low-ish double digit value.
• Honorable (?) mention: changing arrival inclination. Maybe you want a nice equatorial orbit, but your destination has a vastly different orbital inclination than Kerbin, so you're coming in from above or below the equatorial plane. I've seen this confuse and frustrate people a lot, because they typically try to force this with the initial transfer node, or a correction node halfway along the orbit, and it pretty much never works. So it's important to understand that choosing your arrival inclination is a subset of point three. As you refine your intercept, you choose the position of your periapsis. If you place it over the poles, you get a polar orbit reliably. If you position it over the equator, you get... no, not an equatorial orbit. You get the inclination of your transfer orbit. And anywhere between the pole and the equator, you get a mix. It is impossible to get an equatorial arrival orbit with a course correction when you're coming in from out of plane. So what do you do? Well, you make a course correction as in step 3 to place your node periapsis directly over the equator. And then, you fold a classic plane change into your insertion burn. Yes, this will be expensive, but: because the pythagorean theorem is a thing, combining a 1,000 m/s retrograde burn with a 1,000 m/s antinormal burn will cost only 1,414 m/s, not 2,000 m/s. And that's how you "correct your course" for an equatorial arrival orbit.

So what about those course correction nodes halfway along the orbit that people told you to make? Well, they can work, too. You can make intercept refinements from much further away than the edge of the target SoI, and it's typically even cheaper. But the problem with that is that you will be in a different reference frame, so your node handles will be oriented differently, as will your camera. That makes it less intuitive to work with.

Plane change nodes, where they are required, will also typically be around halfway along the orbit, give or take some depending on the individual case. And yes, you can even fix an error in your transfer orbit later than right after your initial burn.

So often enough, a course correction node halfway along he orbit is a "one stop shop" solution. You match planes, and fix the transfer burn error, and reposition the arrival periapsis in one fell swoop. That's a perfectly valid approach, and many people use it. But I still wanted to break it down into its individual components above, so you better understand what one might try to achieve with various kinds of course corrections.

 

Edited July 22, 2020 by Streetwind
Dumb typo

[Pahimarus]

Thanks everyone. Was able to generate intercepts with Eeloo for both spacecraft within my delta-v budget and replicate the process several times. The additional tips on throttle controls make me confident I will be able to successfully complete the burns when the time comes.

I play my game with multiple missions running simultaneously (makes the long missions feel like they are long) so it may be some time before my little craft enter Eeloo orbit. 

[Srpadget]

On 7/21/2020 at 5:59 PM, Pahimarus said:

Would it be accurate to state that Sir Padget's "plane change node" (described in #4 of his step-by-step instructions) is, in essence, doing the same thing as a "mid-course correction"?

That was certainly my intent--for a typical Eeloo transfer, most of the midcourse correction will be plane change.  But that does not mean you necessarily have to match planes/inclinations, as bewing made explicit.

Putting it on the AN/DN and matching inclinations makes the trial-and-error to get an encounter simpler/more forgiving, but as noted it's wasteful of dV.

Putting it elsewhere and 'splitting the difference' (if you will) between the initial transfer-orbit plan and the target plane is generally much more efficient--but when there's a substantial plane change involved and a small destination body with a small SOI, getting the encounter gets fussier than if you've matched planes.

And it's almost never a pure plane change--you will need radial/prograde components in order to make and/or fine-tune the intercept.  I noted that in passing in my giant wall-o-text, but I probably should have emphasized that/been more explicit.

And congrats on a successful transfer!  Eeloo is challenging!  (Dres would be as well, but of course nobody ever goes to Dres.   )