A Degree Makes a Difference (Stock Edition)

[Norcalplanner]

This is a continuation of my previous study on TWR, ascent profiles, and aero drag vs. gravity losses begun in a previous thread, A Degree Makes a Difference - or How I Learned to Stop Worrying About Drag and Concentrate on Gravity Losses. That previous post was performed in RSS, and while many found it interesting, it wasn't directly applicable to a stock game.  To remedy this, I've flown a bunch of ascents on stock Kerbin, and tabulated the results.  As before, I created a basic rocket, but this time used only two Kickbacks because of stock's lower delta V requirements.  The Kickbacks were then thrust limited to 70%, 85%, and finally 100%.  After an initial tip to the east, MechJeb's SmartASS function would hold the craft on a surface prograde vector.  Once the navball switches to an orbital reference, SmartASS is told to hold orbital prograde until the Ap is over 100 km.  MechJeb is then told to circularize at Ap.

Here's a picture of the rocket, which is all stock:

After conducting over 20 runs, I've created albums with the best runs at each thrust level, and pasted links to them below.

View post on imgur.com

 

View post on imgur.com

 

View post on imgur.com

Remember, the only difference in these runs is the thrust level and the ascent profile.  The rocket itself is identical.

Most of the previous tips still apply, so I'll put them here again with some tweaks applicable to stock:

1. Launch with an initial TWR of at least 1.4.  Piloting gets more difficult when it's over 1.7.

2. Don't worry about drag - gravity losses are much larger and more important, unless you reach orbital velocity below 30 km..

3. Ignore the flame effects.  Pay attention to the temperature gauges.

4. At higher thrust levels, crank it to the east immediately after launching, but be precise about it.

5. Try to keep vertical velocity below 500 m/s.  If it's over 700 m/s, you're going to have noticeably higher gravity losses.

6. Getting a rocket to orbit in stock for less than 3,200 m/s of delta V is very doable.  Using less than 3,100 m/s is a harder but still achievable goal.

Edited August 20, 2016 by Norcalplanner
Added a photo

[Norcalplanner]

Here's some comparative data from the best runs at each thrust level.

Comparative TWR Ascent Data

TWR	Turn	Steering	Aero	Gravity	Delta V Spent	Delta V Left
1.25	5 deg @ 40 m/s	-5.2 m/s	-76.0 m/s	-1,123.4 m/s	3,234.4 m/s	2,773 m/s
1.41	5 deg @ 20 m/s	-3.1 m/s	-101.2 m/s	-961.9 m/s	3,102.6 m/s	2,957 m/s
1.57	8 deg @ 10 m/s	-2.4 m/s	-103.4 m/s	-884.2 m/s	3,030.6 m/s	3,061 m/s

Everything is fairly self-explanatory - initial SL TWR, initial eastward turn angle and speed, steering losses, aero losses, gravity losses, total delta V spent, and delta V remaining in the tank of the craft's upper stage.

[suicidejunkie]

1 hour ago, Norcalplanner said:

5. Try to keep vertical velocity below 500 m/s.  If it's over 700 km/s, you're going to have noticeably higher gravity losses.

Bit of a typo there.  At 700km/s physics probably won't even notice the planet.

Do you mean high drag losses at those speeds?  Faster usually means less gravity losses.

[Norcalplanner]

1 minute ago, suicidejunkie said:

Bit of a typo there.  At 700km/s physics probably won't even notice the planet.

Do you mean high drag losses at those speeds?  Faster usually means less gravity losses.

Thanks, typo fixed.

Talking about limiting vertical velocity, visible in the upper left info window in the screenshots.

Really upset with imgur's recent changes...

[foamyesque]

I'm curious as to the best achievable numbers with varying rocket diameters and payloads.

[Diche Bach]

Great stuff Norcal. So even though I've been pretty successful, I've been "doing it wrong!"

