TWR & Fuel Efficiency question

[dognosh]

Is there such a thing as too much TWR?

Scenario : Launching a refueler(it will orbit kerbin at 700k) with a mix of SRBs and liquid fueled engines. My aim is to get it to 700k with as much fuel as possible. I played with the staging and various SRBs to achieve this. I have had TWRs from 1.4 to over 2 for the 1st stage.

Is there an efficient TWR that you guys stick to? (assume I use MJ2 to get to orbit using the same trajectory everytime) or the more the merrier ?

ty

[UnusualAttitude]

High TWR is good in theory since you reduce gravity losses (a slower ascent is less efficient as you are fighting gravity for a longer time). But a high TWR rocket may also be difficult to control and flip out or disintegrate as you gain speed in the thick lower atmosphere. So the ideal TWR is the highest TWR you can get away with without unplanned disassembly. 

A TWR of 1,4 at launch should be fine. 2 might be pushing it a bit, depending on your design. (Is your rocket long and thin? does it have fins?)

[Aegolius13]

I tend to launch at about 1.5, let in creep up to 2, then throttle down to keep it around 2 for the rest of the ascent.  Don't know if that's optimal, but it seems to work well.

[regex]

I usually go for a launch TWR of around 1.2 ~1.3 in stock and RO, works the best for me and generally results in a launch profile I am comfortable with. I do launch with higher TWR, most especially with SRM lowers (Minotaur, for instance), but they require more aggressive turning and control to achieve a flatter trajectory.

[crackerbal]

over 2 twr at lounch is to hight, if u dont moderate your acceleration your rocket will start to flipping around 8-10k coz of to hight acceleration and drag

usually i'm aming for a 1.6 -1.9 twr at lunch, assuming it will raise as long as u brun fuel and gain alitude, i try to keep it between 1.75-2 in flight by reducing the thrust, and around 1.5 after 10km-14km, and when i say i'm aiming for 1.6-1.9 it's mean that usualy everythink that exeds 1.6 is spend as initial booster and drop after less then minute

i belive the worst of all is having to hight thrust compared to the low, i mean it's better to lounch with 1.2 twr then 4.5 anyways the best to calculate this is to estimate the velocity u want at some point of altitude ex:u need to put 200t rocket at 500m/S on 10k ( ksp wiki for the calculations formula ) btw once you did your first 45° gravity turn, continue the turn and just try to aim for "time to PE" and keep it between 40-60sec by adjusting your thrust ( even if it's below 1 twr) and u should be perfect

i seen a post with a great twr discuss i will try to find it and put it here.

btw i'm actually searching for an optimal twr on eve if anyone have an idea, actually with a twr of 1.4 at 2km from sea level i accomplish a twr of almost 4 at 10k, with only 15% thrust at this point... the atmo and gravity on eve are so broken....

[Temeter]

46 minutes ago, UnusualAttitude said:

High TWR is good in theory since you reduce gravity losses (a slower ascent is less efficient as you are fighting gravity for a longer time). But a high TWR rocket may also be difficult to control and flip out or disintegrate as you gain speed in the thick lower atmosphere. So the ideal TWR is the highest TWR you can get away with without unplanned disassembly. 

High TWR is generally bad, because engine cost and weight are important factors. Rather add another fuel tank if you got the engine power. 1.2 TWR is minimum, tho.

One thing to note: SRBs in stock KSP are incredibly cheap, so boosting your original TWR with a few boosters while gaining some delta v is a valid strategy. In comparision, LF engines are far too expensive to spam them.

Edited October 28, 2016 by Temeter

[Streetwind]

59 minutes ago, dognosh said:

Is there such a thing as too much TWR?

Scenario : Launching a refueler(it will orbit kerbin at 700k) with a mix of SRBs and liquid fueled engines. My aim is to get it to 700k with as much fuel as possible. I played with the staging and various SRBs to achieve this. I have had TWRs from 1.4 to over 2 for the 1st stage.

