What is the optimal Thrust to Weight ratio for launching, space travel and landing?

[TheRhinelander]

So I've been playing ksp for a little over a year and I've always tried to keep the TWR just above 1.00 to cram as much fuel and delta V into a rocket as possible while still able to lift its own weight. This works good enough, but now i'm building rockets with TWR of ~1.3 to 1.7 and time to orbit decreased by at least a good minute. Of course having a lower TWR saves a little fuel by not punching your way though Kerban's atmosphere, but with that extra fuel tank I'm still carrying the dry mass to space (or orbit depending on the setup). Once in orbit however a TWR doesn't really matter as much because your not counteracting any notable gravity but I'm not a big fan of long burns to get more delta V. My landing craft for someplace like the Mun usually has like a ~8.00 TWR, the responce time is great So is it really worth sticking on that extra tank?

EDIT: thanks for the replies! I'm going to try this in a little bit!

Edited December 26, 2013 by TheRhinelander

[Murph]

Approx 2.0 from the pad up to orbit, but especially from the pad to at least 25,000m. In practical terms, that might mean around 1.5 at the pad, rising to 2.5 before staging. At least 25,000m because that's where the atmosphere is thickest and atmospheric drag is a major issue. Too low, and you're inefficient from spending too long countering gravity. Too much, and you're wasting fuel fighting drag. Less than 1.0 at the pad means you will not be going to space today, and close to 1.0 means you will be going to space painfully slowly and inefficiently. It's not unreasonable to have a slightly high max TWR off the pad due to SRBs, but to run at reduced throttle (which reduces the actual TWR) until SRB separation.

In space, more or less as low as is practical to complete your burns in a reasonable time (e.g. less than 5 mins). Somewhere around 0.5 is probably good, usually doesn't need to be more than 1.0. Having a high TWR for your orbital burns is mostly wasteful, unless the same stage needs that TWR for something involving ascent or descent.

For landing, somewhere between 1.5 and 2.0 (and that needs to be scaled to the local gravity for the planet or moon you're landing on) is probably sufficient, but possibly more for convenience/safety when coming in with no atmospheric drag. Somewhere around 2.0 for the subsequent ascent after landing, if that's planned, same as from the Kerbin pad.

There is no single answer to the optimal TWR for each situation, it's never an exact thing, due to the difference in mission, standard fuel tank sizes, etc. It's good to be close to optimal, but doesn't need to be obsessively close, as diminishing returns apply once you're moderately close to optimal.

[Nao]

With current drag model lifting the payload to orbit on planets with atmosphere with best mass efficiency is around 1.7 TWR (1.8 for best fuel eff.), that means you start the stage at ~1.4 and end it at 2.1.

Ascent from atmosphere less bodies more depend on ISP of engines used, and its from only 1.2 for LN-N to 1.5+ standard ~350-370 Isp engines.

When in orbit the less engines the better but burns below 0.2 TWR start to suffer from increased dV demands due to Oberth effect and steering losses. I would say 0.1 TWR should be a lower bound for standard burns.

For ascent on Kerbin TWR lower than 1.3 is quite wasteful as you burn much more fuel to fight gravity than you would burn fighting aerodynamic drag. For a rocket with unlimited thrust the best ascent TWR is around 2.1 .

2 for balancing the aero drag with gravity drag one to one, and 0.1 for acceleration of the craft itself. But because engines have weight too, its more efficient to use less powerful engines at 100% constantly than throttling down much heavier engine that can give above 2 TWR at launch.

Oh and a fun fact for atmosphere-less planets: If you want to land and ascent on only one stage with LV-N engines, best starting TWR is actually below 1 (0,8-0,9 depending on planet/moon).

edit: agree with above ninja

[Smidge204]

Obviously more than 1.0 for takeoff. Most people seem to agree that a TWR of 1.5 to 2.0 is ideal for heavier rockets where higher G forces may cause failures. Remember that as your ship uses up the fuel, mass decreases but thrust remains the same... so a TWR of 1.5 on the launchpad might be well over 2.0 just before it burns out.

