Launch "Kerbal 1-5" to LKO

[Rodhern]

8 minutes ago, Reactordrone said:

Well if we can adjust the boosters, then here's one I prepared earlier. 100% thrust on the boosters and don't start the core engine up until booster burn out. No mono used, 72.82 fuel left.

Hi, you might not have noticed, but I actually did list you with 73 units (rounded) in the FAR and GOaP list in the OP earlier today.

Technically you aren't allowed to adjust the booster level, because, for the GOaP challenge, you must wait till you pass 1000 m ASL with the adjustments, and then it is too late.

As it turns out, players like to play around a little extra, to see what happens. Don't we all possess that KSP curiosity (if we got the game in the first place), so therefore I decided to add references to the slightly cheaty attempts as well.

[JAFO]

First entry under the new 1.2 conditions!

(For those wondering, 1.2 tweaked the gravitational constant slightly, and also made changes to the aerodynamic model. Basically, blunt objects generate more drag, and streamlined objects generate less drag, than before. Which makes the Kerbal 1-5 used in this challenge, a real slug.)

When I saw this challenge, and read through the rules for 'Vanilla AP', I immediately realised it was an absolutely perfect test for the Gravity Turn mod.

Unlike MJ, GT doesn't just fly a "good enough" launch profile. It aims to improve over a (limited) number of iterations/launches, and can dynamically adapt to unexpected events during flight. GT is also a tool in the true spirit of Unix programs, in that, rather than trying to be a jack-of-all-trades, it does just one thing, and does that one thing as well as possible. And that one thing is gravity turns. It won't even try to circularise, although if MJ is installed, it will hand off to MJ to auto-execute the circularisation burn. A good indicator of how well GT performs, is that typically, only 20-40m/s is required to circularise the orbit. It's as close as we currently have to a mod that can perform a continuous-burn-to-orbit maneuver.

GT works by setting a launch speed at which to begin the gravity turn, and a starting angle to turn to. After this, it attempts to maintain a set time to Ap throughout the launch, by controlling the throttle, down to a preset minimum throttle setting. Absolutely minimal use is made of pitch and yaw, except to maintain heading, or increasing the time to Ap if necessary, after jettisoning a stage, for example. Also unlike MJ, where GT really excels, is when you want to test for effects of even small changes to your craft. GT gives you the ability to perfectly execute identical, repeatable launch profiles time after time, allowing you to judge your changes by the final fuel consumption, itemised loss data, and dV figures in orbit. While I've yet to try it, I think its consistency is such that you could probably even use it to evaluate the difference that day/night atmospheric pressures make to a launch.

The only change made to the ship itself was the addition of a pair of launch clamps, as I'd noticed in earlier tests that the ship tended to sway around on the pad before launch, and could easily be already pointing at a small angle from the vertical, potentially advantageous, or disadvantageous, on takeoff. There's a saying on these forums about gravity turns, that "a degree makes a difference".. The truth is, as any GT user knows, even a tenth of a degree makes a difference, especially at lift-off. Accordingly, the clamps were used to eliminate this variable from the trials. No use was made of monopropellant in any of the flights, and gimbals were not locked. GT makes no use of SAS or reaction wheels during launch, and disables them whilst in control of the ship.

http://imgur.com/a/LbrSL

For this entry, as the rules allow, I used MJ to circularise to orbit. GT began with the default settings, and the "Improve Guess" button was used on each subsequent launch to try and improve on results. After several launches, GT decided it had done its best, and stopped trying to improve. At that point, I began trying to manually improve on GT's results, by fine-tuning the initial launch parameters, and after a few more launches, succeeded.

When this happened, GT "woke up" again, and attempted to improve on my hand-tweaked results, in which it succeeded handsomely.

