TWR on the launchpad.

[Streetwind]

So I've had a long history of following the mantra of "aim for 1.7 to 2.0 TWR on the launchpad". I know what TWR means for each individual stage - liftoff, ascent, orbital. And I know how to fix TWR issues when I encounter them.

However, I've been thinking lately. 2.0 TWR at liftoff... why, exactly, is that the best way?

I mean, ostensibly you want a thrust profile that lets you accelerate at terminal velocity without too much throttling. Some people even advocate 2.0 as a mnimum for for a hard initial sprint to 110 m/s so you can be flying at terminal velocity as early as possible. It sounds very sensible; clearly you are flying more efficiently at that point.

But now follow this thought experiment: assume you have a rocket with a launchpad TWR of 2.0. This thing can do the sprint-style launch just fine. But instead, you add extra fuel to the liftoff stage only. You add extra fuel until the TWR is only 1.2 or so. As a result, the rocket will not sprint off the pad. It will crawl off the pad, laborously gaining altitude meter by excruciating meter. But, here's the interesting thing: At some point, the rocket will have consumed all the extra fuel you have added, and it will be back to 2.0 TWR and the same amount of fuel that it had at the beginning. But instead of sitting on the pad with that amount of fuel, it is already off the ground a little ways, in slightly thinner atmosphere, with some initial speed under its belt!

Obviously this doesn't work with upper stages. Those either need a certain TWR to not fall out of the sky in the middle of the gravity turn, and don't have the luxury of taking their time to burn off excess fuel; or they are orbital stages and have exceptionally low TWR requirements to begin with. But for the liftoff stage only, doing this is an option. The question I have is: why are we not doing this? Why are we doing the sprint-to-terminal-velocity instead? I'm aware that dry mass is high for tanks in KSP, but we do have decouplers for building drop tanks and asparagus bundles, so that can't be the excuse.

In other words, what am I missing here?

[Francesco]

The 1.7 TWR advice is usually given for asparagus staged rockets, so that the overall TWR of the ship can be kept constant since staging happens frequently during the ascent.

A lower liftoff TWR is, as you noticed, more advantageous on serial staged rockets.

However, there is an advantage in climbing the first kilometer quickly - raising the ISP of the engines.

The biggest example of this is the LV-N: at sea level it's 220, one kilometer up and it's already something like 320-330 - comparable to all the other rocket engines.

[Pecan]

"2" is just a rule-of-thumb and I don't know anyone who recommends a launch TWR of 2 - that's rather an average you should aim for over the first stage(s). As you say, it should start with a lower TWR and, burning fuel, end with more - say 1.5 - 2.5. The trick to that is then; how many stages do you want? The more stages the closer each can start and finish at around the magic '2', which is shorthand for "as fast as terminal velocity is increasing".

ninja'd!

[Alshain]

Yeah, for an SSTO you want to cram in all the fuel you can without going below 1.0 TWR and dropping below minimum dV because your carrying everything with you. However for multi staged rockets, that kind of thinking is generally irrelevant because your dropping the biggest weight, the engine, so TWR increases anyway.

[Streetwind]

Isp can be an argument, but not really against overloading the liftoff stage with fuel. Because again, by the time the rocket returns to its not-overloaded state, it is a little ways into the air, and its engine Isp is higher than if it was sitting on the pad

@Pecan - 2.0 is advocated by those who advocate the fastest possible sprint to terminal velocity, not in general. I should probably have worded that better, apologies.

[allmhuran]

In other words, what am I missing here?

You're not really missing anything, your thinking is correct as far as it goes. The only consideration you've left out is pure efficiency. Using a lot of fuel to go not very far (which would happen at the start of a low TWR ascent) is inefficient, requiring larger tanks (or more of them) and so on.

If you're not going for the most efficient launch but rather the biggest, or if you're launching something unstable, etc, then this isn't something to worry about.

When 0.24 comes with budgets, though, it will be.

[Pecan]

For an initial sprint I think you can consider the 30s burn-time for the smaller SRBs. Moar boosters! Dropped almost as soon as you're off the pad :-)

If that is the way people want to race off the pad then, as I said, the fast/short staging is the reason for doing it.

[Jouni]

Until now, there has been no in-game performance metric, so people have chosen to optimize the payload fraction of their launch vehicles.

