How can I reach kerbin orbit using 3,400 m/s of delta-v or less?

[63Hayden]

I've been playing KSP for years, and for the most part have just guessed my way to orbit, while worrying about delta v for ejection and insertion burns at other bodies. However, I've since taken an interest in being as efficient as possible in all aspects of my gameplay. I was following the suggested 3,400 m/s of delta v found in the common delta v map. I went as far as installing mechjeb to reach orbit automomonously, which you'd expect to be as efficient as possible. However, the delta v requirement for the circularisation would always come up short by 100 m/s or so. As a follow up, is the 3,400 m/s suggested in the map based on vacuum thrust?

[Zeiss Ikon]

MechJeb isn't necessarily perfect for any given maneuver, especially if you're trying to optimize for dV.  In general, the method to launch to LKO with minimum dV (as I understand it) is to set up your gravity turn as early as possible, have the correct thrust-to-weight, and burn continuously until your Pe climbs to your Ap height (which means you've circularized).

Normally, however, we approximate this -- for instance, my TWR with my (curently) standard "almost SSTO" rocket is always too high for my gravity turn, requiring a shutdown of nearly two minutes before my circularization burn.  The approximation that reportedly works best is to start your gravity turn below 1 km or 100 m/s, such that you reach 45 degrees from vertical close to 10 km,  have stages with the correct TWR to keep your Ap marker approximately one minute out from the time you pass 15 km altitude until you shutdown to coast to orbital height, then burn again centered on Ap to circularize -- or steer your climb angle to maintain that time to Ap.  I don't attempt to do this, because I normally fly without fins (unnecessary weight and drag if you have gimballed engines), and turning more than a few degrees from prograde above 100 m/s will result in the rocket tumbling.  Throttling to keep from pushing Ap too far out increases gravity losses.

If, OTOH, you do fly with fins, you may find this difficult in that the rocket will self-right to prograde, and it takes a LOT of pitch or yaw authority to overcome this enough to raise or lower your prograde vector enough to matter.

The subway maps are generally based on vacuum Isp, as you surmised; the problem is, different engines have different levels of Isp increase from sea level to vacuum -- unless you're using the same engine that was used to generate the data for the subway map you're using, "your mileage may vary" (to use an old American phrase from car advertising in the early days of EPA fuel economy ratings).

The best I can offer is to tell you that my favorite mid-game lifter, a Twin Boar with a 6400 tank on top, will SSTO with up to about 15 T of payload; adding drop tanks or boosters will increase that figure some (but then it's not a single stage, is it?), and my launches aren't particularly well optimized either.  i don't use MechJeb or KER, so i don't know the exact dV I consume, but I know that this booster and its variants with tanks or boosters will put the same load into LKO, every single time, and as long as my spacecraft doesn't mass more than that figure, it'll get to orbit (with just a little upper stage fuel consumed to finish circularizing, so the booster will reliably fall back to burn up on reentry).

 

Edited April 5, 2018 by Zeiss Ikon

[63Hayden]

5 minutes ago, Zeiss Ikon said:

MechJeb isn't necessarily perfect for any given maneuver, especially if you're trying to optimize for dV.  In general, the method to launch to LKO with minimum dV (as I understand it) is to set up your gravity turn as early as possible, have the correct thrust-to-weight, and burn continuously until your Pe climbs to your Ap height (which means you've circularized).

Normally, however, we approximate this -- for instance, my TWR with my (curently) standard "almost SSTO" rocket is always too high for my gravity turn, requiring a shutdown of nearly two minutes before my circularization burn.  The approximation that reportedly works best is to start your gravity turn below 1 km or 100 m/s, such that you reach 45 degrees from vertical close to 10 km,  have stages with the correct TWR to keep your Ap marker approximately one minute out from the time you pass 15 km altitude until you shutdown to coast to orbital height, then burn again centered on Ap to circularize -- or steer your climb angle to maintain that time to Ap.  I don't attempt to do this, because I normally fly without fins (unnecessary weight and drag if you have gimballed engines), and turning more than a few degrees from prograde above 100 m/s will result in the rocket tumbling.  Throttling to keep from pushing Ap too far out increases gravity losses.

If, OTOH, you do fly with fins, you may find this difficult in that the rocket will self-right to prograde, and it takes a LOT of pitch or yaw authority to overcome this enough to raise or lower your prograde vector enough to matter.

