Quick question about Transfer Windows

[sadron]

So I'm using Kerbal Alarm Clock to calculate when I need to transfer from Kerbin to my interplanetary destination, Moho, which at least tells me what time I should begin focusing on getting this ship out of Kerbin orbit and off towards Moho for the glory of Kerbalkind. But I still don't have an idea on what else I need to do. Do I need to be outside of Kerbin's orbit? How does the engine I'm using factor into this? I'm using a modded nuclear engine, a "Fatman" engine with 830 isp and around 90 thrust. Fairly nice. I also use MechJeb2, maybe you can tell me what to keep an eye on in that. If this is in the wrong section I apologize.

[tavert]

Here are the basic steps for interplanetary transfers:

1. Start from low Kerbin orbit, set target for destination planet

2. Create a maneuver, add delta-V in the prograde direction until your projected orbit escapes Kerbin's sphere of influence. You'll need around 1000 m/s for Eve or Duna, around 1500 m/s for Dres, around 2000 m/s for Jool, and anywhere from 2000-3000 m/s for Moho or Eeloo.

3. Click and drag your maneuver around until your escape direction is in the retrograde (for inner planets) or prograde (for outer planets) direction of Kerbin's orbit around the sun.

4. Adjust your maneuver delta-V to cross the orbit of your destination planet until a closest approach indicator appears, reduce the closest approach distance as low as you can.

5. For inclined targets, you can either adjust the inclination along with your ejection burn, or wait until you escape Kerbin and add a correction maneuver. Simplest place for correction maneuvers is usually either the ascending node or descending node. If you will cross both before you get to your destination, then pick whichever node is further from the sun to do your correction maneuver at.

If you have low thrust-to-weight ratio, your initial ejection burn may take a long time, so keep that in mind. You can split up the burn into "periapsis kicks" if you want to maximize efficiency, but at the cost of more complicated timing.

MechJeb2's maneuver planner will set up interplanetary transfers for you if you want. Set the destination as target and create an "Interplanetary transfer" maneuver, then after you've executed that you do a "Fine-tune closest approach to target" maneuver for the course correction.

Edited June 1, 2013 by tavert

[sadron]

Well I don't want MechJeb to hold my hand, I use it mostly for data info such as distance to rendezvous target and what not. Still, your steps really helped clarify some stuff for me.

[alexmun]

One thing to note is that Kerbal Alarm Clock's timing is based on Hohmann transfers, which is awesome for coplanar, circular orbits. Unfortunately, Moho's orbit is neither circular nor coplanar with Kerbin's. I got frustrated enough trying to hit Moho without wasting a ton of fuel to put together a calculator to visualize the actual optimal launch windows. You can try it here: http://alexmoon.github.io/ksp

[Vanamonde]

travert's advice is pretty good, but I believe there is an issue with being in TOO low of a Kerbin orbit. If it's a long burn, can't this cause you to come out at the wrong ejection angle? Also, it is EXTREMELY costly in fuel to try to change inclination during a departure burn. That's better left to a half-trip course correction. (The ideal would be to begin the departure burn already on a Kerbin orbit parallel to the target's inclination, but that's extremely difficult to align and time.)

[alexmun]

If it's a long burn, can't this cause you to come out at the wrong ejection angle?

Yes, this is definitely a problem. I usually calculate how long the burn is going to be then start my burn half that time before I hit my ejection angle. It's not perfect but it tends to get me close enough that a few small adjustments give me an encounter.

For inclination changes, making the correction midway through your transfer is a very reasonable approach (you don't even need to do it at the ascending or descending node, you just need to incline your orbit so it intersects the target orbit at the right point). However, there are a couple of advantages to doing the inclination change as part of your ejection maneuver. First, combining two maneuvers into one requires less delta-v in the same way that traveling diagonally is a shorter path than traveling horizontally then vertically to get to the same point. Second, the closer you are to a planet when you perform a maneuver the more you benefit from the Oberth effect. The hard part is figuring out how much inclination you need on your ejection maneuver in order to hit your target. That's one reason I made my launch window calculator.

[Uberick]

Also, if you're using a low thrust engine that has long burn times, and you know your ejection angle, you can start doing a series of small burns to slowly build up to almost escape velocity over several orbits of Kerbin. Then, wait for your phase angle to line up, and then perform your final burn to achieve escape velocity.

This can help you keep everything lined up if you have very long burn times.