By adding more fuel, you can launch larger payloads to orbit with the same engines, but with diminishing returns. Below certain initial TWR, you need more delta-v to reach orbit. Because the rocket climbs slower, its engines have to fight longer against gravity, and the gravity losses are larger. As a result, you need to add even more fuel to reach orbit with the larger payload, and the payload fraction decreases.

[Laie]

However, I've been thinking lately. 2.0 TWR at liftoff... why, exactly, is that the best way?

A TWR of exactly 2 will make you approach, but never quite reach, terminal velocity. As long as you're pointed straight up, at any rate. That rule of thumb possibly comes from times when there was no mechjeb or kerbal engineer to display your terminal velocity, much less throttle automatically.

But now follow this thought experiment: assume you have a rocket with a launchpad TWR of 2.0. [...] add extra fuel until the TWR is only 1.2 or so. [...] At some point, the rocket will have consumed all the extra fuel you have added, and it will be back to 2.0 TWR and the same amount of fuel that it had at the beginning. But instead of sitting on the pad with that amount of fuel, it is already off the ground a little ways, in slightly thinner atmosphere, with some initial speed under its belt!

You got me here. I suspect that all that additional fuel will buy you comparatively little, but I don't know.

[Moonfrog]

if your the type whos a min maxer, then by all means continue with the thread. If however you are not, then i'll tell you it doesn't really matter unless you do something extreme, the margin of error is actually very very high. Even if you get it right your only saving maybe 100 delta V at most.

Its not something that concerns me personally, since the delta V required to reach orbit and the delta V of your rocket will not change whatever u do. Personally I start with a 1.2-1.4 TWR. Don't know why anyone would tell u to start at TWR of 2. The only major air resistance u will face is in the initial 10,000 metres climb. After that I doubt u would ever reach terminal velocity ever. I believe at one point u would have to be traveling at something like 3289 m/s. efficient or not, it doesn't really matter. Its a lot of clever maths for something that isn't worth it in my opinion. As long as you don't try go at a super fast speed in the first 10,000 metres climb, you be just as efficient as anyone else.

[Xavven]

Streetwind is correct. You will actually get more overall delta-v by taking your existing lifter and adding fuel tanks until your TWR is barely over 1.0. Mathematically speaking, this happens because you are increasing your mass ratio. That is, a larger percentage of your rocket is fuel by mass when you add fuel tanks.

However, you will also lose more delta-v during your ascent due to gravity drag because you are not flying at terminal velocity. You spend more time thrusting in direct opposition to Kerbin's gravity. The reason people say you should aim for a TWR of 2 is that you need about that much in order to only use 4500 m/s of delta-v getting to LKO. With a lower TWR, you will use more than 4500 m/s of delta-v.

So, the question is, if you add more fuel until your TWR is just above 1.0, will the additional delta-v you get out of the design be enough to overcome the additional delta-v that it will take you to get to LKO?

Well, let's find out experimentally! I built two SSTO rockets and flew them to a 100km LKO. The first one had a payload of 1.29t, 2 orange tanks, and a mainsail (TWR about 2). The second one had the same payload, but 4 orange tanks, and a mainsail (TWR about 1). Here are their stats:

