What really affects the total delta-v of a rocket?

[HorusKol]

I'm using KER to figure out the delta-v of my rocket, so I can better plan missions - but I noticed that the changes I was making to my rocket stack wasn't really affecting the final delta-v all that much.

So, I started with my probe, put the smallest tank and an LV-909 under it (lander). Under that, I had one of the 400 tanks and an LVT-30 (transition). Under that, I put 4x400 tanks in a stack on top of another LVT-30, and then 4 of the BACC solid boosters on decouplers.

     **
     ||
     /\
     ||
     /\
   | || |
   | || |
   |.||.|
   | || |
   ^ /\ ^ 

I found that if I waited to fire my main LVT-30 until the BACCs had burnt out, I got a significant boost to my delta-v, but then other changes were minimal - for example, adding 4 or 8 of the smaller SRBs added only a couple of hundred m/s. I even changed my quadrupled my central stack (4x the fuel, 4x the thrust) - barely registered any change.

So - what really affects the total delta-v of the rocket - it seems more staging would be better?

Edited October 1, 2014 by HorusKol

[UmbralRaptor]

For a single stage, ÃŽâ€V == 9.82*Isp*ln(wet_mass/dry_mass)

Given the limited advantages of adding more fuel, extra stages help a great deal. Once a stage is vaguely near its "ideal" (no payload) mass ratio, scaling it up further has little effect (not much growth in the raw mass ratio, and it has a sub-linear effect on ÃŽâ€V). But depending on other details, it's possible that the booster represent a minimal fraction of the rocket's overall mass. Pictures may help.

In the case of the LV-T30, I suspect it's the difference in effective Isp of the stages -- the LV-T30's is ~40-50% better! (and depending on booster count, the TWR difference may not be an issue)

(The 9.82 is only in KSP. For a real rocket you'd use 9.8 - 9.81 as the conversion factor, depending on assumptions.)

[Specialist290]

So - what really affects the total delta-v of the rocket - it seems more staging would be better?

Sounds like you're on the right track here.

Delta-v is, all things considered, actually quite simple to figure out. There are only two factors that really affect the delta-v of a given stage: the "mass ratio" (i.e. ratio of a fully-fueled rocket's mass to its mass after all the fuel has been exhausted) and the fuel efficiency of the stage's engines, as measured by its exhaust velocity.

I used to have a whole huge spiel I'd type out to explain it in more detail, but nowadays I can just link to this handy website; the pages under the "Orbital Mechanics - Basics" section are quite useful for explaining delta-v in detail, among other concepts. You may also find these four web pages over at Atomic Rockets to be quite helpful as well, as they explain the idea from a real-world perspective.

[AlexinTokyo]

Essentially, what they said. The only thing I didn't see mentioned is that you must remember to count the stages above the one you're calculating for as part of the dry mass of that stage.

So, for example, if you have a top stage and payload of (total) 10 t, and a booster stage of 20 t fueled and 2 t dry, the ÃŽâ€v of the booster stage would be calculated as 9.82 x I_sp x ln((10+20)/(10+2)).

[Tortoise]

The mass of the craft ,fuel, and engine efficiency.

[el_coyoto]

Just like the guys said.

• The mass of everything on the rocket which is not fuel
  
• The quantity of fuel
  
• Efficiency of engines
  
• Staging
  

So, to improve your rocket delta-v, you have several options :

• Make everything that is not fuel on the rocket lighter (less RCS, less batteries, unmanned, i.e. building small) : this has a huge impact on delta-v.
  
• Add more fuel : diminishing return on delta-v after each added tank because of the so called "tyranny of the rocket equation", use with caution. You've already noticed this when you quadrupled your central stack without seeing much change.
  
• More efficient engines might help BUT if they are too heavy, they can hurt rule #1. For light ships, a light chemical engine might yield a better delta-v then a more efficient but heavier NERVA for example.
  
• Staging removes "dead mass" (empty tanks, engines which thrust is not needed any more), thus improving rule #1 on your rocket the moment you stage. Probably the easiest trick to increase delta-v as you have already noticed...
  

[Laie]

In laymen's terms (my username has to be good for something after all):

You get thrust from throwing stuff out of the back of your rocket. You get more dV if A) you have more stuff to throw out (more fuel) and the faster you throw it (higher ISP). Any weight on your rocket that can not eventually be thrown out will reduce delta-V.

[Moonfrog]

1. Mass

2. Fuel hence your fuel mass raito.

3. Efficiency of engines your ISP

General principle is to avoid over engineering, by building too big a stage. Under engineering, creating a small stage that increase the payload for the second so much that it loses delta V due to extra unneccesary payload.

