Direct descent vs low PE landing

[Kosmo-not]

I decided to do some experimenting with direct descent vs low PE landing on the Mun.

The lander:

pod + parachute

decoupler

ASAS

1 standard fuel tank

3 lander legs

small engine

Starting scenario:

350m/s heading straight down towards the Mun surface (-90°) at an altitude of 1.4Mm.

This scenario is biased towards a direct descent landing.

Direct descent:

Burn started at 45km. I made sure my velocity was a few hundred m/s by 10km, and kept the speed up to avoid gravity losses. I landed, expending a total of 984 delta-V. Fuel left was 275.9295.

Low PE descent:

Burned radially to raise PE to 5km (about a 17° difference equating to a 104m/s burn). Landed, expending a total of 980 delta-V (this includes the course change burn). Fuel left was 276.7127.

Conclusion:

For this combination of mass and thrust, a low PE landing is more efficient than direct descent. More thrust is required to make a direct descent more efficient than a low PE descent. I knew there is a value of thrust vs mass that would make one more efficient than the other. The craft used in this experiment comes pretty close to that value.

*edit*

I tested the two descent methods with mechjeb, with the mechjeb replacing the ASAS on the previous stated lander.

Starting conditions were: 381m/s @ 2Mm altitude and -90° angle.

Direct descent: 1018 delta-V

Low PE descent: 1053 delta-V (this includes an 80 delta-V course change)

Conclusion:

It\'s a pretty close call for realistic scenarios. How often do you end up with near -90° angles when entering the Mun\'s SOI? It\'s safe to say that, for this craft, if you have a PE above ground level, go with the low PE landing.

[sal_vager]

Am I wrong in thinking there is not enough in it to choose one landing over another? Personally I do low PE landings as I prefer them, but after seeing this I may do some Direct Descents just for fun.

Kosmo-not, your tireless pursuit of efficiency is impressive.

[semininja]

Can I recommend using MechJeb to maximize consistency?

[Endeavour]

Although I\'m sure the 4 delta-v makes a difference at some point, I think these are close enough that either should be just fine. I always do mine from a low PE because the safety burns are more iffy in the direct decent (for me at least).

[Kosmo-not]

I guess I\'ll post some results using mechjeb soon.

The 4 delta-V difference doesn\'t really mean much, having a human pilot.

I\'ll get the new results posted soon. Maybe I\'ll hold off on activating the landing autopilot until 40km or so.

[semininja]

With the current version of MechJeb, I think you can just leave the landing computer on during both descents and it will apply burns where necessary without too much extra fuel use.

[Ydoow]

Yeah I\'d prefer a few more tests with MechJeb for consistency.

You showed a difference of 1 fuel unit, which isn\'t much and I feel may be a result of human error (no offense)

[DonLorenzo]

Care to test the scenario that was talked about in some other thread? Can\'t remember if that was with you or someone else. Starting scenario = stable mun orbit, method 1 = kill lateral velocity then \'fall\' and stop. method 2 = lower PE and then brake+land.

[boolybooly]

OK interesting subject and I made a related post about this in another thread FYI. Personally I dont use Mechjeb which is less empirical so I will be interested to see what you find.

From what you said about your direct descent though...

Direct descent:

Burn started at 45km. I made sure my velocity was a few hundred m/s by 10km, and kept the speed up to avoid gravity losses. I landed, expending a total of 984 delta-V. Fuel left was 275.9295.

...it sounds like you throttled down at 10km to keep the speed up? If this is the case then the burn began too early which if true would mean it could be done more efficiently by delaying the burn to the last possible moment so you are doing 100% burn from start of deceleration until <1000m (depending on terrain altitude).

FYI

[sal_vager]

Aren\'t you having to deal with the exact same potential energy in the rocket and the same energy from the Muns gravity either way? Seeing as there is no drag or lift available.

I would not be surprised if Mechjeb returned almost identical results both ways, with any type of rocket.

[UmbralRaptor]

Aren\'t you having to deal with the exact same potential energy in the rocket and the same energy from the Muns gravity either way? Seeing as there is no drag or lift available.

I would not be surprised if Mechjeb returned almost identical results both ways, with any type of rocket.

I think it comes down to minimizing gravity drag. Which isn\'t especially large on the Mün. Also, that Kosmo-not had to raise his Pe probably killed any savings. I tend towards direct descents, but that\'s more for the conceptual simplicity...

I\'m going to second the interest in in seeing how much of a difference there is when descending from various circular orbits (5 km, 100 km, etc)

[Kosmo-not]

Care to test the scenario that was talked about in some other thread? Can\'t remember if that was with you or someone else. Starting scenario = stable mun orbit, method 1 = kill lateral velocity then \'fall\' and stop. method 2 = lower PE and then brake+land.

I would say:

Lower your PE to the surface and conduct a burn that brings you to a stop just as you reach the surface.

Also, for those who didn\'t notice, I did an experiment with mechjeb and posted the results above.

*edit*

I did the experiment with a circular orbit of 100km

Killing all horizontal speed and descending used 1033 delta-V

Lowering the PE to 1km and starting the landing burn at 5km used 763 delta-V

[DonLorenzo]

Thank you

[UmbralRaptor]

Thanks. I\'m guessing that the low Pe method is still better if your initial landing site is full of boulders, though.