Is there an efficient TWR that you guys stick to? (assume I use MJ2 to get to orbit using the same trajectory everytime) or the more the merrier ?

ty

 

I believe you will enjoy a read through this thread It's from a slightly older version, but there were only minor changes to the parts and physics since then, so you should be able to just copy a design and get its advertised performance.

 

Edited October 28, 2016 by Streetwind

[Jonfliesgoats]

I shoot for about 2 at a minimum.  High TWR let's you rapidly accelerate to 300m/s, get above most of the atmosphere by 10-15k and then rapidly ascend.

You can offset some of the flipping through gentle application of throttle, too. 

[Norcalplanner]

It depends on a lot of factors. Designing for highest payload mass fraction in sandbox is very different from designing for lowest funds/ton in career.

There's some good info and discussion in a thread I did a few months ago:

This version is specific to RSS, but there's a link in the OP to another thread with more testing in stock scale.

And to those who are throttling down, I'd strongly encourage you not to. Unless your craft is uncontrollable, burning up, or otherwise having trouble, keep the pedal to the metal for the entire ascent. Ignore the flames - it's purely a visual effect. Pay attention to any thermometers which appear, and slow down only if it's needed to keep the temperature bar from filling up.

[Aegolius13]

4 minutes ago, Norcalplanner said:

And to those who are throttling down, I'd strongly encourage you not to. Unless your craft is uncontrollable, burning up, or otherwise having trouble, keep the pedal to the metal for the entire ascent. Ignore the flames - it's purely a visual effect. Pay attention to any thermometers which appear, and slow down only if it's needed to keep the temperature bar from filling up.

At least in 1.2, I've found pitch control to be the bigger reason for keeping TWR somewhat low.  There seems to be a pretty fine line between having your skip barely turn, and having it pitch too far and get stuck in low atmosphere. This seems even worse when your rocket is going super fast.  Keeping TWR under 2 for the first portion of the gravity turn had worked better for me in this regard.

[zarakon]

High TWR results in less gravity loss, but the amount of delta-v you can gain by adding more fuel is greater (as long as TWR doesn't get too low)

[John Cochran]

TWR is good, but an efficient launch is even better. Instead of using MechJeb for your launch, try using GravityTurn instead. It's rather startling at how overpowered many rockets in KSP actually are. The thread for GT is here at 

Word of warning involving GT. Launches with GT are SLOW. But they are quite efficient and you can reduce the TWR quite a bit once you understand where in the flight profile that you really need the thrust. So lighter engines and more fuel actually transported to orbit. You'll still want to use MJ to circularize the orbit and for its other functions since GT just handles launches as efficiently as possible.

[Norcalplanner]

54 minutes ago, Aegolius13 said:

At least in 1.2, I've found pitch control to be the bigger reason for keeping TWR somewhat low.  There seems to be a pretty fine line between having your skip barely turn, and having it pitch too far and get stuck in low atmosphere. This seems even worse when your rocket is going super fast.  Keeping TWR under 2 for the first portion of the gravity turn had worked better for me in this regard.

Totally agree. If you look at the stock version of the thread I linked above, my recommendation is to launch at an initial SLT TWR of 1.4 to 1.7. But don't be afraid to let the TWR increase as the stage burns fuel. Many of my rockets have  a TWR  of 4 or 5 just before a staging event. If your rocket has the thrust and is shaped properly, there's no reason why you shouldn't already be going 600 to 700 m/s at 10 km, at 40 degrees above the horizon.

[Jonfliesgoats]

Full throttle reduces DV loss to gravity to be sure.  With a bulbous, goofy payload, throttling for control becomes a necessity, but the window where aerodynamics causes problems passes quickly.  So even with control difficulty a high TWR can be used to minimize gravity losses.  When the throttle comes back up, smooth throttling allows a manageable control situation and gives you a chance to slowly approach controllability limits v. bashing your head into those limits with extreme vigor.