3 factors u can change usually. in terms of parts.

1. Mass of parts to make the main ship, can u make it using smaller parts and still work?

2. Do you need that much fuel?

3. Despite what jebidiah and most kerbals think do you need that big of an engine? Also take into account ISP of engines. Personally I find mass often is more important than ISP if your putting an engine on the top stage, since mass effects all bottom stages, from the top.

used to be a time when people thought ant engines were not useful, but ant engine is perfect for Minmus due to its really low gravity. Really don't need a bigger engine than an ant going to that place.

[Red Iron Crown]

Just like the guys said.

• The mass of everything on the rocket which is not fuel
  
• The quantity of fuel
  
• Efficiency of engines
  
• Staging
  

So, to improve your rocket delta-v, you have several options :

• Make everything that is not fuel on the rocket lighter (less RCS, less batteries, unmanned, i.e. building small) : this has a huge impact on delta-v.
  
• Add more fuel : diminishing return on delta-v after each added tank because of the so called "tyranny of the rocket equation", use with caution. You've already noticed this when you quadrupled your central stack without seeing much change.
  
• More efficient engines might help BUT if they are too heavy, they can hurt rule #1. For light ships, a light chemical engine might yield a better delta-v then a more efficient but heavier NERVA for example.
  
• Staging removes "dead mass" (empty tanks, engines which thrust is not needed any more), thus improving rule #1 on your rocket the moment you stage. Probably the easiest trick to increase delta-v as you have already noticed...
  

Came in here to post exactly this. There are only four ways to increase a vessel's delta-V, the quoted post describes them nicely.

[AlexanderB]

Honestly, replace the LV-909 with a 48-7s, the gain in lower mass should offset the slightly lesser ISP.

[benc85]

Sounds like you've got a massive Power to Weight ratio on your early stage, which will waste a lot of your DV during your ascent. You want to be aiming for between 1.6 and 1.8 PtW for anything taking off from the Launchpad. Anything that's noticeably higher will just burn your fuel as you encounter greater friction at greater speeds.

[Yakky]

Honestly, replace the LV-909 with a 48-7s, the gain in lower mass should offset the slightly lesser ISP.

This is a good piece of advice and one that can seem a bit counterintuitive. Some of the high ISP engines can be quite heavy. For example, the LV-N engine weighs 2.25 tons. You only benefit from carrying such a heavy engine when you are planning to burn enough fuel in it to make the extra weight worthwhile.

And be very careful about using multiple engines. For interplanetary stages where thrust to weight ratios don't matter, you'll always optimize delta-V by having just a single engine. Why? Because a single engine will still burn as much fuel as multiple engines (given enough time), but will do so with less mass. In a sense, multiple engines are just dead weight. (This is obviously not true when you require high thrust such as for launch or powered landings.)

Assuming you're like me and are too impatient to burn an ion engine for hours or days on end, you might wonder what engine is optimal for interplanetary burns. I did an optimality analysis and it turns out the only other two engines you really need to consider for interplanetary stages are the 48-7S and the LV-N. Every other choice is always inferior to those two, including the LV-909 (which surprised me because I would have guessed it would have been optimal for certain sizes of spacecraft).

Here's the rule of thumb for interplanetary engine selection: Above 4 tons total wet mass (excluding the engine), you always want the LV-N. Below 2 tons (excluding engine), you usually want the 48-7S. And in between 2 and 4 tons, it depends on how much fuel you're carrying, but if you have a lot of fuel (over 75%) you probably want the 48-7S and otherwise you probably want the LV-N.

Edited October 1, 2014 by Yakky

[HorusKol]

1. Make everything that is not fuel on the rocket lighter (less RCS, less batteries, unmanned, i.e. building small) : this has a huge impact on delta-v.
   
2. Add more fuel : diminishing return on delta-v after each added tank because of the so called "tyranny of the rocket equation", use with caution. You've already noticed this when you quadrupled your central stack without seeing much change.
   
3. More efficient engines might help BUT if they are too heavy, they can hurt rule #1. For light ships, a light chemical engine might yield a better delta-v then a more efficient but heavier NERVA for example.
   
4. Staging removes "dead mass" (empty tanks, engines which thrust is not needed any more), thus improving rule #1 on your rocket the moment you stage. Probably the easiest trick to increase delta-v as you have already noticed...
   

