Launch Efficiency Exercise [Updated for 0.21.1]

[Tarmenius]

*Updated for 0.21.1*

The craft has been updated to reflect the changes in 0.21.1. Instead of the old ASAS, an Avionics Package and four(4) Z-100 radially-mounted batteries provide the equivalent mass (see the picture below) of the previous design. Since the new craft's drag profile may be different, and given the incredibly narrow margin between entries, I will also be adding a new scoreboard to reflect the change in characteristics. As always, input is welcome, and have fun!

With so many new players joining over the recent months, I thought it might be nice to revisit an old challenge focused on a basic and fundamental part of the game: Launching a craft into orbit.

Here we will take an incredibly simple design and aim to put it in an orbit no lower than 74km x 74km. Note that having an Apoapsis greater than 75km is perfectly acceptable, but will obviously result in more fuel being spent and therefore a lower score.

Speaking of which, your score will be the amount of liquid fuel remaining. No math necessary!

Given its very simple design, duplicating it yourself is the fastest option. However, I can also upload the craft file to the SpacePort on request. When re-creating it yourself, please be sure to use the LV-T30 engine and all four radial batteries. You may also want to use launch clamps (I find it sways on the launchpad), though this is not required. And if you’re so inclined, MechJeb is totally allowed. Several participants in the past version of this challenge were able to out-perform the ascent auto-pilot, so I don’t think it gives an unfair advantage to anyone.

The updated craft for 0.21.1:

You will have a pretty tight fuel margin, so good throttle management and a proper ascent profile will be very important. To submit an entry, please include a screenshot of your Pe/Ap from the Map View as well as one showing your remaining fuel and total battery capacity. Of course, if you can include all data points in one screenshot, that's great, too! Additionally, if you want to share the specifics of your gravity turn, be encouraged to do so. As this challenge focuses on the basics, I’m hoping this will provide new players a way to improve their skills and learn from other participants.

Here is my own sample entry:

A score of 6.62! Woo! That beat my original submission! But, I’m sure many of you can do better.

To provide readers with a quick way to gauge the relative performance of MechJeb's auto-pilot versus manual piloting, I will note its use next to entries on the Scoreboard. I will also note entries that used Information Panels only, whether from MechJeb or Kerbal Engineer Redux. Bear in mind that this in no way reflects a less valid or less impressive entry. Programming the auto-pilot effectively requires as much knowledge of an efficient ascent path as any manual pilot.

Scoreboard:

1. Nao: 10.55 (Auto-pilot)

2. Raptor831: 10.44 (Auto-Throttle/ Manual Steer)

3. tavert: 10.29 (Auto-pilot)

4. metaphor: 10.25 (Vanilla)

5. honolululu: 10.10 (Vanilla)

6. Bystander: 10.04 (Auto-pilot)

7. mootavic: 9.56 (Auto-pilot)

8. Danger Will Kerbinson!: 9.50 (Info Panel)

9. totalitor: 8.23 (Vanilla)

10. g00bd0g: 8.13 (Auto-pilot)

11. zarakon: 8.08 (Vanilla)

12. Bobnova: 7.84 (No Assists)

13. Giggleplex777: 7.78 (Vanilla)

14. tssn1611: 7.74 (Vanilla)

15. NeilC: 7.71 (Auto-pilot)

16. Quintic: 7.12 (Vanilla)

17. Tarmenius: 6.62 (Vanilla)

18. Epthelyn: 6.23 (Auto-pilot)

19. PakledHostage: 4.72 (Vanilla)

20. AndreyATGB: 4.44 (Info Panel)

21. Nurph: 4.09 (Vanilla)

22. Goozeman: 4.07 (Attitude-assist)

23. mcirish3: 2.39 (Vanilla)

24. Respawn: 1.43 (Info Panel)

25. Enigma179: 0.00 (Info Panel)

Double-Digit Club

tavert (First Member)

mataphor

honolululu

Raptor831

Bystander

Nao

Have fun!