Results:

Launch 1: Orbit = 80.062x80.138, LF = 31.47, Total dV expended = 3592.8. Score = 28.68%
Launch 2: Orbit = 80.087x80.176, LF = 39.35, Total dV expended = 3477.6. Score = 35.87%
Launch 3: Orbit = 80.082x80.181, LF = 26.85, Total dV expended = 3643.1. Score = 24.64%
Launch 4: Orbit = 80.099x80.187, LF = 37.47, Total dV expended = 3499.6. Score = 34.12%
Launch 5: Orbit = 80.102x80.197, LF = 35.62, Total dV expended = 3532.1. Score = 32.42%
Launch 6: Orbit = 80.118x80.213, LF = 39.04, Total dV expended = 3482.0. Score = 35.58%

At this point GT declared launch 2 the best, and stopped trying to improve. The next three launches were manual tweaks to try and improve on launch 2.

Launch 7: Orbit = 80.113x80.207, LF = 37.31, Total dV expended = 3507.5. Score = 33.97%
Launch 8: Orbit = 80.110x80.202, LF = 39.25, Total dV expended = 3479.0. Score = 35.78%
Launch 9: Orbit = 80.092x80.195, LF = 39.46, Total dV expended = 3475.8. Score = 35.97%

The next time I tried the "Improve Guess" button, GT realised I'd topped its score, and decided it wanted back in the game. Now it began trying smaller tweaks, in line with what I'd been attempting.

Launch 10: Orbit = 80.120x80.212, LF = 39.52, Total dV expended = 3474.8. Score = 36.03%
Launch 11: Orbit = 80.110x80.204, LF = 39.61, Total dV expended = 3473.5. Score = 36.11%
Launch 12: Orbit = 80.129x80.236, LF = 39.67, Total dV expended = 3472.5. Score = 36.17%

So I'm officially submitting 36.17% as my entry in the 'Vanilla AP' category. Whilst it's not hard for a human player to improve on MJ's launches, I'd be quite surprised if anyone can manually beat GT's score in 'Vanilla'.

Final settings for GT were as follows:
Speed to commence turn - 84m/s
Initial turn angle - 13.50°
Hold AP time start - 50 seconds (default)
Hold AP time finish - 50 seconds (default)
Sensitivity (min throttle) - 0.50 (default)
Destination Height (km) - 80.00 (default)

In a normal GT launch, I'd expect a lot of plasma effects, but due to the highly un-aerodynamic shape of the Kerbal 1-5 (frankly, from a drag perspective, this thing's a pig, -as demonstrated by the unusually high proportion of aero drag losses making up total losses-), best results were achieved by a launch profile that avoided the speed and altitude combination where such effects occur. I was similarly surprised by the high(ish) total dV required. I'm more used to seeing GT launches in the 3000-3200 dV range. My best result to an 80x80 orbit with GT (in KSP 1.0.5) is 2924.8m/s.

Edited October 15, 2016 by JAFO
#$%! this #$%&! album posting "system"!

[emaier138]

I found this thread while looking for info about more efficient gravity turns and decides to give this a try.

Version 1.2 

Didn't change a thing, just loaded up and launched off the pad. Used SAS, as well as manually throttling, and manually piloting/tweaking the whole time (no lock prograde and leave it.

 

80,022 x 80,212 with 33.75LF

*unsure how to paste a screenshot in the post, normally its just drag and drop on mac. Will edit when i figure this out*

[JAFO]

Good to see someone else giving it a shot..

 

3 hours ago, emaier138 said:

I found this thread while looking for info about more efficient gravity turns and decides to give this a try.

This thread really could use something about 'gravity turns' in the thread title. I'm sure it would attract a lot more entries, that way.

 

3 hours ago, emaier138 said:

80,022 x 80,212 with 33.75LF

 

So, after running those numbers through the Score Calculator Spreadsheet, that gives you a score of 30.72%

 

Edited October 16, 2016 by JAFO

[Rodhern]

The latest dev. notes have a vague hint that maybe gravity will change in version 1.2.1.