: More TWR (within reason) not LESS

and MOAR acceleration (within reason) not less

So just to clarify: these considerations are specific to an orbital ascent from a body with gravitational factors in the ballpark of Kerbin and with a fairly dense atmosphere. Not very familiar with the other planets in the game and have yet to land on any of them other than Minmus and Mun, but I would think that the few that have atmospheres (the Venus analog, Eve is it?) things are different, and I don't suppose "landing" on the gas giants is possible, though if it were that would also be rather different eh?

[foamyesque]

More TWR results in less deltaV required pretty much everywhere, Eve included; Eve's gravity is so high, in fact, that massive amounts of thrust are basically required in any event. The primary danger there is in exploding from going to fast on the ascent.

Downside is, more TWR also necessarily means less deltaV available, since you're carrying more engine.

[Diche Bach]

1 hour ago, Norcalplanner said:

Thanks, typo fixed.

Talking about limiting vertical velocity, visible in the upper left info window in the screenshots.

Really upset with imgur's recent changes...

If you are disliking Imgur, give this guys site a whirl. I use it regularly, and it has been up for several years.

http://tw.greywool.com/i/

If you do, give 'em a small donation though eh? I gave 'em like 10 or 15 Euro a year or so ago. His Donation meter was pretty much empty at that point, but it has since bumped up a bit. I doubt he'd want it to become a gigantic internet Mecca of forumites from various game communities posting pics there, but as long as he gets a few bucks I doubt he'll mind.

I haven't played Mount & Blade for years, and have not frequented their site for nearly as long . . . but when I noticed many links to that guys site in the Off Topic section of the Tale Worlds forums I bookmarked it and have used it for image posting ever since. The dudes server is probably one of the richest repositories of anthropogenic eye-candy and various strange gamer imagery in existence by now. There are a few sub-forums of the off-topics forums of TW which are veritable treasure troves on any subject you can think of.

[Norcalplanner]

31 minutes ago, Diche Bach said:

Great stuff Norcal. So even though I've been pretty successful, I've been "doing it wrong!"

: More TWR (within reason) not LESS

and MOAR acceleration (within reason) not less

So just to clarify: these considerations are specific to an orbital ascent from a body with gravitational factors in the ballpark of Kerbin and with a fairly dense atmosphere. Not very familiar with the other planets in the game and have yet to land on any of them other than Minmus and Mun, but I would think that the few that have atmospheres (the Venus analog, Eve is it?) things are different, and I don't suppose "landing" on the gas giants is possible, though if it were that would also be rather different eh?

They're specific to this particular 2.5m rocket, with its particular aero and thrust profile.  The takeaways are that a) higher TWRs are generally more efficient, even if mass is exactly the same, and b) you can frequently get higher performance out of a rocket by altering the ascent profile, typically by turning earlier and/or more aggressively.  I'll post some more comparative data shortly where the only variable is the ascent profile.

[Norcalplanner]

Here's some more info showing how much the ascent profile plays a part.  Methodology was the same for each run, as mentioned in the OP.

Comparative Ascent Profiles - Identical 1.57 TWR

TWR	Turn	Steering	Aero	Gravity	Delta V Spent	Delta V Left
1.57	5 deg @ 10 m/s	-3.6 m/s	-60.4 m/s	-1,057.3 m/s	3,160.2 m/s	2,935 m/s
1.57	6 deg @ 10 m/s	-2.8 m/s	-67.1 m/s	-994.7 m/s	3,103.1 m/s	2,990 m/s
1.57	7 deg @ 10 m/s	-2.5 m/s	-77.1 m/s	-941.2 m/s	3,059.1 m/s	3,033 m/s
1.57	8 deg @ 10 m/s	-2.4 m/s	-103.4 m/s	-884.2 m/s	3,030.6 m/s	3,061 m/s
1.57	9 deg @ 10 m/s	-2.4 m/s	-263.1 m/s	-866.9 m/s	3,161.9 m/s	2,932 m/s
1.57	10 deg @ 10 m/s	DNF	DNF	DNF	DNF	DNF

With this TWR, there was a definite crossover point between 8 degrees and 9 degrees.  The latter ascent was so flat that the Ap reached 100 km while the rocket was going sideways at 27 km altitude.  It stayed down in comparatively draggy atmosphere for several minutes, requiring occasional bursts of acceleration to raise the Ap back over 100 km.