The subway maps are generally based on vacuum Isp, as you surmised; the problem is, different engines have different levels of Isp increase from sea level to vacuum -- unless you're using the same engine that was used to generate the data for the subway map you're using, "your mileage may vary" (to use an old American phrase from car advertising in the early days of EPA fuel economy ratings).

The best I can offer is to tell you that my favorite mid-game lifter, a Twin Boar with a 6400 tank on top, will SSTO with up to about 15 T of payload; adding drop tanks or boosters will increase that figure some (but then it's not a single stage, is it?), and my launches aren't particularly well optimized either.  i don't use MechJeb or KER, so i don't know the exact dV I consume, but I know that this booster and its variants with tanks or boosters will put the same load into LKO, every single time, and as long as my spacecraft doesn't mass more than that figure, it'll get to orbit (with just a little upper stage fuel consumed to finish circularizing, so the booster will reliably fall back to burn up on reentry).

 

Thanks for taking the time to answer my question. The ascent procedure you provided is generally what I would aim to follow in the past, so I guess I'm not too inefficient in what I'm doing. However I would typically reduce my thrust to avoid overshooting my target orbit Apokee, and by no longer doing that, I might see an reduction in the delta-v required for orbit. I find KSP to be more enjoyable when I'm guessing my way through things, and more often than not I have ample delta-v to do what I want to do anyway. I suppose the 3,400 m/s is just there as a rough target for reaching orbit, and after reading your post, it's pretty obvious now that that figure can't possibly apply to every vehicle design due to factors like TWR and atmospheric drag. For now, I'll work on improving my ascent profile, and ensure my designs have a sufficient TWR. Since I started using Kerbal Engineer years ago, I always based my TWR on the Saturn V, as I saw that as the staple launch vehicle. Anything with a TWR of 1.15 or higher I thought was more than enough, but after reading some posts I now realize that you should aim for something closer to 1.5. Anyway, I appreciate the help, and have yourself a good day.

 

[CrashyMcCrashFace]

From reading real life orbital process, it seems it's all about the pitch-over maneuver. It you're trying to be an economical as possible this is the only time you should steer. Given your thrust and orbital requirement everything can be calculated backwards to the pitch-over maneuver. After that you just follow pro-grade and optionally moderate throttle. I've been trying this every time and noticed it can be a big difference depending on what you're trying to achieve (orbit, Mun, tiny satellite, beast size Mk1-3 capsule, etc). If true, almost everybody online (tutorials and vloggers) is doing it wrong.

Disclaimer: That's just the 10,000 foot high level since I'm new and not sure I really understand.

Edited April 5, 2018 by CrashyMcCrashFace

[63Hayden]

1 hour ago, CrashyMcCrashFace said:

From reading real life orbital process, it seems it's all about the pitch-over maneuver. It you're trying to be an economical as possible this is the only time you should steer. Given your thrust and orbital requirement everything can be calculated backwards to the pitch-over maneuver. After that you just follow pro-grade and optionally moderate throttle. I've been trying this every time and noticed it can be a big difference depending on what you're trying to achieve (orbit, Mun, tiny satellite, beast size Mk1-3 capsule, etc). If true, almost everybody online (tutorials and vloggers) is doing it wrong.

Disclaimer: That's just the 10,000 foot high level since I'm new and not sure I really understand.

Yeah, right now the biggest thing I can improve on is my ascent profile. I'll take into consideration the suggestions you guys have made here, and I should be more efficient from here on out. Thanks for the input.

[OhioBob]

@63Hayden, don't get too hung up on delta-v.  Minimizing delta-v does not necessarily mean you are being efficient.  Often a vehicle that is very good at getting to orbit using little delta-v is very poor in terms of cost efficiency or payload fraction.  Being efficient is about getting the most payload to orbit using the least amount of rocket.  Delta-v is irrelevant.

 

[blakemw]

It depends a lot on the launch TWR and the ASL vs Vac ISP of your engines. Like for example the Swivel has rather bad ASL ISP vs Vac ISP, a reasonable Swivel rocket can easily take 3600m/s to get into orbit just because it makes significantly worse use of fuel when low in the atmosphere. This is not saying whether or not the Swivel is bad as a launch engine (though it is), just that a large difference between ASL and Vac performance strongly distorts the dV to orbit which is measured in Vac dV.

In contrast, an engine like the Twin-Boar has a practically flat ISP curve, meaning it almost makes equally good use of fuel at all altitudes.