Quote

An example of these bugs people have found is the standardisation in gravity values to 9.80665, making the planet orbits shift ever so slightly and the geosync heights change - they’ll be back in their right positions, and another has been the UI flicker in the editor scene that is seen once you load up enough parts.

What on Kerbin that actually means, I am not able to tell. I have not updated the score calculator spreadsheet yet, neither have I updated the OP with the latest entries (JAFO and emaier138). I will get to that at some point - I might just wait and do it in KSP ver. 1.2.1 instead.

 

@JAFO:

Thank you. Your experiments go a long way to answer the original question about just how much fuel a gravity turn saves you. (To know the potential saving we must search for the 'optimal' ascent - as you are doing).

As for the pure fair competition part, probably launch clamps aren't entirely fair. The launch clamps are there to reduce the randomness, to make it easier to optimize the trajectory. I get why you use them, and indeed a reproduceable result is more useful to the question as a whole. Maybe once KSP version 1.2.1 is here you can try a few launches without the clamps (hint hint).

 

@emaier138:

Thank you. It is always nice to see a vanilla entry. Technically I think you are the only one so far that followed every rule to the letter. One might argue you are in the number one spot of this challenge at the moment.

[Eidahlil]

For comparison I decided to do a tried and true "lock SAS to Prograde, 45 degrees pitch at 12km altitude" approach.

Craft unmodified (gimbals were locked though), KSP version 1.2, new sandbox game on normal, used MechJeb to stage and circularize, otherwise just good old SAS. In fact, I'm pretty sure I did not touch the controls at all between 2km and 50km.

42.23 (corrected number) LF remaining after getting to 80x80km. (80.083x80.014km to be precise, you can see it in the KER readout to the left of the altitude.)

I don't think I saw any heat bars, but Jeb found a brick under the throttle pedal. He was too busy trying to crush it with his foot to think of throwing it out the window (though opening the window might have caused some heating issues, assuming he would open it before throwing the brick out), so no impact on the test. It was later determined that procurement acquired a large batch of discount throttle limiters, and an overzealous intern applied them everywhere liberally. (I'm just trying to invent a somewhat reasonable explanation for having thrust limiters on everything in this ship.)

Here's a more detailed account of the launch, for those interested:

Spoiler

 

EDIT: looked at the wrong number for the score :)), thanks @Rodhern !

Edited October 25, 2016 by Eidahlil
Added exact orbit paremeters

[JAFO]

 

 

8 hours ago, Eidahlil said:

51.62 LF remaining after getting to 80x80km.

 

Wow.. 51.62 L..  That's a score of 47.97%!  That truly is an impressive result..

Edited October 24, 2016 by JAFO
I'm an idiot

[Eidahlil]

18 minutes ago, JAFO said:

 Such a shame you had to break rule 5 of the Vanilla Rules to do it.

Sorry for that paragraph if it's confusing. I did not touch the thrust limiters in any way, as specified above, only locked the gimbals, as rule 6 allows. What I did, was wonder WHY would anyone put them there. But I left them in place.

If you watch closely, you can actually see the thrust limiters are still in place when I'm locking the gimbals at the start of the video.

On 9/2/2016 at 3:30 PM, Rodhern said:

• You may not tweak the stock vessel in any way; in particular you must leave the thrust limiters in place.
• An exception to the above rule is that you are allowed to lock and unlock the engine gimbals (and/or change the gimbal limits). You are also allowed to change, disable and enable the reaction wheel mode to suit your flying style.

[JAFO]

Deleted due to severe case of foot in mouth.

Edited October 24, 2016 by JAFO
Nevermind.. I'm an idiot

[Rodhern]

10 hours ago, Eidahlil said:

51.62 LF remaining after getting to 80x80km.

Spoiler

 

I think you might have confused oxidizer and liquid fuel. 42.23 units of LF Is still a lot though

[JAFO]

8 hours ago, Rodhern said:

I think you might have confused oxidizer and liquid fuel. 42.23 units of LF Is still a lot though

So, (without knowing the precise final Ap and Pe, so assuming 80x80 exactly) that gives a score of 38.59%. Still an excellent achievement.

