Delta-V or TWR on takeoff?

[OhioBob]

The mass fraction of your mass fraction optimized rocket is worse than what I just said low dv rockets are capable of.

When you say “mass fraction†I assume you are referring to the payload mass fraction. I don’t really care that much about payload fraction. I care even less about ÃŽâ€v. What I’m interested in is maximizing the total payload mass that a particular pair of engines can put into orbit. The payload mass definitely goes up when we add on more propellant, though the payload fraction might decrease some. That’s OK – fuel is cheap, engines are not.

If I have one launcher with a TWR of 2.0 and a payload fraction 0.24, and another launcher with a TWR of 1.4 and a payload fraction of 0.21, I’ll take the latter. With the same engines, the second rocket will launch 25% greater payload mass.

I'm curious if you ever flew these?

Yes, I experimented with them. The payload capacities that I could attain in the game were a little less than the simulations. I assume this is mainly because I flew less than ideal ascent profiles. The lowest TWR rocket was especially difficult to fly because its extreme slenderness made it rather wobbly. Otherwise it performed alright. I was definitely able to launch much heavier payloads with the larger, low TWR rockets.

Edited October 20, 2015 by OhioBob

[Alshain]

So, just out of curiosity for you guys that build the really efficient lifters. How many stages do you tend to use on the main stack (not counting any radial boosters)? Do you prefer single or do you prefer to have a small kicker stage or even 3?

Edited October 20, 2015 by Alshain

[GoSlash27]

So, just out of curiosity for you guys that build the really efficient lifters. How many stages do you tend to use on the main stack (not counting any radial boosters)? Do you prefer single or do you prefer to have a small kicker stage or even 3?

Alshain,

Almost always 2 stages. It's very rare that I've found a situation where 3 stages improves my bottom line.

Best,

-Slashy

Of course... SSTO spaceplanes are still the king of "orbit for cheap"....

[waterlubber]

I launch with at least 1.6 TWR. Screw efficiency, I want to get to space fast. The performance hits from mach heating are insane.

[Superfluous J]

So, just out of curiosity for you guys that build the really efficient lifters. How many stages do you tend to use on the main stack (not counting any radial boosters)? Do you prefer single or do you prefer to have a small kicker stage or even 3?

Not sure if I count as someone who makes "really efficient lifters" but for me it's almost always 2. Sometimes 1. Sometimes 3 (though the first of those is frequently just a couple SRBs to help it get underway). Usually 2. And ideally the last lifting stage is dropped when my Ap is at the target altitude (100km these days but I used to shoot for 80) and the Pe is high but still in the atmosphere.

I tend to go for 1.3-1.6 TWR on the launchpad, guide my gravity turn so it's quite shallow, and consider throttling down a personal tragedy. If my rocket flips or has any other speed issues, I fix them in the VAB, not during launch.

Edited October 20, 2015 by 5thHorseman

[Empiro]

I'm not the most efficient, but I think my designs are reasonably so. I almost always use 2.5 stages for medium sized and larger rockets for missions that take me further than LKO. It's good enough to make good progress in Hard career without needing to grind missions.

The upper stage is almost always a Poodle or Terrier. Lower stage is often a Swivel or Skipper. I strap a pair (sometimes more) SRBs to increase the TWR at launch. It feels natural, and makes good use of the strengths of each engine.

Generally, the first 1.5 stages carries enough fuel to get me very close to orbit (sometimes enough to make orbit, but I always stage before it does so that I don't leave litter around LKO). I complete circularization with the upper stage. This leaves me with a near full 3rd stage in orbit, and with an engine that works well in vacuum.

[Red Iron Crown]

Two stages plus SRBs when I'm chasing efficiency (though I'm not nearly in the ballpark with some others in this thread in that regard). More often I chase simplicity and part count, which is usually one stage plus SRBs or two serial stages.

[OhioBob]

My first preference is always for two stages. However, when I can’t find a first stage engine with the right thrust, I generally go a size smaller and augment with strap-on SRBs.

[Alshain]

Well glad to see I've been doing that right anyway, I usually do 2 stage myself. I always seem to get the lowest cost lifter that way.

[OhioBob]

When you say “mass fraction†I assume you are referring to the payload mass fraction. I don’t really care that much about payload fraction. I care even less about ÃŽâ€v. What I’m interested in is maximizing the total payload mass that a particular pair of engines can put into orbit. The payload mass definitely goes up when we add on more propellant, though the payload fraction might decrease some. That’s OK – fuel is cheap, engines are not.

If I have one launcher with a TWR of 2.0 and a payload fraction 0.24, and another launcher with a TWR of 1.4 and a payload fraction of 0.21, I’ll take the latter. With the same engines, the second rocket will launch 25% greater payload mass.

After thinking about what I posted above, I’ve come to realize that I need to invent a new term – payload to thrust ratio (PTR). It is the ratio of payload weight to launch thrust. It is comparable to payload fraction in the following way,

PTR = PF / TWR

I think that cost optimization is largely a matter of finding the maximum PTR. Since engines are the most expensive part of a launch vehicle, it makes sense that you’d want to lift your payload using the smallest possible engines, i.e. you want a high PTR.

