Vertical Ascent vs. To LXO First

[arkie87]

Sorry to bring this up again, guys, but I think we have overlooked something.

GoSlash27 http://forum.kerbalspaceprogram.com/threads/102606-Transfering-to-Mun-First-to-LKO-or-Directly-to-Mun?p=1593793&viewfull=1#post1593793 said that in his experience with stock aerodynamics, it will take approximately 340 m/s extra deltaV to go vertical ascent directly to the mun vs. performing gravity turn into LKO, then periapsis transfer to the Mun. In my own experience with FAR installed, the difference is smaller, around 70 m/s

. The theoretical reason for this difference is that in stock, terminal velocity (which is optimal ascent velocity) is low and easily reachable with low TWR. Thus, there is no benefit of high TWR since it will not be used. In FAR, terminal velocity is essentially unreachable, and thus, there are petential benefits of high TWR craft. Furthermore, the potential for a pilot-induced sub-optimal gravity turn, which could easily bleed off a few hundred m/s deltaV if you raise apoapsis too high upon liftoff or stay too low for too long. Thus, vertical ascent could match gravity turn under the right conditions.

Nevertheless, Kerbin isn't the only planet. My question is if you are on the Mun such that Kerbin is perpetually setting (since the Mun is tidally locked) i.e. on the rear-side of the planet w.r.t. Mun's velocity vector, do you want to get into LMO first or go straight vertical to return to Kerbin.

My preliminary math shows it takes less deltaV to go straight up (807 m/s) vs. hopping up to 10 km altitude (to avoid hitting terrain) and then burning horizontally until escape velocity is achieved (964 m/s). Since Mun's gravity is small, TWR will inevitably be large so gravitational losses will be small/negligible.

What do you guys think?

If i am right, it implies that vertical ascent vs. LXO-to-transfer optimality is dependent on the planet/mun and other conditions, rather than always assuming vertical ascent is bad.

Edited December 18, 2014 by arkie87

[LethalDose]

First and foremost, thanks for recording and posting your flight anyone who's interested can evaluate your evidence.

I think there's some theoretical merit to your statement that direct assent may, conditionally, be a more favorable assent profile than using intermediate parking orbits. However, in practice, I think those conditions are extremely rare, and your example (asending from the backside of the Mun) is one of those rare situations. Also, the dV savings you've demonstrated, while significant (~20%), is minor in a more absolute sense: 150 m/s of dV just ain't that much, although it is if you've only got 850 m/s in the tank. What you're saying also makes sense: Low gravity worlds cause less gravity drag during the escape.

Right now, it looks like you've described one set of conditions where a direct assent would be superior to the parking orbit: an escape burn from the surface of a low gravity body when you're facing the correct direction you want to travel out of the SOI.

[bold for conditions described that make the direct assent superior]

Are there any other situations where you think direct assent would be the better flight path?

[Kesa]

I'm glad you made a new thread with new argument in favor of vertical ascent, which I might have underestimated .

[rkman]

For all i know this is an issue of basic rocketry: the more time you spent 'fighting gravity' - that is to say moving more or less opposite to the direction of gravity instead of more or less perpendicular - the higher the gravity loss (speed loss due to gravity), and the more dV you need to get where you want to go.

That's why getting to LKO is more efficient with a gravity turn (and even something that looks like a gravity turn but isn't really a gravity turn) than by going straight up until you clear the atmosphere and then burning sideways.

Going vertical up to the Mun means spending even more time moving right against the direction of gravity.

[GoSlash27]

Arkie,

You missed the actual reason why your vertical profile showed smaller losses as compared to prograde, and I had specifically told you in that thread:

You had accelerated at a wastefully- high rate, and your drag losses were exorbitant. This gave you a greater DV drag penalty in the prograde gravity turn than in the vertical, and made the vertical escape seem less wasteful in comparison, when in reality both of your launches were terribly inefficient.

How much DV did you expend total in your prograde ascent?

*edit* on the subject of munar escape, prograde acceleration as rapidly as possible. Always and no matter what circumstances. Vertical costs you more in gravity losses, robs you of free velocity from rotation, and kills your Oberth effect gains.

Best,

-Slashy

Edited December 15, 2014 by GoSlash27

[Superfluous J]

*edit* on the subject of munar escape, prograde acceleration as rapidly as possible. Always and no matter what circumstances.

I take it you've never landed in a crater or valley? Prograde when landed is directly East. If you burn directly East without going up at least a *little* bit first you're in for a world of hurt.

