[1.0.5] GravityTurn version 1.3.1 - Automated Efficient Launches (1.1 pre-release available)

[Overengineer1]

@mcirish3 Setting the sensitivity lower mostly just makes the launch take longer, without increasing efficiency.  A zero setting often (always?) causes the engine to kind of flutter on and off during the tail end of the launch, which just feels like a horrible idea even if I've never seen it cause an issue.  The recommended setting is 0.2 which seems like a nice middle ground without sacrificing fuel.

[Jim DiGriz]

I suspect that once you're horizontal and you've cleared the atmosphere enough that you can hit orbital velocity without burning up, that you should just max throttle and lift the apoapsis.  If you let it bleed out slowly you'll ultimately be increasing the altitude that you carry that fuel to before burning it, and that will be slightly less efficient (Oberth basically).  So I think sensitivity 0 should be less efficient...

[Overengineer1]

@Jim DiGriz: The whole point of this mod is that we don't have to have these kinds of hypothetical discussions about a set of principles that interact with each other in overwhelmingly complicated ways.  We can just test them and see what works the best.

I did 4 runs with the same ship, different Sensitivity values: 0.01, 0.1, 0.2, and 0.5.  There was no remarkable difference in any of them, all had around 1000 loss, with a difference between them of less than 1 percent, well under what we'd consider a statistically significant difference.

Anyway, the Oberth effect doesn't work that way.  It's not about being closer to the planet, it's about traveling faster: a rocket can be more effective when traveling at high speed.  If you're in orbit, then the point where you're traveling fastest will be where you're closest to the planet.  But when you're launching, the opposite is true.

[DerekL1963]

Great mod!   Much better than MJ's...

Looking forward to the 'auto-learn' function, because I do tend to re-use a lot of designs.   Question though, will it learn and remember even if you revert back to the launch pad?

Edited January 27, 2016 by DerekL1963

[Overengineer1]

@DerekL1963  Yes, it will learn from that, provided you go all the way to atmosphere exit before you revert.

[fourfa]

Any chance to throw a .version file in the zip package?  Manually managing lots of mods is a lot easier if they all have it

Edited January 27, 2016 by fourfa

[tater]

The staging of fairings is a little odd. Say I have a 2d stage, then a fairing, then a 3d stage. I have the 2d stage and separator together, then the fairing, then the next separator and engine together. It will stage the 2d engine, AND the fairing, often when the ship is still engulfed in flames.

[Overengineer1]

@tater Fairings are almost never worth carrying past about 20km.  It takes more fuel to carry their weight than they save in drag above that point.  You can adjust this behavior by changing the pressure cutoff setting.  You will probably make your launches less efficient if you make a drastic change from the default value.

[Dr.LoveJoy]

On 1/21/2016 at 4:26 AM, Shalfar said:

I'm seeing an issue where shortly after the beginning of the gravity turn the craft attempt to return to vertical flight for just a moment.  This is very pronounced for small rockets.  With stock physics this is troubling but can be overcome, but with FAR it sends the rocket tumbling.

This probably means your rocket is unbalanced, in terms of center of mass vs center of drag/lift. Also that the TWR is likely a bit low, because the program only seems to repoint if when the initial turn is executed, apoapsis starts to decrease or stalls out. I use SMART A.S.S for setups that are like that.

But also I have found that adding TWR keeps things pointed correctly. Fairings help with aero. And you can always add winglets, or wings with control surfaces at the very bottom of the stage to help counter the tip over problem. They have to be big enough to do it though :-) Small ones wont work that well on larger fuselages, unless it is a very small correction that is needed. However, best solution is increase TWR/SLT in  the first stage... that way, it won't try to repoint.

I have found when it works ok with this mod, generally it will work better in any situation. :-) I would turn on Aero effects so you can see the drag and thrust, etc, vectors while the rocket is in motion to determine your course of action.

[tater]

1 hour ago, Overengineer1 said:

@tater Fairings are almost never worth carrying past about 20km.  It takes more fuel to carry their weight than they save in drag above that point.  You can adjust this behavior by changing the pressure cutoff setting.  You will probably make your launches less efficient if you make a drastic change from the default value.

I was under the impression that the fairing mass was carried in the base, not the actual fairing (it should be in the fairing, obviously, but I read otherwise someplace). 

[Overengineer1]

@tater Test it yourself.  Your available delta-v will go up substantially when the fairing pops.  This is well known and well researched, by others and by me.

[tater]

Cool, it's nice to know they work properly.