For motoring around in space, TWR is virtually meaningless and is more a convenience than a requirement in many situations. Higher TWR means maneuvers take less time, which might be necessary if you have limited time to act, but often you have maneuver windows from several minutes to tens of days or more, so a little advanced planning means you can get by with a TWR as low as 0.2 (or even less if you're patient!)

For landing, it depends on where you're landing. I prefer to have TWRs of at least 3.0 or more for landing craft without chutes because you often have VERY limited time to react. You can always throttle down. If you have chutes it's less critical, possibly even unnecessary.

=Smidge=

[numerobis]

Depends what you mean by efficiency. If you want to minimize deltaV, in the vertical portion of atmospheric flight you want TWR about 2; you don't need as much in the turn, but I haven't seen a proof of optimality of anything. On an airless body, you want infinite TWR if you don't have any mountains to avoid -- if there's mountains you need to do a gravity turn first, for which you only need finite TWR. Once in orbit, TWR is irrelevant.

However, if you want to minimize total mass, the engine mass matters and you generally want lower TWR. In practice, around 1.5 through the atmosphere works well with stock parts.

[capi3101]

~1.7-2.2 is optimal for a Kerbin launch; higher than that and you're losing delta-V to drag, lower than that and you're losing it to gravity. That's one of the big advantages of a well-done asparagus setup; you keep your figures in that sweet spot.

That said, I regularly get into space on a single stage booster with a launch TWR of 1.2 just for the hell of it, so really what matters is what works best for you.

Incidentally, there is never any such damn thing as infinite TWR; there's always gravity. There's just not as much of it once you get into space.

[Skorpychan]

Launching, I prefer 1.5-2.5. Higher the better; I prefer to blast off like a firework while dropping small outer tanks, then throttle back once the extra weight is dumped, before going full throttle at the gravity turn.

Space travel, somewhere up 0.2 to 0.5 is optimal for maneuver node timings. For rendezvousing, mechjeb seems to prefer a higher TWR, and dislikes my heavy fuel tanks with tiny radials on them. For planetary encounters, I find ~0.4 is good for a balance of efficiency (lack of dry weight, needing less fuel to drag the fuel) with time making the node (20 minute burns, even on x4 physics warp are a grind.)

[arq]

Launching in an atmosphere you want a TWR in the range 1.7-2.2 (localized for the body's gravity). Landing in an atmosphere depends on the thickness of the atmo and the surface gravity, but anywhere above 1.3 should do fine, unless you have parachutes which mean you need very little or no thrust. Really, doing engines-only is feasible because the atmosphere will do most of your braking (ie on Kerbin you only need to burn off the last 100m/s).

Launching or landing in a vacuum you want as much thrust as possible to get into/out of orbit as fast as possible, though technically anything over 1 will work (and for descent you could even start lower, since it should increase). Something like 5 or greater will make it significantly easier and more efficient.

For moving around in space, a lower TWR implies less engine mass, which increases fuel fraction and thus dV. However, if you have less than 0.5m/s^2 (Kerbin TWR=0.05), you will likely regret it. Not to mention lose efficiency because of long maneuvers.

[blizzy78]

Generally you want to have a launch TWR so high that you can go as fast as possible, but not so high that you exceed terminal velocity. Around 1.6-1.7 should be good.

[Moonfrog]

Its very difficult to give an answer, since no one really provides any numbers to compare. Everyone seems to have there preference. I find the delta V u get from adding fuel usually provides more than enough Delta V to compensate any loss. I once built a rocket with 4600 delta V. It made it into orbit starting at 1.2 TWR for first stage. Then 1.5 TWR for second stage to orbit.

Unless someone tells me u can gain an extra 500 delta V at least from being this optimal on takeoff, I don't really see the point of worrying about it. Twr to me only has to be high enough so that when I reach the point of going into where I push horizontal at the apophasis that its high enough to do that optimally. I only try to be optimal for the first 10km. I only take drag into account. I don't care about reaching infinite speeds to be optimal when atmosphere is nearly gone. By that time its not important anymore.

1.2 to 1.4 works for me at launch.

[Murph]

Unless someone tells me u can gain an extra 500 delta V at least from being this optimal on takeoff, I don't really see the point of worrying about it. Twr to me only has to be high enough so that when I reach the point of going into where I push horizontal at the apophasis that its high enough to do that optimally. I only try to be optimal for the first 10km. I only take drag into account. I don't care about reaching infinite speeds to be optimal when atmosphere is nearly gone. By that time its not important anymore.