[Pecan]

Sorry to bring this up again, guys...

Please don't apologise, I think it's really important that we have at least five different live threads dealing with the same subject.

How many km/s are you saving in the best ascent with mishaps than you are in the 'safe' ascent with perfection? Conversely, how perfect do you have to be in the perfect ascent to gain anything over the worst-case 'safe' approach?

Ascent through atmosphere, with TWR, drag, gravity-turn start, rate, etc. has so many variables it is impossible to define ANY path that is best. Trying to define ONE path that not-only takes that into account but a guaranteed fail. I guarantee you won't reach 75km exactly where, when and at with the vector you expected in the first place. Your continued-burn is therefore wrong and will magnify the error. Now you're thousands of kilometers out and have to make an adjustment-burn. Better to have taken it in steps?

[GoSlash27]

I take it you've never landed in a crater or valley? Prograde when landed is directly East. If you burn directly East without going up at least a *little* bit first you're in for a world of hurt.

Either a failure to clarify on my part or a failure to comprehend on your part...

I'm talking about prograde escape vs. vertical in terms of efficiency. Vertical is not ever going to be more efficient than horizontal.

It should be self- evident that smacking into terrain is *not* an efficient escape profile.

Best,

-Slashy

[Empiro]

Sorry to bring this up again, guys, but I think we have overlooked something.

GoSlash27 http://forum.kerbalspaceprogram.com/threads/102606-Transfering-to-Mun-First-to-LKO-or-Directly-to-Mun?p=1593793&viewfull=1#post1593793 said that with his software, it will take approximately 340 m/s extra deltaV to go vertical ascent directly to the mun vs. performing gravity turn into LKO, then periapsis transfer to the Mun. In my own experience, the difference is smaller, around 70 m/s

. Possible reasons for this discrepancy include my use of FAR aerodynamics (not sure which aerodynamics GoSlash27's software uses, or if it uses aerodynamics at all), and the potential for a pilot-induced sub-optimal gravity turn, which could easily bleed off a few hundred m/s deltaV if you raise apoapsis too high upon liftoff or stay too low for too long. Thus, vertical ascent could match gravity turn under the right conditions.

Nevertheless, Kerbin isn't the only planet. My question is if you are on the Mun such that Kerbin is perpetually setting (since the Mun is tidally locked) i.e. on the rear-side of the planet w.r.t. Mun's velocity vector, do you want to get into LMO first or go straight vertical to return to Kerbin.

My preliminary math shows it takes less deltaV to go straight up (807 m/s) vs. hopping up to 10 km altitude (to avoid hitting terrain) and then burning horizontally until escape velocity is achieved (964 m/s). Since Mun's gravity is small, TWR will inevitably be large so gravitational losses will be small/negligible.

What do you guys think?

If i am right, it implies that vertical ascent vs. LXO-to-transfer optimality is dependent on the planet/mun and other conditions, rather than always assuming vertical ascent is bad.

Thanks for showing the video. I think it's allowed us to see what's going on and give analysis and suggestions.

The thing about your craft is that it has a really high TWR, and as such you're not really performing a gravity turn since the TWR is so high. If you look at some FAR ascent videos with TWR of about 1.4-1.5, you'll see what happens is that at 50 m/s, you pitch over about 5 degrees (depending on your TWR and other factors), and then turn off SAS and basically let the rocket fly itself. This works well for small and large rockets, and makes the likelihood of breaking apart due to aerodynamic stress very low.

Gravity turns actually require very little control input, because you need to pitch over so little. 4 of the controllable winglets is sufficient (and as far as I can tell, in FAR with large crafts, drag is extremely high if you don't have fins of some sort, so you need them no matter what).

I recommend that you try to build the same rocket except with just a skipper engine instead of a mainsail (it will have a TWR of 1.47). Do a gravity turn like I described above. if you do it right, you shouldn't have to give any input after the initial turn, and your trajectory will be about 45 degrees at 10km, and flatten out fairly close to horizontal when your AP reaches 80km or so. Circularize at 80km, and transfer to the Mun. I bet you'll wind up with more delta-V than either of your two existing tests, even though this rocket is cheaper, and has less thrust.

You might have to try a few times to get things just right, but like any skill it takes practice, and once you get it, you'll be able to reliably do it in the future.