While not the most dramatic results, these clearly show that making a small change in the initial tip angle (8 degrees instead of 5 degrees) can save 130 m/s in getting the craft to orbit.

Edited August 20, 2016 by Norcalplanner
Typos

[pandaman]

Nice work Norcalplanner.

I'm still wrestling with the optimums that work for me, to use as a starting point for launching new designs, so this will give me a bit of a pointer at least.

[LN400]

These are very interesting findings but something that should be looked at, imo, is the spread in results for multiple launches for each condition, to get more solid statistics. This isn't something I would ask of one person but if we could have a community "research" group where everyone are testing in the same conditions, with the same ships, same mods etc, each doing a small portion of the tests needed, then it could very well be doable. The data then being collected and compared.

Edited August 20, 2016 by LN400

[GoSlash27]

Norcalplanner,

 Thanks for the valuable data!

I think we figured out back in 1.0 that the cheapest trip to orbit (in terms of DV) was high t/w, a large rocket, and aiming the gravity turn so that circularization occurs at 43 km.

Of course... getting to orbit with the lowest DV expenditure isn't a practical concern. At least not in career. What matters most is getting to orbit cheaply. That generally requires much lower t/w.

Best,
-Slashy

[numerobis]

The best cheap-and-cheerful rocket does best circularizing at 44km...

To achieve that, you need to thrust a bit radial (15 degrees or so) because its TWR is so low. So the overall indication is not that far off for cheap rockets compared to low-dV rockets.

To minimize coat, start with SRBs. They have quite high TWR and low cost, so you zoom through the lower atmosphere. Aim for about 30-40 degrees at 10km. Once they cut out, switch to a low-TWR LFO stage (TWR 0.7-1). Fly surface prograde until you're at pitch about 15-20, then hold that until Ap starts running away from you, then pitch progressively closer to prograde until you reach orbit, then a bit more to exit the atmosphere.

The exact numbers depend on lots of details. Just fly it a few times, with help from MJ.

[Norcalplanner]

8 hours ago, pandaman said:

Nice work Norcalplanner.

I'm still wrestling with the optimums that work for me, to use as a starting point for launching new designs, so this will give me a bit of a pointer at least.

Thanks. I think that you've hit the nail on the head - the main point with these tests (at least in my mind) is this: Once you've got a halfway decent rocket design that can make it to orbit, play around with the ascent profile to see if it can be flown more efficiently. The initial turn angle and speed can affect your ascent delta v expenditure by hundreds of m/s.

[eddiew]

Nice analysis, thanks  I tend to lean things east in the VAB when they're on launch stabilisers, really helps avoid wobbles due to piloting or flexing.

[Norcalplanner]

51 minutes ago, eddiew said:

Nice analysis, thanks  I tend to lean things east in the VAB when they're on launch stabilisers, really helps avoid wobbles due to piloting or flexing.

I've done that in the past, but I've found that it only works well for rockets up to a certain size.  It also breaks immersion for me a little bit to see anything that doesn't look like a Little Joe or Black Brant sitting at an angle on the pad. 

[Norcalplanner]

As part of my testing, I also did runs at what seems to be the most typical advice, which is "up to 50 m/s, then tilt east 5 or 10 degrees".  I wanted to see what sort of results I would get with the three thrust levels on the  test rocket.

As a reminder, this rocket is not flying a true gravity turn.  It's actively steering to remain locked on the prograde vector, first in a surface reference, then in an orbital reference.  A rocket flying a true gravity turn, depending on aero forces to turn the uncontrolled rocket after the initial controlled tilt, is going to have different results, as will a rocket without SRBs.  With those disclaimers out of the way, here are the results.