1.2 to 1.4 works for me at launch.

Launching with a TWR of 1.2 very likely does cost you upwards of 500m/s dV in orbit, to be honest, and is quite unlikely to end up close to optimal in the first 10km, unless it very quickly rises to be close to 2.0.

You're correct about it not being so important in the upper atmosphere, but we're not talking about "infinite speeds". If you do something close to an optimal gravity turn (i.e. gradually and progressively push the nose over as you get higher, not a single sharp turn), it's about chasing terminal velocity up to around 25–30,000m, then aiming to have 2000+ m/s orbital velocity by the time you escape the atmosphere (since that's what you need for successful orbit).

[Psycix]

Its very difficult to give an answer, since no one really provides any numbers to compare. Everyone seems to have there preference. I find the delta V u get from adding fuel usually provides more than enough Delta V to compensate any loss. I once built a rocket with 4600 delta V. It made it into orbit starting at 1.2 TWR for first stage. Then 1.5 TWR for second stage to orbit.

Unless someone tells me u can gain an extra 500 delta V at least from being this optimal on takeoff, I don't really see the point of worrying about it. Twr to me only has to be high enough so that when I reach the point of going into where I push horizontal at the apophasis that its high enough to do that optimally. I only try to be optimal for the first 10km. I only take drag into account. I don't care about reaching infinite speeds to be optimal when atmosphere is nearly gone. By that time its not important anymore.

1.2 to 1.4 works for me at launch.

Try adding an engine and raise the TWR to 1.5 to 1.7 (some radials or a more powerful one).

Yes, your DV will become slightly lower, but the delta-V needed to reach orbit will also become lower as you spend less time countering gravity.

[tavert]

Here's a tool that'll give you some numbers on how to maximize payload fraction for airless-body landing and/or takeoff (currently only looks at single-stage vehicles): http://forum.kerbalspaceprogram.com/threads/61659-Wolfram-Web-App-Optimal-Single-stage-Lander-Design-Tool

[Timo M]

My moho transfer stage had 0.21 twr and i absolutely hated it. I was out of kerbin’s SOI with 700km/s to burn (15 min) it takes much more dv cas no oberth effect

[Kelderek]

As long as you can control your vessel all the way to orbit and not explode from overheating in the process then generally go for as high a TWR as you can.  The only time a lower TWR is really needed on my rockets is when the design is less streamlined and has more drag -- this makes it harder to control and benefits from lower acceleration to keep it stable.  If you ask for "optimal" you're going to get a lot of different answers because everyone has different criteria that they care about.  I once launched a rocket with 1.03 TWR and got to orbit -- that is an extreme example, but it worked for me in a special case where I had a gigantic rocket (something like 10,000 tons) and the only thing that mattered to me was reaching orbit, which I was able to do.  Determine what criteria you care about the most and then choose a TWR that allows you to achieve it.

[Temstar]

Are you two aware that you're replying to a thread from 2 years ago?

Anyway you might have noticed that people's response from back in those early days are clustered around 1.7g to 2.0g for lift off. One of the big reason for this is cost was not a factor. Back then there was no career mode and all parts prices were only place holder and for academic interest only, Reliant (or rather LV-T30, as it didn't have its nickname then) and Mainsail had the same cost. So the only thing that mattered and people could use to compare rocket performance was payload fraction.

If you ask about TWR at lift off nowadays you'll probably find people suggest somewhere between 1.3g-1.5g. The lower acceleration reduces aerodynamic forces so is easier on rockets with dodgy centre of pressure. But more importantly career mode is now a thing and parts prices actually matter. As with real rockets engines are expensive while fuel and fuel tanks are relatively cheap. So nowadays for career mode rockets cost per ton to orbit is much more important than payload fraction. And since engines are expensive you want to hang onto them for longer and get more bang out of the bucks you paid for them. Increase the fuel load for the first stage is a simple way to achieve this. By adding fuel to the first stage you decrease cost per ton to orbit (because your first stage is now doing more work for the same amount of engine) as well as payload fraction (because your rocket now suffers more gravity loss from low TWR first stage).