And be very careful about using multiple engines. For interplanetary stages where thrust to weight ratios don't matter, you'll always optimize delta-V by having just a single engine. Why? Because a single engine will still burn as much fuel as multiple engines (given enough time), but will do so with less mass. In a sense, multiple engines are just dead weight. (This is obviously not true when you require high thrust such as for launch or powered landings.)

These help a lot - thanks

[Laie]

And be very careful about using multiple engines. For interplanetary stages where thrust to weight ratios don't matter, you'll always optimize delta-V by having just a single engine.

Yes and no. The inability to burn at or even near periapsis will cost you dearly if you're a dV miser. If you want to go to Jool, you need about 2000m/s; you can slowly accumulate the first 1000m/s by short-and-efficient burns near your periapsis; but the next 1000m/s have to happen all in one go.

With the lowest-TWR vessel I ever used (a 800mm/s² Ion Probe), I spent like two hours on setting up nodes and waiting for burns, and when I checked my dV counter afterwards I had used like 20% more delta-V than it said on the nodes.

For practical purposes, I find that an acceleration of about 2m/s² doesn't cost all that much nominal delta-V, wastes little effective delta-V, and gets me to Jool in like twenty minutes. Which is still a lot of time if you have nothing to do but watch your prograde marker.

[cantab]

I found that if I waited to fire my main LVT-30 until the BACCs had burnt out, I got a significant boost to my delta-v

Informally speaking, this would be because that way the BACCs do their job and are then discarded, meaning the LV-T30 and its fuel are pushing a lighter rocket thus getting more delta-V.

for example, adding 4 or 8 of the smaller SRBs added only a couple of hundred m/s.

This is because ultimately they're not adding that much extra fuel, especially against the entire weight of the base rocket.

I even changed my quadrupled my central stack (4x the fuel, 4x the thrust) - barely registered any change.

This indicates that by this point your central stack is mostly lifting its own weight rather than the payload, because the payload is too light in comparison to the rocket. If it flies well that quad-stack design might be useful when you need to lift something bigger.

[Yakky]

Yes and no. The inability to burn at or even near periapsis will cost you dearly if you're a dV miser. If you want to go to Jool, you need about 2000m/s; you can slowly accumulate the first 1000m/s by short-and-efficient burns near your periapsis; but the next 1000m/s have to happen all in one go.

With the lowest-TWR vessel I ever used (a 800mm/s² Ion Probe), I spent like two hours on setting up nodes and waiting for burns, and when I checked my dV counter afterwards I had used like 20% more delta-V than it said on the nodes.

For practical purposes, I find that an acceleration of about 2m/s² doesn't cost all that much nominal delta-V, wastes little effective delta-V, and gets me to Jool in like twenty minutes. Which is still a lot of time if you have nothing to do but watch your prograde marker.

This is a very good point. You have to make sure you are still able to harness the Oberth Effect. And of course there is a direct correlation between thrust/weight ratio and Fun Factor, though I haven't yet seen the exact scientific derivation of that relationship. I like your rule of thumb on acceleration.

[Empiro]

The basic general advice for maximizing dV: don't carry any extra weight you're not using. Empty tanks and engines that aren't firing at maximum thrust are extra weight, and should be dropped if possible. If your TWR is too high, it basically means that you're carrying too much engine. This is the biggest reason why asparagus staging is so efficient -- you empty tanks a pair at a time and drop them off as soon as they're empty. You also fire all your engines and drop those as soon as you don't need the thrust any more.

Aim for ~1.5 TWR for the launch stage, 0.7 for your circularization stage, and as low as your patience can stand for anything after that. In .24 career, SRBs are so comparatively cheap and get used quite a bit. One trick trick I do with SRBs is to break 4 of them into 2 pairs, and reduce the thrust setting for one pair to keep the TWR at 1.5 at launch, then when the SRB that's burning faster empties, I drop it off. Sometimes, I need to adjust the thrust setting for both so that the TWR doesn't drop too low. KER is even good enough to calculate the dV and TWR properly after you drop off one pair of boosters (though if you use seperatrons, it tends to mess up the calculation, and you need to empty the seperatrons). It's not quite as efficient as true asparagus staging, but it's quite cheap.

[Padishar]

KER is even good enough to calculate the dV and TWR properly after you drop off one pair of boosters (though if you use seperatrons, it tends to mess up the calculation, and you need to empty the seperatrons). It's not quite as efficient as true asparagus staging, but it's quite cheap.

Do you have an example craft with sepratrons that KER gets wrong as you describe? It should realise that they aren't attached to the vessel when actually used as sepratrons so if they do affect the deltaV then I need to fix something...