Edited September 10, 2013 by Tarmenius
Updated Scoreboard

[Designer225]

I really don't need MechJeb for ascenting to orbit, and I'm sure many people knows how to ascent without MechJeb, but perhaps using MechJeb would give some unfair advantages to MechJeb users. Don't you think that there's a cost involved? Like the score is 80% or 90% of the remaining fuel?

[Tarmenius]

I know that most of us who have been around for a while will know how to get into orbit without any fuss. It's quite routine for me as well. But improving efficiency during the launch can do wonders to reduce the size of the rocket being used and it can take new players a while to learn how to manage their launches. Really, this challenge is for any player who wants to learn how to make a more efficient ascent, new or old.

As for MechJeb, if it could perform the ascent far more efficiently than any human pilot, then I'd see that as an unfair advantage. But last I checked, this wasn't the case. Thus, no score penalty to MechJeb users.

[NeilC]

I managed to do a bit better with MechJeb: 2.60 fuel remaining.

There's not much spare dV on this craft - it took me 4 tries to get an ascent profile that would achieve orbit. That said, I spent less than 10 minutes on this so i doubt my ascent profile is anywhere near optimal - I stopped as soon as I beat 1.33. With a craft this basic and easy to control, I don't think MechJeb offers any advantage except repeatability so we can compare pure ascent profile efficiency without any pilot error messing things up.

Looking forward to seeing others optimize a few more fuel units out of this, but I'd be impressed to see someone hit double digits. This craft is toight like a toiger.

[zarakon]

8.08

[Tarmenius]

I don't think MechJeb offers any advantage except repeatability so we can compare pure ascent profile efficiency without any pilot error messing things up.

That's my thought exactly. And if someone was able to program MechJeb's auto-pilot so well that it beats the rest of us, I will be impressed. In my experience, you simply can't tweak MechJeb's profile minutely enough to compete with manual piloting.

Good work, guys! The scoreboard has been updated accordingly. I can't wait to see what people are capable of with this craft. NeilC, was your first four attempts also with MechJeb, or were you also able to achieve orbit manually? Just a point of curiosity, of course.

[metaphor]

9.23

I didn't use MechJeb's autopilot, just the info. MechJeb tends to accelerate beyond terminal velocity around 10km altitude. I kept it at 2/3 throttle after 10km. At 10km I was at 65 degrees up, and finished gravity turn at ~40 km and 10 degrees up. That seemed to work out best.

It seems like it's definitely possible to get to double digits.

Edited July 6, 2013 by metaphor

[totalitor]

Good challenge! I was thinking almost the same. My result is 7.35

[Tarmenius]

I kept it at 2/3 throttle after 10km.

That's about where I keep it, too. Do you perform your gravity turn in one slow maneuver, or in steps? I do mine in about 4 or 5 small steps, turning to particular pitch settings at specific altitudes. For example, one step is pointing to +30degrees pitch at 30km.

It certainly does look like double digits is a reachable goal. I wonder who will break it first!

[PakledHostage]

Here's my entry. I'm on my way out to go sailing so I'll have to fill in the details later, but here are the screen shots:

[Enigma179]

Apparently I need to work on my efficiency. First try I didn't even manage to get a periapsis above zero, second try I barely managed to pull it out with no fuel left;

Still, as someone who usually sets aside 5000 delta v instead of 4500 for my launches, I'm happy with this result for now.

[koshelenkovv]

PakledHostage, you have a decoupler instead of a heavier ASAS unit.

[Tarmenius]

Apparently I need to work on my efficiency. ...I'm happy with this result for now.

I'm glad you could get it to work! With a couple more tries, I'm sure you could beat at least my own score!

[nhnifong]

Metaphor, I'd be really interested in watching your ascent on video. Are you up for it?

[PakledHostage]

PakledHostage, you have a decoupler instead of a heavier ASAS unit.

OK. My bad. I'll have a look at it again later and give it another try.

[mcirish3]

Well here is my entry 5th place I Think.

Edited July 7, 2013 by mcirish3

[Tarmenius]

mcirish3: Congrats on a successful orbit, and thanks for the video submission! The numbers were kind of hard to discern, but I think I've gotten the score correct. If I have it wrong, please post a screenshot for clarification

And PakledHostage, it's good to see you're still around. I'm looking forward to finding out what you can accomplish here.