[togfox]

Would the ultimate test be that where the craft first enters the Mun SOI at 2 million meters? At that point (press F5 to quick save) you can do a direct descent, which will require an initial course correction, then repeat with a drift to PE, circularise and de-orbit burn before landing.

This for me is the million dollar question. This is the decision I have to make when first entering the Mun SOI. I\'ve tried a direct descent from MUN SOI and it uses a LOT of fuel and no longer even consider it.

[Kosmo-not]

Would the ultimate test be that where the craft first enters the Mun SOI at 2 million meters? At that point (press F5 to quick save) you can do a direct descent, which will require an initial course correction, then repeat with a drift to PE, circularise and de-orbit burn before landing.

This for me is the million dollar question. This is the decision I have to make when first entering the Mun SOI. I\'ve tried a direct descent from MUN SOI and it uses a LOT of fuel and no longer even consider it.

If you\'re asking about a scenario in which PE is at 5km with 381m/s @ 2Mm altitude, refer to the second experiment in the first post. Add the course change delta-V to the direct descent, and subtract it from the other.

For any PE above ground, I would highly recommend a very low PE landing procedure. You save a good amount of delta-V by doing this because you\'re not losing it to gravity drag.

For a PE below ground, you\'ll have to make a judgement call. Do you have enough thrust to make a direct descent more efficient than the alternative? You don\'t have to go straight down from 2Mm. A less elliptical path into the Mun will still bring you down at a nearly vertical angle towards the end. If you think have enough thrust, don\'t make any course changes, even if you\'re coming in at an awkward angle.

For myself, I won\'t be doing any direct descents in the future. Too much worrying about burn altitude while saving a little delta-V (if any). Much easier to do the horizontal (or a little less) approach. If you think you\'re going to impact the ground, you can always point up.

[Alchemist]

In the theoretical case of indefinitely high thrust both trajectories should be equal (+- the correction burn). But in case of rather low TWR low orbit is both more efficient and safer.

[PakledHostage]

If you\'re asking about a scenario in which PE is at 5km with 381m/s @ 2Mm altitude, refer to the second experiment in the first post. Add the course change delta-V to the direct descent, and subtract it from the other.

While I agree with your conclusion that a low Pe landing is more efficient, I don\'t think you\'re making a fair comparison in your test.

As I\'m sure you know, there are an infinite number of hyperbolic trajectories about the Mun that have a speed of 381 m/s at an altitude of 2000 km. The special case where Pe is 5 km is only one of those. When you make a burn to change that hyperbolic trajectory into a radial one, you are adding energy and the tests are no longer equivalent. We\'d really need to compare orbits that have the same initial specific orbital energy.

Given that it is possible with careful planning to target your Munar Pe merely by adjusting the timing of your TMI burn (or at most with miniscule differences in TMI Delta-V), I don’t think it is fair to include significant mid-course orbital corrections in your comparison.

[Kosmo-not]

While I agree with your conclusion that a low Pe landing is more efficient, I don\'t think you\'re making a fair comparison in your test.

As I\'m sure you know, there are an infinite number of hyperbolic trajectories about the Mun that have a speed of 381 m/s at an altitude of 2000 km. The special case where Pe is 5 km is only one of those. When you make a burn to change that hyperbolic trajectory into a radial one, you are adding energy and the tests are no longer equivalent. We\'d really need to compare orbits that have the same initial specific orbital energy.

Given that it is possible with careful planning to target your Munar Pe merely by adjusting the timing of your TMI burn (or at most with miniscule differences in TMI Delta-V), I don’t think it is fair to include significant mid-course orbital corrections in your comparison.

The two tests start with the same specific orbital energy. The correction change burn does not add or subtract orbital energy, since it\'s perpendicular to the velocity vector and does not change the velocity of the craft. The change in altitude between loading the quicksave and changing course should be negligible.

[PakledHostage]

The two tests start with the same specific orbital energy. The correction change burn does not add or subtract orbital energy, since it\'s perpendicular to the velocity vector and does not change the velocity of the craft. The change in altitude between loading the quicksave and changing course should be negligible.

After reading your response to my PM (thanks for that), I understand what you are getting at. What you told me there isn\'t clear from your post above though.

What you are saying is that MechJeb orients the spacecraft perpendicular to its velocity vector throughout the burn. This causes the spacecraft to move in an approximately circular arc during the burn. (I say approximately because the ship\'s velocity is not constant along the undisturbed hyperbolic trajectory.) There is no effect on the magnitude of the ship\'s velocity (only the direction) as a result, and the specific orbital energy is unchanged.

That being said, my point still stands. You are expending 80 m/s Delta-V to change the ship\'s direction of travel and including that 80 m/s in the results for the one case, but you are not including any Delta-V for mid-course corrections in the other case. Had the TMI burn for the two cases been executed slightly differently however, this 80 m/s Delta-V burn could have been eliminated. As I said in my post above, it is possible to achieve a Munar periapsis of 5 km through careful planning of the TMI burn. It is also possible to plan a Munar impact trajectory that starts from the same Kerbin orbit but that requires only very slightly different timing and/or Delta-V at the time of the TMI burn. As a result, I don\'t think it is an appropriate comparison to penalise one case with the Delta-V of a mid-course correction and not the other. The mid-course correction burn should either be ignored or averaged between the two cases.