[Mystique]

I guess I saw question in OP similar to mine, though I can't seem to find a clear answer:

I have SRB and LFE as booster stage. If I stage LFE to start right after SRB decouple in mid-air (SRB fire, rocket flies, SRB stops, SRB decouple, only then LFE starts), I get more total delta-v for booster stages than if I stage SRB to start together with LFE on the launchpad and drop SRB in mid-air while LFE is still working (SRB and LFE fire together, rocket flies, SRB stops, SRB decouples, LFE still working for some time).

Here are screens, notice sequence of engines starting and total resulting delta-v:

According to my understanding of how things work, delta-v should be higher when all engines start at once, because otherwise SRB carry dead weight consisting of payload and not working LFE stage; similar to asparagus concept - all engines start together and drop dead weight as soon as it becomes "dead", i.e. when fuel runs out.

[Red Iron Crown]

Mystique, in your example craft the SRBs deliver slightly more delta V when the LFE is started with them because they end up lifting less liquid fuel during their burn, however that gain is dominated by the loss of dV from spending liquid fuel accelerating the mass of the SRBs. The scenario where you ignite the LFE after the SRBs are dropped produces slightly less dV from the solids but much more from the LFE, resulting in more total dV.

[cantab]

According to my understanding of how things work, delta-v should be higher when all engines start at once

This only really applies for a rocket designed to take advantage of using all its engines by using lighter lower-thrust engines.

In fact, the following two scenarios have the same delta-V:

First stage of two FL-T800 tanks and two LV-T45s. Second stage of one FL-T800 tank and one LV-T45.

Asparagus staging with a core and two boosters, each one FL-T800 tank and one LV-T45.

[DrD]

I guess I saw question in OP similar to mine, though I can't seem to find a clear answer:

I have SRB and LFE as booster stage. If I stage LFE to start right after SRB decouple in mid-air (SRB fire, rocket flies, SRB stops, SRB decouple, only then LFE starts), I get more total delta-v for booster stages than if I stage SRB to start together with LFE on the launchpad and drop SRB in mid-air while LFE is still working (SRB and LFE fire together, rocket flies, SRB stops, SRB decouples, LFE still working for some time).

Here are screens, notice sequence of engines starting and total resulting delta-v:

http://s13.postimg.org/4v53rlq0j/screenshot4.jpg

http://s13.postimg.org/80ppht8mr/screenshot5.jpg

According to my understanding of how things work, delta-v should be higher when all engines start at once, because otherwise SRB carry dead weight consisting of payload and not working LFE stage; similar to asparagus concept - all engines start together and drop dead weight as soon as it becomes "dead", i.e. when fuel runs out.

I can't see the pictures clearly enough to read the stats but what could be happening is a TWR that is too high.

If your TWR firing all 3 engines is too high you'll waste delta V on the atmosphere.

Whenever I add SRB I have to either decrease the tweakable throttle in the SRB or set stages so the LFE starts later as in your ship.

As an experiment look at the TWR of your #1 ship (all 3 engines at once) and of the #2 ship. I'd guess in #1 it's way high and in #2 it's lower. Now go back to #1 and adjust the throttle on SRB down until TWR is the same as in #1 and your delta V may be closer.

The only prob with adjusting throttle on SRB down alot is now you're carrying their weight longer, so often it's still more efficient to burn them at 100% and let your LFE idle until they're done. KER tells you how long engines burn and if the burn time of SRB is close too (or longer htan) the LFE you will lose much of the benefit of staging.

The general rule is keep your TWR around 1.6 while getting rid of stages ASAP. Higher results in losing to atmosphere (or even burning up with DRE), lower means staying in the gravity well longer than needed, and can lead to control problems. With SRB this very often means keeping the LFE for stage 2.

[John FX]

Stage often, stage small.

[Mystique]

I can't see the pictures clearly enough to read the stats but what could be happening is a TWR that is too high.

If your TWR firing all 3 engines is too high you'll waste delta V on the atmosphere.

Nay, this thing on screens was created just to show difference I was asking about, nobody meant to actually launch it

Usually I aim to start lift-off stage (with SRBs, if it's a heavy thingy) with starting TWR about 1.3-1.4 and final TWR up to 2.5 and delta-v enough to push rocket out of thickest atmosphere layers while starting gravity turn, then next LFE stage starts with TWR aroudn 1.1-1.6 for horizontal burn above 30 km of stock Kerbin and for circularization burn.

And on these rockets I noticed small difference of delta-v as described above (though it's usually much less than 100 m/s, maybe because engines and SRBs are selected to fit exact ship), that's why I decided to ask.

Anyway, now I understood it, question closed