I think we lose like ten percent of ours to gravity and only a few pwercent to aerodynamics, so..,you know...get fast as fast as you can.

[dognosh]

WOW ! Amazing knowledge here, thanks a bunch guys. Will read the linked threads when I get back on sunday.

[Jonfliesgoats]

My gravity turns always suck.  The math behind efficient gravity turns looks great, but the effectof me constantly making course corrections, correcting a little too much one way or the other robs me of a fair bit of DV.  Also, if I am interested in saving DV by doing a gravity turn it means I don't have much DV to spare.  

If I don't have much reserve DV, the success of my launch depends entirely on the smoothness of my thumbs.  So for every time I manage to get some extra mass into orbit with a good turn, I wind up jacking up a launch and raining tons of orbiter down on kerbin.

[Nich]

@Norcalplanner you forgot to link your cheap and cheerful challenge.  Tons of good info in there that most people forget about when trying to minimized dv to orbit.  less dv != less cost

Edited October 29, 2016 by Nich

[pincushionman]

Remember that what you actually want is horizontal speed - the only reason to go vertical at all is to get obstacles (terrain, atmosphere) out of your way so you can kick back and relax with that sweeeet horizontal speed. Any effort put into vertical acceleration is eventually lost, while the horizontal speed you keep forever.

With high TWR, you have to dedicate more effort to vertical acceleration early, otherwise you'll be in too thick an atmosphere when you're trying to go horizontal. That means you have to bring more dV to the party high up to make up the horizontal speed you couldn't get earlier - so you end up paying for that speed twice.

With a more manageable TWR, you can manage your turn much more effectively, adding significant horizontal speed while you're still deep in the atmosphere, but not yet going fast enough to risk losing control of your rocket. It's quite a bit more efficient to add both horizontal and vertical speed at the same time if you can.

If you're good, you can be going nearly horizontal by 30-50 km and still be going fast enough to raise your apoapsis, not having to worry about vertical acceleration for the rest of the ride. If you're really good, you can thrust continuously from takeoff and make your orbit circular as you reach your target apoapsis. And no thrust is wasted. These stunts, however, are not possible if you have too high a TWR - you'll be forced to go vertical, coast to your apoapsis (which will rise really quickly), and make up for lost time when you get there.

That said, though, it's easier to overpower your TWR somewhat and leave yourself some wiggle room. But don't overdo it, so you can get some significant horizontal speed on the way up, or you're going to be all kinds of screwed.

[Vanamonde]

This question about how to play the game has been moved to Gameplay Questions. 

[AeroGav]

Any idea on how this relates to Spaceplanes?

My grossly simplified hypothesis, is that  "If an ideal rocket has a TWR of 1.7 and my spaceplane has a lift:drag ratio of 2.7,  then the ideal spaceplane TWR is   1.7 divided by 2.7   or about 0.6."      

This is because the ideal rocket TWR minimises gravity losses whilst not weighing the  craft down with excess engine mass.   In a rocket, engine power is used directly to counteract gravity, but in a plane this is done indirectly via the wings, which create lift in return for drag.  Hence, having a lift:drag ratio over 1 reduces the amount of power you need.

Is this anywhere close to the truth? 

[GoSlash27]

I agree with what the others have said above; getting to orbit with minimal DV expenditure isn't the same thing as getting to orbit with the least stage mass, fuel consumption, or cost.

I design my boost stage to have 1.4 t/w off the pad if liquid fueled, and 1.2 if SRB. Upper "transstage" is initially 0.7 t/w.

If your t/w is too high while in atmosphere, it won't want to cleanly follow the gravity turn. You may also run into heating issues.

I taper my thrust in the boost stage so that t/w= 2cos(pitch). It costs a little in DV, but makes up for it with controllability and precision.

10 hours ago, AeroGav said:

Any idea on how this relates to Spaceplanes?