[Vanamonde]

First, combining two maneuvers into one requires less delta-v in the same way that traveling diagonally is a shorter path than traveling horizontally then vertically to get to the same point.

I thought so, too, until Maltesh explained that this is incorrect. I believe that if you try both methods and compare them, the mid-course inclination correction saves so much dV that the sum of it and the transit burn is significantly smaller than the dV of a single burn for transit/inclination combined. But I'm afraid I must leave it to some of our more mathematically-inclined forum fellows to check my recollection and explain why that's the case. Is there anyone who'd like to help out?

[tavert]

We should be able to do the math on this by adding an "ecliptic projection" mode to alexmun's transfer calculator. If you replace the actual orbits of the planets with the projection of their orbits onto the ecliptic plane, then you'd get the delta-V of a purely in-plane ejection burn. Then we determine the required course correction burn as a function of when it is performed and minimize for delta-V, and compare the sum for the optimal 2-burn transfer to the 1-burn transfer.

[CalculusWarrior]

Kerbal Alarm Clock has two options for transfer calculations--it either takes them from a chart of optimal burn times (I forget where the chart is located, but it's pretty comprehensive), or from this website: ksp.olex.biz.

Naturally, taking the calc from the chart will be more accurate (the website assumes that every planet is moving in exactly circular, zero inclination orbits), but it only goes up to a certain in-game time, I think it's Year 99.

If you chose the calculation method, there is a chance that you will not be in the exact proper position when the time comes for your burn. Fortunately, it is likely that you won't be too far off (unless you're trying for Eeloo, of course).

Basically, all you have to do is burn prograde at the angle specified by the above website. This is independant of planet positions, it merely takes into account the amount of burn you will have to do. It specifies an angle that, after burning up the required delta-v, will put you on a prograde trajectory around the sun. Of course, if you don't want to mess around with fancy orbit angles, you could put down a maneuver node and adjust its position and amount of burn required until it shows an escape trajectory that is either facing the direction of Kerbin's orbit, or the opposite direction, depending on whether you want to visit an exterior/interior planet.

Of course, another option is to simply burn at any point in the orbit until you get an escape trajectory, but stop as soon as you do, and then coast out into an interplanetary orbit that is essentially the same size as Kerbin's orbit. Then, you can treat it as trying to get from Kerbin to the Mun, and burn prograde/retrograde when you're in the right position.

Finally, another suggestion to get yourself familiar about interplanetary burns, try sending a small probe to orbit the Mun and try to get from there to Minmus. The website I linked you to can also calculate angles and delta-v for intra-system transfers too.

I hope this helps!

[sadron]

One thing to note is that Kerbal Alarm Clock's timing is based on Hohmann transfers, which is awesome for coplanar, circular orbits. Unfortunately, Moho's orbit is neither circular nor coplanar with Kerbin's. I got frustrated enough trying to hit Moho without wasting a ton of fuel to put together a calculator to visualize the actual optimal launch windows. You can try it here: http://alexmoon.github.io/ksp

Maybe I'm just stupid but I don't really understand your calculator's chart. Ugh, I feel like such an idiot... I should be able to understand these concepts x-x

[alexmun]

We should be able to do the math on this by adding an "ecliptic projection" mode to alexmun's transfer calculator. If you replace the actual orbits of the planets with the projection of their orbits onto the ecliptic plane, then you'd get the delta-V of a purely in-plane ejection burn. Then we determine the required course correction burn as a function of when it is performed and minimize for delta-V, and compare the sum for the optimal 2-burn transfer to the 1-burn transfer.

That's a good idea. I'm out of town this weekend, but I might work on that after I get home.

Maybe I'm just stupid but I don't really understand your calculator's chart. Ugh, I feel like such an idiot... I should be able to understand these concepts x-x

Don't feel stupid, this stuff is not easy. I'm still figuring a lot of it out myself. What the chart shows you is how much delta-v (change in velocity) it takes to get from point A at a specific time and arrive at point B at a specific time. So the blue areas of the chart represent the most efficient times to leave and arrive. Also have a look at olex's calculator (http://ksp.olex.biz) which does a better job of walking you through the mechanics of doing a transfer.

[sadron]

Still, I'm a smart person, I hate that I need "pathways" in wording to understand a concept for some stuff when other stuff comes out as easy to understand. Still, thanks for making it a bit easier for me to understand.