[Gordon Fecyk]

[12 FEB 2017] Regarding the wonky thrust limiter settings, this is from the RealPlume Stock Configs thread:

Known Issues:

• Thrust limiting on engines get reset upon install. BEWARE IN THE KSP TUTORIALS
• Switching to a vessel with an engine as a root part will cause the plume to not show.

So it appears this was my culprit. Any vessel loaded with thrust limiting has their limiting reset on load when RP Stock and SmokeScreen are installed. Now I know for future challenges regarding stock craft, tutorials and scenarios.

[resume original post]

If video entries are acceptable, here are two. I figured after six months of faffing about, this was a challenge I could actually take up.

In both of these I relied on the SRBs alone to get started, just setting the throttle to zero until they were about to burn out. This is the same method I use for most of my SRB-assisted launches, where I use the cheap boosters to get to thinner air and let the higher ISP engines take over from there. I then let Jeb do his job of staying prograde, using the throttle to adjust my ascent profile instead of trying to steer. I crashed many a Kerbal X before realizing that the pilot can do a better job than me of steering.

The first one (Might be glitched by a bad YouTube upload):

LF was 63.32, OX was 77.39, AP was 86.398 km, PE was 80.166 km. No visible shock heating, but mach effects were observed before SRB separation. Midday launch, start of new sandbox game.

The second one:

LF was 68.27, OX was 83.45, AP was 82.059 km, PE was 80.608 km. Shock heating observed at 28 km, but nothing tripped the temperature sensors. Midday launch; pretty much reverted to launch from first attempt.

(Added) I reviewed Ediahlil's video entry and noted my solid fuel consumption rate was almost three times that of Ediahlil's. Either I messed with the SRB thrust limiter setting, or a mod I installed did. Let's file my entries under GOaP or outright disqualification until I get back to check.

--

Edited February 13, 2017 by Gordon Fecyk
Added cause of my thrust limiters being weird

[tg626]

Used MechJeb

Settings shown on screenshot

[JAFO]

2 hours ago, Gordon Fecyk said:

In both of these I relied on the SRBs alone to get started, just setting the throttle to zero until they were about to burn out.

LF was 63.32, OX was 77.39, AP was 86.398 km, PE was 80.166 km.

The second one:

LF was 68.27, OX was 83.45, AP was 82.059 km, PE was 80.608 km. 

Going by the numbers, this is clearly a winning strategy.. well done!

[Rodhern]

@Gordon Fecyk: Is it a GOaP entry or a vanilla one? If it is a vanilla one I think something have gone wrong; the Kerbal 1-5 is supposed to have an initial TWR below 1 [edit: when the main engine is off] until at least some booster fuel is burned.

@tg626: Thank you. I am willing to call that a stable 80x80 orbit. I am not trying to pick on you, but to me it is a little bit disappointing to see a few meters of periapsis missing. The scoring method allows periapsis between 80 km and 100 km. The idea was to avoid forcing a bang on circularization. I realize that there is some advantage to do an exact 80x80 orbit; that is intentional on my part. Feel free to submit new entries - maybe you can even improve on your current score.

P.S. I am not updating the OP leader boards at the moment, I plan to get to that at some point, maybe next week, but possibly even later.

Edited October 25, 2016 by Rodhern

[Gordon Fecyk]

2 hours ago, Rodhern said:

Is it a GOaP entry or a vanilla one? If it is a vanilla one I think something have gone wrong; the Kerbal 1-5 is supposed to have an initial TWR below 1 [edit: when the main engine is off] until at least some booster fuel is burned.