Rocket 1

Payload 1.29

Engine 1 Large Mainsail 1

Engine 2

Tank 1 Large 2880 Tank 2

Tank 2

Fuel Remaining 303

Delta-V Remaining 645

Initial Mass (m0) 79.29

Final Mass (m1) 15.29

Mass Ratio (m0/m1) 5.19

Total Thrust 1500

Acceleration Full 18.9

Acceleration Dry 98.1

Delta-V in Atmosphere (ÃŽâ€v) 4526

Delta-V in Vacuum (ÃŽâ€v) 5334

Rocket 2

Payload 1.29

Engine 1 Large Mainsail 1

Engine 2

Tank 1 Large 2880 Tank 4

Tank 2

Fuel Remaining 730

Delta-V Remaining 968

Initial Mass (m0) 151.29

Final Mass (m1) 23.29

Mass Ratio (m0/m1) 6.50

Total Thrust 1500

Acceleration Full 9.9

Acceleration Dry 64.4

Delta-V in Atmosphere (ÃŽâ€v) 5145

Delta-V in Vacuum (ÃŽâ€v) 6064

The first rocket ended up with 303 units of liquid fuel left, equating to 645 m/s delta-v remaining once in orbit. The second rocket ended up with 730 units of liquid fuel left, and 968 m/s delta-v left (yes, accounting for the extra 8t of empty tanks it has to haul around still). Isn't that interesting? The second rocket theoretically had between 619 m/s and 730 m/s more delta-v on the pad, and ended up with 323 more delta-v in orbit. This means that the inefficient ascent of the second rocket used up an extra 300-400 m/s delta-v. It's also interesting that it took twice as much fuel (5,760 units) just to get another 323 delta-v.

Here are screenshots:

First rocket

Second rocket

[Kryxal]

Screenshots are missing ... and I thought the concept here was to use the extra fuel in DROP tanks. Taking the tanks with you is a bit painful...

[Claw]

... and I thought the concept here was to use the extra fuel in DROP tanks.

I was going to say something similar to what Xavven posted (but without all the fancy test flying).

Taking off with lower TWR will result in a less efficient climb. However, as he pointed out, you will end up in orbit with more dV overall.

While Xavven's example was possibly taken to an extreme, I think it illustrates the point (and maybe more simply without adding the complexity of staging). Even if you are dropping the tanks, it still takes longer to get up to speed and will ultimately use more fuel because you are lifting more mass.

Really I think it all boils down to what tradeoffs you want to make, and what goal you have in mind.

[LethalDose]

Yeah, for an SSTO you want to cram in all the fuel you can without going below 1.0 TWR and dropping below minimum dV because your carrying everything with you. However for multi staged rockets, that kind of thinking is generally irrelevant because your dropping the biggest weight, the engine, so TWR increases anyway.

Just for the record, SSTO spaceplanes can make orbit without ever having a TWR > 1 because they're never relying on thrust alone to counteract gravity.

I have a few "low-tech" (pre-RAPIER) SSTOs that have TWR that range from 0.6 - 0.8 based on fuel load and engine system that can comfortably make a circular 125 km orbit and return without refueling. (I use FAR, so stock's aero model may not allow this, but I kinda doubt it...)

[Claw]

Yep, you can also make it to orbit with a space plane SSTO in stock.

Edited July 13, 2014 by Claw

[LethalDose]

Yep, you can also make it to orbit with a space plane SSTO in stock.

Sorry if I was ambiguous: I meant I wasn't sure if you can make it to orbit with with a spaceplane in stock with a TWR that never exceeds 1. Since FAR changes the aero model (including, I assume, the details of aerodynamic lift), I wasn't sure if that precluded planes with TWR <1 to make orbit.

I'm guessing it doesn't, but the academic in me insists on listing all the conditions of my observations.

[Claw]

I guess I was also ambiguous. You can make it with a TWR that never exceeds 1. Although it can be pretty painful. Mostly that's because you have to reduce the takeoff TWR so much because of the TurboJet thurst curve. Takeoff TWR needs to be less than 0.5, and the climbout takes forever.

[Alshain]

Just for the record, SSTO spaceplanes can make orbit without ever having a TWR > 1 because they're never relying on thrust alone to counteract gravity.

I have a few "low-tech" (pre-RAPIER) SSTOs that have TWR that range from 0.6 - 0.8 based on fuel load and engine system that can comfortably make a circular 125 km orbit and return without refueling. (I use FAR, so stock's aero model may not allow this, but I kinda doubt it...)

I'm aware of this. He said Launchpad, not Runway.

[ola]

If you use FAR you should have a lower TWR IMNHSO, around 1.5 or so. Preferably you shouldn't go over ~2.0 until max Q, which seems to occur around 7000-8000 meters. After that, Jeb's your uncle.

A lower TWR than, say, 1.4 seems to burn lots of delta V just to keep the rocket flying.

[Streetwind]

Thank you Xavven, that perfectly illustrates what I was thinking about! You hit all the relevant points.

So basically, it is indeed favorable for total dV to overload the liftoff stage until it can barely take off. However at that point you're already running into diminishing returns and burning fuel very inefficiently. In a world where tanks and fuel cost nothing and you're out to min/max, doing this can give you extra dV, but with money being an issue soon it might require consideration whether or not it's worth it.