Keeping MN direction for VERY long burns

[marcast]

I've built a ship to send a lander to Laythe, with no aerobraking and no gravity assists. It has three nuclear engines but it is quite heavy, so it needs a 42 minutes burn to reach Jool. I have two problems:

1 - I can't start the engines 21 minutes before the maneuver node because in this case my periaxis should fall below 50000 mt. inside the kerbin atmosphere (even if I start at 200000 mt.). But ok, I could solve this by starting closer to the MN and making some later course correction on the way to Jool.

2 - The real problem is that the ship can't keep the MN direction for such a long time. After about 15 minutes the remaining burning time stops decreasing and starts increasing, the orbit becomes very narrow and everything goes in a mess.

What can I do?

[The Aziz]

You can separate the burn into few. Let's say first one to get into elliptical orbit around Kerbin, second to complete the transfer.

Though, 42 minutes for a burn is a lot for nukes, what TWR you have?

Also, you say no gravity assist, but using Tylo for MUCH cheaper approach is relatively easy. Just fine tune your encounter with Jool - do it as early as possible - so that the trajectory crosses Tylo orbit at the right time, and if properly set up, it should put you in Jool orbit without using a drop of fuel. Think of nearly half of delta v saved for Laythe maneuvers.

[paul_c]

With a 42 minute burn, I'd do it in 3 stages:

1. Burn at -3 mins to +7mins of the ideal ejection angle to get the orbit looking more/less pointing in the right direction. I'm going to guess your Jool burn is ~1900m/s so that's (say) 500m/s, which means it will take you, I dunno, Ap 1000km? You can shut off the burn early or keep going a bit longer, to get the angle right

(Now, as the orbit gets more and more eccentric, the fact you're burning not at the Pe will have less and less influence on the ArgPe).

2. Do another burn to get 'just' before the Mun, you don't want an inadvertent Mun encounter! So, its going to be be about another 350m/s.

3. Do a burn to leave SOI of Kerbin plus whatever else is needed to send you on your way. I think its going to be more efficient to keep burning here, even though its not at the Pe, ie the Oberth effect outweighs the cosine losses. But if its looking bad, just stop and wait until you're out of Kerbin's SOI.

Once you're out of Kerbin's SOI and on the way to Jool, the low TWR isn't really an issue - you have a very wide margin of when to do an inclination change burn (like, weeks.....) and any mid-course correction to tighten up the approach nice and low over Jool.

[18Watt]

You've discovered there is a practical limit to how much you can use gravity to assist you when your TWR is very low.  The solutions posted above would be my first suggestion.

Personally, here's what I do when my TWR is too low.

• I may need to raise my orbit around the departure body (Kerbin).  Obviously the lower your orbit, the more gain you get from the Oberth effect.  But if your burn is too long, you end up wasting fuel anyway.
• For burns roughly  4 minutes, I use an orbit of 100-120km around Kerbin.
• For 5-10 minute burns, I may use 120-200km orbits to depart.
• For a 40+ minute burn, I wouldn't waste my time trying to get a boost from Kerbin's gravity.  I doubt you will gain anything, or if so it will be very insignificant.  Instead I'd just burn to barely leave Kerbin's SoI, then do the transfer burn from Solar orbit.

Here's an example-  I have a large ship powered with Vector engines.  I normally refuel at Minmus, then hop just outside of Kerbin's SoI for the transfer burn to Jool (or wherever).  However, occasionally it ends up being more convenient to depart from a lower orbit around Kerbin.  The ship has a TWR that gives me about a 5-6 minute burn to Jool, using a Kerbin orbit of ~120km.  For that ship, the total DV difference between leaving from a 120km orbit around Kerbin and just hopping outside Kerbin's SoI for the transfer burn is. . . . about 90m/s.  Again, with a 40+ minute burn I wouldn't bother with trying to get a gravity gain from Kerbin.  

Another way to look at it:  You are spending too much effort to gain perhaps 50m/s with Oberth at Kerbin.  However, you can easily save at least 2,000m/s at Jool by using Tylo's gravity.  (Or even Vall's or Laythe's)  With a 40+ minute burn, you are never going to see a significant advantage using Oberth at Kerbin- but you can easily get a huge boost from Tylo.  I'd focus on getting assists from Tylo instead of the minimal gains you may or may not get at Kerbin.

[Empiro]

How many m/s is your maneuver? If it's significantly more than 2000m/s, consider tweaking your maneuver or waiting for a better window.

If you still have a 40-minute burn, try a higher orbit to start with. For Kerbin to Jool, the gate orbit is at 250km. This altitude minimizes the delta-V to get to Jool (not counting the delta-V required to get into that orbit). This will reduce the burn you have to do and give you more time to do it. 

You can also split the maneuver into pieces. Do half of the burn so that your AP is high but within Kerbin SOI still, and then complete the burn the next time you come around. You still need about 950-1000 above Kerbin escape so you're still looking at a 20 minute burn.

One other thing you can try to do is to ignore the maneuver node once you're significantly past it. Once you're about 5 minutes past the maneuver point, just burn prograde. It will cause your ejection angle to be a bit off. One way to help plan around this is to split your maneuver in KSP -- place one maneuver that you can do within ~5 minutes at Kerbin's PE. Place another one about 5 minutes after that one that you can do in 5 minutes, and so on. It's a pain to do any corrections in timing (you'd need to move all your maneuver nodes around), but it will tell you with good accuracy where you'll end up.

[ajburges]

Everyone has sound advice, but I want to draw your attention to the root of you problem:

The maneuver node system assumes instantaneous velocity changes. When burns are short (a fraction of an orbital period) the difference between your node and your burn is negligible, but longer burns will compound error. Long burn craft need integration of velocity over the course of the burn. That's a much more complicated problem to solve and KSP does not have tools to help solve it.