My grossly simplified hypothesis, is that  "If an ideal rocket has a TWR of 1.7 and my spaceplane has a lift:drag ratio of 2.7,  then the ideal spaceplane TWR is   1.7 divided by 2.7   or about 0.6."      

This is because the ideal rocket TWR minimises gravity losses whilst not weighing the  craft down with excess engine mass.   In a rocket, engine power is used directly to counteract gravity, but in a plane this is done indirectly via the wings, which create lift in return for drag.  Hence, having a lift:drag ratio over 1 reduces the amount of power you need.

Is this anywhere close to the truth? 

AeroGav,

 It really doesn't translate to spaceplanes. There, your mission is to get as much velocity out of the air breathers as possible before transitioning to rockets. Rules for the rocket portion of the flight is the same as conventional rockets; 0.7 t/w minimum so Apoapsis stays 45 seconds ahead.

 For the air breathing portion, you simply want to get hypersonic. If you have enough thrust to overcome your drag and accelerate in level flight at 360 m/sec, that's all you need. Minimizing drag is far more important than raw thrust because you're not following a gravity turn. I've built successful SSTO spaceplanes with an initial t/w as low as .32. Once it gets past the Mach wall, drag drops and thrust increases, allowing you to top-out after that.

Best,
-Slashy

Edited October 30, 2016 by GoSlash27

[magnemoe]

20 hours ago, pincushionman said:

Remember that what you actually want is horizontal speed - the only reason to go vertical at all is to get obstacles (terrain, atmosphere) out of your way so you can kick back and relax with that sweeeet horizontal speed. Any effort put into vertical acceleration is eventually lost, while the horizontal speed you keep forever.

With high TWR, you have to dedicate more effort to vertical acceleration early, otherwise you'll be in too thick an atmosphere when you're trying to go horizontal. That means you have to bring more dV to the party high up to make up the horizontal speed you couldn't get earlier - so you end up paying for that speed twice.

With a more manageable TWR, you can manage your turn much more effectively, adding significant horizontal speed while you're still deep in the atmosphere, but not yet going fast enough to risk losing control of your rocket. It's quite a bit more efficient to add both horizontal and vertical speed at the same time if you can.

If you're good, you can be going nearly horizontal by 30-50 km and still be going fast enough to raise your apoapsis, not having to worry about vertical acceleration for the rest of the ride. If you're really good, you can thrust continuously from takeoff and make your orbit circular as you reach your target apoapsis. And no thrust is wasted. These stunts, however, are not possible if you have too high a TWR - you'll be forced to go vertical, coast to your apoapsis (which will rise really quickly), and make up for lost time when you get there.

That said, though, it's easier to overpower your TWR somewhat and leave yourself some wiggle room. But don't overdo it, so you can get some significant horizontal speed on the way up, or you're going to be all kinds of screwed.

This, on an body with no atmosphere you simply want an vertical component large enough to avoid hitting ground before your vertical speed is so high you get an Ap on the other side of the body. 
Now let take the extremes, an small craft with an vector engine will not be able to turn fast enough so an mechjeb launch will go mostly right up then circulate.
An better way to launch this would be to do an short vertical burn to get off ground then an horizontal burn to get orbital speed and circulate on the other side. 
On the other hand something with low TWR like many tankers and bases who might have an TWR of 1.5 on Minmus or Mun will waste lots of fuel burning vertical while building up speed, this is easier to see at Mun because of the higher dV cost.
On Kerbin you also have drag however this is less important than most think, someone did an test and found that both an higher twr and an shallower gravity turn than most uses was beneficial, 
An twr closer too 2 was nice during launch. yes it might become too high higher up as your stage become lighter. 

Another issue is that high twr require you to lift extra engines who are heavy and cost funds, for upper stage high twr is pretty pointless, If I launch something with engines and fuel itself its smart to use it as upper stage 

[AeroGav]

35 minutes ago, GoSlash27 said:

AeroGav,

 It really doesn't translate to spaceplanes. There, your mission is to get as much velocity out of the air breathers as possible before transitioning to rockets.