Results of 5-10 degrees at 50 m/s

TWR	Turn	Steering	Aero	Gravity	Delta V Spent	Delta V Left
1.25	5 deg @ 50 m/s	-8.6 m/s	-54.2 m/s	-1,295.2 m/s	3,395.9 m/s	2,614 m/s
1.25	10 deg @ 50 m/s	-5.4 m/s	-106.5 m/s	-1,062.4 m/s	3,212.1 m/s	2,794 m/s
1.41	5 deg @ 50 m/s	-26.3 m/s	-47.9 m/s	-1,403.8 m/s	3,525.4 m/s	2,541 m/s
1.41	10 deg @ 50 m/s	-7.1 m/s	-55.4 m/s	-1,199.5 m/s	3,302.1 m/s	2,762 m/s
1.57	5 deg @ 50 m/s	-54.4 m/s	-46.3 m/s	-1,444.4 m/s	3,598.2 m/s	2,501 m/s
1.57	10 deg @ 50 m/s	-16.9 m/s	-51.2 m/s	-1,289.8 m/s	3,404.4 m/s	2,668 m/s

So the takeaway from these results is that if you're not performing a true gravity turn, where you need to get going a certain speed so that your fins can "bite" the air effectively, waiting until 50 m/s is going to increase your delta V expenditure to orbit.  For an actively steered rocket, turning earlier in the ascent should be more effective, particularly at higher TWRs.

Edit: So in looking at this chart, compared to the one in the original post, I now see that for a TWR of 1.25, 10 degrees @ 50 m/s (second line) is actually more efficient than what I thought was the best entry, which was 5 degrees @ 40 m/s.  I'll need to re-run some tests and see how well larger turns do at slightly faster initial velocities, then update the results accordingly.

Edited August 20, 2016 by Norcalplanner
Revelation regarding 1.25 TWR

[eddiew]

1 hour ago, Norcalplanner said:

I've done that in the past, but I've found that it only works well for rockets up to a certain size.  It also breaks immersion for me a little bit to see anything that doesn't look like a Little Joe or Black Brant sitting at an angle on the pad. 

Lol, yeah, I admit it's ugly   But it's a sacrifice I'm prepared to make, and thus far, anything up to a 3.75m has worked ^^

Spoiler

It's not obvious from this camera angle, but the rocket here is a Tweakscaled orange tank on top of a Rhino engine, canted over at about 3-4 'clicks' eastwards. I reckon you can still make a decent mission report without making it look too dodgy on the launch pad  

From your analysis, I should have given it maybe 8-10 clicks... but that really is hard to hide. With even just a few, you can go SAS-prograde from 50-100m/s upwards to have a smooth, and fast, yaw over to the east ^^

 

[Norcalplanner]

OK, I've run some more tests, and it looks at this point like only the 1.25 TWR runs were affected by an initial turning speed higher than what was previously done.  After I run a few more I'll have some more data to post, although it might be tomorrow.

[GoSlash27]

And FWIW, I've found that even with my "cheap" lifters (which follow a true gravity turn), I find substantial DV savings by following an aggressive trajectory like you demonstrate here. I necessarily have to hit a higher speed before the kick, but the path winds up being the same.

Best,
-Slashy

Edited August 20, 2016 by GoSlash27

[LN400]

One question: When there is talk about dv, is there a standard as reference? I mean, is it all amt, or all vac, or a specific combination? This question really goes out to all who use dv benchmarks of any kind but it would be relevant to know what is being used here..

[Norcalplanner]

Typically it's vacuum delta V.  In this specific case, there's a parameter you can have MechJeb display, which is where I'm getting that bit of data (along with the losses) from.  In the screen shots, it's the second window in the upper left corner.

[GoSlash27]

LN400,

 When speaking of first stage DV, I calculate it at 1/2 ATM. Everything else is vacuum.

Best,
-Slashy

[LN400]

I gather then that there is no standard but here's a request to all who reads this topic, that everyone reporting on dv also mention which dv they are talking about since it can mean several hundreds of m/s difference.