I was attempting a vanilla entry. I didn't do anything to the craft itself; this is the stock 1-5 in the 1.2 build. At least I think it is; I'll double-check when I get home later. I've read in this thread that the boosters were thrust-limited on the stock craft.

Would RealPlume Stock be messing with the parts? If it's a problem with RealPlume or other visual mods, I'll try a completely unmodded install. Or does time of day really matter as you suggested in the original post, where the sea level air may be ever-so-slightly thinner? I'd be surprised if that was actually modelled in the stock game.

In the meantime, if you still want to put the entry in, you could do it as a GOaP entry subject to verification. Though under those rules I could probably save more liquid fuel by using RCS to close my orbit, as others here have suggested.

(Added) I reviewed Ediahlil's video entry and noted my solid fuel consumption rate was almost three times that of Ediahlil's. Either I messed with the SRB thrust limiter setting, or a mod I installed did. Let's file my entries under GOaP or outright disqualification until I get back to check.

--

Edited October 25, 2016 by Gordon Fecyk
Noted differences in solid fuel use

[JAFO]

1 hour ago, Gordon Fecyk said:

Or does time of day really matter as you suggested in the original post, where the sea level air may be ever-so-slightly thinner? I'd be surprised if that was actually modelled in the stock game.

Yes, it really has been modeled in the stock game, and has been for some time. Don't have any links handy to point to, but if you search for night/day atmospheric pressure differences in KSP, you should find it. To the best of my knowledge, nobody has yet tried to actually quantify the differences or their extent, but it would be nice to see it attempted.

[Gordon Fecyk]

I just tried this again. I found out that the stock RT-10 Hammer SRB was not changed by RealPlume Stock or another add-on; its file date is unchanged from when I installed the 1.2.0.1586 build. Also the Kerbal 1-5 stock craft in this build seems to have the RT-10 thrust limiters set to 100%. Just took a look at the craft files from my 1.2 and 1.1.3 installations, and noted that the 1.2 version has 100% where the 1.1.3 version has 40%. I'm pretty sure I didn't change the stock craft files in my 1.2 installation.

Also, I stuck a barometer on a probe sitting on the launch pad and left it for six game hours, leaving the pressure display turned on. 1000x time warp to six hours later, and the pressure value didn't change over that time. Assuming Kerbin doesn't have seasons because of its zero degree axial tilt, this seems like it wouldn't change ever. Is anyone up for hacking a craft file that suspends a probe on a 10 km launch clamp for a repeat test? Otherwise it doesn't appear time of day would have an impact on fuel economy.

I can try again with the SRBs set to 40% thrust and see what happens. Here we go:

The solid fuel consumption rate should match this time... I didn't show the resources tab during the ascent... ugh, but I ended up separating the main stage at a much lower altitude this time.

This result had 45.29 fuel and 55.36 oxidizer remaining. Apoapsis was 81.258 km and periapsis was 80.269 km. This one had shock heating effects at around 23 km, but nothing that tripped a temperature bar. This one should be fine for a vanilla entry.

Edited October 25, 2016 by Gordon Fecyk
Added attempt three with SRB limiters set

[JAFO]

54 minutes ago, Gordon Fecyk said:

I just tried this again. I found out that the stock RT-10 Hammer SRB was not changed by RealPlume Stock or another add-on; its file date is unchanged from when I installed the 1.2.0.1586 build. Also the Kerbal 1-5 stock craft in this build seems to have the RT-10 thrust limiters set to 100%. Just took a look at the craft files from my 1.2 and 1.1.3 installations, and noted that the 1.2 version has 100% where the 1.1.3 version has 40%. I'm pretty sure I didn't change the stock craft files in my 1.2 installation.

Very strange.. in my 1.2.0.1586 stock install (from a fresh clean download), the RT-10s on the Kerbal 1-5 have their thrust limiters set to 40%.