Edited December 15, 2014 by Empiro

[arkie87]

For all i know this is an issue of basic rocketry: the more time you spent 'fighting gravity' - that is to say moving more or less opposite to the direction of gravity instead of more or less perpendicular - the higher the gravity loss (speed loss due to gravity), and the more dV you need to get where you want to go.

That's why getting to LKO is more efficient with a gravity turn (and even something that looks like a gravity turn but isn't really a gravity turn) than by going straight up until you clear the atmosphere and then burning sideways.

Going vertical up to the Mun means spending even more time moving right against the direction of gravity.

You basically ignored the concrete mathematical numbers i gave in favor of vertical ascent for this case....

[arkie87]

Arkie,

You missed the actual reason why your vertical profile showed smaller losses as compared to prograde, and I had specifically told you in that thread:

You had accelerated at a wastefully- high rate, and your drag losses were exorbitant. This gave you a greater DV drag penalty in the prograde gravity turn than in the vertical, and made the vertical escape seem less wasteful in comparison, when in reality both of your launches were terribly inefficient.

How much DV did you expend total in your prograde ascent?

*edit* on the subject of munar escape, prograde acceleration as rapidly as possible. Always and no matter what circumstances. Vertical costs you more in gravity losses, robs you of free velocity from rotation, and kills your Oberth effect gains.

Best,

-Slashy

I thought we resolved this?

Your argument is 100% false with FAR installed.

In stock, i would have FAR (pun intended) exceeded terminal velocity, so i would have wasted too much deltaV fighting drag.

With FAR installed, I am always wayyyy below terminal velocity, as per this video (

), and so, I am actually wasting more fuel fighting gravity than drag...

[GoSlash27]

Just noticed that the OP was mistaken about a critical item.

My numbers were not derived from some software. They were empirical results from actual flights.

So when I reported a 340 m/s difference, I wasn't saying that I plugged some numbers into a piece of script and it said "340".

I *personally* flew these profiles in a very efficient rocket in KSP and recorded how many m/sec it took.

Best,

-Slashy

[arkie87]

Please don't apologise, I think it's really important that we have at least five different live threads dealing with the same subject.

Sarcasm much?

[GoSlash27]

No, we had not resolved this, which is why you now have 3 different threads arguing the same fallacious theory.

How much DV did it take you to achieve apoapsis in your tests?

*edit* sorry, 5? We're up to 5 threads now? I'd lost count... /edit

[arkie87]

Thanks for showing the video. I think it's allowed us to see what's going on and give analysis and suggestions.

The thing about your craft is that it has a really high TWR, and as such you're not really performing a gravity turn since the TWR is so high. If you look at some FAR ascent videos with TWR of about 1.4-1.5, you'll see what happens is that at 50 m/s, you pitch over about 5 degrees (depending on your TWR and other factors), and then turn off SAS and basically let the rocket fly itself. This works well for small and large rockets, and makes the likelihood of breaking apart due to aerodynamic stress very low.

Gravity turns actually require very little control input, because you need to pitch over so little. 4 of the controllable winglets is sufficient (and as far as I can tell, in FAR with large crafts, drag is extremely high if you don't have fins of some sort, so you need them no matter what).

I recommend that you try to build the same rocket except with just a skipper engine instead of a mainsail (it will have a TWR of 1.47). Do a gravity turn like I described above. if you do it right, you shouldn't have to give any input after the initial turn, and your trajectory will be about 45 degrees at 10km, and flatten out fairly close to horizontal when your AP reaches 80km or so. Circularize at 80km, and transfer to the Mun. I bet you'll wind up with more delta-V than either of your two existing tests, even though this rocket is cheaper, and has less thrust.

You might have to try a few times to get things just right, but like any skill it takes practice, and once you get it, you'll be able to reliably do it in the future.

I will try with lower TWR. I assume you will be correct-- since vertical ascent is detrimental for low TWR...

A fairer comparison would be for the same payload, what is cheapest (in terms of Kerbucks) way to get to a certain apoapsis?

I imagine a skipper would be best for LKO-to-mun approach but strapping loads of SRB's (which are cheap) would be best for vertical ascent approach.

I will actually test this and post results. Perhaps, i will also try reducing thrust on SRB's and going for a LKO-to-Mun transfer as well.

- - - Updated - - -

No, we had not resolved this, which is why you now have 3 different threads arguing the same fallacious theory.

How much DV did it take you to achieve apoapsis in your tests?

So you disagree that terminal velocity is optimal ascent velocity? Or what?