The next factor is launchpad TWR. A rocket with a TWR of 2.0 will require about 200-300m/s less than a rocket with a TWR of 1.2 - again this isn't to say it's better to have a higher TWR, it requires going very engine-heavy, but it does result in lower dV to orbit numbers (not necessarily lower cost per kg to orbit though). The cheapest way to get a high launchpad TWR is adding small SRBs like Hammer or Thumper, and actually happens to be pretty economical.

Next up is streamlining, which is very important if you're going to make good use of a higher TWR. Every m/s lost to aerodynamic drag is another m/s to orbit. Streamlining is usually a fair investment, though in challenges to launch mass as cheaply as possible streamlining usually isn't used because it's cheaper to buy more boosters than to add more streamlining parts. But if you want a consistently low dV to orbit, then streamlining is great. It should be noted that popping the fairing and thus ditching that weight mid-ascent usually isn't properly accounted for by dV estimators so fairings are another distortion factor.

Finally is trajectory, which should simply be "as horizontal as possible, without getting trapped in the atmosphere", generally a gravity turn that involves pitching over right after launch and leveling out somewhere around 30000m.

If you optimize for all of these factors you can get into orbit for about 2800m/s! In practice though, the most economical rockets tend to do it for between 3050-3300m/s if using solids to give a kick off the pad, or around 3300-3400m/s if using liquid only.

Edited April 6, 2018 by blakemw

[GoSlash27]

What OhioBob and Blakemw said.

 If it takes 3500 m/sec for your payload to reach orbit, then that's just what it takes. Add another 100 m/sec DV. The importance of the DV figure lies in it's consistency. It's just there to let you know how much rocket you need. Don't use it as a mark of efficiency... at least during launch. The true measure of efficiency is the price tag.

Best,
-Slashy

[63Hayden]

4 hours ago, blakemw said:

It depends a lot on the launch TWR and the ASL vs Vac ISP of your engines. Like for example the Swivel has rather bad ASL ISP vs Vac ISP, a reasonable Swivel rocket can easily take 3600m/s to get into orbit just because it makes significantly worse use of fuel when low in the atmosphere. This is not saying whether or not the Swivel is bad as a launch engine (though it is), just that a large difference between ASL and Vac performance strongly distorts the dV to orbit which is measured in Vac dV.

In contrast, an engine like the Twin-Boar has a practically flat ISP curve, meaning it almost makes equally good use of fuel at all altitudes.

The next factor is launchpad TWR. A rocket with a TWR of 2.0 will require about 200-300m/s less than a rocket with a TWR of 1.2 - again this isn't to say it's better to have a higher TWR, it requires going very engine-heavy, but it does result in lower dV to orbit numbers (not necessarily lower cost per kg to orbit though). The cheapest way to get a high launchpad TWR is adding small SRBs like Hammer or Thumper, and actually happens to be pretty economical.

Next up is streamlining, which is very important if you're going to make good use of a higher TWR. Every m/s lost to aerodynamic drag is another m/s to orbit. Streamlining is usually a fair investment, though in challenges to launch mass as cheaply as possible streamlining usually isn't used because it's cheaper to buy more boosters than to add more streamlining parts. But if you want a consistently low dV to orbit, then streamlining is great. It should be noted that popping the fairing and thus ditching that weight mid-ascent usually isn't properly accounted for by dV estimators so fairings are another distortion factor.

Finally is trajectory, which should simply be "as horizontal as possible, without getting trapped in the atmosphere", generally a gravity turn that involves pitching over right after launch and leveling out somewhere around 30000m.

If you optimize for all of these factors you can get into orbit for about 2800m/s! In practice though, the most economical rockets tend to do it for between 3050-3300m/s if using solids to give a kick off the pad, or around 3300-3400m/s if using liquid only.

Thanks for the very informative post. I now realise that I wasn't giving enough thought to the variation in ISP at sea level vs vacuum. I'm very surprised to hear that you can reach orbit using 2,800m/s, as you say; I'm typically a close to a full 1k higher than that with my launch vehicles.

Despite having well 1,000 hours in gameplay, I still clearly have a lot to learn.