[FiiZzioN]

12 hours ago, Overengineer1 said:

@tater Test it yourself.  Your available delta-v will go up substantially when the fairing pops.  This is well known and well researched, by others and by me.

That may be the case, but what about when players go about it in an RP aspect? Some players, such as myself, will carry fairings higher not only for aerodynamics but as "payload protection". If you go about it that way, you tend to jettison fairings, either stock or procedural, around 35km (or higher) when you are just about to enter or are in the vacuum that is space. Of course, this depends on the type of player as I've seen some people jettison at 30km, some as late as 45km or 50km.

Well, that's my take on @tater's original question anyways. 

[Overengineer1]

@FiiZzioN  That's why the pressure cutoff setting is available, a lower value will wait until the atmosphere is thinner at higher altitudes.  If you don't want to pop the fairing at all, either put it higher on the staging order or disable automatic staging.

[Hairy Al]

Any chance of putting a version number somewhere in the file? Either in a .version file or in the .zip file name would be good. It's difficult to tell whether I have the latest version otherwise. Thanks.

[DStaal]

All this talk of fairings...

This mod never pops my fairings off - even on single-stage rocket.  (Well, two stage: The engine, and the fairings as the next stage.)  Maybe because I'm using procedural fairings?  Is that something that could be fixed?

[Nothalogh]

Could you add a pre and post staging delay value, like mechjeb has, so we can tune it for a clean separation

[Geoclasm]

I love this  - is there a way to make use this to gravity turn into a polar orbit?

[mcirish3]

5 minutes ago, Geoclasm said:

I love this  - is there a way to make use this to gravity turn into a polar orbit?

Already in the mod.  It is called Inclination.  Set it = 90^o

[Birdco_Space]

Awesome mod.  It takes a lot of the drudgery out of launches, and allows a good deal of optimization of designs by taking human error out of the launch equation.

Question, is it possible to change the default values for the plugin?  I don't use the best guess, instead i've found a group of setting that reliably gets me to orbit for between 3300 to 3400 m/s regardless of the rocket,  I use this consistency to ruthlessly optimize my designs.  it would be nice if I didn't need to change the 4 values I routinely change every time I launch

Specifically, I've found a starting time to AP of 30 sec, and an ending of 60 sec is a pretty efficient curve, is forgiving of low TWR designs, and gets me out of the really flamey parts of the atmosphere for not too much of a dV penalty.

 

 

[mcirish3]

2 hours ago, Birdco_Space said:

Specifically, I've found a starting time to AP of 30 sec, and an ending of 60 sec is a pretty efficient curve, is forgiving of low TWR designs,

For low TWR designs these are very near @Overengineer1's recomendations.  I am not sure that he has worked out how to optimize his best guess for low TWR ships, but he is aware of the issue.  he built this mod primarily for KSP players who tend to build rockets with too much TWR.  That said his latest DEV version does "learn" with each launch for each ship design.   In other words the more you repeatedly launch the same ship design the better it will do at predicting the optimal trajectory for that ship.
 

@Overengineer1 in regards to your new "learning system"  you mentioned that it needs to make it out of atmo before it remembers the launch profile.  Have you found a way for it to detect if the profile caused the ship to burn up?  This may be an important second condition you may want it to learn.

[Overengineer1]

@mcirish3 I'm having it build a kind of "database" of launches for a particular ship design.  It is tracking things like total loss and max critical heat, so yes, burning up is something that it will learn from.

Messing around with the AP time settings is a little hard to quantify for the guess.  Right now I'm thinking of adding a special case for SRBs to extend the APTimeStart, so that the rest of the launch fades gradually down to a 40 second AP Time, instead of currently where the SRB stage will push the AP time out to a minute or more and the mod will throttle all the way back.  This will make a big improvement at least for a lot of my personal ship designs.

Version 1.3 will include this SRB feature as the only adjustment to AP Time settings.  The 1.3 guess improvements will be made only to the Start Time and Turn Angle for this version.  Later on maybe I can better quantify when/how/where/why to make other changes to the AP Time guess, but I have to take it one step at a time.

[mcirish3]

Dude, thank you and as always remember you have (at least in my opinion) the best  auto assent mod available for KSP, so any improvement you make will just raise the bar for any potential competing mods that come along.

[Geoclasm]

18 hours ago, mcirish3 said:

Already in the mod.  It is called Inclination.  Set it = 90^o

I figured it had to be I just didn't know how. Awesome - thanks.

[mcirish3]

54 minutes ago, Geoclasm said:

I figured it had to be I just didn't know how. Awesome - thanks.

It was not in the very first release but I do believe it was the very reason for the first update.