[honolululu]

Trying a couple new things today. At least one was successful. My first challenge! I hope I did it correctly. 6.75 units left.

Click for embiggening.

[NeilC]

After many attempts and much tweaking, here's my latest score: 7.26

Tarmenius: I do just about all my flying with Mechjeb. I enjoy KSP more for the engineering and planning aspect than the piloting aspect. Perhaps that's mostly because in my case, pilot error swamps all other factors.

metaphor: I second the request for a video, and/or an explanation of your strategy. When you're flying manually, what's your thought process? I lack the "feel for flying" you must have, and I'm at a loss for how to do better than this.

This ascent profile took me out of the prograde reticle a LOT more than I would have liked (hence my higher steering losses), and yet when I tweak the profile and/or max accel very slightly so my steering losses are down to 5m/s I wind up with LESS fuel remaining due to higher gravity losses. I know this is because I'm pointing "up" for too long and gravity is fighting my accel... so I try to tweak down and wind up with LESS fuel remaining due to higher drag losses! What's the sweet spot?

Also, metaphor, from my experience lurking these challenges it's usually either you or tavert who defines what's possible around here. If you think double digits is possible, put your money where your mouth is.

Edited July 7, 2013 by NeilC

[Tarmenius]

honolululu, Congrats on your first Challenge! You certainly did it correctly and performed admirably.

NeilC, I totally understand the engineering draw of this game. It's where I spend most of my time. And a very nice improvement on your previous score! My experience with MechJeb is pretty limited, but I wonder whether adjusting the "Turn Shape" setting will help reduce steering losses while also keeping gravity losses down. Unfortunately, it might translate into slightly higher drag losses. I suppose it's a matter of "Which force applies the greater influence over the duration of the launch?" and focusing on bringing that down.

[mcirish3]

Sorry about the bad Video I will submit a much better quality one tomorrow with a much better result. Probably not till tomorrow evening though.

[NeilC]

My experience with MechJeb is pretty limited, but I wonder whether adjusting the "Turn Shape" setting will help reduce steering losses while also keeping gravity losses down. Unfortunately, it might translate into slightly higher drag losses. I suppose it's a matter of "Which force applies the greater influence over the duration of the launch?" and focusing on bringing that down.

Yeah, that's the crux of it. Throttle control is also very important to minimize drag losses, since they're a cubic function in velocity (IRL anyway, I know KSP's current aero model is lacking) accelerating too hard too early means a lot of wasted energy to drag. The rule of thumb is to coast at terminal velocity (drag=1G) until you're above the thick stuff, but there's lots of room for optimization with any specific craft. I spent a couple hours today tweaking the gravity turn start altitude (1-6km), turn shape(50-100%), and max accel (22-30m/s, but always limited to terminal). I'm at the point now where my gut isn't giving me any more guidance - short of running a giant test matrix with 100s of flights until I can back out some trends, I'm not sure how to improve.

[metaphor]

Metaphor, I'd be really interested in watching your ascent on video. Are you up for it?

Sure if I can find a good way to record video.

I use MechJeb for info, so that might be a big advantage. The terminal velocity is pretty important, I try to move as close to terminal velocity as possible, at least for the first part of the flight. Go straight up until 6km or so, start tilting rocket then, get to about 65 degrees by 10km, and then start changing the angle on the navball a little faster until you're pointed at around 10 degrees by 40 km altitude (and that's around the cutoff time until apoapsis). I cut the throttle down to 2/3 after 10km instead of accelerating then. I used an initial apoapsis at 70km before bringing my orbit up to 75km (that's why my screenshot is on the night side), but I don't know how much of a difference that makes. Had about 5 attempts at around 7 or 8 fuel left before I got one with 9.

If you're having too many steering losses from pointing too far ahead of your prograde, try lowering your throttle so the prograde marker has time to catch up with you. I don't really know what else I can say. I'll do some more attempts to maybe get a 10.

[mcirish3]

Sure if I can find a good way to record video.

Check this out it works great and is free.

ScreenCapture

http://sourceforge.net/projects/screencapturer/files/

[Guest doughbred]

Well I can tell you're better than me.