In the example quoted above, the PTR of the first rocket is 0.12 and the PTR of the second is 0.15. Although the first has a higher payload fraction, the second is the more cost effective launcher.

[Alshain]

After thinking about what I posted above, I’ve come to realize that I need to invent a new term – payload to thrust ratio (PTR). It is the ratio of payload weight to launch thrust. It is comparable to payload fraction in the following way,

PTR = PF / TWR

I think that cost optimization is largely a matter of finding the maximum PTR. Since engines are the most expensive part of a launch vehicle, it makes sense that you’d want to lift your payload using the smallest possible engines, i.e. you want a high PTR.

In the example quoted above, the PTR of the first rocket is 0.12 and the PTR of the second is 0.15. Although the first has a higher payload fraction, the second is the more cost effective launcher.

But your payload to thrust ratio doesn't scale with cost. I could build two rockets with equal "PTR" that have very different costs.

In preparation for 1.0.5 and a new career I was just re-designing some early tier lifters. For my 2 ton lifter I ended up with a 1.04 TWR, which is a little low for my preference so I strapped 2 flea booster on the side and now I hat 2.53 TWR. By the time I discard them my main stack is up to 1.3 TWR. All it cost me was 2080 funds (inc. decoupler and nose cone). I could have put liquid fuel boosters on it and ended with the same TWR, same payload mass, but significantly more expensive. Later I realized the boosters were discarded so early the nosecones were kinda pointless, so that's 480 funds less, without changing really the 'PTR' significantly (TWR raised by .01). So you see the PTR calculation doesn't accurately reflect cost optimization.

Edited October 20, 2015 by Alshain

[OhioBob]

But your payload to thrust ratio doesn't scale with cost. I could build two rockets with equal "PTR" that have very different costs.

In preparation for 1.0.5 and a new career I was just re-designing some early tier lifters. For my 2 ton lifter I ended up with a 1.04 TWR, which is a little low for my preference so I strapped 2 flea booster on the side and now I hat 2.53 TWR. By the time I discard them my main stack is up to 1.3 TWR. All it cost me was 2080 funds (inc. decoupler and nose cone). I could have put liquid fuel boosters on it and ended with the same TWR, same payload mass, but significantly more expensive. Later I realized the boosters were discarded so early the nosecones were kinda pointless, so that's 480 funds less, without changing really the 'PTR' significantly (TWR raised by .01). So you see the PTR calculation doesn't accurately reflect cost optimization.

My intent is that for a given engine we want to maximize PTR. As you correctly identify, it doesn’t necessarily compare across different engine configurations. However, for a specific engine configuration, we want to maximize the payload we can lift, i.e. we want to maximize PTR. In that case, PTR is a more important metric than payload fraction.

That last point is where I think Requia and I differ. If I correctly understand Requia’s argument, he is saying that his high TWR designs are better because they have a higher payload fraction. My argument is that payload fraction is not the best metric to look at for cost optimization. By adding on more propellant tanks we lower the TWR, lower the payload fraction, but increase the PTR. We deliver a heavier payload using the same engines and with only a small marginal increase in cost.

(ETA) From my initial theoretical investigations, it appears that the lowest unit cost occurs when the liftoff TWR is about 1.2. However, from a gameplay perspective, I think that is a bit too low. I find that when I pile on enough fuel tanks to bring the TWR down that low, the rocket becomes too wobbly, sluggish, and difficult to control. With a simple in-line, two-stage design, I’ve been targeting a TWR of about 1.3-1.4 with good results, though I'm not opposed to going as high as 1.5.

Edited October 20, 2015 by OhioBob

[Alshain]

I agree payload fraction is not the best thing to look at, but neither is your PTR. I don't think there is any silver bullet metric to look at for lowering cost, you just have to try different things and see which works best at the lowest cost. That's what I do. I do try to keep my TWR low (1.2 area, unless I'm using very short duration boosters, then it can be higher), but that is more because I have difficulty steering high TWR, but I also want it high enough not to crawl off the launchpad.

[OhioBob]

I agree payload fraction is not the best thing to look at, but neither is your PTR. I don't think there is any silver bullet metric to look at for lowering cost, you just have to try different things and see which works best at the lowest cost. That's what I do. I do try to keep my TWR low (1.2 area, unless I'm using very short duration boosters, then it can be higher), but that is more because I have difficulty steering high TWR, but I also want it high enough not to crawl off the launchpad.

I agree that PTR as a specific number doesn’t have any meaning across different designs. However, I think that maximizing PTR for a particular engine configuration does have meaning. For instance, if I’m designing a launcher using a Twin-Boar first stage and a Skipper second stage, then I want to launch the biggest payload I can with those engines. If I achieve that goal, then I would have maximized PTR and, most likely, minimized unit cost.