54 minutes ago, Gordon Fecyk said:

Also, I stuck a barometer on a probe sitting on the launch pad and left it for six game hours, leaving the pressure display turned on. 1000x time warp to six hours later, and the pressure value didn't change over that time. Assuming Kerbin doesn't have seasons because of its zero degree axial tilt, this seems like it wouldn't change ever. Is anyone up for hacking a craft file that suspends a probe on a 10 km launch clamp for a repeat test? Otherwise it doesn't appear time of day would have an impact on fuel economy.

I suspect that, like so many other things, once you have time warp running, the pressure value runs on rails, which would be why you didn't get to see any changes..  you'd have to go back to normal time at regular intervals and take readings, I think.

Edited October 25, 2016 by JAFO

[Gordon Fecyk]

44 minutes ago, JAFO said:

I suspect that, like so many other things, once you have time warp running, the pressure value runs on rails, which would be why you didn't get to see any changes..  you'd have to go back to normal time at regular intervals and take readings, I think.

Like this?

I paused on-rails time warp every hour and used physics warp at these points, then ran through another three hours and used no time warp to be sure. But if my craft files are different (coming from the Steam version) could other things be different as well?

--

[JAFO]

7 minutes ago, Gordon Fecyk said:

But if my craft files are different (coming from the Steam version) could other things be different as well?

Even if some craft files are different, I don't think the core executable would be.. and that's where the atmo pressure would be dealt with.

Re atmospheric temp/density/pressure changes from day to night-time, see this thread:

 

[Gordon Fecyk]

On ‎10‎/‎25‎/‎2016 at 6:27 PM, JAFO said:

atmospheric temp/density/pressure changes from day to night-time

With an actual craft and an RT-10 booster, this is what I ended up measuring:

Craft and sensors at 76 m on launch pad

Time	Apoapsis (m)	Pressure (kPa)	Temp (K)
0:01	62732	100.2166	310.66
1:01	62721	100.2174	311.39
2:01	62458	100.2169	306.91
3:01	62200	100.2166	302.74
4:01	62211	100.2175	299.37
5:01	62475	100.2173	304.72

The measured numbers don't always jive with the flight results, but the flight results do say time of day does have an impact. I am surprised Squad bothered to model this! More testing is required, but that's getting off-topic for this challenge.

--

[Rodhern]

KSP version 1.2.1 is here. The Kerbin gravity changed (again). I will at some point update the OP. Future entries are supposed to be flown in version 1.2.1.

[Rodhern]

Hello, I have updated the scoring to fit KSP version 1.2.1.

The changes turn out to be pretty insignificant. The Kerbal 1-5 have lost some engine performance (compared to KSP ver. 1.1.3) but is considered more aerodynamically streamlined now. As a result, the straight vertical ascent is a few seconds faster, which almost cancels out the degraded engine performance.

 

@Eidahlil:

On 10/24/2016 at 5:20 AM, Eidahlil said:

... (I'm just trying to invent a somewhat reasonable explanation for having thrust limiters on everything in this ship.) ...

Maybe the structural parts and the engines on this relatively basic level orbiter cannot handle a higher thrust setting. I imagine that even the 100 % setting on an RT-10 signifies the point just before breaking the thing.

 

@JAFO: Is GravityTurn up and running for KSP ver. 1.2.1? In that case maybe you want to fly a ver. 1.2.1 entry (at least once without the launch clamps).

 

[Gordon Fecyk]

A series of attempts (skip to five minute mark for submission) in KSP 1.2.1 with the stock and properly thrust-limited Kerbal 1-5. And sorry for all of the other apparently invalid attempts.

Final attempt has an apoapsis of 80.347 km, periapsis of 80.105 km, and liquid fuel remaining of 40.58 units. This would give a score of 37.04% but the base variables may have changed for KSP 1.2.1.

Please, if I missed something or invalidated the entry somehow, let me know and I'll try again. This is my first KSP challenge attempt ever and I don't want to get a bad rep by breaking rules.

--