Edited April 6, 2018 by 63Hayden

[OhioBob]

Just to elaborate a bit I what I said earlier.  For any particular rocket, it is good practice to fly it on a trajectory and in a manner that will get it to orbit utilizing the least amount of propellant, i.e. the least amount of Δv.  That just makes sense because you can then either use less propellant, or you have more propellant left over for subsequent maneuvers after getting to orbit.  So if you have a rocket capable of getting to orbit using 3400 m/s ΔV, then that's what you should try to do.  If it takes you 3500 m/s, then you are not flying the rocket as efficiently as it is capable of doing.

My original point, however, is that lowering Δv should never be your goal when designing a rocket.  In other words, let's say your last design could get to orbit using 3400 m/s.  So now you say to yourself, I want to design a rocket that can get to orbit using 3300 m/s because that's going to be more efficiency.  THAT IS AN INCORRECT WAY TO GO ABOUT IT.  Efficiency of design is not about Δv.  And if you design a launch vehicle to minimize Δv, you're probably going to end up with a bad design from an efficiency standpoint.  You should design to maximum payload versus cost, and the heck with how much Δv it takes.  If it takes 3500 m/s but gets a large payload to orbit cheaply, then it's a good design.

[blakemw]

This is an example of a rocket which can do 2800m/s to orbit:

I put it in orbit with 1544m/s leftover, a dV to orbit of about 2760m/s.

Heating actually becomes the most major problem at this point, turn off heating and it can do like 2600m/s.

Edited April 6, 2018 by blakemw

[63Hayden]

8 minutes ago, blakemw said:

This is an example of a rocket which can do 2800m/s to orbit:

I put it in orbit with 1544m/s leftover, a dV to orbit of about 2760m/s.

Heating actually becomes the most major problem at this point, turn off heating and it can do like 2600m/s.

Is the main advantage to this design it's high TWR. I've built similar rockets to this in the past when messing around in sandbox mode, but it can't be very cheap if you were using it regulalry in career mode.

[blakemw]

29 minutes ago, 63Hayden said:

Is the main advantage to this design it's high TWR. I've built similar rockets to this in the past when messing around in sandbox mode, but it can't be very cheap if you were using it regulalry in career mode.

Exactly, a combination of extreme TWR and streamlining.

But you only need that TWR for like the first 20 seconds to get it moving fast. This initial kick can be provided cheaply with lots and lots of hammers (or thumpers), you can easily get a TWR of like 3.0 this way, it's wasteful to use LF engines to get a high TWR but with solids the nozzle isn't the expensive part, it's the solid fuel, so burning lots of solid fuel really fast is just fine. I sometimes actually use designs along these lines, like starting with a Twin-Boar core with a TWR of 1.0, then I add Hammers or Thumpers until the thing has a launchpad TWR of like 2.5. It's very economical, doing so can increase the cost of the rocket by like 20%, but increase the payload by like 50%. Using solids to give a good hard kick off the pad is great.

Edited April 6, 2018 by blakemw

[63Hayden]

5 hours ago, blakemw said:

Exactly, a combination of extreme TWR and streamlining.

But you only need that TWR for like the first 20 seconds to get it moving fast. This initial kick can be provided cheaply with lots and lots of hammers (or thumpers), you can easily get a TWR of like 3.0 this way, it's wasteful to use LF engines to get a high TWR but with solids the nozzle isn't the expensive part, it's the solid fuel, so burning lots of solid fuel really fast is just fine. I sometimes actually use designs along these lines, like starting with a Twin-Boar core with a TWR of 1.0, then I add Hammers or Thumpers until the thing has a launchpad TWR of like 2.5. It's very economical, doing so can increase the cost of the rocket by like 20%, but increase the payload by like 50%. Using solids to give a good hard kick off the pad is great.

That makes sense. I will incorporate this information in future designs. Thanks.

[Curveball Anders]

The good thing with MJ (and GTC) is that you can run the same launch over and over again, watching the data and tweaking the design or tweaking launch parameters.

Doing that a number of times and you'll get a good feel for an efficient launch and can do it manually and winging a design together.

And as has been repeated, efficiency is about funds/tons cargo to target orbit (and since it's a game, without being too tedious).

 

Edited April 7, 2018 by Curveball Anders
gramar

[maculator]

The sooner you escape the gravity the better. Just barley not  burning up proofed to be the most efficient approach to me. Positive sideffect are that it seems to be pretty efficient too and makes it less tedious.

I also do aim for a orbit that is juuust above the 70km mark, you can raise it later at any point but like I said first goal shuld be to escape gravity asap.