 

For transfer stages I like to keep the TWR above 0.3g, for landing I prefer 1.5g-2.0g of local gravity.

[GoSlash27]

Well... I suppose since it's been necro'd it is worth commenting on since so much has changed.

There are my preferred design criteria for KSP 1.05:

Booster: 1.2 (cheap) to 1.4 (light)
Transstage: 0.7 minimum
Transfer: 0.5 minimum
Lander: 1.5 minimum

It should be noted that there are a few different ways to define "efficiency". The most commonly used is minimum DV expenditure, and this is rarely an ideal goal to shoot for.

Some others would be  minimum stage mass, minimum stage cost, and minimum fuel consumption.

Best,

-Slashy
 

 

Edited January 25, 2016 by GoSlash27

[Snark]

10 hours ago, Kelderek said:

As long as you can control your vessel all the way to orbit and not explode from overheating in the process then generally go for as high a TWR as you can.

Bearing in mind that engines are heavy, and having more engine than you actually need is just dead weight that will cost you in terms of dV.

The main thing that high TWR buys you is lower gravity losses.  That's a valid concern, gravity loss is expensive.  However, gravity loss is only a really serious consideration for a relatively small percentage of your total dV to orbit.  By the time you're at 20 km altitude, you should already be pitched to just 30 degrees or so above the horizontal, at most; in which case gravity losses are less than 1/7th what they were at launch.  When you're at that stage, you really don't want to be lugging more engine mass than you need.

[ineon]

4 hours ago, GoSlash27 said:

Some others would be  minimum stage mass, minimum stage cost, and minimum fuel consumption.

Another importantly overlooked one is time*. If you have time pressure on your Kerballing time (work, school, family, boy/girlfriend, ...) then having a highly time efficient launch/transfer/landing can mean more time for other contracts ( = more money).

*Although not relevant in sandbox mode.

[fourfa]

16 hours ago, GoSlash27 said:

Lander: 1.5 minimum

Local to the gravity of the world being landed on?  Just thought it might help to clarify, as the others are probably local to Kerbin

[GoSlash27]

7 hours ago, fourfa said:

Local to the gravity of the world being landed on?  Just thought it might help to clarify, as the others are probably local to Kerbin

Yes, absolutely. Local surface gravity, not Kerbin.

Best,
-Slashy

[RizzoTheRat]

I tend to favour a bit more than that for landers, I've had too many unscheduled kinetic penetrator experiments with unexpected high ground

[fourfa]

Me too - 2.0 local TWR gives a lot more flexibility for landers. That's my personal minimum for safety.  But 1.5 is certainly possible with luckskill!

[SanderB]

when getting from Kerbin's launchpad to orbit I primarily think in terms of what takes the least time and secondly funds. For aerodynamic payloads this usually means making a 1st stage with a TWR of ~0.9 (before SRBs) and several solid boosters on the side. The solid boosters usually provide enough thrust to give my vessel 300-400m/s of vertical velocity before staging. When I need the mammoth or rhino engines, I leave the SRBs and go with a more conventional ~1.5 TWR.

(My formula of $ / t to LKO is (hangar cost - recovered funds) / tonnage to orbit.)

In terms of $ / payload(t) ie. cost per ton of payload to orbit I get about 1,300-1,500$ / ton for every launch, as long as that payload is aerodynamic. If we go with conventional 1.5 TWR LFO 1st stages without solid boosters, you're paying about 25 to 50% extra per tonnage to orbit. The cost of launching 

If you add reuseability, the cost efficiency of getting to orbit massively increases. Non airbreathing vehicles can go as low as 500$ / t to LKO. Airbreathing launch vehicles usually don't cost as much but they cost a hell of a lot more time to re-use and they usually don't lift large or clumsy payloads. Reuseable LFO rockets are a bit harder to use (than expendables) but with some mod support (mechjeb's landing module or the trajectories mod) you can get >96% re-useability reliably. Reuseable rockets take way less time than spaceplanes and are a lot less of a headache. All they need is a flat landing site, airbrakes, parachutes and enough fuel to do a suicide burn if needed.