For a given engine, PTR is directly proportional to payload mass, thus it is really just a redundant metric. Saying that PTR is 0.12 vs. 0.15 is no different than saying that one vehicle can launch a 24 t payload and the other a 30 t payload. The only slight advantage that I would give to PTR is that is PTR, payload fraction, and TWR are interrelated by the equation, PTR=PF/TWR. Some people might find it easier to use a ratio rather than dealing with actual mass figures. PTR definitely has very limited use.

[GoSlash27]

Looking strictly at PTR is good for the lower half of a stage (it should always be SRBs), but the upper half really needs to focus on payload fraction in order to minimize the workload of the booster stage.

The booster is the most expensive part of the launch to orbit.

Best,

-Slashy

[Norcalplanner]

There's some good discussion here which tracks very closely with my WIP tutorial, Rules of Thumb for Building Cheap and Cheerful Rockets. If any of you fine folks wanted to go over and provide some constructive criticism or alternative viewpoints, I'd greatly appreciate it.

[MaverickSawyer]

Personally, I target about 3300-3500 m/s dV, according to Mechjeb, and a sea level TWR of 1.3 or higher. (MechJeb's been reading a little high on the TWR lately, so aim a little higher than your targeted figure...) I can usually eyeball it, though that comes from years of experience.

[LostOblivion]

From what I've been seeing, for TWRs around 1.3 or less you start to spend more dV to get to orbit, quite independently of one's piloting skills.

In my experience in situations where the TWR would otherwise be low yet enough for liftoff, it's often better to replace a Skipper engine with a Mainsail, even thought the latter adds 3t (and thus reduces dV), because you end up using less dV to get to orbit.

I suspect this is the result of a similar mechanism to why in 0.9 one tried to get a TWR that resulted in the ship ascending at close to Terminal Velocity: to get out of the gravity well (and from paying the fuel price of simply fighting gravity) as fast as possible but not so fast that the losses due to aerodynamic drag were larger than the gains from it.

In 1.0 and the new drag model, means that Terminal Velocity is near impossible to reach, but in the lower atmosphere there is a vast increase in aerodynamic drag losses at hypersonic speeds (when there are flames around the ship) so one tries to ascend through the first 30km or so just below the level at which flames start to appear but otherwise as fast as possible, and that sits at around the 1.3 - 1.5 range for TWR in the launchpad.

This is my experience also. 1.3 to 1.5 in TWR is best for Kerbin.

[Saltless Lemons]

Okay.... It might be worthwhile mentioning that I'm in grade 9, and don't really understand rocket science . I get the basics, understand dV and TWR, I can get to Minmus pretty easy, but that's about it. It might also be worthwhile mentioning that the 'boosters' were not SRBs, just liquidfuels, and that they were drawing from the same fuel as my main engine. (I did have SRBs attached as well.) So. The stats:

Without LF Boosters:

TWR: 2.22

dV: 2027 m/s

Mass: 130.5 tonnes

With LF Boosters:

TWR: 2.47

dV: 2360 m/s

Mass: 133.8 tonnes

Hmm... That's not what was happening before... A few days ago, attaching the boosters would increase TWR but decrease dV. Not sure what changed, I haven't touched it since.

Either way, thanks for all the help guys, I appreciate it!

[Acet]

Hmm... That's not what was happening before... A few days ago, attaching the boosters would increase TWR but decrease dV. Not sure what changed, I haven't touched it since.

Either way, thanks for all the help guys, I appreciate it!

Maybe the order in which you are firing the engines has changed or the order in which you are dropping spent components is different.

Remember that the efficiency of the engines (i.e. Isp) affects dV - all other things being equal, the more the Isp of an engine the more dV you get for the same amount of fuel - and solid boosters are notoriously inefficient (they're only useful because of being dirt cheap), which is why they are usually only used on the 1st stage to be fired.

Also liquid engines in KSP have lower Isp values the higher the atmospheric pressure (the thicker the atmosphere around them, the less efficient the process by which they work is) - so that's why you see them listed with two Isp values, one for vacuum and one for Kerbin atmosphere at sea level - and their Isp at any point during launch depends on atmospheric pressure at that height (the higher altitude the lower the pressure and closer their Isp is to the vacuum one). What this means is that the later you fire a liquid engine (when going up in atmosphere), the more its Isp, so if for example you order your stages so that you LFO engines only get fired after the SRBs are spent, then you get more dV (since when they do start using their fuel, they are higher and thus under less atmospheric pressure, so are more efficient).

Last but not least, you should strive to drop useless mass as soon as possible. For example, dragging SRBs along once they are spent is a waste of fuel (since the fuel left in the stages still in use is now having to drag up not only the mass of the bits of the ship still useful but also the mass of the empty and entirely useless solid fuel boosters).

So purely by changing the order by which things fire (even simply from using LFO engines at the same time as the SRBs versus using them separately) or the time when spent sections get dropped, your dV will change.

Edited October 22, 2015 by Acet