+1,   though it's a compromise between getting a good velocity and not carrying too many Rapiers with you to Duna.  They ain't light.  It's a case of diminishing returns because after mach 4.5 thrust starts to fall off a cliff.

Quote

Rules for the rocket portion of the flight is the same as conventional rockets; 0.7 t/w minimum so Apoapsis stays 45 seconds ahead.

It's not quite the same as a rocket upper stage though, at least it doesn't have to be.   Wings still produce lift,  obviously not as much as they do in thicker atmosphere,  but they don't really need to because orbital free fall effect increasingly supports the plane's weight.  Drag also falls with decreasing pressure : lift:drag ratio doesn't change much from what you saw in the rest of the flight once you crossed the sound barrier  So, as you get closer to space, you need less lift, and as l/d is constant, this means you have less drag :

Watch the first couple of minutes of this Duna shot -

Just after engine separation @ mach 4.2 (1274 m/s)

Grav force (Weight)  about 150kn  (about 15 tons)

Lift about 100kn (climb rate slowly increasing, so orbital effect must already be supporting 1/3 of our mass)

Lift drag ratio 3.34

Drag 30kn

Thrust 60kn (one nuke)

By mach 4.9 (1533 m/s) @ 29.7km climb rate starts trending upward again,  with 63.5 kn lift  vs a grav force of 144 kn (we've burned off a couple percent mass in fuel).   Lift drag ratio has improved slightly to 3.5 , so drag is down to 18.1kn.

[GoSlash27]

15 minutes ago, AeroGav said:

+1,   though it's a compromise between getting a good velocity and not carrying too many Rapiers with you to Duna.  They ain't light.  It's a case of diminishing returns because after mach 4.5 thrust starts to fall off a cliff.

It's not quite the same as a rocket upper stage though, at least it doesn't have to be.   Wings still produce lift,  obviously not as much as they do in thicker atmosphere,  but they don't really need to because orbital free fall effect increasingly supports the plane's weight.  Drag also falls with decreasing pressure : lift:drag ratio doesn't change much from what you saw in the rest of the flight once you crossed the sound barrier  So, as you get closer to space, you need less lift, and as l/d is constant, this means you have less drag :

Watch the first couple of minutes of this Duna shot -

Just after engine separation @ mach 4.2 (1274 m/s)

Grav force (Weight)  about 150kn  (about 15 tons)

Lift about 100kn (climb rate slowly increasing, so orbital effect must already be supporting 1/3 of our mass)

Lift drag ratio 3.34

Drag 30kn

Thrust 60kn (one nuke)

By mach 4.9 (1533 m/s) @ 29.7km climb rate starts trending upward again,  with 63.5 kn lift  vs a grav force of 144 kn (we've burned off a couple percent mass in fuel).   Lift drag ratio has improved slightly to 3.5 , so drag is down to 18.1kn.

AeroGav,

 All true, and an ideal space plane ascent will try to circ at 45km before coasting up to 70+ for the final orbital burn. I've done this with as little as 0.5 t/w, but it's rare that I ever have to add an engine to hit 0.7. I prefer to have the higher t/w so I don't have to dawdle for 20 minutes in the high atmosphere worrying about the heat.

 In the case of RAPIERs, you should never have a problem with achieving the necessary closed cycle t/w if you've got enough thrust to get supersonic in open cycle.

 As for "carrying RAPIERs to Duna", I have no experience with that, as I would never design such a ship. I believe space planes make lousy rockets, so I never design them to do anything more than ferry supplies and passengers between a surface base and low orbit.
 There are lots of folks around who do interplanetary missions with space planes. They may have some insight into this.

Best,
-Slashy

Edited October 30, 2016 by GoSlash27