[arkie87]

Just noticed that the OP was mistaken about a critical item.

My numbers were not derived from some software. They were empirical results from actual flights.

So when I reported a 340 m/s difference, I wasn't saying that I plugged some numbers into a piece of script and it said "340".

I *personally* flew these profiles in a very efficient rocket in KSP and recorded how many m/sec it took.

Best,

-Slashy

That's even better since you do not use FAR... and as we've established, terminal velocity in stock is much lower than FAR, so a vertical ascent will suffer the consequences...

[Kesa]

Just a note to say that if you really want to compare the two methods of ascent, you need two different types of craft.

Saying "my 4 TWR ratio rocket failed to make orbit" as well as saying "kerbal X fail to make vertical ascent" aren't valid arguments.

[arkie87]

Just noticed that the OP was mistaken about a critical item.

My numbers were not derived from some software. They were empirical results from actual flights.

So when I reported a 340 m/s difference, I wasn't saying that I plugged some numbers into a piece of script and it said "340".

I *personally* flew these profiles in a very efficient rocket in KSP and recorded how many m/sec it took.

Best,

-Slashy

I have revised OP to reflect this information. By the way, it only enhances my argument... since you are using stock aerodynamics...

[GoSlash27]

So you disagree that terminal velocity is optimal ascent velocity? Or what?

This would be resolved by you going back to one of your other threads on this subject where I've already debunked that notion.

*this* thread is dealing with airless bodies where "terminal velocity" doesn't exist. And vertical escape isn't more efficient here under any circumstances either.

-Slashy

[Wanderfound]

Sarcasm much?

Yes, it's sarcastic, but it's not entirely unjustified. One thread to discuss this is fine; five is overkill. Keep the discussion going, sure, but there's no reason to not keep it all in one thread. Duplicating threads just results in endless re-hashing of the already settled points.

[GoSlash27]

I have revised OP to reflect this information. By the way, it only enhances my argument... since you are using stock aerodynamics...

Your argument is incorrect, so it doesn't matter how much it's "enhanced".

This is a very simple matter to correct if you'll just answer the question I've asked you twice: How much DV did it take for you to establish your periapsis? <-- that makes 3 times

[arkie87]

Either a failure to clarify on my part or a failure to comprehend on your part...

I'm talking about prograde escape vs. vertical in terms of efficiency. Vertical is not ever going to be more efficient than horizontal.

It should be self- evident that smacking into terrain is *not* an efficient escape profile.

Best,

-Slashy

It's not always possible to burn prograde immediately. Examples include if you are in an atmosphere or deep in a crater, you have to burn vertically first to escape dense atmosphere or rise up out of the crater. That said, unless you have infinite TWR, you also cannot burn horizontally since you need to first clear the ground, which requires some wasted deltaV to climb. On the Mun, Ive heard safe altitude is 10 km, so assuming you have to fly up to 10 km before you can turn sideways (or even if you burn 45 degrees or less but still rise up to 10km), it appears vertical and horizontal methods are close enough, that vertical might win under the right conditions.

[arkie87]

Your argument is incorrect, so it doesn't matter how much it's "enhanced".

This is a very simple matter to correct if you'll just answer the question I've asked you twice: How much DV did it take for you to establish your periapsis? <-- that makes 3 times

If you are asking me how much deltaV vertical ascent takes? You can look it up in my video? Otherwise, not sure what you are asking....

Also, rather that asserting it is "incorrect", I would appreciate if you explain why, rather than cockily asserting your knowledge trumps mine.

[arkie87]

Yes, it's sarcastic, but it's not entirely unjustified. One thread to discuss this is fine; five is overkill. Keep the discussion going, sure, but there's no reason to not keep it all in one thread. Duplicating threads just results in endless re-hashing of the already settled points.

The reason i started a new one is because I wanted to start out with a more refined statement of the problem, so that people dont chime in with information they think is new or will resolve the issue, despite it being mentioned 5 pages ago, and we are currently debating something else...

[arkie87]

This would be resolved by you going back to one of your other threads on this subject where I've already debunked that notion.

*this* thread is dealing with airless bodies where "terminal velocity" doesn't exist. And vertical escape isn't more efficient here under any circumstances either.

-Slashy

I dont think you've debunked that notion at all.

So by all means, debunk away.

Terminal velocity here is infinity. Terminal velocity with FAR installed, is essentially infinity as well...

Once again, stop asserting your answer is correct and PROVE it.